Martec2010 160329072110

Paper List
Table of Contents
Keynote Papers
Prospects of shipbuilding in Bangladesh
Dr. Abdullahel Bari (Ananda Group)
Van Gogh, Gauguin, Fluidity and the gap in engineering education
Prof. Dr. Sander Calisal, Turkey
On the prediction of ship performance in actual sea conditions using computational fluid dynamics
Dr. Hiromichi Akimoto, Japan
Hull form parameterization technique with local and global optimization algorithms
Prof. Dr. Ho Hwan Chun, Korea
History, sate-of-the-arts and future trend of research on ship manoeuvrability and controllability
Prof. Dr. Kazuhiko Hasegawa, Japan
A holistic view on ballast water management
Prof. Dr. Ehsan Mesbahi, UK
xiii
xxiii
xxxi
xxxv
li
Technical Papers
Numerical study of free surface effect on the flow around shallowly submerged hydrofoil
Md. Ashim Ali & Md. Mashud Karim
Experiment of bended hydrofoil-supported ocean structure pillar
M A Djabbar & Juswan
Investigation on the stability of a trimaran wing in ground effect (WIG) craft with endplate
Adi Maimun, Agoes Priyanto, Noverdo Saputra, Saaed Jamee, Ike Suharti, Md.Mobassher
Development of a ship weather-routing algorithm for specific application in north Indian ocean
region
Debabrata Sen & Chinmaya P Padhy
Computational study on the fish-like underwater robot with two undulating side fins for various
aspect ratios, fin angles and frequencies
Md. Mahbubar Rahman , Yasuyuki Toda & Hiroshi Miki
Assessment of marine propeller hydrodynamic performance in open water via CFD
Mehdi Nakisa, Mohammad Javad Abbasi, and Ahmad Mobasher Amini
Sustainable growth in inland waterways of Bangladesh through Innovation: A Way Forward
Arun Kr Dev
Aerodynamic characteristics of a compound wing during ground effect
Saeed-Jame, Adi-Maimun, Agoes-Priyanto & Nor-Azwadi
Dynamic response of offshore triceratops under environmental loads
Srinivasan Chandrasekaran, Madhuri Seeram, A. K. Jain & Gaurav
Preliminary feasibility study and development Coastal region of Ambon city as Ambon water front
city in supporting potentials and regional competitiveness
Pieter Berhitu, Wilma Latuny & Marthin Efruan
http://www.buet.ac.bd/name/martec2010/2010/contents.htm[3/29/2016 1:19:10 PM]
1-6
6-11
13-19
21-27
29-34
35-44
45-52
53-60
61-66
67-72
Paper List
Modeling of gas turbine co-propulsion engine to ecotourism vessel to improved sailing speed up to
35 knots
O. Sulaiman1, O.S. Han, W.B. Wan Nik, A. H. Saharuddin, A.S.A.kader & M.F. Ahma
Qualitative method for antifouling long life paint for Marine facilities or system
O. Sulaiman, C.M.Sam, W.B. Wan Nik, A.S.A.kader & A. H. Saharuddin
Hatch cover analysis of capsize bulk carriers and suggestions for alternate materials
K. Kunal, C. Kannan & S. Surendran
A study on numerical analysis of unsteady cavitaing flow over two dimensional hydrofoils
N. Mostafa, M. M. Karim & M. M. A. Sarker
Static and dynamic analysis of semi-submersible type floaters for offshore wind turbine
A.C.Mayilvahanan & R.Panneer Selvam
Experimental investigation of slamming loads on a wedge
Tanvir Mehedi Sayeed, Heather Peng & Brian Veitch
Numerical investigation of bow slamming on ships with large flare
Rahaman Md. Mashiur, Kun Zheng & Akimoto Hiromichi
Drag analysis of different ship models using computational fluid dynamics tools
Salina Aktar , Dr. Goutam Kumar Saha & Dr. Md. Abdul Alim
Dynamic response behaviour of multi-legged articulated tower with & without TMD
Srinivasan Chandrasekaran, Bhaskar K, Lino Harilal and Brijith R
Investigation of the response of a boxed shaped free floating buoy in linear water waves
Muammer Din Arif, M. Rafiqul Islam, A. K. M. Sadrul Islam &, Enaiyat Ghani Ovy
Role of bank protective works at upstream in improving siltation at downstream: case study in the
Jamuna with application of numerical model
Masum Ur Rahman & Sarwat Jahan
Thermomechanically treated advanced steels for structural applications
M.A. Islam
Computational drag analysis over a car body
M. M. Islam & M. Mamun
An experimental analysis of the internal flow effects on marine risers
Motohiko Murai & Marcio Yamamoto
Wind sail analysis using computational fluid dynamics simulation
A Sulisetyono, A Nasirudin & FA Wibowo
Corrosion behaviour of aluminum alloy in seawater
W. B. Wan Nik, O. Sulaiman, A. Fadhli & R. Rosliza
Hydrodynamic calculation and effects of deadrise angle on the planing crafts
Hassan Ghassemi, Saman Kermani & Abdollah Ardeshir
Max-plus approach to scheduling problems of block assembly lines
Hiroyuki Kajiwara, Yasuhiro Hitoi & Khairul Hassan
Optimization of the preliminary design parameters and cost evaluation of container vessel for the
inland waterways of Bangladesh.
Goutam Kumar Saha, Golam Mortuja & Md. Emdadol Haque
Passing ship effects on a moored ship: A numerical Study
Md. Nazrul Islam, M. Rafiqul Islam & Md. Sadiqul Baree
Effect of diameter of heated block on MHD free convection flow in a cavity
Salma Parvin & Rehena Nasrin
An environment friendly method of driving ships with air jet propulsion
Mohammad Jiaur Rahman
73-80
81-88
89-94
95-100
101-106
107-112
113-121
123-129
131-136
137-142
143-147
149-153
155-158
159-165
167-173
175-180
181-187
189-194
195-200
201-207
209-213
215-220
Paper List
Motion responses and incident wave forces on a moored semi submersible in regular waves
Hassan Abyn, Adi Maimun, M. Rafiqul Islam & Dr. Allan Magee
Prediction of manoeuvring behaviour of an offshore supply vessel by using simulation program
Che Wan Mohd Noor, Abdul Majeed Muzathik, Wan Mohd Norsani Wan Nik, Mohammad Fadhli
Ahmad, Khalid Samo & Rajoo Balaji
Prediction of manoeuvring behaviour of an offshore supply vessel by using simulation program
Various aspects of sound ship recycling in South Asia: A compromise not a confrontation
Arun Kr Dev
Role of acids in the degradation of alkanolamine during CO2 and H2S removal process
Md. Sakinul Islam, Rozita Yusoff & Brahim Si Ali
Assessment of ship manoeuvrability in shallow waterways
Osman Md. Amin & Kazuhiko Hasegawa
Probabilistic evaluation of plate effectiveness in terms of effective width of attached plating
S. C. Vhanmane & B. Bhattacharya
Techno economic study of potential using solar energy as a supporting power supply for diesel
engine for landing craft
O. Sulaiman, H.Aron, , A. H. Saharuddin ,W.B. Wan Nik, A.S.A.kader & M.F. Ahmad
Numerical investigation on natural convection from an open rectangular cavity containing a heated
circular cylinder
Sheikh Anwar Hossain & Dr. M. A. Alim
A numerical study on hydrodynamic interaction for a small 3-D body floating freely close to a large
3-D body in waves
Mir Tareque Ali, Md. Mashiur Rahaman & Tasnuva Anan
An investigation of ship performance in seas
M.S. Baree
MHD-conjugate free convection flow from an isothermal horizontal cylinder with stress work
NHM. A. Azim, ATM. M. Rahman & M. K. Chowdhury
Ship recycling prospects in Bangladesh
Kh. Akhter Hossain, K. Shahriar Iqbal & N. M. Golam Zakaria
Conceptual framework for the evaluation of container terminals expansion by marginal approach
K.B.Loke, A.S.A.Kader & A.M.Zamani
An alternative structural design for tankers
Mohammad Ashickur Rahman Khan
Development of G-code from CAD model for rotational parts
Md. Deloyer Jahan
Effects of temperature dependent thermal conductivity and viscous dissipation on conjugate free
convection flow along a vertical flat plate
ATM. M. Rahman, NHM. A. Azim, M. M. Rahman & M. K. Chowdhury
Effect of temperature dependent viscosity inversely proportional to linear function of temperature
on magnetohydrodynamic natural convection flow along a vertical wavy surface
Nazma Parveen & Md. Abdul Alim
Development of an intelligent system for monitoring the ship engine room using mobile phone
Badrun Nahar , Md. Liakot Ali & Md. Sharfaraj Hussain
Cross-shore sediment transport due to wave: A laboratory study
Badal Mahalder and Umme Kulsum Navera
An object oriented approach to beam analysis
M. S. Islam, G. M. Khalil & N. A. Nahian
221-226
227-232
233-240
241-247
249-253
255-263
265-269
271-278
279-284
285-290
291-296
297-302
303-310
311-316
317-322
323-328
329-334
335-339
341-346
347-350
Paper List
IMO ballast water convention and enforceability of G2

Oihane Cabezas-Basurko & Ehsan Mesbahi
Three dimensional MHD flow with heat and mass transfer through a porous medium with periodic
permeability and chemical reaction
N. Ananda Reddy, S. Vijaya Kumar Varma & M. C. Raju
Estimation of sway, heave and roll of a moored floating breakwater due to interaction with waves
Md. Ataur Rahman & Silwati Al Womera
Safety strategy for Ro-Ro passenger ferries operating on Indonesian waters
Sunaryo, Antoni Arif Pribadi & Tri Tjahjono
Effects of temperature dependent thermal conductivity on MHD free convection flow along a vertical
flat plate with heat generation and joule heating
A.K.M. Safiqul Islam, M. A. Alim & ATM. M. R. Sharker
Selecting of diesel engine for heavy weather service in inter islands trade in Indonesia seawater
Yusuf Siahaya
A practical approach to triangulation of planar surface
Md. Shahidul Islam, Naem-Ibne-Rahman & Md. Mezbah Uddin
Numerical-experimental modelling of a planning vessels dynamics in regular waves
A.Najafi, S.Hamzeh, S.Kalanari & M.S.Seif
Study of incorporating kite or balloon technology to UMT research boat
M.Azlan Musa, S.S. Ilmin, W.B. Wan Nik & A.M. Muzathik
Study on some competitive parameters for shipbuilding industry in Bangladesh
N. M. Golam Zakaria, M.M. Rahaman & Kh. Akhter Hossain
Numerical prediction of wave-making resistance of tetramaran hull in unbounded water
Md. Shahjada Tarafder, Gazi Md. Khalil1 & Md. Shahriar Nizam
Sustainable container ship for inland shipping in Bangladesh
K. Shahriar Iqbal, N. M. Golam Zakaria, M. Imran Khaled & M. Milon Khan
Effect of curvature on entry flow
M. A. Masud & Md. Mahmud Alam
Response of a containership AT outer anchorage
B.Rajesh Reguram, Sajjan S.C & S.Surendran
Reconstruction of capsize type of accidents by fault tree analysis
M.T. Hossain, Z.I. Awal & S. Das
A study on pullout capacity of horizontal and inclined plate anchors for offshore installations
Baleshwar Singh, Birjukumar Mistri & Ravi Patel
Effect of FGM coating on thermoelastic characteristics of a rotating circular disk
A. M. Afsar & F. A. Sohag
An experimental study of two-phase looped thermosyphon
A. K. Azad, S. K. Sarkar, C. M. Feroz & M. M. Alam
Optimization of ship hull parameter of inland vessel with respect to regression based resistance
analysis
Goutam Kumar Saha & Asim Kumar Sarker
A theoretical method for the design of marine tandem propellers
Dr. Muhammad Refayet Ullah
Design and dynamic behaviors of marine structures
Yoshiyuki Inoue
Analysis and improvement in performance characteristics of Francis turbine
S. Hari Narayanan & Sri Saila Mallikarjunan K V
351-357
359-364
365-369
371-375
377-385
387-391
393-398
399-406
407-411
413-418
419-425
427-432
433-438
439-444
445-452
453-458
459-464
465-469
471-476
477-482
483-491
493
Paper List
Application of kite and its effect on reduction of fuel consumption and marine environmental
protection in ships
M. Sayehbani, M. Ghiasi & S. Dehesh
A concept selection of strait- crossing systems : a case in Sunda strait
Daniel Mohammad Rosyid
494
495
Proceedings of MARTEC 2010

The International Conference on Marine Technology
11-12 December 2010, BUET, Dhaka, Bangladesh
POTENTIALS, PRIORITIES AND PROSPECTS OF SHIPBUILDING

IN BANGLADESH
Dr Abdullahel Bari, FRINA
Chairman
Ananda Group
10/1, 9th floor, City Heart 67, Naya Paltan
Dhaka-1000, Bangladesh
Email: ananda@anandagroup.biz
PREFACE
Bangladesh has been since 2005 building and exporting ships to owners from Denmark, Mozambique,
Germany, The Netherlands and Finland. Germanischer Lloyd has declared in September 2008 Bangladesh as a
shipbuilding nation of international standards. Bangladesh is traditionally a sea fairing nation and has a rich
heritage of timber shipbuilding of many hundred years. Ibne Batuta traveled back in 14th century in a timber
ship built in Sonargoan, Bangladesh. Caesar Fredrick, the European traveler, reported Chittagong as one of the
best places for building timber commercial vessels in 15th century. Sultan of Turkey had his fleet built from
Bangladesh in 17th century. Bangladesh built timber warships fought in Trafalgar in 1805. Bangladesh is
presently contributing to the shipbuilding industries globally through its exported workforce. These facts do not
speak only of a heritage but of an inbuilt ability of shipbuilding of people of this region which had been for ages
dependent on waters. Bangladesh could not keep pace with the evolvement of new technologies such as, use of
man made materials replacing natural building material timbers and fuel burned power pushing away wind
power, i.e. sails. Bangladesh did not have well developed fleet for seaborne commerce nor adequate inland or
coastal carriers at her independence in 1971. There was nearly no serious shipbuilding in steel in Bangladesh
then. However only over a period of 40 years after independence, Bangladesh has developed a large fleet of
about 20,000 vessels comprised of inland & coastal commercial vessels, and various types of working and
fishing craft. There are indigenous shipyards, more than 200 in numbers (Fig.-1) building and repairing these
vessels and in turn have provided a strong prepared base for building ships of international standards.
Shipbuilding industry, to grow, needs among others, demand for shipbuilding, skilled low cost workforce,
elevated engineering skills, deep draft basins with access to the sea, a reasonable infrastructure, reasonably
good weather, long term government policy support, an attractive investment climate, allied forward &
backward linkage industries, a strong financial system, access to the shipping world and courageous
entrepreneurs. This presentation will establish that most of these Bangladesh already has and for a few
remaining, the environment is conducive for accelerated availability or a solution. National and international
issues and factors which may apparently not encourage shipbuilding in Bangladesh will be addressed to,
impacts of them evaluated and measures towards finding a solution, if there is any, incorporated.
1. POTENTIALS
1.1 Workforce
Shipbuilding, an ancient assembling industry
producing tailored products, accordingly having the
largest human input per unit of produce, is always
moving to countries with lower wages of required

skills. Bangladesh has comparatively a lower cost of
human inputs and can offer the best combination of
cost, quality and productivity with its fast growing
young workforce.
Table-1: Comparison of labour costs

Bdesh
0.5
India
1.0
China
1.5
Singapore
3.0
S.Korea
6.0
xiii
Japan
12.0
Italy
13.0
France
13.0
Norway
14.0
Finland
15.0
Proceedinggs of MARTEC 20010
Figu
ure 1: Maximum
m Available waater depths
1.2 Produ
uctivity
1.4 Quallity
A compparison of prodductivity (Table-2) shows thaat

while Chinna may be well
w
ahead of India in totaal
shipbuildinng, its producttivity is almost the same aas
India. Pro
oductivity of Bangladesh workforce iis
expected very
v
similar to that
t of India.
Quick
k learning ability of the worrkforce is a keey
element to
t help Banglaadesh move innto internationnal
shipbuildiing faster. Tim
me required foor an individuual
artisan off a category to move to the hiigher level, succh
w
has beeen
as, a semii-skilled weldeer to a skilled welder,
found onn an average to be 3 mon
nths. Indigenouus
shipyards, dockyards annd workshops as
a a whole in its
i
present sttatus is large inn numbers and
d proving a veery
large poool of practicaally experienced shipbuildinng
workforcee such as, sshipwrights, gas
g
cutters annd
welders. These artisans may be quiickly trained to
undertakee building shhips following classificatioon
societies standards annd rules. Thee workforce is
conversan
nt with Engllish and it is possible to
communicate in Engliish nearly in
n all levels. In
w
are veryy diligent.
general workers
1.3 Lower Effective Cost

C
Table-33 speaks for th
he strength off Bangladesh too
grow as a shipbuilding nation in term
ms of effectivve
cost of worrkforce.
A draw
wn comparisonn of the cost annd productivityy
of a few shipbuilding
s
coountries vis--v
vis Bangladeshh
shows Indiia the nearest competitor is 2.5
2 times awayy
and Koreaa, 4 times cosstlier in termss of workforce
alone.
Taable-2: Compaarison of produuctivity
Country
Japan (20004)
Korea (20
004)
China (20004)
India (200
06)
Bangladeesh
Compleetion M DWT
Employees
E
23.2
23.0
8.8
0.6
80,000
71,800
158,000
12,000
xiv
Producctivity DWT
Perrson/year
290
320
56
50
40
Table-3: Cost-effectiveness
Country
Weighted labour rate
Weighted productivity
Weighted average output cost
Bangladesh
0.5
1.0
0.50
India
1.5
1.2
1.25
China
1.5
1.4
1.07
Korea
6.0
3.0
2.00
Singapore
4.0
2.0
2.00
Germany
15.0
5.0
3.00
on a reasonably attained even keel of 4 m may be

built and taken to the sea from any place in between
the sea and the capital, Dhaka. Multipurpose cargo, oil
tankers and container vessels having light weight of
say, 7000/8000 tons i.e. DWT of 20000 25000 tons
may be built for export in shipyards to be established
on the river frontage shown in Fig.-1. There is a
seasonal variation of water depths in rivers due to
monsoon and flood as shown in Fig-1. A variation of
5.0m is expected in the worst case only due to flood.
There is no bridge and the electricity cable is at a
height of 150m above the highest water level. Air
draft is thus unrestricted.
1.5 Higher Level Skills

There are quite a many engineering universities
under Public and Private sector producing graduates
of various disciplines, such as, engineering, industrial
production, metallurgy, management, finance &
marketing. Bangladesh University of Engineering &
Technology (BUET), one of them, has been producing
Naval Architect and Marine Engineers since 1974.
Chittagong Marine Academy has been producing
entrant marine engineers and captains. Quite a few
universities and institutions are producing graduates
in the field of management, finance and marketing.
Indeed Bangladesh has a prestigious position
worldwide for its young white colour workforce.
2.2 Meghna basin

Keeping in view the market segment which turns
to be the most suitable for Bangladesh at least in the
beginning, Meghna basin may be the best place to
start with. Meghna basin is near to the capital
connected to the seaports and the capital by highways.
It has blessings of calm weather and sweet water.
Being inland it offers an opportunity of using
slipways at a reasonably lower investment. Tidal
variation is 400mm at Meghnaghat. Meghnaghat is
only 230 Kms from the sea and therefore needs only a
journey of 10/12 hours for a new built vessel to go to
the sea for tests & trials.
1.6 Expatriates
Since 1974, more than 50% of BUET produced
graduate naval architects and marine engineers have
been working abroad and about 60% of them in
shipyards. Ex-mariners can be great contributors in
the future shipbuilding industries of Bangladesh. In
Singapore, more than 30% engineers and workers are
from Bangladesh in shipbuilding and ship repair. In
Dubai and in shipyards in the Persian Gulf,
Bangladeshi workers represent more than 20% of the
total workforce. Shipbuilding engineers and workers
from Bangladesh are engaged in countries like USA,
Australia and Canada and numbers are increasing.
Many of them on return may join the shipbuilding
sector and inject their experience and add new
exposure to accelerate faster healthy growth of
shipbuilding in the country.
2.3 Mongla basin

Mongla basin with available water depth of
minimum 8.0 m may be considered one of the best
places for shipyards and dry-docks for building and
repair of ships of reasonable sizes. The basin enjoys a
tide water rise of more than 3.0m. The basin is very
near to the Mongla port and not far from the city and
outside the cyclone zone.
2. GEOPHYSICAL CONDITIONS
2.1 Deep draft basins
2.4 Karnofuli basin
Excluding sand bars and shoals here and there,

rivers are very deep in Bangladesh. Water frontage
which may be considered suitable for organized
planned shipyards is shown in Fig.-1. Available water
depth maintained guaranteed up to Dhaka port is
3.66m. However commercial vessels navigate with
4.25m loaded draft. Thus vessels having a light draft
The Karnofuli River has been unfortunately

handicapped by bridges of low height. Beyond the
bridge, the Karnofuli River up to the sea may be kept
outside of any planned shipbuilding activities because
other better important national cause may use, such
as, for port facilities etc.
xv
2.5 Long coast line & river falls
3.5 Backward linkage industries
The long coast line and mouths of rivers falling to

the sea offer excellent places for development of
shipyards and shipbuilding zones. The topography of
rivers and river banks bestow special advantages for
building small and medium size ships. Shankoo basin
and Meghna fall are good examples.
Bangladesh is now exporting a few items meeting

quality requirements, which may be used in
shipbuilding. Of them, furniture, pipes, iron chains,
wire and cables, light fittings, glass sheets, synthetic
ropes are mentionable. The total export earnings from
these items was US$ 32.04 million in 2008-2009.
Smaller industries such as for
(a) ship outfitting i.e. manufacture of doors,
windows, port holes, bollards, fairleads, stairs,
ladders & gangways, ventilation grills and mouths,
container fittings, cathodes, manholes covers, bell
mouths & suction strainers,
(b) safety accessories and signs i.e. life jackets,
buoys & boats with davits, survival suits, rescue
nets & baskets, maritime signs, symbols &
posters,
(c) anchor & anchor cables,
(d) shafts, propellers & stern tubes,
(e) switch gears, panels etc.
may be quickly developed in Bangladesh. Indeed the
size of investments required for these industries suit
most to the present state of industrialization in
Bangladesh. When these industries develop, the value
addition may exceed 55% of the export value of a
ship. These industries being known through feeding
national yards for their quality may export to
shipyards around the globe. The export earnings and
employment generations of these industries then may
even equate those of shipbuilding itself.
2.6 Weather conditions

Apart from the heavy long monsoon and therefore
high humidity and rains, Bangladesh has a very good
weather allowing 14 working hours a day for most of
the year. Temperature is very reasonable and cold
protection is hardly needed.
3. OTHER FACTORS
3.1 Infrastructures
Infrastructures required for shipbuilding industry
to grow are not superhighways but super fast internet
system which Bangladesh has. Bangladesh offers
excellent telephone communications. Captive power
supply is the norm of the day and a large shipyard
may easily have its own power generation.
Bangladesh
has
reasonably
developed
gas
transmission network and electricity distribution
system.
3.2 Training institutions

There are vocational training institutions financed
and managed by the government with an intake
capacity of some 35512 annually producing skilled
manpower, a portion of which may flow to
shipbuilding industries. In addition, there are
institutions under NGOs and the private sector which
impart trainings for manpower export. On demand,
private sector training institutions may feed
shipbuilding industries. Any shipyard may have
license to initiate its own training programme and get
trained personnel certified of a particular skill in
various fields of artisans.
3.6 Access to the shipping world

Bangladesh built ships enjoy duty free market
access. Bangladesh has earned confidence of shipping
world of the quality of her built ships. It is known to
shipping world through its sea farers, marine
engineers and captains and to the shipbuilding
through its work force and engineers working in
shipyards around the globe. It is also well known to
the shipping world through its ship breaking industry.
Identifications of potentials of a deep sea port reflect
the significance of its important geophysical maritime
location.
3.3 Linkage industries
3.7 Courageous entrepreneurs
Forward and backward linkage industries both of

services and hardware are essential for shipbuilding.
SME, small and medium enterprises, is a slogan in
Bangladesh. The Government, development partners
and financial institutions are jointly and individually
working to strengthen the sector.
Courageous entrepreneurs have created a

reasonably strong industrial base in just over three
decades from a purely agrarian economy of
Bangladesh which suffered extreme commercial
exploitations and social & technological oppression
including blockade to the outside world in the age of
colonialism and neo colonialism. These entrepreneurs
will be playing a commendable role in the
development of modern shipbuilding industries. The
growth will take place in the private sector, an
advantage compared to many competing countries.
3.4 Forward linkage industries

Forward linkage industries such as design firms,
model testing basins may quickly be established in
Bangladesh. For shipbuilding consumables, such as
welding materials & gases already there are a few
plants and new plants may quite easily be set.
xvi
3.8 Statuttory quality control

Indian Register off Ships (IRS
S) and Chinna
Societty,
both
governmennt
Classification
organizatioons, discharge the function likke internationaal
classificatiion society in shipbuilding.
s
T
These
countriees
are develooping substantiially and bothh are becomingg
major playyers in internattional shipbuild
ding arena. Foor
Bangladesh
h to enter intto internationaal shipbuildingg,
rigid and appropriate control
c
shall be
b ensured onn
shipbuildinng industry, ship desiggn and shipp
constructioon. Bangladeesh is plannning to havve
Bangladesh
h Register of Ships
S
(BRS).
4. WOR
RLD SHIPB
BUILDING
Shipbu
uilding demandd is a functioon of shippingg
activities and
a in turn, of seaborne comm
merce, nationaal
waterbornee trades, any water
w
borne activities, such ass,
fishing, ex
xploration, extrraction and reseervations, coasst
guarding and
a naval aspirrations. Shipbu
uilding demandd
creates shippbuilding capaacity.
In war-shaaken world, Jap
pan took shipbuuilding in 19522
as its majoor investment sector
s
and became one of thhe
leading ecoonomies in 19
970 in 20 yearrs. South Koreea
took shipb
building as a thrust sector in 1980 whenn
worldwidee recession occcurred and beccame the leadeer
in 2000, also in 20 years..
China too
ok shipbuildingg in Asian Finnancial Crisis in
1990 seriously and is a major econom
mic power todaay
in about 20 years. IIndia started providing 300%
d is earning US
S$
incentivess to shipbuildinng in 1999 and
5.00 billion now annuaally. Bangladeesh has declareed
shipbuildiing as a thrustt sector in 20009 and hopefullly
by 2019 Bangladesh
B
wiill be a rich ecconomy buildinng
& exportiing ships.
4.1 Worlld shipbuildiing capacity

Globaal shipbuilding capacity (Fig..-4) was only 75
7
MDWT against
a
a foreccasted demand of 196 MDW
WT
at the endd of 2006. Shhipbuilding faccilities had beeen
faster deeveloped in nnewly emergeed shipbuildinng
nations, such
s
as, in Chhina, Vietnam
m and India annd
sophisticaated large faciliities came up in
i Korea.
The supply equating deemand was 1200 MDWT at thhe
008. Forecastedd newbuilding demands of 1996
end of 20
MDWT could be meet adequately by created &
planned facilities
f
by ennd of 2010. Fig.-5
F
shows thhe
global eco
onomy enteringg recession at the end of 20007.
The recession started pushing the world shippinng
d
so also the demand for
f newbuildinng.
demand down
The demaand continued to fall and reaached a level of
nearly 000 MDWT by tthe end of 20009. A few shipps
ordered here
h
and theree, such as Iraan to Korea are
a
ignored. Quite
Q
a many shipyards speecially in Chinna
fell and so
ome planned shhipyards discontinued.
F
Fig.-2:
New bu
uilding Orders Korea
K
Fig.-4: Supplyy and demand world
shippingg/shipbuildingg
4.2 Capa
acity develop
pment in Ban
ngladesh
F
Fig.-3:
New bu
uilding orders China
C
The demand
d
in thee areas of serrvice & suppoort
craft, suppply vessels, port servicingg craft, vesseels
playing inn national waaters, fishing fleets
f
as well as
smaller shhips and vesssels for short haul and feedder
services has been consistent.
c
Thhe replacemeent
requiremeent of a larrge number of
o small shipps
operating in European waters will crreate a growinng
demand. Enforcement
E
oof use of doublle skin tankers,, a
strong deesire of fuel eefficient ships and restrictionns
including emission lim
mits to hold goood environmeent
and usuall need of replaacement will keep
k
the demannd
growing.
xvii
These vessels are of great importance

i
too
h export shipbbuilding, not only
o
because oof
Bangladesh
their sizes that Bangladeshi yardss may handlle
b
of thhe
reasonablyy satisfactorilyy but also because
higher reqquired human input in thesee categories oof
craft. Hum
man input is our strength
h and many a
countries, these days, do not like to get in thhe
constructioon & supply of these vessels. Europeann
leftover bu
uilders are grradually beingg pushed awayy
from the supply chain of these smalll vessels/craft
ft.
Korea andd even Chinaa are losing interest
i
in thiis
section of market. An oppportunity is thhus on the sighht
f Bangladeshh.
to evolve for
A shipp built in Banngladesh will be about 15%
%
cheaper annd that if a siingle ship is in
i demand andd
Bangladesh
h has the techhnical know-hoow and facilityy
to build the order shall come to Banglaadesh. It shouldd
be an interrnational intereest to develop Bangladesh
B
as a
shipbuildinng nation for a greater beneffit to the worldd
shipping in terms off supply of vessels at a
comparativvely lower costt. May this be one
o of the mosst
crucial issuues in view of the
t recession.
Bangladesh
h may withoutt much difficuulty become thhe
4th largest ship
s
exporter and
a some day, not
n far off, mayy
supersede Korea.
K
Compaared to China, a ship built in Bangladesh
B
will be 12.5% cheaper.
Two shhipyards are buuilding ships for export, onne
of them haas been operaating now for more than twoo
decades annd the other has been develooped only oveer
last two years. Three other yardds are eitheer
ng their existinng facilities andd or offing new
w
modernizin
and plann
ning to start buuilding ships in
i less than tw
wo
years. In next two yearrs i.e. by the end of 2012, at
b building shipps
least 5 shipyards in Banngladesh will be
b
for exporrt and by 20133 at least 10 shhipyards will be
exportingg ships built by them.
5. WEA
AKNESSES
S TO BE OV
VERCOME
E
Bangladesh is a very new entrant in thhe
a
weaknessses
international shipbuildding. There are
me by appropriaate steps to meeet
which muust be overcom
challenges and be ccompetitive. A glimpse of
weaknesses is pictured bbelow:
Scarcity of capittal
H
High
financing cost
B
Backwardness
in technology
In
nadequate electtric supply
W
Weak
diplomacyy
In
nadequate mannagement pool for expandinng
shhipbuilding inddustries
L
Lack
of basic ddesign abilities
L
Longer
lead tim
me in material mobilization
m
L
Lack
of compreehensive skill development in
vaarious shipbuillding trades
N policy bodyy for advising government on
No
o
shhipbuilding
Su
ubsidies/suppoort in variouus shipbuildinng
co
ountries.
Fig..- 5 : Change inn economic baackground

Coost in Million US$
U
T
Table-4:
Compaared building costs
c
S No.
Sl.
1
2
3
4
5
Cosst item
Maaterials, Machinnery &
Equuipment
Lab
bour
Oveerhead
Fin
nancing
Misscellaneous
Tottal
xviii
Chinna
4.255
Banglaadesh
4.550
1.755
0.500
1.000
0.500
8.000
1.000
0.220
1.10
0.220
7.000
development. Bangladesh aims to attain middle

income level and two digit annual growth rates by
2020. This is possible on the one hand by
industrialization and on the other hand by increasing
foreign currency earnings. Both ends may be met
through a faster healthy growth of shipbuilding, which
creates scope for employment and earns huge foreign
currency through export as well as bless the economy
with various spills over benefits.
6. GOVERNMENT POLICIES
Although shipbuilding is a global industry and
international bodies are monitoring various aspects of
shipping and shipbuilding including quality etc., there
is no global umbrella to create an even price platform
for shipbuilders to compete. The governments, such
as, Korea, China, Vietnam, India and other
shipbuilding nations provide 20-30% of total ships
prices as subsidies to shipyards in one form or
another. The shipbuilding industry has its own
distinctive features as compared to other industries. It
is unique in a way that it sells first and construct later
unlike other industries where one manufactures first
and sells later. Further shipyards get orders only if
they are credible i.e., deliver quality ships and on time
and it can be credible only after successfully
executing
consistently
under
international
competition. Further it has to be globally competitive
against the best yards in the world. Unfortunately the
shipyards in Bangladesh are facing exorbitantly high
financing charges as compared to foreign yards.
Unlike other manufacturing industries the product
takes two years to be delivered and requires high cost
finances over a long period. This weakens the
competitiveness of Bangladeshi yards. However in
general, government policy in Bangladesh is to
encourage and attract foreign investments.
Bangladesh has a quite well defined FDI (foreign
Direct Investment) attracting policy. All benefit
available for a national entrepreneur is equally
available to foreign investors. Bangladesh is looking
forward to having investment from abroad side by
side to that of national private sector investments.
Diversifications of exports and expansion of export
basket is an integral part of the long term adopted
industrial and economic policy. Bangladesh
government is keen to support a fast healthy growth
of shipbuilding industries.
Shipbuilding has been declared as a Thrust
Sector. Green channel port clearance system of
imports of raw materials for export shipbuilding has
been accorded. A 5% incentive has been provided.
Bangladesh government is already keenly initiating
various action plans to overcome the aforementioned
weaknesses.
To compete with other Asian shipbuilding giants
in the sector, Govt. may allow Bangladesh ships
exporters a subsidy to the tune of 20%. Subsidy will
help to get a foothold in the international market.
Bangladesh is willing to extend all necessary
supports.
7.1 National shipping cost

Bangladesh today is paying billions of dollars as
freight for her sea borne commerce, less than 6% of
which only is carried by Bangladeshi ships. Both
import and export trades are growing nearly 8%
annually but often under various undue pressures
from foreign shipping companies. India, to its
shipping, has been supporting local shipbuilding as
well as shipping by financing 30% of ships prices by
providing 30% cash incentives to local shipyards for
building any ship longer than 80m. Only a proper
development of shipbuilding sector may bestow the
nation with shipping companies at reasonable costs.
Such flagged ships will carry her commerce both
ways and also create a platform for marine cadets and
seamen training who in turn may be employed in
other flagged ships to earn foreign currencies. Inland
waterway and coastal shipping are playing very vital
roles in the overall national development in
Bangladesh by rendering transport services.
7.2 Training centers

Shipyards are training grounds and impart through
on job practice very high skills turning the society, on
the one hand, technologically advanced and on the
other hand enriching the manpower export of the
country. Shipyards personnel such as electricians,
pipe fitters, shipwrights, welders & cutters are highly
valued in all industrial fields. Shipbuilding is
subsidies in various countries including the developed
ones taking into consideration, among others, this
need.
7.3 Other benefits

For her industrial growth, infrastructure
development, exploration and extraction of mineral
recourses, production & harvest of sea resources and
for a strong navy Bangladesh needs her shipbuilding
industry to grow.
8. PROSPECTS
7. PRIORITIES
Shipbuilding may be defined to encompass the

design & drawings, engineering, the new building,
repair & rehabilitation of ships and various linkage
industries producing any item that may go on board a
ship in the process of building and repair. The
Building ships for export is a priority to the

business community, the government and financing
institutions in Bangladesh. Over ages, shipbuilding
has served as the stepping stone for national
xix
prospects have
h
been guesstimated in money
m
terms foor
a planned period of ten
n years. A heuuristic approachh
pplied.
has been ap
8.3 Contrributions of lin

nkage industrries
50 ind
dustrial units ((Fig.-8) totalinng both forwaard
and back
kward links, have been conservativeely
estimatedd. These industrries will feed the
t new buildinng
and ship repair industtries resulting in an increaase
d earnings in thhe
value adddition, i.e. increeasing retained
country and
a reducing leead time. Thesse industries will
w
earn foreeign currenciees by exportting to foreiggn
shipyards as well and suuch earnings may
m well exceeed
US$ 250 million per yeear. An employyment of 10,0000
is envisag
ged.
8.1 New bu
uilding earnin
ngs
There will
w be 10 shippyards exportinng ships by end
of 2013. Considering
C
th
he present trennd continued, a
growth paattern has been shown in Fig.-6 whichh
although has
h no real dataa support but the
t spirit of thhe
author who
w
has experienced commendable
developmeents and chan
nges in the marine
m
field inn
Bangladesh
h over the last three decades. With 50 yardds
developed by 2021, Bang
gladesh will exxport 500 shipps
on an averrage each of 100,000 dwt earnnings US$ 12.5
billions annnually retainning more thhan US$ 7.000
equivalent per year. A diirect employm
ment of 150,0000
may be attaained.
G
pattern of repair yardss in Bangladeshh
Fig.-7: Growth
* A repaair facility (sslipway/dry-doock) capable of
docking up
u to Panamax with 50 dockiings in a year on
o
an averag
ge.
Fig.-6: Groowth pattern off shipyards in Bangladesh

B
* A unit shipyard is a facility witth employmennt
strength of
o 3000 and capable
c
of buuilding 10 noss.
10,000 dw
wt MPC of an average
a
value, over projected
time, of US
S$ 25,000 milllion per piece.
8.2 Repairr earnings
In ship repairs, materrial input accouunts for usuallyy
25%-30% of value of a repair, rem
maining is foor
services i.ee. financing coosts, facility coost, utilities andd
manpower. Thus Banglaadesh has specific interests inn
developingg ship repairr facilities. However
H
suchh
facilities need
n
comparatiively higher in
nvestments andd
high leveel skilled maanpower. Rep
pair industriees
generally follow a reasonably grownn new buildingg
sector.
Havingg 20 docks/yarrds (Fig.-7) eaach serving 500
ships annu
ually and billinng US$ 5.00 million
m
for eachh
ship serveed, Bangladessh may earn 5 Billion US
S
dollars rettaining about US$ 3.50 biillion. A direcct
employmennt opportunityy of 80,000 workforces iis
foreseen.
G
patterrn of linkagee industries in
Fig.-8: Growth
Bangladesh
* A linkaage industrial unit with 2000 employs will
w
produce US$ 10.00 million a year,
y
all beinng
considered as export thhrough ship buuilding/ repair or
directly accross the bordeer.
8.4 Beneffits of developed national sh

hipping lines
One of
o the major spills
s
over bennefits of a weell
grown shhipbuilding seector is the evolvement of
xx
national shipping lines. Assuming an addition of 10

ships a year to national shipping lines, a fleet of 200
ships i.e. 2 million tons carrying capacity is seen
developed by 2021. International sea borne commerce
of Bangladesh may well exceed double of the volume
of today i.e. reach to 100 million tons. 25% of these
commerce will be borne by own fleet by 2021 earning
about US$ 2.00 billion a year. For a healthy growth of
exports and supplies of raw materials to local
industries, national fleet carrying the commerce is a
must. Apart from the retained earnings, employment
benefits and created training scope, the role which
will be played by evolved nation shipping lines in the
national economy is invaluable.
in 10 years and the number one in the field in 30

years, I foresee the required investments from home
and abroad in shipbuilding and repair industries as
well as in forward and backward linkage industries in
Bangladesh.
Bangladesh has all the ingredients for labour
intensive shipbuilding and allied industries. With the
encouragement
of
Government,
dynamic
entrepreneurship,
infrastructural
advantages
shipbuilding may take the lead and stand out as the
flagship sector of our economy. Soon, Bangladesh
will enjoy the ripple effect of shipbuilding in the form
of expansion in heavy industries. A change in lives of
our people, through industrial employment
generation, is expected and inevitable.
9. INFERENCE
Bangladesh may lookout for the market segment
of small & medium sea going vessels for new
building and repair and vessels operating in national
waters in many countries. This market is more than
US$ 100.00 Billion and a US$ 10.00 Billion portion
of this market is not difficult to be secured by
Bangladesh.
Cherishing the firm goal of Bangladesh to be the
4th largest ship exporters and repair services renderer
xxi

VAN GOGH, GAUGUIN, FLUIDITY AND THE UNDERGRADUATE

ENGINEERING EDUCATION
Professor Dr. Sander Calisal
Faculty of Engineering
Piri Reis University
Istanbul, Turkeys
Email: calisal@pirireis.edu.tr.
ABSTRACT
There seems to be a significant difference or gap between the preparedness of graduating engineering students
and the value of their skills to North American industry. This fact started to surface in about mid eighties in US
as a result of a survey done at that time and also later by the National Society of Professional Engineers in the
United States. This paper presents some thoughts as to why such a large gap has appeared during the past
decade or so, and what needs to be done to reduce it. A digital curriculum concept will be discussed, as well as
student-based teaching and learning. The requirements for change appear to be challenging, as they may
include the need for additional space and teaching equipment, thus increasing project and laboratory costs.
Recent applications of the various attempts in various campuses in USA and at a Canadian campus are
explained.
1. INTRODUCTION
The engineering professions, and also engineering
education, have been changing rapidly. At present,
few theories exist on how the changes to the
profession should be handled in terms of modifying
the undergraduate curriculum. The objective of this
paper and of this presentation is to share with you
some of the information I have gathered and questions
I have considered during my recent work at UBC and
UBCO.
Changes in engineering education are not new;
they have occurred before in North America. The
name of a small Soviet satellite, Sputnik, is usually
associated with the beginning of these changes.
Sputnik is credited with bringing about changes to the
engineering curriculum in the late 1950s and early
1960s. A large number of courses in mathematics,
science and advanced engineering were added to the
undergraduate engineering curricula of that time.
Aerospace departments became very popular as well.
Many of the professors now teaching engineering in
North America were educated with and inherited that
curriculum.
However, if we stop to think about it, other
machines have also changed the world in the past,
and with it, engineering. The internal combustion
engine, the camera, and the computer have brought

about extensive changes in our lives and also in the
field of engineering. Each of these machines has
changed our expectations and also the quality of life
on earth.
We will briefly look at several short examples
concerning these machines, followed by some ideas
on education, some of which were influenced by a
display on fluidity at the University of Michigan in
2002.
2. THE INTERNAL COMBUSTION

ENGINE
The internal combustion engine has changed the
nature of farming, transportation, cities and suburbs,
and has brought about new industries such as tourism,
along with the drive-in concept. It is reported that
about forty per cent of the U.S. economy now
depends on the automobile. If we examine the effect
of the internal combustion engine on only farming and
the farmer, we can say that as tractors came into use in
North America, horse-based agriculture was replaced
by tractor-based agriculture. Farmers now require
less knowledge about horses but more about
machines, fuel, selection of materials for their repairs,
lubrication, and so forth. Motorized agriculture has
not only changed the nature of farming, but also the
xxiii
farmer in terms of the skills and education he requires.

He can now irrigate his fields using a pump powered
by a tractor, and repair his equipment in an
emergency. Forward-thinking farmers have adopted
the new machines; others have resisted change.
However, the result of motorized farming may be seen
both in the fact of larger farms today as well as in the
increased efficiency of their production.
effective use of computers in education, of which the

Computer Assisted Engineering Network (CAEN) is
an example. Around the same time, the mechanical
engineering department contacted the employers of its
graduates, after which it reviewed its curriculum and
revised it in a measurable way to include new
technology. Later, the department published the
results of its findings and the subsequent changes that
it made.
3. THE CAMERA
To have a better understanding of the changes brought

about by the computer, we need to know how
computers affect human activities. I think that what
my friend David Malaher once said in one of our
discussions about computers might be a good
representation. He said that computers are like
microscopes; they give us new dimensions.
Still cameras and movie cameras have changed the

concept of painting and also the style of presentation
of course material in the classroom. We have learned
to freeze time and to explain and lecture in a visual
way. As for painting, the invention of the camera has
changed it completely. Gauguin wrote to Van Gogh in
1886, in loose translation, The camera has been
invented and will be in color soon. We do not have to
copy nature; we should paint from our imagination,
our dreams, our perception. These men had the
vision and the energy to change their style of painting,
and their paintings now hold very respectable places
in world museums. The camera changed the concept
of painting, the artist, and art as we now know it.
Painters such as Van Gogh, Gauguin and Matisse did
not imitate their old masters. Instead, they developed
a new wave in painting, which required a tremendous
amount of courage and determination.
4. THE COMPUTER
Computers have also changed our lives greatly. The
way in which I am able to type this paper is an
excellent example of how both the way we type and I
have changed during the past ten years. Computers
have resulted in new hardware and software industries
and products that depend on computers, such as the
Internet. We depend extensively on the Internet for
communication, and the access to and collection of
data. I believe that engineering, and undergraduate
engineering education in particular, have also changed
drastically as a result of the computer.
My concern now is mainly of the impact of the
computer on engineering practice and on
undergraduate engineering education. We need to
quantify this change in the engineering profession, so
that the impact of the computer on it can be measured
and evaluated. But before we consider that, I would
like to share with you a proverb I learned some time
ago, which says that to progress we need to change,
to change we have to first change our minds.
During my study leave at the University of Michigan I
observed that the mechanical engineering department
in that university had done just that. The university
had built a new North Campus and had included in
every building the infrastructure necessary for the
We can try to determine what those additional

dimensions are. We know that engineers of my
generation specialized in one area, working as a
bridge engineer or marine engineer, and that they
remained in one area for a long time. Design
methodologies and codes did not change a lot. Not
any more. Engineers are now expected to be more
flexible, to be team players, and to be more creative
and synthesis oriented. They are also expected to
have more skills.
Knowing specific design
procedures and the codes is no longer sufficient, as
both can easily be programmed or computerized. It is
not enough to know how to select a pump and do the
pipe head loss calculations as an engineer, as this type
of selection and calculations are relatively easily
available in a small computer package.
The equations used in engineering and engineering
methodology have, however, not changed much over
time.
We still use force equals mass times
acceleration, or Newtonian mechanics. Even the large
systems of equations which we now use easily, were
known earlier. Companies such as Messerschmidt,
the German airplane manufacturer, was said to have
had, in the 1930s, six women whose job it was to
invert a 6 by 6 matrix for the design of airplane wing
profiles, a task which required a day to complete.
Similar positions for matrix inverters appear to still
exist at some European universities; the current
version of the job, however, seems to be to input and
handle matrices properly in computers. We can thus
observe that engineering tools have changed greatly
over the past two decades. Some educators claim that
engineering science is dead, but this remains to be
seen. New engineering products still need to be
developed, such as renewable energy machines, and
for these we need engineering science for modeling,
research development and production.
Digital drawing equipment, including rendering for
three-dimensional
presentation,
and
quick
prototyping are new engineering tools.
Hand
xxiv
drafting is not only rare today; it is also not

considered professional by todays standards. We
are now able to do numerical rather than analytical
designs, and to collect digital rather than analog
signals.
Changes in engineering education are also visible.
Books now contain at least a CD with problemsolving software, numerical programs for engineering
applications, and various animations, and they refer
the reader to a web site for additional information as
well. Electronic versions of books are also available.
While students still prefer paper books, they are able
to follow web-based lectures with color graphics and
animations using their laptops in class. Professors are
able to communicate with students efficiently using email and course web sites.
The results obtained by the National Society of

Professional Engineers in the United States, as shown
in figure 1, add a measurable result to these
observations. This chart might be a good starting
point for thinking about the changes which have
occurred in engineering and the corresponding
modifications that will have to be made in the
undergraduate engineering curriculum. Except in the
areas of mathematics and science, there is a definite
shortfall in the readiness of new engineers.
Teamwork, leadership, and integrative thinking are
not the major teaching targets in current North
American undergraduate education.
6. ASSUMPTIONS AND POSSIBLE

SOLUTIONS
Classrooms have also changed; they now have

computers and digital projectors, and chalk-based
blackboard presentations are rare. Students seem to
be more motivated seeing animations of calculations,
such as the output from CFD software, visual
examples that one can create using a special program.
As a result, professors are now able to achieve a
higher level of communication using wireless
networks available on campus than was previously
possible, and to try student-based teaching using case
studies and project-based curricula.
While we take books for granted information in
general and books in particular were not always easily
available. Religious works were the first books to be
printed and, prior to that, most books were handcopied by scribes. In the library of the old Roman city
Pergamum (Bergama in Turkey), it is reported that
about 5000 scribes copied books for various libraries
around the world during the time of the Roman
Empire.
In the past, in order to acquire new knowledge,
students had to travel long distances. For example,
Pythagoras traveled to Babylon (Baghdad) to learn
how to calculate the square root of integers, and to
learn his theorem. Most of us learned this theorem
from books; we did not have to go to Baghdad. And
today of course, it is possible to obtain even more
extensive information using the Internet. With the
availability of the Internet and e-mail, we can quickly
access any information we need. A relatively new use
of the Internet for delivering information that of
distance education is already beginning to change the
concept of the university. Is it possible that e-books
and e-colleges will one day be the norm?
5. THE GAP IN ENGINEERING

EDUCATION
Most of the above observations can be seen as
tangible results of the development of the computer.
National Society of Professional

Engineers
Respondents were asked to rate new engineers preparedness in
eight areas and then indicate the value their organization places on
preparation in that area. [NSPE]
Figure 1 Value vs Preparedness for new engineering

graduates
My assumption is that this measurable gap between
value and preparedness, as seen in figure 1, is the
result of the changing expectations and practices in
the engineering workplace, brought about primarily
by the use of the computer. In addition, teamwork,
leadership and integrative thinking are new
dimensions to engineering, and appear to be important
to industry. The gap between value and preparedness
is similar to the gap experienced by farmers before
and after the advent of the internal combustion
engine, and by painters after the invention of the
camera. In the field of engineering, computers have
taken the load of analysis and hand calculation off the
engineers, but at the same time have brought new
requirements to the profession. Farvardin N gives the
following figure in Ulsoy (2007) as a comparison of
the expectations from a modern versus classical
engineer. This he does without referring to the root of
the changes. One can speculate that engineering skills
brought by computers such as CAD, FEM took space
from the curriculum that was available for soft
subjects such as English, psychology etc. We can
conjecture that computers allows more detailed design
resulting with optimization and requiring team work
possibly international in definition.
xxv
Figure 2 Traditional versus Modern Engineer from

Ulsoy 2007
One possible solution for the removal of the gap could
be obtained by identifying the role of the computer in
the engineering work environment and by building
new curricula around this observation. The likely
result of such an exercise will be the realization that
undergraduate engineering education in the future will
have to take place in an environment similar to that of
a practicing engineer. That is to say, the type of
computer support which the practicing engineer has in
his workplace must also be available to the
undergraduate engineering student. This has already
been accomplished to some degree, as mentioned
earlier, at various campuses and possibly first at the
University of Michigan.
There seem to exist mainly two basic educational
methods. One is the student-centered approach,
where the student is evaluated on what he has learned
and not on what he does not know; and the teachercentered approach, where the student is evaluated
with quizzes and final examinations to establish the
limits of his knowledge, or what he does not know.
Some educators find that teacher-centered instruction
assigns a passive role to students. These educators
would much rather see that a students evaluation
determine whether a person can think in a disciplined
way. They also add that a students abilities cannot
be measured by the type of short-answer questions
commonly used in quizzes and examinations.
While most educators may agree that student-centered
teaching is more suitable for teaching creative work,
including engineering, the main difficulty seems to be
administrative. This seems to be related to the fact
that the evaluation of the students knowledge is still,
in general, based on standardized tests such as the
Graduate Record Examination (GRE). Standardized
examinations are still the basis for acceptance of
engineering students into graduate studies and into
professional engineering status. This seems to form
the basis of accountability of the educational
procedure as far as the administration is concerned.
Other educators claim that teacher-based instruction

equalizes the educational process by providing
exposure to the same material to all students, and to a
more organized education through an agreed-upon
curriculum. In addition, teacher-centered instruction
seems to be preferred by administrative organizations,
as already mentioned, for the additional reasons that it
is cheaper and more accountable, and requires
fewer infrastructures such as classroom and laboratory
space, educational and laboratory materials, and
teacher time for projects. Student-centered education
may also be at odds with the current emphasis on
accountability, which tends to focus on the
educators role and responsibility for students
learning, rather than on the role of the student. The
main point of discussion for engineering education is,
of course, what the engineering curriculum should
contain in order to enable students to learn leadership,
creative thinking and synthesis.
There are obviously other changes that need to be
made in order to simulate an engineers workplace on
campus. These include providing an environment
conducive to teamwork, integrative thinking and
leadership training (or TIL in short). The objectives
to be met are similar to those found in the teaching of
team sports, and require continuous preparation and
coaching, plus competitions.
Some of the results I observed at the University of
Michigan were most impressive. With the availability
of CFD code in CAEN, referred to above, a
University of Michigan fourth year student was able
to calculate the three-dimensional flow around the
keel of a sailboat. This was for a term paper worth
twenty per cent of the course mark in an
undergraduate course on sailboat design. At UBC
similar attempts have been made to increase the
exposure of students to various new engineering
design tools in various courses. In fluid mechanics
we have begun to include some of the concepts of
CFD in undergraduate courses, and recently, CFD
laboratory content was included in the third year fluid
mechanics course, using FLOWLAB software.
However, we still insist on teaching many concepts
requiring extensive analysis, such as boundary layer
flow, in a lecture-based format. One can question
whether there are more effective ways to introduce
such concepts, such as with the help of computers
physical experiments. Do we really need extensive
analysis such as perturbation methods at this level in
order to introduce mainly concepts? To answer this
question, we need to know if the engineer of the
future will need to use an extensive amount of
analysis or will be using a computer-based approach
such as a CFD code for design. We can ask if he will
be using software provided by salespeople to select
fans, blowers, pumps, etc., for his design.
We need to estimate what type of concepts an
xxvi
engineer will need and what type of educational tools

we must use to teach such concepts in the most
educationally efficient way. Among other things, we
need to know if we require extensive, in-class analysis
in order to teach basic concepts or if we can use new
digital tools. A very good and effective approach now
with the computers is the usage of simulation for the
instruction engine operations or navigational
problems. The student is expected to think and solve
problems associated with ship operation but the
problem is visible and realistic. The procedure also
offers a student based learning procedure and
valuation.
effectively for the education of future engineers. In

addition, we may not have sufficient space and the
necessary infrastructure for teaching a digital
curriculum.
In my opinion, a student-centered approach is more

appropriate, especially for TIL, but an integrated
approach with proper checks and balances is also
required for professional studies. A recipe for change
does not seem to exist, and there may be a number of
solutions for a new digital curriculum and the
associated environment for the teaching of TIL. In
short, we need to answer the following questions:
Computers also offer the opportunity for studentbased teaching, thus enhancing teamwork.
Engineering case study packages, such as those
developed for NSF under the LITEE program, offer
the opportunity to introduce students to TIL. I have
observed that students do an excellent job of learning
when computer-based tools are freely available to
them as part of the curriculum.
How can we respond to the existing value versus

preparedness-gap in engineering education?
How can we effectively use digital (computer-based)
general engineering and design tools in education?
7. THE CHANGE REPORTED SO FAR
My suggestions may be summarized as follows. First,

we need to bring student-based teaching and modern
engineering practice into the classroom and, second,
we need to modify engineering education to stress
teamwork, design, creativity and synthesis (rather
than analysis at this level), possibly leaving analysis
to computers. This will require that we bring
numerical analysis tools into undergraduate education.
This will not be an easy task and we will need
fluidity or change.
In order to change our
educational methods, we must first change our minds.
This refers to the fact that change is made by
individuals first and then by the institutions and is a
highly personal and long experience. This required
change of mind may possibly be the most difficult
task for educators. We must also decide not to
compete with the computer, and we should stop
imitating our own old masters, that is, our professors,
much as Van Gogh, Gauguin and Matisse stopped
painting like their masters. Instead, we should give
new digital tools to our students so that they can
design the machines they can imagine or dream about.
We need to accept the fact that part of the reason for
the existence of this gap is that industry is way ahead
of universities and colleges in the development and
use of computer-aided engineering and design. New
engineering tools have not been developed at
universities; they are not products of universities; and
we really do not know how to use them properly in
universities. An additional problem is that even if we
have these products, we do not know how to use them
All of the above will require that we increase the

numerical and digital components of our courses. I
recommend integrating computers into our
engineering courses and curriculum. In this way the
classical, analysis-based component of undergraduate
courses could be reduced significantly, and
engineering concepts could then be explained through
numerical results rather than by extensive analysis.
We all observed that most universities are connected

to internet in a wired or wireless way. They extended
the computer capacities but not necessarily for
education and without an integrated approach. In USA
national Science Foundation spent substantial amount
of money to restructure engineering education. New
concepts in teaching were encouraged to increase the
quality of engineering education.
Michael Bernitsas 2002 reported that in the Naval
Architecture curriculum at the University of Michigan
the following subjects and educational procedures
were added to the curriculum.
The new curriculum (1994) has courses in
manufacturing, life cycle cost, industrial design
course (second year) before 3rd year core courses.
Team
work
communication
skills
ethics,
environmental awareness, included in (1997-2000)
Simulation based environments to test virtual
prototypes (Computer aided design)
These all seem to be in the direction to close the gap
discussed above.
An MIT initiative that could revolutionize learning
started in September 2003.
The Massachusetts Institute of Technology (MIT),
which decided in 2001 to put all its courses on the
Internet, in September 2003 moved its Open Course
Ware (OCW). Electronically downloadable books
and course ware is now available to all students
around the world. This initiative may not reduce the
gap but is surely digitising the undergraduate
xxvii
education and opening new frontiers opportunities to

educators.
Latorre (1997) reported the modifications made to the
Naval Architecture program at the University of New
Orleans. He lists similar short comings and lists
modifications made to the curriculum to remove the
deficiencies. He also provides a project based future
view of the naval architecture program at UNO.
At the University of British Columbia BC Canada
(UBCO) a new engineering school started in 2005.
With that a completely revised curriculum was
designed in 2004. Some of the fist year first term
courses are listed below.
Statics APSC 180 (3)
3-0-2
Engineering Analysis I APSC (3)
3-0-1
Engineering computation and
Instrumentation APSC (3)
2-2-0
Social Electives (3)
3-0-0
Engineering Fundamentals (3)
2-0-2 *
APSC170
Matter and Energy (3) 3-0-1 APSC
4-2-2
(18) Credits 17-4-5 contacts
Two of the features for this set of courses is rather
unique in engineering education and its application
was very successful. One of them is the inclusion of a
full course on Statics during this term. The second
more radical is the development of the Engineering
fundamentals course that included emphasis on team
development, design, project based learning,
budgeting etc. The student groups designed, built and
competed with their designs. The judges for the
competition were the invited professional engineers of
the community not the instructors. The students
learned the benefits and the difficulties of team work
and designed, planned, built a working device to
perform a well defined task. This was a difficult
course to teach but very valuable to students as the
skills learned during this course were used in most of
the later project based courses Labun (2009). During
the development of this new engineering program it
was agreed upon that all courses and labs would be
designed assuming that the students have a laptop
computer.
In addition a second year design course used a project
based on an air cushion vehicle. Students tested the
pressure field under the cushion, optimized the weight
of the structure, measured the electrical energy
consumed by programming a board. A weighted
formula was used to evaluate the transportation
performance based on the weight carried along a
basketball court and the energy consumes. We
believed that these two courses helped the student
develop team work, communication design,
integration skills.
Figure 2 One of the first year design projects for

APSC 170 at UBCO
Figure 3 UBCO second year Air cushion design

competition
equally
stressing
team
work,
communication.
Similar course developments at the mechanical
engineering department at UBC resulted with an
award winning program called Mech 2 for the second
year Ostaficuk et al (2008). This is a worth studying
program as it offers a very creative a project based
environment to the students. Various very effective
and valuable course designs are underway. They still
remain at individual course or project level and a
more comprehensive look at the general engineering
undergraduate education is required and forthcoming.
We see rather interesting educational changes and
developments in Physics and Medical schools now.
Bronsart R. Clauss G (2006) reported a very
interesting inter university usage of computer aided
education in engineering.
8. CONCLUSION
I believe that in order to progress we need to change,
and in order to change, we must first change our
minds. We need to integrate the use of computers into
our courses for teaching engineering and TIL. For
this we need a digital curriculum that includes the
xxviii
computer as a teaching and an engineering design

tool. When we have reviewed all the courses in our
curriculum and made the required changes to them,
the result will be a modern, successful engineering
program that is more relevant to the requirements of
industry, and the preparedness gap between
engineering education and industry will have been
removed. As change is a continuous process we
should be prepared for the long term and continuous
nature of the energy required for such a change.
[6] Latorre R. Student Development in Naval

Architecture, Marine and Ocean Engineering
Education at UNO Int. J. Ed. 1997 Vol. 13,No
5,p.363-368
REFERENCES
[8] SNAME, Ship Production Committee Curricular

needs of Shipyard Programs June 1984
University of Michigan DTMA 91-82-C20022
[1] Bernitsas M 2002 Engineering for Marine

Environment at an American Research
University
Int. Journal of Engineering
Education Vol. 18 pp 22-31
[2] Bronsart R. Clauss G. mar-ing - the Network of
German Universities for a joint NAOE Master
Program OMAE 2006 Berlin.
[3] Committee on National needs in Maritime
Technology Ship Building Technology and
Education National Academy Press Washington
DC 1996
[4] Glaser R. Education and Thinking ; The Role of
Knowledge learning Research and Development
center University of Pittsburg June 1983 technical
report No PDS-6
[5] Labun A. personal Communication, University of
British Columbia Okanagan School of
Engineering 2006
[7] Ostafichuk P, Croft E,Green S,Schajer G,Rogak

SAnalysis of Mech 2: An Award Winning
Second
Year
Mechanical
Engineering
Curriculum Innovation, Good Practice and
Resaerch in Engineering Education EE, 2008
[9] Ulsoy A.G. Edited by The 5XMEWorkshop:

transforming mechanical engineering Education
and Research in the USA May 10-11 2007
National Science Foundation Arlington VA.
[10] G. Tryggvason, M. Thouless, D. Dutta, S. L.

Ceccio, and D. M. Tilbury, "The New Mechanical
Engineering Curriculum at the University of
Michigan," Journal of Engineering Education,
90(3), pp. 437-444, July 2001.
[11] D. M. Tilbury, S. L. Ceccio, and G. Tryggvason,
"Restructuring the Undergraduate Curriculum of
the Mechanical Engineering and Applied
Mechanics Department at the University of
Michigan," Proceedings of the ASEE Conference,
Milwaukee, Wisconsin, session 2266, June 1997.
xxix

ON THE PREDICTION OF SHIP PERFORMANCE IN ACTUAL SEA CONDITIONS

USING COMPUTATIONAL FLUID DYNAMICS
Hiromichi Akimoto1
1
Department of Systems Innovation, Graduate School of Engineering, The University of Tokyo, Japan
E-mail: akimoto@sys.t.u-tokyo.ac.jp
ABSTRACT
Towing tank test is a common tool for the research and development of ships. However, experimental
conditions of a towing tank are usually limited to head wave conditions and symmetry motion of the ship.
Although there is a strong demand for higher economical efficiency of commercial vessels, we do not have
effective tool for predicting the performance of planned vessel in the actual operating conditions. In this
research, the author use computational fluid dynamics (CFD) simulation to reproduce the unrestricted motion of
the ship in oblique waves to investigate the influence of the freedom of motion to the total resistance of ship. The
results indicate that additional degrees of freedom in motion (roll and sway) and the drag from rudder action
might have significant influence on the performance of ship in oblique wave conditions.
Key words: Computational Fluid Dynamics, CFD, actual sea condition, sea margin, oblique wave, 6DOF
1. INTRODUCTION
Although ship is a transportation mode of low
energy consumption, reduction of green house gasses
emission and fuel consumption of sea transportation is
an important issue because of its vast transportation
volume. However, prediction of the performance of a
planned ship in actual sea environment is still a
difficult task.
The main tool for evaluating the resistance of a ship
before its construction is towing tanks. We can
generate waves to imitate the sea environment and
obtain the motion and resistance of the vessel.
However, their experimental conditions are limited to
head wave conditions and symmetric ship motions,
i.e., heave, pitch and surge because of the mechanical
restriction of towing tests. Since we do not select head
sea course in actual operations, the experimental data
of towing tanks might provide inaccurate information
for the improvement of hull design. Although
maneuvering tanks provide more freedoms in ship
motion, it is still difficult to obtain reliable
performance data because of the mechanical
restriction and the limited size of model ships.
When a ship takes oblique course to incident wave,
it experiences asymmetric wave load and shows
asymmetric motions in roll, sway and yaw direction.
In addition, the ship has non-zero leeway and rudder
angles to cancel the asymmetric force and moment to
keep its course on oblique waves. However, we are
using empirical sea margin to predict the increase of
drag in actual operation only because we do not have
alternative tools for the purpose. Tsujimoto et al. [1]

and Hulskotte et al. [2] showed the increase of fuel
consumption in actual operations and the procedure of
its prediction. Tsujimoto et al. proposed ten-mode for
marine transportation for the fuel economy indicators
of ships in actual operating condition [3]. It is similar
to the fuel economy test cycle of automobiles. It
means there is increasing demand of predicting
practical fuel economy of ships.
With the recent progress of computational fluid
dynamics (CFD) technique, we can perform Reynolds
averaged Navier-Stokes (RaNS) simulation of a ship
in waves. However, applications of RaNS simulation
are usually imitations of towing tank experiments.
Since CFD simulation does not have mechanical
restrictions in obtaining ship motion, it has good
potential for predicting performance of ships with
unrestricted motion on waves. This paper describes an
attempt of numerically evaluating sea-margin of ships
using CFD simulations.
2. COMPUTATIONAL METHOD
2.1 RaNS solver
We use Reynolds averaged Navier-Stokes (RaNS)
simulation code WISDAM-X that is developed by
Orihara et al. [4] in The University of Tokyo. The
code is designed for solving time-dependent
interaction between a ship and free surface flow with
finite volume method (FVM). Summary of the
characteristics of code is as follows.
xxxi
Governing equations
Reynolds-averaged Navier-Stokes
(RaNS) equation
Continuous equation
Pressure solution algorithm
Pressure solution of Marker-and-Cell
(MAC) method.
Spatial discretization
3rd order upstream (advection term)
2nd order central (other terms)
Free surface treatment
Marker Density Function method
Time discretization
1st order Euler explicit
Adams-Bashforth (2nd order for marker
density)
Arrangement of variables in the control
volume
Staggered arrangement of velocity
components
Turbulence model
Baldwin-Lomax 0-eq model
Dynamic subgrid scale (DSGS) model
Figure 2: Free surface and the exterior surface of the

inner domain.
Computation in the inner domain is on the moving

coordinate fixed on the ship in motion. The motion of
ship with six-degrees of freedom (6DOF) in motion is
solved using the hydrodynamic forces and moments
provided by the flow simulation. On the exterior
surface of inner mesh, trilinear interpolation of outer
flow variables gives boundary condition of inner
domain, vice versa.
2.3 Model of rudder
We employ Baldwin-Lomax model and Dynamics

SGS (DSGS) model to evaluate the turbulent viscosity
in the flow field. Local turbulent viscosity is
determined by the weighted average of these two
models. Free surface is modeled by the density
function method.
2.2 Overset grid technique

To provide adequate spatial resolution both on free
surface and in the boundary layer around the hull, we
use structured overset grids as shown in Fig. 1. The
outer domain has fine spatial resolution around the
water surface to simulate incident waves. The mesh of
the inner domain is boundary fitted one to provide
fine resolution in the boundary layer. Fig. 2 shows the
exterior boundary and hull surface of inner mesh on
the free surface.
In oblique wave conditions, a ship requires

adjustment of rudder angles so that it has appropriate
leeway angle and cancel the asymmetric load from
waves (Fig. 3). However, simulating the rudder
control in CFD is time consuming. Therefore, in the
present analysis, we perform a series of CFD
simulations without rudder by changing leeway angle,
and determine the balanced condition of the ship
using the interpolation of simulation results. The
action of rudder is modeled as a finite span
NACA0018 wing with aspect ratio 1.66. The force on
the rudder is determined by experimental data of finite
span wing [5] with correction of the aspect ratio. The
present analysis does not include interference between
the rudder and flow field around the hull.
Wave direction
Y
Z
Inner domain
Free surface
Ship Body
Rudd
Oblique waves
( beam wind )
V'
V
Figure 3: Condition of a ship in oblique wave

condition.
Outer domain
Figure 1: Overset mesh layout (in symmetry

condition)
xxxii
3. RESULTS AND DISCUSSION

3.1 Four DOF motion on oblique waves
We conduct simulations of SR221 (VLCC,
Lpp=320m) and SR108 (container carrier, Lpp=175m)
on oblique waves. Conditions of simulation are shown
in Table. 1. The direction of incident wave is 120
degree (waves come from starboard ahead), wave
amplitude is 1% Lpp and wave length 0.5Lpp. It is
equivalent to Beaufort wind force scale 9 for SR221
and 7 for SR108. Fig. 5 shows a visualization sample
of SR108 with roll and sway motion.
Figure 5: Time histories of roll angle in different
leeway angles.
Table 1: Condition of simulations

Num. grids in inner and
Inner: 130x21x139
outer domains
Outer: 146x61x41
Minimum grid spacing
0.001
Reynolds number
1.0x106
0.151 (SR221)
Froude number
0.275(SR108)
Wave direction
120 deg.
Wave length
0.5 Lpp
Wave amplitude
1% Lpp
Figure 6: Time histories of roll angle in different

leeway angles.
In Fig. 6, lateral force does not balance and the ship
starts drifts. To obtain the appropriate leeway and
angles for keeping the course of ship, we solve the
following system of equations with simplification:
Fyhull ( ) Fyrudder ( ) 0
hull
rudder
M z ( ) M z ( ) 0
Figure 4: Overview of the computational domain
and front view of the ship: SR108 ship with 4DOF
in motion.
Figures 5 and 6 show the time histories of roll angle
and sway velocity of SR221 with different leeway
angles ranging from -3 to 0 degree. In the time
histories, incident wave starts at non-dimensional time
T=8. The ship shows oscillating motions of roll and
sways in addition to heave and pitches ones.
(1)
Here, and are leeway (yaw) and rudder

angles of the ship respectively. Fyhull ( ) and
M zhull ( ) are lateral force and yaw moment acting on

the hull obtained from CFD. The rudder model
determines the force Fyrudder ( ) and yaw moment
M zrudder ( ) . The solution of (1) gives the
approximated set of leeway and rudder angles in the
given wave condition.
Figures 7 and 8 show the increase of total drag in
different DOFs and leeway angles. In these figures we
are using abbreviated names of DOF conditions. For
example, HP means that heave and pitch motions of
the ship are free in the simulation, and HPRS means
xxxiii
heave, pitch, roll and sway motions are free. Hull and
rudder components of drag are separately shown as
the relative increase from the basis of HP condition.
Although the simulation is capable of solving all 6
DOF motion of the ship, we restricted sway and surge
motions for simplicity.
Fig. 5 (SR221) shows the slight increase of hull
drag with the increased freedoms in motion. In HPRS
condition, increase of both hull and rudder drag
components are significant. The increase of hull drag
in HPRS condition is 5% from the base resistance of
HP while the increase of rudder drag is 8%. In total,
added motions of roll and sway lead to 13% increase
of total drag. In the case of SR108, tendency of drag
is less significant. However, it shows complicated
dependency to leeway angle. With the increase of
leeway, hull drag component decreases while drag
from the rudder increases. The present results show
complicated behavior of total resistance in different
mode of motion and the large influence of rudder drag
in oblique wave conditions. It also implies that
reducing lateral force and the change of yaw moment
in oblique wave condition lead to improvement of
economic performance when modification of the hull
shape is possible.
4. CONCLUSION
To investigate the increase of resistance in actual
operating condition, we conducted CFD simulations
of a ship with different degrees of freedom in motion
on oblique waves. The results show that the increase
of resistance depends on the asymmetric motion of the
ship and rudder action. It also means that the
improvement of performance in usual towing tank
tests does not assure higher performance of the ship in
actual operation.
Present simulation does not contain yaw and surge
motions those might have higher influence on the total
resistance. For further investigation, we are going to
include time-dependent steering model in the present
simulation and extend the wave conditions to two
directional and irregular waves.
REFERENCES
[1] M. Tsujimoto, M. Kuroda, K. Shibata, N.

Sogihara and K. Takagi. On a calculation of
decrease of ship speed in actual seas.
Proceedings of the Japan Society of Naval
Architects and Ocean Engineers, vol. 7, pp. 77
80, 2008. (in Japanese)
[2] J. Hulskotte and H. D. van der Gon. Fuel
consumption and associated emissions from
seagoing ships at berth derived from an on-board
survey. Atmospheric Environment, 44:1229
1236, 2010.
[3] M. Tsujimoto, N. Sasaki, T. Fujiwara, M. Ueno,
N. Usui, M. Kado, D. Nomura and K. Takagi,
Calculation Method of 10 mode Index for
Ships, Journal of the Japan Society of Naval
Architects and Ocean Engineers (10), 97-104,
2009-12-01 (in Japanese)
[4] H. Orihara and H. Miyata. Evaluation of added
resistance in regular incident waves by
computational fluid dynamics motion simulation
using an overlapping grid system. Marine
Figure 7: Increase of total drag on oblique waves in
Science and Technology, 8:4760, 2003.
different DOF and leeway angles: SR221 VLCC
[5] H. Orihara and H. Miyata. Numerical
model.
simulation method for flows about a semiplaning boat with a transom stern. Journal of
Ship Research, 44:170-185, 2003
[6] H. J. Goett and W. Bullivant. Tests of N.A.C.A
0009, 0012, and 0018 airfoil in the full scale
tunnel. Technical Report 647, National
Advisory Committee for Aeronautics, USA,
1939.
Figure 8: Increase of total drag on oblique waves in

different DOF and leeway angles: SR108 container
carrier model.
xxxiv

HISTORY, SATE-OF-THE-ARTS AND FUTURE TREND OF RESEARCH ON SHIP

MANOEUVRABILITY AND CONTROLLABILITY
Professor Kazuhiko Hasegawa
Department of Naval Architecture and Ocean Engineering
Osaka University
Japan
ABSTRACT
From the author`s over 30 years experience of engaging in research on ship manoeuverability and
controllability, he will review brief history, state-of-the-arts and future trend of the research on ship
manoeuvrability and controllability Manoeuvrability has less history than other area such as resistance and
propulsion. Due to some historical and mainly economical reasons, the research and development on this field
was influ-enced as well as other fields. Nowadays, the environmental influence is imperative factor to be
considered. Moreover, manoeuvrability itself is special field different from other fields such as resistance and
propulsion, seakeeping or structural problems, be-cause we cannot discuss it without taking account of the
performance of human operator or controller. Ship handling simulator is one of the important tools for this
purpose. For mariners training and education, ship handling simulator is now widely used in the world.
However, the key issue of the ship handling simulator is not on the computer graphic technique nor bridge
equipment installation. The most important and difficult technology is to provide suitable mathematical model
for each subject ship. Safety issue in congested waterways is another important factor to be considered related
to the ship manoeuvrability. In port and harbour area, safety issue is much more serious, because of the nature
of degraded ship manoeuvrability in low ship speed and geo-graphical restriction such as shallow or confined
waterways. These problems will be summarised and discussed with some latest research results.
li

NUMERICAL STUDY OF FREE SURFACE EFFECT ON THE FLOW AROUND

SHALLOWLY SUBMERGED HYDROFOIL
Md. Ashim Ali1 and Md. Mashud Karim2

1
Department of Naval Architecture &Marine Engineering (NAME)

Bangladesh University of Engineering and Technology(BUET)
E-mail: ashim.name@gmail.com
Department of Naval Architecture &Marine Engineering (NAME)

Bangladesh University of Engineering and Technology(BUET),
E-mail: mmkarim@name.buet.ac.bd
ABSTRACT
The wave generation due to the presence of a body moving at steady forward speed beneath a free surface
has been the subject of extensive research work in marine hydrodynamics. In this study, the free surface effect on
the flow around shallowly submerged hydrofoil is numerically computed. Finite Volume Method (FVM) based on
Navier-Stokes equations is used for this purpose. The standard NACA 0012 hydrofoil section is used for ease of
comparison with available experimental data. The k- turbulence model has been implemented to simulate
turbulent flow past the foil surface. To get the free surface elevation, Volume of Fluid (VOF) method is
incorporated in numerical simulation. Grid independency is checked using four grids of different sizes. To
validate the computational results, the free surface wave generated by the flow around hydrofoil at submergence
depth ratio h/c = 0.91 is compared with experimental results published by Duncan. The computed results show
satisfactory agreement with the experimental measurements. Finally, free surface effects on wave profile are
computed for six submergence depth ratios, h/c ranging from 0.911 to 4.0 and maximum amplitude at two
different Froude nos. (Fn = 0.5672 and 0.70)
Keywords: Free Surface Effect, Hydrofoil, CFD, Viscous Flow, Wave Breaking, Volume of Fluid method.
surface wave elevation, the breaking and nonbreaking wave resistance of a two dimensional
hydrofoil. Coleman (1986) tried to simulate the
breaking process for the foil using finite-difference
method by the application of an artificial pressure
distribution on the free surface and a Kutta condition
at the trailing edge. He showed rather phase-shifted
and over predicted wave profile. Liu (1991) calculated
the wave pattern for the foil using an NS solver with
zero-equation turbulence model, but did not show
good agreement with experiments. Hino (1993)
introduced finite-volume method with an unstructured
grid for free surface flow simulations. The method
was based on Euler equations and showed good
results.
1. INTRODUCTION
Numerical prediction of wave pattern, lift force
and drag force of submerged hydrofoil has much
importance to the researchers. The hydrodynamics of
hydrofoils has become a matter of renewed interest in
recent years because of their potentialities in the
design of small craft. Such hydrofoils differ from
conventional hulls in that they obtain their lift by
hydrodynamic action rather than by hydrostatic
buoyancy. In this regard they are closely related to
planning surfaces and airfoils. When deeply
submerged, in fact, the hydrofoil has a flow pattern
which is identical to that of an airfoil of the same
geometry moving at the same Reynolds number. It
seems logical, therefore, to start any theoretical or
experimental study of hydrofoils by using methods
which are well known and widely utilized in
aerodynamics.
Duncan (1983) carried out
experiments for NACA0012 for various submergence
depth, angle of attack and velocity. He obtained free
The main objective of this research is to study the free

surface wave generation for submerged hydrofoil at
different submergence depths and observe the effect
of submergence depth on maximum amplitude of
generated wave. The numerical study is carried out
using CFD software namely STAR CCM+.
of the momentum equation is broken up into two

steps. In the first step, a velocity field is computed
from incremental changes resulting from viscosity,
advection, gravity and body forces. In the second step,
the velocity field is projected onto a zero divergence
vector field resulting into a single Poisson equation
for the pressure field which is solved by using an
incomplete Cholesky conjugate gradient solution
technique.
To get the free surface elevation Volume of Fluid
(VOF) method is used in numerical simulation. This
method is pioneered by Hirt and Nichols (1981). The
VOF technique provides a means of following fluid
regions through an Eulerian mesh of stationary cells.
The basis of the VOF method is the fractional volume
of fluid scheme for tracking free surface boundaries.
The governing equation in this method is given by:
Figure 1: Geometry of flow past a hydrofoil.
2. MATHEMATICAL FORMULATION
The Navier-Stokes equations are the basic
governing equations for a viscous, compressible real
fluid. It is a vector equation obtained by applying
Newton's Law of Motion to a fluid element and is also
called the momentum equation. It is supplemented by
the mass conservation equation, also called continuity
equation. The instantaneous continuity equation and
momentum equation for a compressible fluid can be
written as:
+
uj =0
t x j
( u i ) + u i u j + p ij
t
x j
r
r
DF F (x , t ) r
=
+ V F (x , t ) = 0
Dt
t
)= 0
(2)
i = 1, 2, 3
For numerical study of turbulence, k- model is

used. A k- turbulence model is a two-equation model
in which transport equations are solved for the
turbulent kinetic energy k and its dissipation rate .
Various forms of the k- model have been in use for
several decades, and it has become the most widely
used model for industrial applications. Since the
inception of the k- model, there have been countless
attempts to improve it.
If the fluid is Newtonian, then
i j = 2S * i j ,
1 u i u j
+
2 x j x i
1 u k
3 x k i j
Where, S* is the rate-of-strain tensor.

S i* j =
(4)
Where F is defined whose value is unity at any

point occupied by fluid and zero elsewhere. When
averaged over a computational cell, it is equal to the
fractional volume of the cell occupied by fluid. In
particular, a unit value of F corresponds to a cell full
of fluid, whereas a zero value indicates that the cell
contains no fluid. Cell with F values between 0 and 1
contains a free surface. In addition to defining which
cells contain a boundary, the F function can be used to
define where fluid is located in a boundary cell.
(1)
ji
(3)
In the present study a submerged hydrofoil

moving with a speed U(t) in calm water is considered
as shown in Figure 1. A right-hand Cartesian
coordinate system o-xyz is adopted with x-axis
pointing forward, the z-axis pointing upward. The x-y
plane is coincident with the undisturbed free surface.
c is the chord length of hydrofoil. h is the distance
between the center of the chord length and the free
surface.
The standard k- model is a standard version of the

two-equation model. The basic transport equations of
which are:
d
dt
+
k dV + k (v v ). da = +
g
[G k
t
k
+ G k (( 0 ) + Y M ) + S k ]dV
k . da
(5 )
d
dV + (v v g ). da = + t
dt V
A
A
(
)
+
+
C
G
G
C
G
1 1 k
3 b
nl
+
dV
(
)
+
+
C
Y
S
T
0

M
V
2
3. NUMERICAL SIMULATION
The Navier-Stokes equations are solved by using a
fixed staggered grid. The basic algorithm is the two
step projection method in which a time discretization
. da
(6)
0.2 m coontaining the free

f
surface (zz=0) is defineed.
Four grids namely Grid 1, Grid 2 Grid
d 3 and Grid 4 as
shown inn Figure 4 aare used to check the grrid
independeency. The coarrse hexahedral mesh is refineed
to 0.1 m in
i Grid 1, 0.088 m in Grid 2, 0.05m in Gridd 3
and 0.04 m in Grid 4 as shown in Tablle 1.
Wheree Sk and S are

a the user-sppecified source
terms. 0 is
i the ambientt turbulence value
v
in source
terms, Gk is the generration of turbbulence kinetic
energy duee to the mean velocity graddient, Gnl is thhe
nonlinear production, Gb is the generation oof
turbulence kinetic enerrgy due to buoyancy, YM
on of the fluctuuating dilatationn
represents the contributio
in compresssible turbulennce to the oveerall dissipationn
rate, C1, C2 and C3 arre constant, k and are thhe
turbulent Prandtle
P
numbeers for k and , respectively.
The turbuleence kinetic en
nergy, k is giveen by
k=
3
U avg I
2
)2
f
velocity.
Where, Uavvg is the mean flow
The turbulence intensity,, I and the turbbulence length l
can be foun
nd from the folllowing equations:
Figure 3: C
Close up views of A.
l = 0.07
0 L and I = 0 . 16 (Re ) 8
Also the tuurbulence dissippation rate, defined
d
as:
3
= C
3
4
k2
l
The constaants in Standaard k- model are consideredd

as:
Grid 1
C1 = 1.44,, C2 = 1.92, C = 0.09, k = 1.0 and = 1.33
Grid 2
Grid 3
Grid 4
Fiigure 4: Mesh refinement.

T
Table
1: Grid sppecification
Grid no
Grid 1
Grid 2
Grid 3
Grid 4
Figurre 2: Mesh distrribution in fluiid domain.
Size of refined mesh

0.110 m
0.008 m
0.005 m
0.004 m
4. RESU
ULTS AND DISCUSSIO
D
ONS
s
packaage STAR CCM+ is used foor
CFD software
numerical simulation. In this simulationn, VOF methodd
is employeed for the freee surface traccking. Figure 2
shows the fluid domain which is used
d for numericaal
a
simulation and Figure 3 is the closse up view at
w
shows the
t mesh distrribution aroundd
hydrofoil which
hydrofoil surface.
s
To chheck the grid independency
i
a
numerical model set up with
w four differrent mesh sizees
are used, which
w
are partlyy shown in Fig
gure 4. Trimmeer
meshing model
m
is used to generate vo
olume mesh inn
STAR CC
CM+. The genneral surface settings for alll
meshes aree the same. The inlet is set ass velocity Inlett,
Outlet as pressure
p
Outlett, top and botto
om wall as slipp
wall so thaat there will bee no friction. To
T get the wavve
profile, a refined
r
region between z = 0.2
0 m and z = -
As a teest case, a sim

mulation was carried out for
f
NACA 00012 having h/cc = 0.911 and Fn
F = 0.5672, Re
R
= 1.7695 X 106 and anggle of attack 50. To check grrid
independeency, the waave profile iss computed for
f
different grids and plootted in Figurre 5. From thhis
milar results wiith
figure it iss seen that all ggrids make sim
a little diiscrepancy. Hoowever, the reefined Grid 4 is
chosen foor the simulatioon as it gives reesults very cloose
to experim
mental ones. Inn Figure 6, the numerical resuult
with Gridd 4 is comparedd with the expeerimental resullts
of Duncaan(1983). Up to x/c = 4.0, the numericcal
results shhow good aagreement witth experimenttal
results. After
A
x/c = 4.00, the numeriically calculateed
wave is gradually daamped out but
b
in case of
experimenttal results, thee wave amplitu

ude is observedd
nearly the same. Howeveer, the numericcal results agree
well with those of Muuscari (2003) and Pascarelli
(2002).
0.16
Grid
Grid
Grid
Grid
0.12
0.08
1
2
3
4
0.04
z/c
0.00
-0.04
Fiigure 7: Conveergence historyy for Grid 1.
-0.08
-0.12
-0.16
-1
x/c
Figure 5: Grid indepenndency

0.16
Num
merical
Dunccan (1983)
0.12
0.08
z/c
0.04
0.00
f Grid 2
Figgure 8: Converrgence history for
-0.04
-0.08
-0.12
-0.16
-1
x/c
Figure 6: Comparisson of Expeerimental andd

Numericaal results
May be more refinned mesh couuld give morre
accurate reesults but forr the limitationn of computeer
resources, Grid 4 is chosen here. Figurees 7 to 10 show
w
the converrgence history of the numerrical simulationn
with differrent grid sizes. Simulation is carried ouut
using timee step size off 0.05s. It is seen that thhe
residual foor all parametters is getting constant afteer
6000 iterattion. With the decrease in thhe cell size, thhe
fluctuationn in residual vaalue is gettingg higher. But as
a
the decreaase in the celll size makes the numericaal
results cloose to experim
mental results, the small celll
size is chossen for further simulation.
Fiigure 9: Conveergence historyy for Grid 3.
gure 10: Conveergence history

y for Grid 4.
Fig
Figure 11 to 14 shows the wave generation for a

submerged hydrofoil having h/c = 0.911 at different
times of the simulation. Figure 15 shows the effect of
submergence depth ratio h/c on the free surface
elevation. The rate of change in maximum amplitude
is decreasing with increase in h/c. The variation of
maximum amplitude with respect to h/c is shown in
Figure 16. It is seen that the maximum amplitude
increases exponentially with decrease in submergence
depth ratio, h/c. It means that for small value of h/c,
maximum wave amplitude, Am is very large and there
may be wave breaking on generated wave.
Figure 11: Visualization of free surface at t = 0.0s
0.12
h/c
h/c
h/c
h/c
h/c
h/c
0.08
=
=
=
=
=
=
0.911
1.50
2.00
2.50
3.00
4.00
z/c
0.04
0.00
-0.04
-0.08
-0.12
-1
x/c
Figure 15: Effect of h/c on free surface elevation.

-0.20
Fn = 0.5672
Fn = 0.7000
Maximum Amplitude, Am
-0.16
-0.12
-0.08
-0.04
0.00
0
h/c
Figure 16: Effect of h/c on maximum wave

amplitude.
Table 2: Force coefficients for hydrofoil at =5, Fn
= 0.5672
Submergence
depth ratio,
h/c
0.91
1.50
2.00
2.50
3.00
4.00
Deep water
(Experimental
value)
Lift
coefficient
CL
0.6032519341
0.5493059158
0.5341051221
0.5312077403
0.5296801329
0.5258696675
0.5180
Drag
coefficient
CD
0.02192726918
0.01405391004
0.01186849736
0.01301105693
0.01120258588
0.01115290634
0.0113
Table 2 shows the lift and drag coefficients of

hydrofoil at =5 for different submergence depth
ratios. It also shows the experimental values taken
from the book by Abbott & Doenhoff (1959). As the
submergence depth ratio increases, lift and drag
coefficients decrease except the drag coefficient at h/c
=2.5. At h/c =4 coefficients are almost close to the
experimental values at deep water. So submergence
depth ratio more than 4.0 can be treated as the deep
water case and free surface effect becomes almost
zero.
REFERENCES
[1] Abbott, I. H. and von Doenhoff, A. E., "Theory
of Wing Sections," Dover Publications, New
York, 1959.
[2] Coleman R. M.,Nonlinear Calculation of
Breaking and Non-Breaking Waves Behind a
Two-Dimensional Hydrofoil, 16th Sym. On
Naval Hydro. Berkeley, USA, July 1986.
[3] Duncan J. H., The breaking and non-breaking
wave resistance of a two-dimensional
hydrofoil, J. Fluid Mech. Vol. 126, 1983.
[4] Hino T.,A Finite-Volume Method with
Unstructured Grid for Free Surface Flow
Simulations, 6th Int. Conf. on Numerical Ship
Hydro, Iwoa, USA, Aug. 1993.
[5] Hirt, C.W., Nichols, B.D. (1981), "Volume of
fluid (VOF) method for the dynamics of free
boundaries", Journal of Computational Physics
39 (1): 201-225.
[6] Liu H. and Ikehata M., Computation of Free
Surface Viscous Flow with High Reynolds
Number around a Submerged Hydrofoil, 2nd
Osaka Int. Colloquium on Viscous Fluid
Dynamics in Ship and Ocean Tech., Japan, Sep.
1991.
[7] Muscari R, Di Mascio A (2003) A model for the
simulation of steady spilling breaking waves. J
Ship Res 47:1323.
[8] Pascarelli. A , Iaccarino. G & Fatica. M,
Proceedings of the Summer Program, Center for
Turbulence Research. , 2002.
5. CONCLUSION
In the present work, finite volume method is used
incorporating VOF method and k- turbulence model
for simulation of the turbulent flow around shallowly
submerged hydrofoil. From the above study, it can be
concluded that the present method is successfully
applicable to compute free surface wave generated by
the hydrofoil moving beneath the surface. As the
submergence depth ratio increases, the wave
amplitude decreases. At the submergence depth ratio,
h/c > 4.0, the effect of free surface vanishes.
Similarly, the lift and drag coefficients decrease with
increase in submergence depth ratio and become
constant after h/c = 4.0. This implies that for
submergence depth ratio greater than 4, shallow water
effect does not exist and the deep water condition can
be considered.

The Internatiional Conference on
o Marine Technoology
11-12 Decem
mber 2010, BUET, Dhaka, Bangladessh
EXPE
ERIMENT OF BENDE
ED HYDRO
OFOIL-SUP
PPORTED
CEAN STRU
UCTURE PILLAR
P
OC
M
M.A.DJABBA
AR1 and JUSWAN2
1
Professor,,
Departmeent of Ocean Enginneering,
Faculty off Engineering,
Hasanudddin university, Makkassar, 90245, Indo
onesia.
Email: alhham.djabbar@yahooo.com
Associate Professor
Departm
ment of Ocean Engiineering,
Faculty of
o Engineering,
Hasanudddin university, Maakassar, 90245, Ind
donesia
ABS
STRACT
Ocean stru
ucture may be hit by strong waves
w
or tsunaami. One of some ways to redduce the horizzontal force is as
proposed here
h
by fixing up side down tip bended hyddrofoil. To obsserve the effectt of the bended
d tip in reducinng
the force, an experimennt has been performed
p
at the flume tannk of Ocean Engineering Study
S
Program
m,
Departmennt of Naval Arrchitecture, Faculty of Enginneering, Hasan
nuddin Universsity, Makassar,
r, Indonesia. The
Th
specimen used
u
here is circular cylinderr pipe, represeenting the pillaar of the oceann structure withh down ward tip
t
bent, upside down hyd
drofoil.Previoussly similar exxperiment wass done withouut bending th
he hydrofoil tiip,
indicating only small eff
ffect of reducinng horizontal fforce. The ressults have beenn compared with
w the previous
results of thhe test. It is seeen that the testt results could improve the eff
ffect of reducinng horizontal lo
oad on the pillaar,
based on thhe experimentaal oscillation.
Keywords:: Flume tank, Ocean
O
structuree pillar, Tip bennded hydrofoill
Figure 1 [4].
[
1. INTRO
ODUCTION
N
p1 + gh
g 1+
Climate change
c
affectss many part of the worldd
including ocean
o
and beacch. Coast or beeach have beenn
used for many functions, needed to fulfill thhe
nt of human being includinng recreationaal
requiremen
and sportin
ng activities.
Indonesia as an islaand countries has very longg
o about 80.0000 km, second to
t Canada.
coast line of
Beach and coastall structure most
m
probablyy
destroyed by strong or high
h
amplitudee waves. Therre
are some ways
w
to prevennt them. One of the way, w
we
proposed is using upside down hydrofooils.
The ob
bjective of thhis experimenttal study is too
observe thhe effect of th
he position of the strips oof
hydrofoils on vertical cyllinder motion..
Previouus studies werre performed
d by Mustakim
m
[6], Hamzaah [3], Nur Aeenoon [1], and Juswan and M
M.
A. Djabbaar [5] relating the up side do
own hydrofoills
without strrip (no bendinng) to strengtheen the pillar oof
the coastal structure.
1 2
1
v1 = p 2 + gh2 + v 22
2
2
(1)
Figure 1. Lifting of the Foil (Wikipediia website 20099)
3. EXPE
ERIMENT
The expeeriment was peerformed at thhe wave tank of
Ocean Enngineering Stuudy Program, Department of
Faculty off Engineerinng,
Naval Architecture,
A
Hasanudddin University. The tank (Figgure 2 ), withoout
specimen is 18.4 m long, 1.2 m wiide, 1.0 m higgh.
Water dep
pth of 0.65 m was
w set for the experiment,
Waves weere generated bby existing wavve maker for
regular wave,
w
measurinng about 6 cm
m high and 2000
cm long. Figure
F
3 shows the wave tannk with specimeen
hanging by
b spring (pegaas )
2. HYDR
ROFOILS
Siince long timee ago hydrofo
oils were beingg
used for marine
m
vehicle, the main functtion which is too
lift the bod
dy of the vehiccle by lifting the
t body (hull)).
So the resistance of thhe vehicle beccomes smallerr,
peed increases. The theory iis
which meaans that the sp
based on Bernoulli`s eq
quation (Eq.1)) as shown inn
Unlike previous experiment (regular and high

amplitude waves), only regular waves was used with
5 configurations (Figure. 4) of the specimen making
at least 15 runs were observed. Configuration of strip
position 3 was enlarged for detailed size. One of the
runs is shown in Figure. 5. The hydrofoil measuring
1.15 m uniform long, 0.095 m cord, NACA 2412 was
located 0.6 m from the cylinder bottom. In the runs
the readings were the water (wave crest and through)
level on the mm scaled pipe, visually and plotted in
oscillation versus strip position, shown in Figure. 7.
Figure 2. Wave Tank
Figure 3. Tested material (side view)
4. RESULT AND DISCUSSION
Table 1 Specimen vertical motion

NUMBER
Both Table 1 and Figure 7 show the

experiment results. Configuration of strip position 4
produced shortest oscillation. Since we consider that
the shorter the oscillation the higher the negative lift,
the position 4 produced more down ward lift (force),
however there was slightly unstable of turning. This
was believed due to unsymmetrical moment. The
most stable condition was configuration of strip
position 2. Configuration of position 1 was the same
as previous reading in Fig.6 [5] for regular waves
therefore the present bended tip comparison with no
tip bending produced higher negative lift (small
oscillation) force as we expected.
STRIP
POSITION
no strip
Strip at tip
Strip 15 Cm
from tip
Strip 20 cm
From tip
Strip 30 cm
from tip
4
5
OSCILLATION
( cm )
6
3
4
Figuure 4 Configuurations of the specimen.
Figure 5 Running experiment
Upanddownmovement(cm)
10
9
8
7
6
5
RegularWave
bigwave
3
2
1
0
50
55
60
65
Foillocationfromthepipebottom
Figure 6 Experimental results (motion forces location)
10
Oscillation ( cm )
6
5
4
3
2
1
0
0
1
2
3
4
5
Strip position (hydrofoil at 60 cm from cilynder bottom)
Figure 7 Specimen motion
5. CONCLUSION
There is a possibility, based on smaller
oscillation of using hydrofoil on the coastal structure,
fixing on its pillar (circular cylinder) to reduce the
horizontal wave forces with strip (bended foil).
The experiment used only regular waves and

compared the oscillation of the specimen between
with and without tip bending.
[4]
Juswan And M. A. Djabbar, 2010.

Hydrofoilaided
Coastal
Strukture
Experiment.,First Makassar Internasional
Conference an Civil Engenering ( MICCE
2010 ) ISBN 978-602-95227-0-9, Makassar
[5]
Hamzah, ST. 2009 Bangunan Pantai tahan

Tsunami dengan menggunakan foil terbalik,
Program Studi Teknik Kelautan Jurusan
Perkapalan Fakultas Teknik Universitas
Hasanuddin, Makassar.
[6]
Mustakim, ST. 2008 Bangunan Pantai

Tahan Gelombang Dengan Menggunakan
Foil Terbalik, Program Studi Teknik
Kelautan Jursan Perkapalan Fakultas Teknik
Universitas Hasanuddin, Makassar.
[7]
http://www.wikipedia.org, Bernoulli based

pressure and foil, September 2nd2009
[8]
http://www. wikimedia.org, NACA 2412.svg,

August 19th , 2009
[9]
http://www.airfieldmodels.com, Calculating
airfoil ordinates Naca 2412, August 31st
2009
REFERENCE
[1]
[2]
[3]
Aenun, Noor. ST. 2010 Bangunan Pantai

Tahan Gelombang Dengan Hidrofoil Untuk
Menahan Gelombang Besar.Program Studi
Teknik Kelautan Jurusan Perkapalan
Fakultas Teknik Universitas Hasanuddin,
Makassar.
Albert Pasorong, dan Samaluddin, 2006
Efektivitas
system
kerja
pembangkit
gelombang di saluran gelombang , Program
Studi Teknik Kelautan Jursan Perkapalan
Fakultas Teknik Universitas Hasanuddin,
Makassar.
John D.Anderson. JR. 1985. Fundamentals
of Aerodynamics, Megraw Hill International
Edition, mechanical engineering series.
11
ProceedingsofMARTEC2010
TheInternationalConferenceonMarineTechnology
1112December2010,BUET,Dhaka,Bangladesh
INVESTIGATION ON THE STABILITY OF A TRIMARAN WING IN GROUND

EFFECT (WIG) CRAFT WITH ENDPLATE
Adi Maimun1, Agoes Priyanto1, Noverdo Saputra1, Saaed Jamee1, Ike Suharti1, Md. Mobassher1
1
Department of Marine Technology,

Faculty of Mechanical Engineering,
University Technology Malaysia
Email: adi@fkm.utm.my , verdokapal@yahoo.com
ABSTRACT
Longitudinal stability of Wing in Ground Effect Craft (WIG) is still the main concern of WIG craft designers
and solutions are being sorted out to reduce this effect. In this research, investigations were conducted to
determine the effect of flat ground and end plate proximity on the aerodynamic characteristics and stability
criteria of NACA 6409 rectangular wing. The investigation were performed using vortex lattice method and
examining the effects of flat ground and endplate on the performance of a trimaran WIG for relative ground
clearances of 0.06 < h/c < 0.3, ratio endplate he/c = 0.1 on angles of attack between 0 and 8, aspect ratio
1< AR < 2 Data is presented for lift coefficient with or without endplate, and static stability margin (SSM)
versus angle of attack, and ground clearance. The result of the computation shows that the SSM was
significantly affected with changing ground clearances with or without endplate. The results were also
compared with experimental & CFD data from another research work.
Keywords: WIG trimaran, Longitudinal Stability, NACA 6409, vortex lattice method, endplate, ground
effect
NOMENCLATURE
X,Y,Z
A
Aspect Ratio ( b/c )
b
Wing Span ( m )
c
Chord length ( m )
CM Moment Coefficient (=L/0.5AU2 )
CL Lift Coefficient (=L/0.5AU2 )
CD Drag Coefficient (=D/0.5AU2 )
CDi Induced Drag Coefficient
h/c Ground clearance
he/c endplate ratio
N
Maximum number of element panel
cc
cord along trailing leg of elemental panel (m)
Angle of attack ( 0 )
Dihedral angle ( 0 )
Ground Influence coefficient
Endplate Influence coefficient
0
sweep Angle ( )
air density
vortex strength
F
Influence function geometry of single horshoe
S
wing area (m2)
U
free stream velocity (m/s)
u
backwash velocity (m/s)
v
sidewash velocity (m/s)
w
downwash velocity (m/s)
r1 , r2
vector distance
body-axis system for plan form
wind axis system
x,y,z
X ac
edge
X cg distance center of gravity from leading edge
X h distance center of height from leading edge

Subscript
u
backwash
v
sidewash
w
downwash
n
index for elemental panel
OGE
Out Ground Effect
IGE
In Ground Effect
E
With Endplate
WE
Without Endplate
1. INTRODUCTION
WIG craft is relatively new concept of
transportation, can have a fruitful future as it is more
efficient than equivalent aircraft and quicker than
equivalent marine vessels. The advantages of the wing
in ground effect craft come from additional lift
provided
by
ground
effect
phenomenon.
Wieselsberger [1], Reid [3] and Carter [4],
theoretically and experimentally analyzed the effect of
13
axis system for horsoe vortex

distance along
distance along X,Y,Z
distance center of aerodynamic from leading
the ground on wings. Hemke [2] , analysis of

experimental drag of wing with endplate shown effect
endplate in aerodynamic characteristic. Absolutely,
ground clearances and endplate ratio have influence in
static stability margin (SSM). Kumar [5, 8], Irodov
[6], Zhukov [9], and Staufenbiel [10, 11], Chun &
Chang [13] all addressed the issue of pitch stability
in ground effect, determining that the relative position
of the aerodynamic centre in height and pitch
influenced the nature of the pitch stability. Numerous
studies have been conducted for analyzing the
influence of the ground on wing performance.
However, few largely overlooked. Rozhdestvensky
[12, 14] presents a synopsis of research examining
the influence of wing profile and platform on the
positioning of the two aerodynamic centers.
In this paper, a trimaran WIG as shown in Figure 1
will be analyzed. The main wing of the WIG craft is
of NACA 6409 section and the tail is of NACA 0012
section. The particulars of the craft are shown in table
1.
Figure 2.
w cos U sin = 0
(1)
The boundary condition is to be extended to

represent wing with dihedral as in Figure 3.
Figure 3.
w cos cos v sin U sin cos = 0 (2)
The vortex lattice is located in a plane parallel to free

stream as shown in Figure 4.
Figure 1.
Table 1. Principle dimension of WIG Catamaran

Length over all (LOA)
7.22 m
Breadth over all (BOA)
5.4 m
Hull breadth (B)
0.8 m
Side Hull breadth (B)
0.15 m
Span length (s)
5m
Chord length (c)
4m
2. NUMERICAL FORMULATION
Figure 4.
Related from 0, equation (2) can be

expressed for downwash influenced coefficient:
... (3)
Fw ( x ' , y , z , s , , ' ) =
( y tan ' x ' ) cos

( x ' ) 2 + ( y sin ) 2 + cos 2 ( y 2 tan 2 + z 2 sec 2 ) ' 2 z cos sin ( y + x ' tan ' )
x ' + s cos tan ' cos tan ' ( y + s cos ) cos + ( z + s sin )
x
2
2
2
+ ( y + s cos ) + ( z + s sin )
x ' + s cos tan '
[(
x ' s cos tan ' cos tan ' ( y s cos ) cos + ( z s sin

2
2
2
+ ( y s cos ) + ( z s sin )
x ' s cos tan '
[(
2.1 Vortex lattice Method

The traditional representation for flat wing [7] is
shown in Figure 2.
( y + s cos )
x ' + s cos tan '
2
2
+ ( z + s sin ) ] (x ' + s cos tan ' )2 + ( y + s cos )2 + ( z + s sin )2
[( y + s cos )
14
( y s cos )
2
2
+ ( z s sin ) ] (x '
[( y s cos )
x ' s cos tan

s cos tan
'
) + ( y s cos )
' 2
+ ( z s sin )
Sidewash influenced coefficient:
N /2
2
N / 2 2
C L ( OGE ) = 2 n ccU u ,v , w + n 2b
n S
n S
... (4)
Fv ( x ' , y , z , s , , ' ) =
[(1 U ) + (U
x ' sin z cos tan '

( x ' ) 2 + ( y sin ) 2 + cos 2 ( y 2 tan 2 + z 2 sec 2 2 yx ' tan ' ) ' 2 z cos sin ( y + x ' tan ' )
(x
(x
tan
+ s cos
'
[(x
'
tan
s cos
'
[(x
(y
[( y + s cos
[( y s cos
'
(y
)2
)cos
'
+ ( z + s sin
s cos
)2
tan
)cos
'
+ ( z s sin
)2 ]
'
'
(y
tan
tan
s cos
+ s cos
'
tan
+ s cos
(y
+ s cos
'
(y
(y
s cos
s cos
)2
+ s cos
)2
) cos
(z +
) cos
(z
(z
s sin
(z
s sin
)2
)2
)2 ]
[(x
x ' s cos tan

'
s cos tan
'
+ ( y s cos
s sin
)2
)2
'
)2
+ ( z s sin
u ,v , w
tan )]cos )
... (9)
+ s sin
x ' + s cos tan '
2
2
x ' + s cos tan '
+ ( y + s cos ) + ( z + s sin
[(
u ,v , w
2.2 Effect of flat ground and endplate on

aerodynamic characteristic
The theoretical treatment of ground effect presented
by Wieselsberger [1] or predicting the reduction in
induced drag for wing at various heights of the quarter
chord of the wing above the ground. Experiments
have conducted by Reid [3] & Carter [4]. The induced
drag in ground effect is given by the equation
2
C Di ( IGE )
Backwash influenced coefficient:

Fu ( x , y , z , s , , ) =
'
'
... (5)
z cos y sin
( x ' ) 2 + ( y sin ) 2 + cos 2 ( y 2 tan 2 + z 2 sec 2 2 yx ' tan ' ) ' 2 z cos sin ( y + x ' tan ' )
(x
+ s cos
'
(x
[(x
'
s cos
[(x
tan
)cos tan
2
tan ' ) +
'
+ s cos
'
tan
'
s cos
'
)cos
tan
tan
'
(y
(y
'
(y
(y
'
+ s cos
+ s cos
s cos
s cos
) cos
)2 + ( z
) cos
)2 + ( z
(z
+ s sin
+ s sin
(z
s sin
s sin
)sin
)sin
]
]
r .r r r
1
n 1 2 2 r0 1 2
4
[r1 .r2 ] r1 r2
1
n Fu , v , w ( x, y, z , s, , )
4
C Di (WE ) = (1 )
(7)
1
1
............................ (13)
C L S ..... (14)
b 2
2he
c
When =
.. (15)
2he
1 + 1.66
c
1.66
....... (8)
Just like equation (13) coefficient with endplate

could be shown:
The coefficient along an elemental length of

chordwise and coefficient divided by free stream
dynamic pressure can be expressed to implement
equation (8) in Kutta-Joukowski theorem which gives
the following equation :
CL ( E ) = CL (WE )
15
1
........................... (12)
1
The frictional drag of the endplates can reduce

induced drag, is sufficiently large to increase the
efficiency of the wing. Analysis of experimental drag
of wing with endplate is done by Hemke [2]
From equations (3), (4), (5) and (7), downwash,

sidewash and backwash velocity can be expressed as :
U u ,v, w =
b2
s
C L ( IGE ) = C L (OGE )
(6)
Equation (10) can be expressed:
Using Equations (9), equation (12) can be rewritten

as :
Equation 6 can be integrated to give the induced

velocity for a vortex segment of arbitrary length. This
is done by Bertin & Smith. It takes the form of
equation ( 7 )
U=
h
c .. (11)
When =
h
1.05 + 7.4
c
1 1.32
C L ( IGE ) = C Di
The velocity induced by a straight vortex element can

be calculated by Biot-Savart law which takes the form
as :
1
dl.r
U=
n 3
4
[r ]
C S
= (1 ) L 2 .... (10)
b
1
1
.. (16)
2.2 Longitudinal static stability
C Lh < 0 ............................................................... (26)
WIG should be stable in pitch like an airplane, if

rigid body have been disturbed should be return to
undisturbed position, and mathematically given as:
Referring toequation 21 derivatives could be written

as:
C M < 0 ............................................................... (17)

CM =
CM X cg X h
=
>0
c
C L
dCm dCm dCl

=
< 0 ...................... (18)
x
d
dCl d
And derivatives could be written as:
CM =
if
Where
From equation (27) we can describe the position

center of gravity (COG) craft located downstream of
aerodynamic center in height (ACH) (Rozhdestvensky
[12]).
dCM
CL < 0 ....................................... (19)
dCL
C L > 0
, so
.................................... (27)
After substituting equation (22) to equation (27),

height stability equation can be derived
dC M
< 0 ........................ (20)
dC L
X ac X h > 0 .................................................. (28)
dC M X cg X ac
.............................. (21)
=
dC L
c
The distance between the aerodynamic center in

pitch and aerodynamic center in height are referred to
as Static Stability Margin (SSM) Rhodes and Sayers
[16]
dCM X cg X ac
=
< 0 ............................... (22)
dC L
c
SSM = X ac X h > 0 .................................... (29)

3. Computational result and discussion
From equation (22) it can be deducted the

position of center of gravity (COG) of the craft should
be located upstream of the aerodynamic center of
pitch (ACP) (Rozhdestvensky [12]).
Figures 5-10 draw a comparison of lift

coefficient (CL) for wing only carried out by VLM
simulations, CFD simulation Maimun et al. [17], and
experimental data Jung et al. [15]. The magnitude of
lift coefficient increases with increment of aspect
ratio, angle of attack, and ground clearance.
According to Figure 5-10 lift coefficients of VLM at
0o - 2o angle of attack is less accurate than CFD
results, where CFD is closer to experimental results.
At 4o - 8o VLM results better than CFD results,
because its result near experimental results.
WIG also needs to have stability in the vertical

dimension. The mathematical can be expressed as
Kumar [5, 8]:
C Mh < 0 ............................................................... (23)
C Mh =
dCm dCm dCl

=
< 0 .................... (24)
x
dh
dCl dh
C Mh =
CM
C Lh < 0 ........................................ (25)
C L
The derivatives of the lift coefficient with

increasing height must be negative, mathematically
have been expressed by Irodov [6] & Staufenbiel
[10,11].
Figure 5. Lift Coefficient (CL) versus angle of attack

()for h/c= 0.1 and AR = 1
16
Figure 6. Lift Coefficient (CL) versus angle of attack ()

for h/c= 0.1, AR = 1.5

for h/c= 0.3, AR = 2
Figure 11-15 shows the comparison of lift

coefficient (CL) for wing only with endplate due to
VLM simulations, and experimental data Jung et al.
[15] The magnitude of lift coefficient increases with
increment of aspect ratio, angle of attack, and ground
clearance. According to Figure 11-15 lift coefficients
of VLM at 0o - 2o angle of attack is less accurate than
experimental results. At 4o - 8o, VLM results near
experimental results.
for h/c= 0.1 and AR = 2

for h/c= 0.1 and AR = 1, endplate he/c = 0.1
Figure 8. CL versus angle of attack for h/c= 0.3 and AR = 1

for h/c= 0.1 and AR = 1.5, endplate he/c = 0.1
Figure 9. Lift Coefficient (CL) versus angle of attack () for

h/c= 0.3, AR = 1.5
17
ground clearance. At h/c = 0.06, SSM in stable

condition, but when entering at h/c = 0.1, SSM
becomes unstable, and after entering h/c = 0.3, SSM
becomes stable again. The phenomenon occurred
again in WIG with endplate (see 0 ). From Figures 17
and Figure 18 it can be concluded that there is not a
stable phase between h/c = 0.06 and h/c = 0.3 and
SSM decrease.

Figure 17. SSM wing w/o endplate, h/c= 0.06 - 0.3

Figure 18. SSM wing with endplate, h/c= 0.06 - 0.3
4. Conclusion
In this work, traditional Vortex Lattice
Method (VLM) has been modified by using much
complex Flat Wing Theory [7] and then
Wielselsberger Ground Effect Theory [2] has been
used to calculate lift coefficient In Ground Effect
(IGE). The results show values of Lift Coefficient
(CL) are more close to experimental results for Angle
of Attack (AOA) larger than 2 deg for different
Aspect Ratio (AR) and Ground Clearances.
Researchers suggest this procedure can be followed
for initial estimation of Lift Coefficient for complex
wing platform WIG.
It can be shown that using End Plate on the
tip wing increases lift coefficient but reduces SSM.
On this note, perhaps smart configuration of tail can
solve this paradox which could be investigated later.

for h/c= 0.3 and AR = 1.5, endplate he/c = 0.1

Acknowledgements
Regarding equation (29) the distance between
the aerodynamic center in pitch and aerodynamic
center in height must be greater than zero. 0 shows
stability condition wing without endplate in variation
The authors would like to thank the Ministry

of Science, Technology, and Innovation (MOSTI)
18
Sciences, 42, (3), 211-283 (2006).
Malaysia for funding this research under vote number

79344.
[15]
Jung, K. H., Chun, H. H. and Kim, H. J.,

Experimental investigation of wing-in-ground
effect with a NACA 6409 section, Springer
Jasnoe (2008).
[16]
Rhodes, S. C. and Sayers, A. T., Experimental

Investigation: Stability criteria of an
Uncambered Airfoil in Ground Effect, R & D
Journal of The South African Institution of
Mechanical Engineering, 25 (2009).
[17]
Maimun, A., Jamei, S.,

Priyanto, A. and
Azwadi, N., Influence of Twin Hulls
Geometry on Aerodynamic Characteristics of
WIG Catamaran During Ground Effect, The
9th WSEAS International Conference on
Application
of
Computer
Engineering
(ACE10), Penang, Malaysia, (2010).
REFERENCES
[1]
Wieselsberger, C, Wing Resistance Near the

Ground, NACA, TM-77 (1922).
[2]
Hemke, P. E., Drag Of Wing With End

Plates, NACA, Rep 267 (1927).
[3]
Elliot G.R., Full Scale of Investigation

Ground effect, NACA, Rep-265, (1937).
[4]
Carter, A. W., Effect Ground Proximity On

The Aerodynamic Characteristics Of AspectRatio-1 Airfoil With and Without Endplates,
NASA,TN-D-970, (1961).
[5]
Kumar,
P. E.,
An
Experimental
Investigation
into
the Aerodynamic
Characteristics of a Wing, With and
Without Endplates in Ground Effect, Rept.
Aero 201,College
of Aerodynamics,
Cranfield, England, (1968).
[6]
Irodov, R. D., Criteria of longitudinal

stability of ekranoplan, Ucheniye Zapiski
TSAGI, 1 (4), 63-74 (1970).
[7]
Margason, R. J. and Lamar, J. E., Vortex

Lattice Fotran Program For Estimating
Subsonic Aerodynamic Characteristic of
Complex Planforms, NASA TN D-6142(1971).
[8]
Kumar, P. E., Some Stability Problems of

Ground Effect Vehicle in Forward Motion,
Aeronautical Quarterly, 18, 4152 (1972).
[9]
Zhukov, V. I., Some matters of longitudinal

stability of ekranoplans, Trudy TSAGI (1974).
[10]
Staufenbiel, R.W. and Schlichting, U.J.,

Stability of airplanes in ground effect,
Journal of Aircraft, 25, (4), 289-294 (1988).
[11]
Staufenbiel, R. and Kleineidam, G.,

Longitudinal motion of low-flying vehicles in
nonlinear flowfields, Proceedings of the
Congress of the International Council of
the Aeronautical Sciences, Munich, 293308
(1980).
[12]
Rozhdestvensky, K. V., Aerodynamics of a

Lifting System in Extreme Ground Effect,
Springer, Moscow, (2000).
[13]
H.H. Chun, and C.H. Chang, Longitudinal

stability and dynamic motions of a small
passenger WIG craft , Elsevier journal of
ocean engineering, vol. 29, p. 1145-1162
(2002).
[14]
Rozhdestvensky, K. V., Wing-in-ground

effect vehicles, Progress in Aerospace
19

DEVELOPMENT OF A SHIP WEATHER-ROUTING ALGORITHM FOR

SPECIFIC APPLICATION IN NORTH INDIAN OCEAN REGION
Debabrata Sen1 and Chinmaya P Padhy2
Department of Ocean Engineering & Naval Architecture

Indian Institute of Technology Kharagpur, India
E-mail: deb@naval.iitkgp.ernet..in
Product Engineering Services, Wipro

Technologies, Kolkata, India
E-mail: chinmaya.padhy@wipro.com
ABSTRACT
This paper presents development of a ship weather routing algorithm for determining minimum-time route.
The optimization is based on a form of Dijkstras algorithm. The developed algorithm is investigated using
various realistic wave data for the North Indian Ocean region obtained from the 3rd generation WAM model.
Illustrative minimum-time sea routes on the Arabian sea and the Bay of Bengal have been determined and
presented. All relevant practical and realistic constraints such as presence of land boundaries, consideration
of non-navigable water, effects of wind and current, voluntary speed reduction etc. can be incorporated
within the framework of the algorithm.
Keywords: Ship-routing, added resistance, wave-modeling
trajectory, Zoppoli [19] who formulated the minimal

time algorithm as a N-stage discrete process subjected
to stochastic and dynamic conditions, and Chen [2]
who developed an adaptive open-loop feedback
optimization procedure. Other works include Mitchell
and Papadimitriou [12] who investigated shortest path
through a weighted planar sub-division, and Hagiwara
[7] [8] who proposed the isochrone method for the
solution of the minimum-time route (MTR) problem.
More recent work on the ship routing problem can be
found in Perakis and Papadakis [16], Chen and Lacey
[3], Ulusoy [18] and Montes [13]. Although most of
these works focus on the optimization algorithm using
advance mathematical models, their use for
developing an operational algorithm for practical
application are not often demonstrated and thus their
actual application remains somewhat unclear.
1. INTRODUCTION
Ship routing is defined as a procedure to determine an
optimal track for ocean voyages based on forecasts of
weather, sea conditions, and a ships individual
characteristics for a particular transit. It is concerned
with the choice of the most suitable strategic
trajectory or route and the corresponding control
options from the voyage origin to destination so that a
desired objective function or performance index is
optimized. There are a number of different
approaches to calculate the shortest path/route
between two nodes representing the start and
destination ports. For a given problem, the
implementation of this path finding algorithm may
have to be executed several times due to the varying
ocean wave conditions during the passage of the ship.
It is therefore important to choose an algorithm that is
as efficient as possible. Within specified limits of
weather and sea conditions, the term òptimum may
be defined to mean maximum safety and crew
comfort, minimum fuel consumption, minimum time
underway, or any combination of these factors.
From the above it can be seen that although ship

weather routing is a fairly old problem, due to its
commercial value, the exact operational algorithm is
rarely available in open literature. In this paper, taking
advantage of these modern developments in wave
modeling, an optimum track ship routing algorithm
for ships operating in the Indian Ocean region is
described by using a network graph for the Indian
Ocean. A binary heap version of modified Dijkstra's
algorithm has been used to determine the optimum
route given an input from atmospheric/wave model
and basic ship response functions. The algorithm is
then applied for determining optimal routes for
voyages in the North Indian Ocean region.
The problem of obtaining an optimal trajectory of

ship has attracted attention of many researchers in the
past. This includes Hanssen and James [10] who
demonstrated an optimum ship route under stationary
weather conditions, Haltiner et al. [9] who developed
a routing algorithm using variational methods
assuming ship speed under maximum power to be
independent of time, Faulkner [4] who developed
models based on numerical methods for optimal
21
2.3 Ship Track (Route) Optimization
2. THE ROUTING ALGORITHM
Without loss of generality, the objective which needs

to be optimized for the track optimization problem
can usually be related to a reduction in speed under
the following heads:
2.1 Wave Climate

The most important environmental factors relating to
ships safety and performance on the high seas are
surface winds and waves. For the wave climate, the
final input that is needed for evaluating the vessel
behavior is the spectral representation of the waves.
Involuntary speed reduction
Voluntary speed reduction
Significant advances have been made over the

past few decades in both quality and quantity of
satellite-generated wave data as well as in the oceanwave prediction models. Presently up-to 3rd
generation wave-prediction models are available
producing high-quality wave-specification of
significant wave height, typical representative time
periods (zero-upcrossing or peak period) and the
dominant wave direction when driven by high quality
wind fields. These models can also produce the wave
energy spectrum, both long crested and directional.
Involuntary speed reduction is due to the increased

resistance in a seaway while voluntary speed
reduction is the deliberate reduction in speed by the
ships captain in order to ensure that the ships waveinduced responses are within acceptable safety limits,
since it is found that in general a reduction in speed
(and also heading to some extent) improves seakeeping (i.e. reduces motions). For example, there
may be a maximum limit to the bow slamming, or a
maximum acceptable roll angle etc.
To determine ship behavior in a realistic wave

field, the usual procedure is to use a theoretical wave
spectrum. The mathematical formulations of these
normalized uni-directional (i.e. long-crested or twodimensional) wave energy spectra are based on one or
two parameters: the significant wave height, and/or
average wave periods.
2.3.1 The Track-optimization algorithm

Although many optimization methods have been
developed for the ship routing problem, a survey of
these works shows that most of these methods are
somewhat limited by their inability to handle the part
associated with involuntary speed reduction. The
optimization algorithm to be used should be general
enough so that it can handle a variety of multiple
objective criteria. In general, the optimal-path
problem is a problem of interest in many fields of
study, e.g. traffic engineering. A literature search on
general path-optimization algorithm reveals that one
of the available and easy to implement optimization
technique in this context is the Dijkstras algorithm,
which tries to minimize the distance between any two
node points in a given mesh/grid. The distance can
also be replaced by any `weight function. In the case
of the ship routing problem, the weights can be
viewed as an òbjective or àchievement function,
which can be obtained by combining weather
information along with sea-keeping characteristics of
the hull. It was thus felt that this algorithm could be
successfully applied for the present ship-routing
problem.
In the present work, as is usual for ship

applications, we apply the ITTC spectrum based on
the 3rd generation WAM generated data on significant
wave height and representative periods. This means
over the applicable ocean region WAM produces the
mentioned wave parameters which are then used with
the relevant theoretical ITTC formula to generate the
wave energy spectrum.
2.2 Ship Response in Waves and Added
Resistance
The method to determine ship behaviour in irregular
waves defined by means of a wave spectrum, using
the response parameters in regular waves or socalled RAO (Response Amplitude Operator) is well
known and is available in any standard text on naval
architecture (e.g. Lewis [11]). What is most
important here is the determination of the relevant
RAOs. In this respect, as far as the fundamental
six-modes of motions are concerned, we use the
well-known Salvesen-Faltinsen-Tuck version of the
strip theory for these computations (Salvesen et al.
[17]). As for added resistance, which is a second
order force, there are several formulae for its
determination in terms of the heave and pitch forces
response parameters, see e.g. Bhattacharya [1],
Faltinsen [5]. Here we use the formula of Gerritsma
and Buekelman [6].
Dijkstra's algorithm finds the shortest path from a

point in a graph to a destination. It finds routes by
cost precedence. The algorithm begins at a specific
node and extends outwards. It can be used to calculate
the shortest path between any two vertices in a
weighted graph, where each edge has non-negative
edge weight. Although most applications of shortest
path involve graphs with positive edge weights,
modified algorithms are also available where such a
constraint on weights is not necessary. For our case,
the weight will never be negative (it can at best be
very small), and therefore the original form of
Dijkstras algorithm appears adequate.
22
Based on this algorithm, a code has been

developed, where in a given mesh (grid) with
associated weights, if a start and end (destination)
nodes are defined, then the algorithm finds the
òptimal path joining these two nodes through grid
points such that the total weight is minimized.
setting V is known. Therefore the objective will be

to determine VR given its calm water speed V at a
given engine setting (i.e. revolution per unit time
N ).
2.3.2 Determination of the `Weights

In order to apply Dijkstras algorithm for the ship
routing problem, an area of the sea-surface
encompassing the possible route of the ship needs to
be discretized by means of a grid formed by latitude
and longitude lines. The nodes or vertices can be
taken as a central point in the grid or alternatively the
intersection of each latitude and longitude line can be
considered as a node. Typically the weather
information
(significant
wave
height
and
characteristic wave period) will be available from
satellite generated data and advance wave modeling
methods like WAM at each grid. As in WAM output,
these values are assumed constant over the
corresponding grid.
Figure 1. Illustration for self-propulsion point

The thrust T produced by a propeller depends on its
propulsion characteristics and is usually represented
by means of
curves, where
KT K Q J
The next and most important task is now to

determine an appropriate weight function for each
grid or node by combining the wave conditions
(significant wave height and direction) with ship
response parameters. Once these weights are known
at each node/grid, the problem is now transformed to
a state where Dijkstras algorithm can be applied to
find the optimal path for minimum (or maximum)
weight.
KT = T / (r N 2 D 4 )
is
the
thrust
coefficient,
KQ = Q / (r N D ) is the torque coefficient, and

2
J = Va / ( ND ) is the advance coefficient. Here Q is

torque, D is propeller diameter and Va is the
velocity of advance, which is the velocity of water
past the propeller. Va is related to ship speed by the
relation Va = (1 w)V where w is the wake fraction
for the hull.
The weights wi , j between the path lines joining
For a given ship in operation at a given engine

setting, the parameters N , D, t , w are fixed. Thus J
is a direct function of V . Therefore effective thrust
T (1 t ) = KT r N 2 D 4 (1 t ) as a function of V can be
plotted if the propulsion characteristics, i.e.
KT K Q J is available. The ship speed will then be
adjacent nodes i, j will depend on the parameter that

need to be optimized. There can be several possible
parameters such as minimum travel time, minimum
fuel consumption, safe and comfortable travel etc. In
the present work, we will consider the optimum ship
routing for minimum travel time and discuss the
methods to determine the weight function based on
involuntary speed reduction. For this we first need to
determine the reduction in speed due to the added
drag.
given by the intersection of the effective thrust line

and the resistance line, as shown in figure 1.
This procedure is the way to determine the speed
given the total resistance. However, for the present
purpose of ship routing application, it has the
following difficulties. It needs to be recognized that
the speed has to be determined between each adjacent
grids for all possible path lines connecting the grids.
This will a very large number, and this number will
increase rapidly if the navigational domain is large
(i.e. the possible area of the sea within which the
route is expected to lie, e.g. for large ocean
crossings). The other issue is that for each path line,
the value of RT will be different depending on the
prevailing environmental conditions even if RSW is
constant. Thus for each possible path line, one has to
The speed of the self propelled ship in calm water

is given by the so-called self propulsion point. This is
the speed at which the thrust T , after taking into
account the thrust deduction factor t , i.e.
T (1 t ) equals the calm or still water resistance RSW .
The speed at which the ship will travel for the total
resistance RT including the additional resistance from
all other environmental sources (wave, wind, current)
will similarly be at the speed where RT equates
T (1 t ) . This is schematically shown in figure 1
where V and VR are the calm water and reduced

speed. For a ship in operation, for the given engine
23
determine the curves (1 t )T and RT against speed,

and then determine the intersection of these two lines
to find VR . This procedure, besides needing the
propulsion characteristics in terms of KT KQ J
and therefore we can take it as constant as far as

determination of reduced speed VR is concerned. We
note that the procedure to determine the total added
resistance are all based on approximate theoretical
and/or semi-empirical formulations, and therefore
there is quite some uncertainty in these values.
Additionally, there will also be some uncertainty in
the forecast weather information like wave data and
wind data. Further, it also needs to be noted that
prevailing weather conditions are all assumed
constant over a grid, which itself is an averaging
process introducing some inaccuracy. It also needs to
be noted that the error introduced in assuming thrust
to remain constant will have similar order of error for
all possible paths, and therefore the determined
optimal path will still remain optimal or near optimal.
Thus determination of the reduced speed based on the
assumption that effective thrust is constant over the
small range of speed around the prevailing calm water
speed may be acceptable for practical calculations.
curves, is therefore extremely time and computer

intensive, and may not be feasible for the present
application. We therefore need to consider some
simplified and approximate methods to determine wij ,
which is fast and yet gives the weights within
practically acceptable limits of accuracy.
The reduction in speed due to environmental
factor is expected to be not very large, at least for
most part of the voyage. As a result, it may be
reasonable to assume that over the applicable speed
range (i.e. over the range VR to V ), the propulsion
characteristics remain same. This means, the effective
power PE is assumed constant. In calm water, PE is
given by RSW (V ) V , but in the presence of waves,
winds and current, the effective power will be given
by RT (VR ) VR = ( RSW (VR ) + Radd (VR ))VR . Here Radd is
the resistance due to winds, waves and current. This is
illustrated in figure 2. In order to determine the
reduced speed from this, we note that the calm water
resistance can be expressed in the form aV 2 where
a = 0.5 SCSW . As regards the additional drag, if an
assumption is made that the power required for this
component at speed VR is same as the power required
at speed V , i.e. AB = CD in figure 2, then we get,
aVR 2VR + Radd (V )V = aV 2V
Based on the above assumption, the reduced speed

will be as depicted in figure 3. If we write
Radd (V ) / RSW (V ) = k
(3)
then
RT (V ) = RSW (V ) + Radd (V ) = RSW (V ) (1 + k )
(4)
(1)
From above, we can get the reduced speed as:

1/3
R (V )V
VR = V 3 add
(2)
Figure 3. Procedure to get reduced speed assuming

constant thrust
If a database is created by computing RT (V )
against V for different values of k , then for any
given k , one can determine the corresponding speed
V . Hence by a process of interpolation, for a given
additional resistance value, one can determine the
reduced speed VR , as shown in figure 3. This
procedure is however based on expressing the
additional resistance as k RSW . This means, the
additional resistance due to environmental factors are
all a function of calm water resistance regardless of
speed, which means it has the implicit assumption that
the speed dependence of Radd is same as that of RSW .
Figure 2. Procedure to obtain reduced speed for

constant effective power
To make further simplification, we now made an
ad-hoc assumption that the effective thrust, i.e.
(1 t )T does not vary much over the range VR to V ,
24
Instead of this, it may be more reasonable to assume

that Radd (VR ) = Radd (V ) , as illustrated in figure 2. In
Figure 4 show the route between Damman and

Mumbai, and the return route of Mumbai-Damman
for ship 1, while for the same routes for ship 2,
results are shown in figure 5. From these results and
from comparison of the routes between the two
ships (not shown here for brevity), it is clear that the
routes depend on the ship type, and also on their
direction with respect to the waves.
such a case, if we have RSW expressed as aV , we

can directly get the reduced speed as:
1/ 2
R (V )
VR = V 2 add
a
(5)
Both the approximate procedures given by (2) and

(5) will require determination of the additional
resistance Radd arising from wind, waves and currents
for each possible path line between adjacent grids
using the prevailing environmental conditions at that
location, but for the given calm water ship speed V .
Once the reduced speed for each possible path line
joining nodes i and j is determined, i.e. VRi , j is
found, it is straightforward to determine the weight
functions as the time ti , j taken to cover the distance
between the nodes i and j given by Li , j :
wi , j ti , j =
Li , j
VRi , j
Figure 4. Route between Dammam - Mumbai and

Mumbai - Dammam by Ship1
(6)
Thus, by optimizing (minimizing) wi , j for the

complete path, the minimum time travel route for the
ship can be found. Here the route optimization due
only for involuntary speed reduction, that is, due to
additional drag arising from environmental factors
have been considered. It may be noted that in case of
calm water, w(i , j ) reduces to t0 (i, j ) = Li , j / V , which
is the travel time taken at constant calm water
resistance, and the minimum-time route becomes the
shortest route.
In the present work, both (2) and (5) are used, and
it is found that the difference in the predicted path
from these two is marginal. Therefore finally the
approximate formula (5) is retained, as it is found to
be the most convenient and easy to apply.
Figure 5. Route between Dammam - Mumbai and

Mumbai - Dammam by Ship2
In order to demonstrate the ability of the
algorithm to generate route circumnavigating a land
mass, in figure 6 we show the route from Mumbai on
the west coast of India to Visakhapatmen on the east
coast of India. The result is for ship 1. The algorithm
here simply assigns very large wave heights to those
grids which are land masses. As can be seen, the
algorithm could produce a route going around the
land mass. However, there is a problem with this
track: this route goes through the Palk Strait, the
water between India and Sri Lanka. This Strait is not
open for navigation, and therefore ships are not
allowed to go through it. Such constraints are also
easily handled by the algorithm, by simply `blocking
selected areas of ocean. This is achieved by simply
taking such areas as part of land mass. Thus isolated
areas on open ocean which are not open to navigation
can be treated as islands by simply assigning very
A large number of results have been previously

published for demonstrating the working of the
algorithm, and its correctness in achieving the
minimal time route (e.g. Padhy et al. [14] [15]), and
thus these are not repeated here. Here we apply the
algorithm to determine minimum-time optimal route
for two ships termed Ship1 and Ship 2 for routes
lying in the North Indian Ocean region, for which
realistic wave data are generated using the 3rd
generation WAM. The wave model is run using the
NCMRWF (National Centre for Medium Range
Wave Forecast, at Delhi) wind field data. Ship 1 is a
relatively small vessel of length 60m, breadth 11m
and draft 2.9m, while ship 2 is of length 160.93m,
breadth 23.1m and draft 9.07m.
25
large wave heights associated with those grids (see

figure 9 below)
One problem for this algorithm is that the route

produces often shows a zig-zag saw-tooth type
nature lacking smoothness, e.g. see figure 6. This is
a result of generating the route by adding the nodes
at the centre of the grids by straight lines. A
`smoothing-scheme is therefore devised to handle
this problem, and the results of applying this scheme
is shown in figure 9. It can be seen that the
generated track is now smooth.
Figure 7 and 8 show the result from the algorithm

where the monthly average wave-data is used.
Although the algorithm is run for all 12 months, here
for brevity we show results for only two months,
January and July, the former being a winter month
with relatively calm sea and the latter being a
monsoon month with high sea conditions. Note that
here Palk Strait is treated as navigable water. As can
be clearly seen, the routes generated are quite
different. Such studies are therefore for great use for
planning routes based on time of the year of the
voyage.
The algorithm, as developed is also capable of

handling presence of current and wind fields: these
can be considered by simply taking the additional
drags induced by these effects. Formulations are
available for estimating current drag and wind
loads. Figure 10 shows a minimum-time route in a
wave and current field. In a similar way, the effect
of wind drag can also be considered.
Figure 6. Route between Mumbai-Visakhapatnam
Figure 9. Smoothed track between Cochin and

Port Blair
Figure 7. Route between Adan - Chennai based on

WAM climatological wave data for January
Figure 10. Optimum Cochin - Port Blair route with

waves and current
4. CONCLUDING REMARKS
In this work, an algorithm has been developed for

ship weather routing application considering the
prevailing weather (wind-wave-current) conditions
and ship behavior in waves. The specific application
in this work has been confined to routes lying in the
Figure 8. Route between Adan - Chennai based on

WAM climatological wave data for July
26
North Indian Ocean region, primarily because of the

availability of information from ISRO satellites and
from sites such as NCMRWF of India providing
ocean wave related data over this region. The
optimization procedure is based on a variant of
Dijsktras algorithm, which has been suitably
modified for application to the ship-routing problem.
The algorithm is general enough to consider a variety
of optimization parameters since it is based on a
concept of a weight function associated with the path
that needs to be minimized/maximized. Therefore as
long as such a weight function can be determined
combining weather conditions and ship performances,
the routing algorithm can in principle work. In this
work, the optimization criterion is chosen to be the
time of travel, so that the obtained path represents a
minimum-time-travel route. For this, the procedure to
determine the appropriate weight functions has been
explained. It is found that the algorithm is versatile
and robust enough to handle all constraints that are
usually present in practical application of ship routing
by ship operators.
[9] Haltiner, G.J., Hamilton, H.D. and Arnason, G.

`Minimal time ship routing, Jour. of Applied
Meteorology, Vol. 1, pp. 1-7 (1962).
[10] Hanssen, G.L. and James, R.W. Òptimum ship
routing, Journal of Inst. of Navigation, Vol. 10,
No. 3, pp. 253-272 (1960).
[11] Lewis, E.V., Principles of Naval Architecture,
Vol. 3, SNAME Publications (1989).
[12] Mitchell, J. S. B. and Papadimitriou, C. H. The
weighted region problem. Technical Report,
Dept. of O.R., Stanford University, Stanford,
Calif. (1986).
[13] Montes., A.A., Network Shortest Path
Application for Optimum Track Ship Routing,
Naval Postgraduate School Monterey CA, MS
Thesis (2005).
[14] Padhy, C.P., Sen, D., Bhaskaran, P.K.
Application of wave model for weather routing
of ships in the North Indian Ocean, Natural
Hazards, Vol. 44(3), pp. 373-385 (2008).
[15] Padhy, C.P. and Sen, `Development of a ship
weather routing algorithm and its application
for Indian coastal sea routes, International
Journal of Ecology and Development. Vol. 12,
W09, pp.59 (2009).
REFERENCES
[1] Bhattacharyya, R., Dynamics of

Vehicles, Wiley Interscience (1978).
Marine
[16] Perakis, A. and Papadakis, N. Minimal time

vessel routing in a time-dependent environment.
Transportation Science, Vol.23, pp. 266-276
(1989).
[2] Chen H. A stochastic dynamic program for

minimum cost ship routing. Ph.D thesis,
Department
of
Ocean
Engineering,
Massachusetts Institute of Technology (1978)
[17] Salvsen, N., Tuck, E.O., and Faltinsen, O. Ship

motions and sea loads. SNME Transactions,
Vol. 78, pp. 250-287 (1970).
[3] Chen, H. and Lacey, P., Ùse of Operation

Support Information Technology to Increase
Ship Safety and Efficiency, Transactions of the
Society of Naval Architects and Marine
Engineers Vol. 106, 105-127 (1998).
[18] Ulusoy, T., Stochastic Optimal Control of

Seakeeping and Routing of Ships, Department of
Ocean Engineering, MIT, Doctoral Thesis
Proposal (2004).
[4] Faulkner, F.D., Numerical methods for

determining optimum ship routes. Jour. of Inst.
of Navigation, Vol. 10, No. 4, pp. 351-367
(1964).
[19] Zoppoli, R. `Minimum-time routing as an Nstage decision process Journal of Applied

Meteorology, Vol. 11, pp. 429-435 (1972).
[5] Faltinsen, O.M. Seal loads on ship and offshore

structures, Cambridge University Press (1990).
[6] Gerritsma, J. and Beukelman, Ànalysis of the
resistance increase in waves of a fast cargo
ship, International Shipbuilding Progress, Vol.
19, No. 217, pp. 285-293 (1972).
[7] Hagiwara, À study on the minimum fuel
consumption route-II: Simulation in the North
Pacific Ocean, Journal of Japan Institute of
Navigation, Vol. 72, pp. 87-96 (1985).
[8] Hagiwara, H. Weather routing of (sail-assisted)
motor vessels. Ph.D. thesis. University of Delft,
Netherlands (1989).
27

COMPUTATIONAL STUDY ON THE FISH-LIKE UNDERWATER ROBOT WITH

TWO UNDULATING SIDE FINS FOR VARIOUS ASPECT RATIOS, FIN ANGLES
AND FREQUENCIES
Md. Mahbubar Rahman 1, Yasuyuki Toda2 and Hiroshi Miki 3
Department of Naval Architecture and Ocean Engineering, Graduate School of Engineering,
Osaka University, 2-1 Yamadaoka, Suita City, Osaka 565-0871, Japan.
1
rahman@naoe.eng.osaka-u.ac.jp, 2toda@naoe.eng.osaka-u.ac.jp, 3Hiroshi_Miki@naoe.eng.osaka-u.ac.jp
ABSTRACT
The undulating fin propulsion system has been studied by many researchers including authors group as a
kind of bio-inspired propulsion system. In this study, the swimming motion of a Fish-like body with two
undulating side fins similar to a Stingray and a Cuttlefish was investigated through flow computation around
the body. In CFD technique, the Finite Analytic Method for space discretization and Euler implicit scheme
for time discretization were used along with the PISO algorithm for velocity pressure coupling. A body-fitted
moving grid was generated numerically using the Poisson equation at each time step. The features of the flow
field and hydrodynamic forces acting on the body and fin were discussed based on the computed results. A
simple relationship among the fins principle dimensions was established. To investigate this relationship, the
numerical computation was conducted for various aspect ratios, fin angles and frequencies. The relationship
was examined base on the distribution of pressure difference between upper and lower surface and the thrust
force distribution. Finally, the computed fin open characteristics was compared with the experimental results
of the authors previous studies.
Keywords: Biomimetics, Fish robot, Undulating side fins, Hydrodynamics, Thrust force, Computational
Fluid Dynamics (CFD).
finned robots (Sfakiotakis et al., 1999 and Kato et al.,
2004). Unfortunately, till date, the development of
efficient and environmental friendly underwater
vehicle with undulating side fins is not up to the mark
and it offers adequate opportunity to contribute to this
field.
1. INTRODUCTION
Natural selection is often believed to be the best
possible selection because each species has its own
unique and optimum way of interacting with its
environment obtained through the evolutionary
natural selection process that favored the best.
Robotic engineers borrow the sense and structure
from the animals of the nature to optimize their
designs. This study was based on a fish-like robot that
used the undulating side fin for propulsion. In the
field of underwater robotic research, undulating-fin
robot offers exceptional advantage over propeller in
preserving an undisturbed condition of its
surroundings for data acquisition. Though the
movement of this biomimetic type of robot is slow, it
has interesting areas of applications such as in covert
operations, in experimentations where minimally
disturbed surrounding is of prime importance.
In our laboratory, a fish like underwater robot

with undulating side fins similar to Stingray and
Cuttlefish has been studied for many years (Figure 1).
The aim of the project was to develop an efficient and
environment friendly underwater vehicle. To begin
with, the computation of the flow field around the
simple model with undulating fins and resistance
body was presented (Toda et al., 2002A). The feature
of flow field and hydrodynamic forces acting on the
body and fins was elucidated based on the computed
results to augment the understanding of the complex
fluid mechanics that fishes use to propel themselves.
Preliminary experiments reported Model 1 with a strut
operated at very low frequency (Toda et al., 2002B).
Next the experimental results using the second model
(Model 2) were presented (Toda et al., 2004). It was
followed by report using Model 2
Since the demand for high-efficiency underwater

vehicle is increasing, many researchers are interested
in studies in this field. Many new concepts of
biologically inspired underwater propulsion system
have been developed including several undulating-
29
simulated to save the computational effort. The width

of the resistance body and the fin were denoted by b
and bm respectively. Also u, v and w were the velocity
components for x, y and z directions respectively.
Computation for the similar fin geometries used in the
experiment has been conducted for comparisons.
However, the resistance body was considered as flat
plate and the sides were considered as straight for
simplicity in the grid generation. At first, the
numerical grid was constructed around a flat plate.
Then two fins were produced at the lateral sides of the
flat plate.
Figure 1. Photo of (a) Cuttlefish (b) Fish-like robot
consisted of 16 servomotors for both sides. The strut
of model 1 and the big cables of Model 2 hindered
their free movement for which in Model 3 was
constructed without strut and big cable (Toda et al.,
2006). Model 3 had 17 servo motors for both sides to
produce any fins motion with the servo controller
and the microcomputer being housed inside the
model. Though Model 3 demonstrated the motion in
the 6 degrees of freedom direction with greater
control it had some limitations in adjustment of
vertical gravity center and in communicating over a
large distance. In order to overcome these
discrepancies, Model 4 was constructed by using
wireless LAN communication device and the gravity
center adjustment system in three directions inside the
model. The fin width was kept the same as the Model
3 but the fin length was increased a little resulting in a
change in aspect ratio. Recently the preliminary
experimental results demonstrated the ability of the
new model (Toda et al. 2009). The simple prediction
for different aspect ratios of fin was also shown and
the new model shows improved performance than
before.
Figure 2. Simple model sketch and coordinate

system
All variables were non-dimentionalized using
resistance body length (L), advanced speed (U),
density () and their combinations. The time history
of fin angle in x-direction
was expressed in
Eq. 1~2.
= m sin(2 nt 2 x)
(1)
m = arcsin 1 0.905 ( x 0.5 ) sin
(2)
where,
: The deflection angle from flat plate (0 < x < 1)
The experimental and numerical studies are

complementary to one another. Experiments can give
the total force characteristics but for the detail
analysis of force the numerical study is essential. So
a numerical study was also conducted to recognize the
flow physics around the undulating fin and to
investigate the mechanism of thrust generation
(Rahman et al., 2010A). But that study was limited to
one fin angle and few numbers of frequencies; the
present paper discussed the similar phenomenon but
for two fin angles and large number of frequencies.
The objective of this study was to establish a simple
relationship among the undulating side fins principal
particulars and hydrodynamic forces.
: The angle amplitude (0 < x < 1)

n : Frequency (non-dimensional)
t
: Time (non-dimensional)
x : Distance along fin from leading edge; and

: Maximum fin angle from the flat position
A moving grid was constructed around the body
to handle unsteady motion. A numerical grid was
generated around half of the body at each time step
using Poisson equation. At the first period the flat
plate became fin by gradually increase its amplitude
and other periods were symmetric. The grid
independency was confirmed by checking different
number of grids. The enlarged view near the fin of the
grid is shown in Fig. 3. The body was covered by 41
and 16 grid points in x and y direction respectively.
2. METHODOLOGY
The Model sketch and coordinate system is shown
in Fig. 2. The body is symmetric with respect to xaxis, so only half of the body (shaded region) was
30
Minimum grid spacing was 0.0015 non-dimensional

lengths in the y and z direction.
progressive wave based swimming mode. In this

mode, to propel forward, the progressive wave
Figure 4. Frictional force for a flat plate in

acceleration stage and constant velocity stage.
must be faster than the uniform flow. This is one
kind of lift-based swimming mode, where no
recovery stroke is necessary because lift force can
be generated during the upstroke and downstroke of
the fin.
Figure 3. Computational grid near the fin

The computational domain and the number of grids
for half domain are shown in Table 1.
Table 1. Computational domain and grid points
for half domain
Axis
Computational
domain
Number of grid
points
-1 ~ 4
90
0~3
40
-3 ~ 3
51
The method of computation and boundary

conditions were briefly discussed by Rahman
(2010B). The Navier-Stokes equations and continuity
equations were discretized by the 12- point FiniteAnalytic method in space. Euler implicit scheme
were used for time discretization along with the PISO
algorithm for velocity pressure coupling. The laminar
flow computation was carried out at Re = 10,000.
Figure 5. Time history of x-directional forces for

AR = 0.1, n = 4 and = 300
In this fashion, x, y and z direction forces were
produced during the movement of the undulating
side fins (Fig. 6). The z-direction force is large but
the body doesnt go up or down because of its flat
shape; the mean of z-component force is zero. The
y-component is also generally cancelled out due to
the symmetrical movements of two side fins. So,
only the x-direction force acts on the body. In this
figure, it is also observed that, the x-component
3.1 Forces
For validation of the code, the computation was
first conducted on the flat plate having similar area of
the model. In the simulation, the velocity of the flow
was gradually accelerated from 0 to 1 and then kept
at steady state. The computational frictional force of
flat plate agreed well with the well-known Blasius
solution (B/L 0.01328) in the steady state region
(Figure 4).
The core aim of the thesis was to analyze the
hydrodynamic forces produced by different fins. So
the computation was conducted at different aspect
ratios and fin angles for different frequencies.
Sufficient convergence was ensured by taking
adequate iteration in all the cases. The time history
of x-directional forces at aspect ratio (bm) of 0.1,
frequency (n) of 4 and maximum fin angle of 300 was
drawn (Fig. 5). The robot we used mimicked the flat
fishes with undulating side fins which usually use the
Figure 6. x, y and z direction pressure forces

produced by undulating side fin.
force shows two peaks in one period which
indicated that the fin produced force in both
31
upstroke and downstroke movements. The total

forces for one period of aspect ratio 0.1 at different
frequencies and maximum fin angles are shown in
Fig. 7. From this figure, it is mentionable that for
maximum fin angle of = 300 and n = 3 the total
force is very near to zero. This is in agreement with
the experiment as the self propulsion frequency of
the model was similar to this value.
The contour of x-directional velocity distribution on

a flat plate, is shown in Figure 8. The figure shows
the boundary layer very clearly around the flat plate.
It also shows the x-directional velocity distribution
around the undulating side fins at different time steps
in one period. From the figure, the velocity gradients
at different positions of the fin are clearly visible.
Z
P: -0.5 -0.3 -0.1 0.1 0.3 0.5
X
Y
Figure 7. Time history (one period) of the xdirectional total force for AR = 0.1 at different
frequencies and maximum fin angles.
Figure 9. Pressure distribution on the upper and

lower surfaces of the fin at y = 0.275
When the fin pushed the surrounding water during
its movement, the velocity increased in that region.
The fins undulations reduced the boundary layer,
resulting in increased velocity gradient and, hence,
shear stress. The pressure difference between upper
and lower surface during the fins undulations is
shown in Fig.9. The prediction of pressure difference
between two surfaces is important because it is the
main reason behind the thrust force generation.
The amount of thrust and the propulsive

efficiency depend mostly upon the aspect ratio and
fin angle. High aspect ratio fins are more efficient
as they induce less drag per unit of lift or thrust
produced.
3.2 Contours
Different contours were drawn to visualize the
surface distribution of forces and velocities.
Figure 10. The method of thrust calculation

The thrust was computed by the product of
pressure difference between upper and lower surface
with the normal vectors to x (Fig. 10). The pressure
difference distribution and the thrust distribution on
the body and the fin at four time steps of one period is
shown in Fig. 11. Two important phenomena should
be discussed using this figure. Firstly, both for the
negative and the positive pressure differences, the xdirection forces have maximum value. This confirmed
the previous discussion of Fig. 6, which showed that
the fin produces thrust during both of its upstroke and
downstroke movements. Secondly, there is also
Figure 8. u-velocity distribution on a flat plate and

around a fin (AR = 0.1) for n = 4 at four time steps
in one period
32
pressure difference on the body part, but the body

cannot produce thrust force because the normal
vectors are zero here.
and 0.1 in Figure 12 and Figure 13 respectively. Each

graph apposed in the expected tendency with the
change of aspect ratio and fin angle. It is also found
that the thrust produced by fin at maximum angle of
450 is nearly twice compared to that of 300.
Figure 11. (a) Distribution of pressure difference

(p upper p lower); (b) Distribution of x-directional
thrust force on the full body of AR = 0.1, n = 4 and
= 300 at four time steps in one period
Figure 13. Fin open characteristics, = 450

4.2 Verification
The relation among thrust coefficient, aspect ratio
and fin angle was derived as:
4. FIN OPEN CHARACTERISTICS

4.1 Computed Result
b U
Kx = 2 m 1 01 01 4 2Cpb 2 cos 2 (2 k ( s ct )) dsdb
(5)
L n
The hydrodynamic force produced by the

undulating side fins was computed to examine the
relation between the thrust coefficient (Kx) and
advance coefficient (J), similarly as in propeller chart.
The thrust force (Tx) was estimated by subtracting the
flat plate resistance from the total x-directional force
in the similar manner as the experiment.
From this equation is seen that the thrust

coefficient is proportional to the cube of aspect ratio
and square of maximum fin angle. These two
relationships were established by the computation
(Fig. 14). We know, (AR 0.1/AR 0.075)3 = 2.37; so if
we divide the AR 0.1 result by 2.37, it goes very near
to 0.075s result. In the similar manner, as ( 45/
30)2 = 2.25, the thrust generated by fin at maximum
fin angle of 450 is about 2.25 times of that at 300
irrespective of fins aspect ratio.
Figure 12. Fin open characteristics, = 300

The results for the force in axial direction (xforce) and advanced coefficient were computed in
following non-dimensional procedure.
T
T
U2
Kx =
=
= Tx J 2
n2 L4 U 2 L2 n 2 L2
J=
1
n
U
=
NL
Fig. 14: The relation between kx & AR and kx &
(3)
4.3 Comparison
The computed and calculated results were compared
with the experimental ones. The experimental results
for Model-3 and Model-4 with aspect ratios 0.136 and
0.09 are shown by blue and green dots, respectively
(Fig. 15). The violet and the red lines show the
computed results for AR 0.075 and 0.1. As the aspect
(4)
The fin open characteristics for maximum fin angle of

300 and 450 are shown for aspect ratios of 0.05, 0.075
33
ratios of computed and experimental models were

different, for more comparison the results for AR
similar to the experiments was calculated using the
between the relative velocity of fluid and the fins

surface were investigated based on the distribution of
pressure difference of upper and lower surfaces and
thrust force distribution. The computational model
exhibited a considerable agreement with the empirical
observations. A simple relationship among the fins
principle particulars and hydrodynamics was
established using the computed result. The
comparison between the force measurement and
integrated force using CFD result for simple model
showed that the CFD techniques can predict the
hydrodynamic force produced by the undulating fin.
REFERRENCES
[1] Kato, N., Ayers, J. and Morikawa, H., Biomechanism of Swimming and Flying, SpringerVerlag, Tokyo, Japan (2004).
Figure 15. Comparison among the computed,

calculated and experimental results, = 300
[2] Rahman, M. M., Toda, Y. and Miki, H., Study

on the Performance of the Undulating Side Fins
with various Aspect Ratios using Computed
Flow, Pressure Field and Hydrodynamic
Forces, Proc. of the 5th Asia-Pasific Workshop
on Marine Hydr., Osaka, Japan (2010A).
previous formula, which is shown by red and violet

dash lines. The calculated result shows good
agreement with the experimental result of
corresponding aspect ratios. A little disparity is seen
between the empirical and simulated results at first
sight. But, this is due to the dissimilarities of the
models.
[3] Rahman, M. M., Flow Computation Around

Fish-like Robot with Undulating Side Fins,
M.Sc Thesis, Graduate School of Eng. Osaka
University, Japan (2010B).
[4] Sfakiotakis, M., Lane, D.M. and Davies J.B.C.,
Review of Fish Swimming Modes for Aquatic
Locomotion, IEEE Journal of Oceanic
Engineering, vol. 24, No. 2 (1999).
[5] Toda, Y., Hieda, S. and Sugiguchi, T., Laminar
Flow Computation around a Plate with Two
Undulating Side Fins, Journal of Kansai Society
of Naval Architects, Japan, No.237 (2002A).
[6] Toda, Y., Fukui, K. and Sugiguchi, T.,
Fundamental Study on Propulsion of a Fish-like
Body with Two Undulating side Fins, Proc. of
The 1st Asia Pacific Workshop on Marine
Hydrodynamics, Kobe, Japan (2002B).
Fig. 16: Comparison among the computed,

calculated and experimental results, = 450
[7] Toda, Y., Suzuki, T., Uto, S. and Tanaka, N.

Fundamental Study of a Fishlike Body with
Two Undulating Side-Fins, Bio- Mech. of
Swimming and Flying, Springer, pp. 93-110
(2004).
The fin open characteristics for = 450 are also

shown in Fig. 16. This result shows better conformity
than the 300 results. This is because; the same area of
the fin can make more pressure force when the
maximum fin angle is increased. This is also
consistent with the previous discussion. Over all, it is
mentionable that, the computed result approached the
expected trend with the changes of the aspect ratio
and maximum fin angle.
[8] Toda, Y., Ikeda, H. and Sogihara, N., The

Motion of a Fish-Like Under-Water Vehicle with
two Undulating Side Fins, The Third Inter.
Symposium on Aero Aqua Biomechanisms ,
Ginowan, Okinawa, Japan (2006).
[9] Toda, Y., Danno, M., Sasajima, K and Miki, H.,
Model Experiments on the Squid-Like UnderWater Vehicle with two Undulating Side Fins,
The 4th International Symposium on Aero Aqua
Biomechanisms , Shanghai, China (2009).
5. CONCLUSION
The swimming motion of a fish-like body with two
undulating side fins was investigated through flow
computation around the body. The relationship
34

35

ASSESSMENT OF MARINE PROPELLER HYDRODYNAMIC PERFORMANCE IN

OPEN WATER VIA CFD
Mehdi Nakisa1, Mohammad Javad Abbasi2, and Ahmad Mobasher Amini3
1
Islamic Azad University
Boushehr Branch, Iran
mnakisa2003@yahoo.com
IGS
Mood582003@yahoo.com
3
Assistan Professor
Persian Gulf University, Iran
mobasheramini@pgu.ac.ir
ABSTRACT
In general, marine propeller blades have complicated geometries and as a consequence, the flow around
these propellers is complicated. Propeller tests in open water are commonly used to obtain the graphs of the
hydrodynamic performance. The aim of this paper is to approach the propeller hydrodynamic performance via
numerical modeling using a finite volume commercial code such as Fluent. Modeling of a general B-Series
propeller is based on RANS equations in steady flow with unstructured mesh. The effect of mesh density, type of
turbulent model and numerical solution algorithm on modeling are investigated. In order to get the open water
test performance coefficients for the considered propeller (KT, KQ, ) different advance coefficients (J) are
imposed as boundary condition for the numerical model. Finally, the results of the simulations are compared
with available experimental data for the selection of the best modeling methodology for open water tests of the
propeller.
Key words: RANS, CFD, marine propeller performance
1. INTRODUCTION
NOMENCLATURE
The efficiency of the propulsion system is strongly

dependent on propeller performance, thrust force,
torque of propeller and its efficiency. Therefore, the
simple method for investigation of assessment of
marine propeller hydrodynamic performance is used
to graphs of the propeller performance coefficient
(KT, Kq, ) with respect to advance coefficient (J).
Different numerical methods for the treatment of the
RANS (Reynolds Average Navier Stokes) equations
have been widely developed.
D: Diameter of Propeller, [m]

Z: Number of blades
AE
= Expanded area ratio
A0
P
= Pitch ratio
D
T
n 2 D 4
Q
Kq: Torque coefficient, K Q =
n 2 D 5
Kt: Thrust coefficient, K T =
Although predictions on the flow field are always

valuable, the numerical simulation of marine
propellers is difficult. These difficulties lie in the
usual CFD difficulties such as: turbulence modeling,
flow separation, boundary layer etc. Kodama (1992)
and Funeno (1997) used a RANS equation method
with structural mesh technique for simulation of the
current around a ships hull and propeller [3]. This
method has some difficulties such as: restriction of
results only in steady state and open water, also
because of blades geometry complication. They
couldnt predict the current near the tip of blades
accurately. Funeno (1999 and 2002) simulated the
current around a highly skewed propeller via
unstructured mesh [4]. The results had good
q: Torque at the propeller axis, [n.m]

Br: Boss ratio (br=b/d)
r: Rake angle, [deg.]
Va: Advance velocity, [m/s]
n: Rotational speed, [rpm]
: Density of the water,[kg/m2]
t: Thrust of the propeller
VA
nD
K J
: Propeller efficiency, = T
K Q 2
J: Advance coefficient, J=
35
Proceedings of MARTEC2010
correspondence experiment data for steady state and

unsteady flow, but this method was complicated and
time consuming. Also Martinez (2002) simulated the
propeller via k-epsilon turbulent method in open
water in steady state condition. Totally, the results
were acceptable but there was approximately 30%
error in prediction of torque coefficient [6]. Takekoshi
(2003) simulated the propeller via standard k-omega
turbulent method in open water and steady state
conditions. For simulation of propeller geometry and
its surrounding, he used the propeller symmetry and
simulated only one blade and compared the results of
experiments with his simulated result; 15 percent
error in the simulation results. [7]
2.1
BOUNDARY CONDITIONS AND
NUMERICAL SOLUTION CONTROL:
For simulation the propeller in steady state flow,
we divided the calculation zone in two cylindrical
divisions:
- Moving zone: length and diameter of this zone
are depended on diameter and boss of the propeller.
The aim of moving zone is simulating the movement
of propeller and boss rotation and applying the
Coriolis acceleration term in the Navier Stokes
equations.
Rhee simulated 2D propeller model in unsteady

state flow. He used the sliding mesh method for
displaying the propeller rotation and turbulent flow of
standard k-omega. Its error in comparison with
experimental method was approximately 13 percent
[8].
2. MODEL AND EXPERIMENTAL

MEASUREMENTS:
B-Wageningen series is one of the well-known
standard propeller series which is used in propulsion
system of merchant ships.
Fig.1: The scheme of propeller, Boss and numerical

domain
The range of P/D in these series of propellers is

0.6 to 1.4 with Four to Seven blades.
The water in our simulation is incompressible flow;

two different approaches have been tested, for the
inlet boundary condition:
- Inlet flow velocity
- Inlet total pressure
Four blades one are used commonly for merchant

ships. Therefore in our study a four blades BWageningen standard propeller simulated with the
following characteristics
The inlet flow velocity condition is selected,

because it is considered with more physical meaning.
model name: B-Wageningen

Z: 4
D: 0.275
Br: 0.1670
For modeling the inlet flow velocity, we used the

stationary zone and velocity of flow.
We used two options for simulation of the outlet
boundary condition:
- Outflow
- Pressure outlet
P
at 0.7R: 0.7
D
AE
: 0.4
A0
Though outlet flow condition, increases calculation

time and decreases convergence rate. We used it for
modeling of stationary zone, since the outlet flow
condition has more physical meaning than outlet
pressure condition. For simulation the boundary
condition around stationary zone, has been used the
slip boundary condition, in order to decrease the wall
influence on stream flow as shown in Fig.1.
R: 10 deg.
We have used the experimental results of
hydrodynamic coefficient performance in open water
test [5].
- Stationary zone: this zone is cylindrical and includes
the boss, propeller and moving zone. According to
Takekoshi [1], upstream length is taken as 2D,
downstream length is 5D and zone diameter is 3D, as
shown in Fig.1. where D is the diameter of the
propeller.
2.2 MESHING AND CALCULATION

ZONE:
A geometrical discretization of the propeller is
made for the treatment. Unstructured tetrahedral cells
36
are used to define a control volume (symmetric with

respect to the propeller axis).
1- First order upwind

2- Second order upwind
3- Power law
4- Third order QUICK method
A mesh refinement zone is defined near the

propeller surface in order to capture the phenomena
better at the propeller blades.
We have used third order QUICK method for

discretization of the momentum equation and second
order upwind method for discretization of the other
equations.
We must exert the momentum, conservation of

mass and energy equation on flow field. We can not
say that small mesh is good mesh because we must
consider the time and cost calculation.
2.5
INTERPOLATION METHOD OF
THE PRESSURE:
Final model has around 1,100,000 cells. Grids on

the propeller surface are triangle shape that of various
sizes because the cells near the root, blade edges and
tip of propeller are smaller than other parts.
There are several interpolation methods for pressure

in the segregation method as follows:
- Standard method
- Linear method
- Second order method
- Body force
- Presto method
The standard method is popular. So we have used
the standard method for interpolation of pressure.
Finally, all of the calculation zone and domain are

meshed by tetrahedral meshes as shown in Figs. 2, 3.
2.6
PRESSURE-VELOCITY COUPLING
METHOD:
There are three methods for pressure-velocity
coupling in the segregation method:
- SIMPLEC
- SIMPLE
- PISO
Fig.2: Final mesh for the calculation
The algorithm of PISO is applicable in transient

flow but SIMPLEC and SIMPLE algorithms are
applicable in steady state flow, therefore we have used
the SIMPLE algorithm for pressure-velocity coupling.
2.7
In FLUENT there are two options for increasing

the solution speed:
- Algebraic multi-grids (AMG)
- Full approximation storage (FAS)
We must select the AMG algorithm in FLUENT,
because we solved our equations by segregation
method.
Fig.3: Perspective view of the domain
2.3
LINEARIZATION METHOD:
2.8
INTRODUCTION OF PROPELLER
ROTATION:
There are two methods for linearization the equations:

-implicit method
-explicit method
For simulating the equations we used the segregated
and implicit method.
2.4
MULTI-GRIDS METHOD:
To simulate the rotational region around the

propeller in FLUENT, including the Coriolis
acceleration in the RANS equations, we have two
options:
- MRF (moving reference frame method)
- SLD (sliding mesh method)
The above methods were investigated with several
turbulence models for selecting the best method of
propeller simulation [2].
DISCRETIZATION METHOD:
There are several methods that depended on the type

and accuracy of problems for discretizing the
equations:
37
2.9
Using the both criteria together is suitable for

convergence test in CFD results.
Histories of convergence results the flow field
around the propeller has been shown in Fig 5.
TURBULENT SIMULATION:
There are several methods for turbulent flow

simulation and determination, e.g. the Reynolds stress
model. These models are divided to three divisions:
-zero equation model
-One equation model
-Two equation model
The zero equation models are solved by algebraic
equations. The one and two equation models are used
by one and two extra PDE, respectively. Selection of
the turbulence model depend on:
- Physics of flow
- Accuracy range
- Computation facilities (RAM, CPU, etc.)
- Duration time for solution
There are several methods for modeling the flow
around the propeller. These are as follows [10]:
- In-viscid Model
- Spalart - Allmaras Model
- Standard K-epsilon Model
- RNG K-epsilon Model
- Realizable K-epsilon Model
- Standard K-omega Model
- SST K-omega Model
2.10
FLOW SIMULATION
THE SOLID WALL:
Fig.4: distribution of Y+ on blade surface
AROUND
The solid walls impress on the turbulent flows,

severely. This is clear that the solid walls impressed
on average flow velocity without SLD mesh
condition, thus, simulation the flow near the solid wall
has two ways:
- Wall function
- Two equation models with low Reynolds Number
The main problem of wall functions is limitation of
using in various flows, but in high Reynolds No. flow
fields, we could use the wall functions because the
physical variable parameters near the wall, need not to
solve with CFD. Wall functions are commonly used in
industrial simulation.
Y+ is normal orientation on blade surface for
description of the flow condition that it is within the
range of 50 to 500 [Fig.4], often the flow has high
Reynolds No. and the viscid region of tip blades is
very narrow, therefore we simulated the flow behavior
near the tip blade by wall function.
2.11
CONVERGENCE TEST:
There is no law for general convergence test. We

can use two criteria convergence tests for accuracy the
tests:
-Mathematical criteria test: convergence of
residual conservation equations to certain quantity
(10-4).
-Physical criteria test: uniformly and fluctuation
of momentum and drag force. [9]
38
Spalart-Allmaras Model
with MRF
0.7
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
Num. Efficiency
0.4
Num. 10KQ
Num. KT
0.3
Exp. 10KQ
0.2
0.1
Exp. KT
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.7
0.8
Advance Coefficient (J)
Spalart-Allmaras Model
with SLD
0.7
Fig. 5: Convergence history for Va=1.650m/s,

n=900rpm with standard-epsilon turbulence
model
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
Num. Efficiency
0.4
Num. 10KQ
Num. KT
0.3
Exp. 10KQ
0.2
0.1
Exp. KT
3 PROPELLER PERFORMANCE GRAPHS

AND EXPERIMENTAL RESULTS:
0.0
0.1
Num. KT
Exp.10 KQ
Num. 10 KQ
0.6
with MRF
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
Num. Efficiency
Num. 10KQ
0.4
0.3
Num. KT
0.2
Exp. KT
0.1
Exp. 10KQ
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.7
0.8
0.7
0.8
Standard-k,epsilon Model
with SLD
0.7
0.7
Exp. KT
0.5
Standard-k,epsilon Model
with MRF
0.6
0.4
0.7
Inviscide Model
Num. Efficinecy
0.3
The aim of this paper to model the propeller

performance curves by CFD and compare them with
experimental results.
The rotational speed of propeller is constant
(n=900 rpm) and we change the inlet velocity of flow
between 0.4125 m/s and 3.3 m/s in order to reach the
advance coefficient between J=0.1 and J=0.8 .
The Computed values of thrust and torque
reached the measured values for each velocity. The
rotational flow fields around the propeller have been
simulated by MRF and SLD methods as shows in
Fig.6.
Exp. Efficinecy
0.2
0.6
Num. Efficiency
Exp. Efficiency
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
Num. Efficiency
0.5
Num. 10KQ
0.4
0.4
0.3
Num. 10KQ
Exp. 10KQ
0.3
Num. KT
0.2
Exp. 10KQ
Num. KT
0.2
Exp. KT
0.1
0.1
Exp. KT
0.0
0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.2
0.3
0.4
0.5
0.6
0.8
RNG-k,epsilon Model
Inviscide Model
with MRF
0.7
with SLD
0.7
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
0.6
Num. Efficiency
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
Num. Efficiency
0.5
0.5
0.4
Exp. Efficiency
0.4
Num. 10KQ
Num. KT
Num. 10KQ
Num. KT
0.3
0.3
Exp. 10KQ
Exp. 10KQ
0.2
0.2
0.1
Exp. KT
0.1
Exp. KT
0.0
0.0
0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.2
0.3
0.4
0.5
0.8
39
0.6
RNG-k,epsilon Model
SST-k,omega Model
with MRF
with SLD
0.7
0.7
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
0.5
Num. Efficiency
Num. 10KQ
0.4
0.4
Num. KT
Num. 10KQ
0.3
Num. Efficiency
0.3
Exp. 10KQ
Exp. 10KQ
0.2
0.2
Num. KT
0.1
0.1
Exp. KT
Exp. KT
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.1
0.8
0.2
0.3
0.4
0.6
0.7
0.8
SST-k,omega Model
Realizable-k,epsilon Model
with SLD
with MRF
0.7
0.7
0.6
0.5
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
Num. Efficiency
Num. Efficiency
0.5
0.5
0.4
0.4
Num. 10KQ
Num. 10KQ
Num. KT
0.3
0.3
Exp. 10KQ
0.2
0.2
Num. KT
0.1
0.1
Exp. 10KQ
Exp. KT
Exp. KT
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.1
0.2
0.3
with SLD
0.6
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
0.5
0.6
0.7
0.8
Fig.6: simulation with various turbulent flow

models by MRF and SLD methods
Realizable-k,epsilon Model
0.7
Exp. Efficinecy
0.4
Exp. Efficiency
The turbulent flow model of SST-k,omega with

SLD technique appear to yield the best numerical
solution of flow around the propeller.[Fig.7]
Num. Efficiency
0.5
0.4
Num. 10KQ
Num. KT
0.3
Exp. 10KQ
0.2
SST-k,omega Model
with SLD
0.1
0.7
Exp. KT
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.6
0.8
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
Num. Efficiency
0.5
0.4
Standard-k,omega Model
with MRF
0.3
0.7
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Num. 10KQ
0.2
Exp. Efficiency
Num. KT
0.1
0.5
Exp. 10KQ
Exp. KT
Num. Efficiency
Num. 10KQ
0.0
0.4
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8

0.3
Exp. 10KQ
Num. KT
Fig.7: Simulation of turbulent flow model by SLD

technique
0.2
Exp. KT
0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
3.1
INVESTIGATION OF VELOCITY
AND PRESSURE FIELD AROUND THE
PROPELLER:
Standard-k,omega Model
with SLD
The theoretical flow behavior around any

propeller blade is often calculated and explained using
the momentum theory [4]. According to this theory,
the pressure in the fluid will suffer a continuous
decrease as it approaches the propeller, due to the
suction generated. In parallel, with the pressure
decrease, the speed of the fluid will increase as it
comes closer to the propeller. This pressure decrease
is kept up to the propeller itself, where an abrupt
pressure increase happens due to the tangential shear
produced by the blades.
0.7
0.6
Exp. Efficinecy
Num. Efficinecy
Exp. KT
Num. KT
Exp.10 KQ
Num. 10 KQ
Exp. Efficiency
0.5
Num. Efficiency
0.4
Num. 10KQ
0.3
Exp. 10KQ
0.2
Num. KT
0.1
Exp. KT
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
40
Thus, the axial velocity component has relatively

small values in the vicinity of the blade tip and
propeller hub. Velocity deficit dominates the region
around the blade tip [11].
For verification of the numerical simulation the
pressure and velocity fields around the propeller have
been obtained. The pressure and velocity contours
around the propeller in J=0.1 and J=0.8, are shown in
Fig. 8
The flow simulation results are verified by
considering the contour results.
The propeller performances influence the flow
field around it decreases with increase in the advance
coefficient (J), also the propeller does not influence
the pressure and velocity profiles, behind and front of
propeller.
Contours of x velocity in J = 0.1
Contours of dynamic pressure in J = 0.1

Contours of total pressure on suction surface in
J = 0.1
Contours of total pressure on pressure surface in

J = 0.1
Contours of total pressure on suction surface in

J = 0.1
Contours of velocity magnitude on surface in
J = 0.1
41

Contours of total pressure on pressure
in J = 0.1
surface
Fig.8: contours of pressure and velocity
propeller
3.2
COMPARISON
EXPERIMENTAL AND
RESULTS:
around the
BETWEEN
THEORETICAL
For assessment of marine propeller hydrodynamic

performance, we have used the SST-k, omega
turbulence model with SLD technique so that its
numerical results are near the experimental results.
In error bar of numerical results, for thrust
coefficient force (KT) and torque coefficient (KQ),
average of values are around 8 and 13 percent,
respectively. Prediction results of error in KQ is
greater than that. [Figs.9, 10]
In Fig .11 we can see the prediction results of error
in propeller efficiency (), around 11 percent, and the
errors are decreased with increasing the advance
coefficient (J).
Contours of velocity magnitude on

surface in J = 0.1
Error Bar Diagram of Trust Coefficient

25
Error Bar
Percent of Error
20
Velocity vectors by x velocity in J = 0.8
15
10
0
Error Bar
2.01
5.63
7.56
9.54
4.97
11.24
21.91
Advance Coefficient
Fig. 9: Error bar for prediction of KT

Error Bar Diagram of Torque Coefficient
25
ErrorBar
Percent of Error
20
15
10
Contours of dynamic pressure in J = 0.8
ErrorBar
13.76
10.59
11.21
4.53
3.87
16.93
21.24
Advance Coefficient
Fig. 10: Error bar for prediction of (KQ)
42
Error Bar Diagram of Efficiency

18
Percent of Error
16
ErrorBar
The numerical method based on RANS equations

is useful and suitable method for analysis of the flow
behavior around the propeller .
14
12
10
8
These results are compatible to the experimental

results. The benefits of this method are:
6
4
2
0
ErrorBar
13.86
14.67
16.88
13.46
8.51
4.87
7.38
- Decreases time and cost
Advance Coefficient
- Explains the flow around the propeller with

pressure and velocity contours,
- No limitation in velocity flow and model size.
- Simulation with simple assumption
Fig. 11: Error bar for prediction of
3.3
THE WAYS OF APPROACHING TO
ACTUAL CONVERGENCE ANSWERS
ACCORDING TO PHYSICS OF FLOW:
ACKNOWLEDGMENTS
The authors gratefully acknowledge the Professor
Migeotte and Professor Rad Sharif University of
Technology, Islamic Azad University, Bushehr branch
for their technical suggestion.
We can approach to the suitable convergence

answers by considering:
- Boundary condition
- Mesh density and spatial distribution
REFERENCES
- Solution procedure
INSTABILITY SOURCES OF
SOLUTION:
3.4
Instability show itself, by nonphysical quantities

and negative k, data. Some instabilities show
themselves as not actual data of length scale and time
scale in turbulent flow.
[1]
Gunzburger,
M.D.,
Nicolaides,
R.A.,
Incompressible computational fluid dynamics.
Trends and advances, Cambridge University
Press, 1993
[2]
Soltani, M. and Rahimi, R. computational fluid

dynamic by FLUENT software, Tehran, 2003
[3] Uto, S. and Kodama, Y., Application of CFD to

the Flow Computation around a Marine
Propeller Grid Generation and Inviscid Flow
Computation using Euler Equations- , J. Kansai
Society of Naval Archtects, No. 218, pp. 171180, 1992
Causes of these instabilities are in k, equations;

therefore it must be accurate selection of turbulent
parameters.
3.5
CONCLUSIONS:
STABILIZING TECHNIQUES:
Two techniques can be used for stabilization and

decreasing the instability:
[4] Funeno, I. and et al., Analysis of Steady

Viscous Flow around a Highly Skewed Propeller
(in Japanese), J. Kansai Society of Naval
Architects, No. 231, pp. 1-6, 1999
- Up winding
- Clipping
[5] Funeno, I., On Viscous Flow around Marine

Propellers HubVortex and Scale Effect-,
Proceedings of New S-Tech 2002 (Third
Conference for New Ship and Marine
Technology), pp. 17-26, 2002
The aim of upwind technique is stabilization of

advection terms in high advection zones that using by:
streamline method
first order method
[6]
hybrid method
Clipping is the process that it prevents of
instabilities in the first and third groups of above
mentioned source errors. In this method, it is to
determine the positive minimum quantity for k, . We
can prevent decreasing of the minimum determined
quantity.
Martnez-Calle, Julin, AN Open Water

Numerical Model for a Marine Propeller: a
Coparsion with Exprimental Data, Proceedings
of the ASME FEDSM02 2002 Joint USEuropean
Fluids
Engineering
Summer
Conference July 14-18, 2002. Montreal, Canada
[7] Yoshihisa TAKEKOSHI, Simulation of Steady

and Unsteady Cavitation on a Marine Propeller
Using a RANS CFD Code, Fifth International
Symposium on Cavitation (CAV2003) Osaka,
Japan, November 1-4, 2003
43
[10] Shojaeefard, M.H., Noor poor hashtroodi, A.R,

Introduction of computational fluid dynamic,
Tehran, 2000
[8] Shin Hyung Rhee *, EvangelosKoutsavdis,

Two-dimensional simulation of unsteady
marine propulsor blade flow using dynamic
meshing techniques, Fluent Inc., 10 Cavendish
Ct., Lebanon, NH 03766, USA, 2004
[11] Paik, B.G. , Lee, C.M. , Lee, S.J. ,PIV analysis

of flow around a container ship model with a
rotating propeller, Experimental in fluids 36 ,
p833-846 , 2004.
[9] Freitas, C,J., Journal of fluids engineering

editorial policy statement on the control of
numerical accuracy, ASME J. Fluids Eng., 115,
pp. 339-340. 1993
44

SUSTAINABLE GROWTH IN INLAND WATERWAYS OF BANGLADESH

THROUGH INNOVATION: A WAY FORWARD!
Arun Kr Dev1
1
School of Marine Science and Technology, Newcastle University, United Kingdom

Newcastle University Marine International (NUMI) - Singapore
E-mail: a.k.dev@ncl.ac.uk
ABSTRACT
Bangladesh is covered by a network of about 24,000 km of rivers, canals, creeks and haors that together amount
about 7% of the country's surface. At present, the navigable river routes in Bangladesh during the monsoon
stretch about 6,000 km, although this is reduced to about 3,800 km in the dry season. What else can be so
important other than these precious inland waterways of Bangladesh? It is now time for a planned and
systematic future roadmap for these valued waterways in various diversified uses with new and innovative
ideas. The author highlights a few important milestones that Bangladesh Government, through its various
agencies expertise along with academics and marine industry, may take due interest and thus initiate a long
term planning and later execute them over a few stages. In this paper, an innovative idea of designing large
passenger vessels is presented. It clearly demonstrates that even with overcrowding of passengers, the new type
of vessels will have less possibility of capsize compared to existing monohull vessels. Finally, the author has
described many challenging ideas where innovation and new technology are required but all these will help
achieving them in low cost not only at the beginning but also throughout the life cycle management of the
developed products. It may not be possible to pick up all these new and innovative ideas at one time but over the
years, in a systematic panned manner, the goals can be achieved.
Keywords: inland waterways, navigable routes, safety against capsizing, landing station pontoons, floating
warehouses, inland cargo fleet, maritime education & training, carbon foot print, climate change, shipbuilding
materials, dredging, fuel efficient, carbon footprint
1. INTRODUCTION
effectiveness in this sector require closer

attention. There might be possibility of exporting
sands to countries, which do not have such
resources.
The author of this technical paper, who comes

from Bangladesh and spent his early carrier in
BIWTA, keeps him well informed about development
of maritime activities in Bangladesh including
accidents of passenger ferries happening almost every
year. The author highlights a few important
milestones that Bangladesh Government, through its
various agencies expertise along with academics and
marine industry, may take due interest and thus
initiate a long term planning and later execute them
over a few stages. The important features for growth
in inland waterways are as follows:
Navigable waterways to be classified in both dry

and monsoon based not only water depth but also
on the sea state conditions clearly indicating
suitability of type of vessels especially passenger
vessels plying at night.
Dredging of rivers and estuaries are already

ongoing efforts in Bangladesh. Efficiency and
45
Innovative design of passenger launches and

passenger vessels on an idea based on KISS
(keep it simple but not simpler) to be introduced.
Introduction of inland and coastal classification

of vessels under Bangladesh Government perhaps
under a proposed name Class BG may be
implemented soon having its branch offices in
major rivers and coastal ports in Bangladesh.
In addition to existing ports, many other new

river and coastal ports to be opened. Port
facilities like landing station pontoons for both
passenger and cargo vessels are to be introduced
again with innovative designs. Building them
with steel materials and maintaining them yearly
is a huge expenditure for any Government.
high side. Following every floods, large sand banks

accumulated in rivers needs to be taken care regularly
by systematic dredging activities.
Large floating warehouses can be constructed

with innovative designs and placed in major
ports.
Similar to passengers, business
entrepreneurs will be encouraged to use new type
of inland and coastal cargo vessels to transport
goods including refrigerated ones from ports to
ports resulting less pressure on road transport,
which only increases carbon footprints and even
more expensive. If necessary, Bangladesh will
have its own standard containers (geometry &
size) to suit its needs in terms of cost, handling,
etc.
Innovation has to be applied even in designing

self propelled cargo vessels, dumb barges, etc.
with alternative materials as steel is not available
in Bangladesh and even so, it is expensive.
There are even possibilities of renewable energy

in marine propulsion technology and also using
gas (CNG) in marine diesel engines. Such
initiatives will certainly reduce carbon footprints
and help climate change.
Different housing schemes and cyclone shelters

in the form of floaters may be made available to
people living in offshore islands where cyclones
have devastating effects in terms of tidal waves,
etc.
All mechanized country crafts or even various

types of wooden boats may be built from
different materials alternative to timber and steel;
both of which are in fact scarce in Bangladesh.
Usage of wood in building of marine mechanized
boats, launches and vessels should in fact be
banned to reduce logging.
2.3 Innovative Ship Design

Traditional ship design based on long time
experiences should now be re-assessed and
deficiencies to be pointed out. With introduction of
extensive IT based ship design methodologies, new
efficient hull to be introduced for fuel efficiency
including better manoeuvring capability. Monohull
passenger launches or passenger vessels where
overcrowding, collision accidents, etc. cause very
often loss of passengers lives should now be given a
new idea so that such accidents are reduced to
minimum. The author will later show an example how
stability can be increased with a different hull concept
and can save capsizing of passenger launches and
passenger vessels in Bangladesh.
2.4 Inland and Coastal Ports

More inland and coastal ports are to be constructed
and existing inland and costal ports to be redesigned
for safe and efficient loading and unloading of both
passenger and cargoes. Transportation of consumer
goods should now be used by inland and coastal
vessels reducing dependence on road transport.
Landing station pontoons of all kinds should be built
of concrete materials and NOT of steel any longer.
This will save huge revenue by Govt every year from
maintenance of steel plates. Inland and coastal
container ships may be introduced and all new and
existing ports should be supported with container
logistics. Bangladesh can even invent its own standard
containers and handling devices. Innovations are
required in this sector.
2. GROWTH IN INLAND WATER WAYS

2.1 Classification of Inland Water Ways
2.5 Floating Warehouses

The present classification of various rivers may
require further review in terms of water depth, and sea
state condition in monsoon and dry season. Such
information is necessary for designers and operators
for smooth operation of inland and coastal fleet of
Bangladesh.
Concept of large floating ware houses may easily

be introduced for transhipment from large inland and
coastal ports to smaller ports. These large floating
warehouses can also be built with concrete materials
instead of steel. Floating warehouses have an added
advantage of not being affected by rising flood waters,
which is a natural phenomenon in Bangladesh.
2.2 Dredging
To keep rivers and estuaries including large canals,
constant dredging is essential in Bangladesh every
year. Especially after monsoon, large sand banks
require to be dredged out. When liquid fuel powered
dredger is expensive OPEX, CNG powered diesel
engines may be used in all dredgers. Instead of always
using cutter suction dredgers, TSHDs (Trailer Suction
Hopper Dredger) may be introduced so that dredged
materials could be recycled to reclaim new lands or to
increase existing small islands (chars). But this
requires a master planning for a country like
Bangladesh where population density is on extremely
2.6 Shipbuilding Materials

Traditionally, now steel is being used for
shipbuilding mainly for self propelled barges,
passenger launches or vessels and also coastal vessels.
Time has now come that a new and cheap alternative
material and also in terms of less or zero maintenance
be introduced. All barge shaped floating structures
now can easily be built with concrete. Even there is a
possibility of building self propelled barges using
concrete materials. Smaller boats should start using
Ferro cement in place of steel and wood.
46
trades for both home and overseas demands can also

be met by giving simple certificate and advanced
certificate courses, which are not offered in local
polytechnics and vocational training institutes.
2.2 Renewable Energy

In the past, many country boats have used sails
instead of manual rowing. With the introduction of
smaller engines (main source from irrigation engines),
suddenly this tradition has gone away. Even these
boats are diesel powered; sail should be used as a
hybrid propulsion technique. Even large self propelled
vessels can use sail power when such favourable wind
is available. Owners and operators should be
encouraged by innovative ideas. All these efforts will
make inlands boats or vessels more fuel efficient and
at the same time reduce emission.
2.10 Vessel Classification

Bangladesh Inland and Coastal vessels may now
come under a new classification called BG. In line
with this initiative, rules and regulations are
formulated for design, building and survey of these
ships so that class BG can be assigned. For smaller
say less than 24m boats, a more basic class is also
introduced from the point of view of safety of
navigation as well as loss of life and properties. In
order to achieve such initiative, DG Shipping can start
working
with
any
reputable
International
Classification Society as well as Flag Administration.
One of the possibilities would be to talking to LR and
MCA in UK.
2.7 Offshore Islands

Large or small offshore islands population should
start using floating houses made of concrete materials
or Ferro cement. Planned embankments along with
canals inside such offshore islands can save both
farming and human life when there is a tidal surge
with cyclonic storm. Man has to co-live with the
nature and NOT fighting the nature. Pilot project for
one or two islands may be tried to see its effectiveness
practically.
3. PASSENGER VESSEL SAFETY

In this paper, an innovative idea of designing large
passenger vessels is presented. It clearly demonstrates
that even with overcrowding of passengers, the new
type of vessels will have less possibility of capsize
compared to monohull vessels.
2.8 Mechanized Country Crafts

As said earlier, all mechanised country crafts
building materials should now be changed to Ferro
cement and sail power be added again to make them
fuel efficient and maintenance free. Additionally,
basic safety gear for their crew and safe navigational
means to be fitted for these boats. At present, they
pose serious dangers in river traffic especially where
there is a densely population of criss-cross country
boats.
3.1 Present Design

There are a lot of small monohull passenger
launches/ferries in Bangladesh, which can easily
capsize when over-loaded in inland water (moderate
or light wave, or even in calm waters) due to
overcrowding passengers or crowded at one side.
Figure 1 shows us how serious crowding in
Bangladesh inland water area. The safety of such
vessels or any small passenger vessels can be
dramatically increased by very inexpensive means, by
simply installing two outriggers. Higher initial
stability of the modernized ship means sufficiently
smaller probability of overturning. A shorter natural
roll period should be compensated by greater roll
damping by outriggers of a special form.
2.9 Maritime Education and Training

While new and innovative ideas can be generated
and timely applied, a large new workforce is needed
too. Hence education and training is important to
educate new players in the game and also to train the
existing players. It is time that Bangladesh
Government may now start thinking in establishing a
large and comprehensive Maritime University having
its HQ at Dhaka and satellite campuses at major cities
like Chittagong, Barisal, Khulna, Sylhet and Rajshahi.
Under one umbrella, all marine people of different
disciplines like Naval Architecture, Marine
Engineering, Marine-Mechanical, Marine-Electrical,
Marine Transportation, Marine Logistics, etc. can be
dealt with. Crews for inland waters and coastal waters
including sea going can share common resources.
This will certainly be more effective than what now
exist as fragmented. Depending on the students
backgrounds, Diploma, Advanced Diploma and
Degree could be offered including postgraduate
degrees for creating an environment of research and
development. Even manpower development in marine
Most of the passenger vessels in Bangladesh can be

classified broadly in two types: open type (sunken
deck; often termed as one and a half, 1, Decker) and
closed type (freeboard deck; double Decker and now
some of them may be triple Decker). The overall
height of the open type vessels are relatively low
compared to closed type vessels, some of which have
even a continuous deck, which was not allowed say
two decades ago. The length of the closed type vessels
have also become longer compared to many such
vessels in eighties and nineties. Again, their
propulsion is mainly twin screws with twin rudders
(manual/mechanical type with no rudder stoppers).
47
Totally 2 sets of conditions for comparison for the

stability: 560 passengers (42 tonnes) and 1120
passengers (84 tonnes) have been set.
Engine rooms are again open type with no or little

safety measures. These are designed and built without
any international classification societys rules. There
have been many instances that these both types of
vessels capsize either due to overcrowding or in bad
weather or compounded by both.
From the TRIBON Calc and Hydro get the

lightweight and deadweight LCG, VCG and TCG.
Righting Lever (GZ) Curve, Drafts at equilibrium
angle, Hydrostatics at equilibrium angle, Intact
Stability criteria (checked whether the condition
complies with the regulations or not). Use all the
information to analysis the stability of each condition.
Figure2.MonohullandMultihullshipmodels
3.3 Monohull vs. Multihull

For solving the problem of inland passenger ship
capsizing due to overcrowding, the Author thinks the
multi-hull is one of the best choices. The additional
hulls provide stability, typically to hold the vessel
upright against the sideways force. This is in contrast
to monohulls which typically use a keel and/or ballast
for this purpose.
Figure 1. Serious crowding condition in inland

passenger launches and vessels in Bangladesh
Multihulls are typically either catamarans, which

have two similar hulls, or trimarans, which have a
larger hull in the centre and two smaller outriggers on
either side. In either case they are typically much
wider than the equivalent monohull, it means that
multihulls are less prone to sink than monohulls when
their hulls are compromised.
Accordingly for safe the inland navigable waters

of Bangladesh, designing a new inland passenger
launch/ferry with multihull seem a good solution.
From different kinds of multihull, finally choosing a
concept based on vessel with outriggers as the core
of new design. vessel with outriggers is convenient
and economic to manufacture. Based on the previous
research and based the main hull dimensions, it is
decision making process to calculate the dimension
and position of outriggers.
Multihulls could be faster than monohulls, because

the absence of ballast reduces their weight and the
amount of drag through the water considerably. For
inland waterways the speed of ship is not very high,
the maximum ship speed is 12 knots. Multihulls have
a shallower draft which is able to be utilized in
shallower waterways, along with shallower moorings
and reduces the risk of underwater collision. It means
the multihull design is fit for inland waterways.
3.2 Design Methodology

The monohull vessel is designed to carry 560
passengers (42 tons): Deck A (red colour), Deck B
(orange colour) and Deck C (yellow colour) could
carry 200 persons (15 tonnes), 200 persons (15
tonnes) and 150 persons (12 tonnes) respectively.
Use of TRIBON Calc and Hydro is made to get the
GZ curve of this condition and initial heel angle due
to the overcrowding (when all passengers are
crowding to one side).
Most multihulls are positively buoyant. Should hull

penetration occur the vessel is unlikely to sink thus
enabling time for rescue or repair. In fact, most
modern production multihulls are officially rated as
unsinkable, by various regulatory agencies around the
world. Multihulls give also a larger deck area.
For the multihull also carries same amount of

passengers, then compare with the monohull under the
same condition to compare whether the multi hull
vessel has the higher initial stability.
Figure 3 shows the difference between monohull

and multihull in respect of degrees of heel due to
wind.
48
Table 2. Main Dimensions MULTIHULL

Beam overall / main hull, m
15 / 9
Outrigger length / beam, m
20 / 2
Overall draft / outrigger draft, m
1/1
Distance between CL's of

the main hull and outrigger, m
Advance of outrigger's midship to
that of main hull, m
FP of outriggers are placed at
aft of main hull FP
Main hull depth, m
7
-15
25
2
Figure 3. The difference between monohull and

multihullsheelingangles[1]
4. STABILITY ANALYSIS
4.1 Intact Stability Analysis
Use Tribon to build a monohull vessel and a vessel
with outriggers and run the Tribon Calc and Hydro to
do the analysis of ship static stability.
Table 1. Main Dimensions MONOHULL
Length Overall
50.00 metres
Length B.P
54.00 metres
Breadth mld.
9.000 metres
Depth mld.
2.000 metres
Design Draft (moulded)
1.000 metres
Summer Load Draft (extreme) 1.000 metres

Displacement at Load Draft
300
tonnes
Lightship Weight
160
tonnes
Deadweight at Load Draft
140
tonnes
Subdivision Length (Ls)
54.00 metres
Aft end of Ls aft of AP
0.000 metres
Subdivision Load Line (ds)
1.000 metres
Lightest Service Draft (d0)
0.700 metres
Figure 4. Various Sections of the monohull

Condition 1: Basic Monohull for 560 passenger
Dimensions of Decks
Length (m)
Deck A
(red )
44
Deck B
(orange)
42
Deck C
(yellow)
40
Breath (m)
Height (m)
Table 3. IMO A167 Intact Stability criteria
49
# Criterion
1 Area under GZ curve up to 30
degrees > 0.055
2 Area under GZ curve from 30 to
40 deg. or down flood > 0.03
deg. or down flood > 0.09
4 Maximum GZ to be at least 0.20
metre at 30 degrees or above
5 Maximum GZ to be at an angle >
25 degrees
6 Initial GM to be at least 0.15
metres
Actual Critical
Value Value
0.665 0.055
0.291 0.030
0.956 0.090
0.179 0.200
25.049 25.000
6.388 0.150
Figure 5. Various Sections of the multihull
Condition complies with the regulations
Condition 3: Basic Multihull for 560 passengers

Condition 2: Basic Monohul for 1120 passengers

# Criterion
degrees > 0.055
2 Area under GZ curve from 30
to 40 deg. or down flood >
0.03
4 Maximum GZ to be at least
0.20 metre at 30 degrees or
above
5 Maximum GZ to be at an angle
> 25 degrees
metres
# Criterion
Actual Critical
Value Value
0.540 0.055

degrees > 0.055
5 Maximum GZ to be at an angle >
25 degrees
metres
0.212 0.030
0.752 0.090
0.142 0.200
22.095 25.000 F
Actual Critical
Value Value
0.150 0.055
0.630 0.030
0.213 0.090
0.368 0.200
27.421 25.000
1.997 0.150
0.529 0.150

** Condition does not comply **
# Criterion
degrees > 0.055
> 25 degrees
metres
Actual Critical
Value Value
0.131 0.055
0.528 0.030
0.183 0.090
0.316 0.200
25.076 25.000
1.722 0.150
50
4.2 Heel Angle
When ship with outriggers even the double amount of

passengers on each decks, she still can suffer. And the
Under same loading conditions if the entire

passengers crowd at one side, the hell angle of the
monohull is 2.16 degrees much larger than the heel
angle of multihull 0.6 degree. And it shows that the
stability criteria for monohull are not complied also
when the multihull still complies with sufficient
margin.
stability result is satisfied the criteria.

# Criterion
degrees > 0.055
> 25 degrees
metres
Actual Critical
Value Value
0.125 0.055
0.505 0.030
1.750 0.090
0.301 0.200
25.050 25.000
1.622 0.150
Figure 8. Sectional view showing overcrowding and

the associated stability curve (monohull)
Figure 6. Stability Comparison between monohull vs.

multihull for 560 passengers
Figure 9. Sectional view showing overcrowding and

the associated stability curve (multihull)
Figure 7. Stability Comparison between monohull vs.

multihull for 1120 passengers
51
4.3 Discussion
vanishing stability is larger than monohulls; the

maximum GZ is larger than monohulls and the initial
metacentric height (GM) is larger than monohulls.
Based on a monohull vessel geometry to design a

vessel with outriggers, the dimension and position of
outriggers are very important and will affect the
resistance and stability of ship. There are many
aspects which need to be improved and perfect in this
concept design project. The adequate water plane area
of outriggers will provide the transverse stability of
symmetric outrigger ships. Here just assumed the
outrigger/main hull displacements ration is 5%.
The considerable amount of stability information

shows higher initial stability of the ship with
outriggers which has sufficiently smaller probability
of overturning. However, damage stability of the
multihull especially for the outriggers damage to be
studied also carefully. If necessary, the outriggers to
have small compartments to reduce flooding and thus
reducing heeling moment.
Table 1 and Table 2 show dimensions of the

monohull and multihull respectively. Besides, Table 3
through Table 7 shows the details of the IMO intact
stability criteria for both monohull and multihull
vessels for various passenger numbers including
taking account of overcrowding.
5. CONCLUSIONS
Bangladesh having its enormous inland waterways
has tremendous potential to harness its growth in long
term development plan. Introducing new initiatives,
there is ample opportunity to get most of the benefits
out of these inland waterways.
The two type ships GZ curves under same loading

condition have an identical characteristic which is GZ
curve increase sharply during heel angle 0 to 20
degree. And maximum GZ values can all be read
around heel angle 25 to 30 degree. But usually GZ
curve increases gently until the maximum GZ which
occurs at the heel angle larger than 30 degree. The
flat hull form may be one of the reasons, which leads
to the situation. See Figure 6 and Figure 7
respectively.
New policies along with research and development

augmented by innovative ideas can see the better
results out of these vast inland waterways.
Immediate attention could be drawn to problems
like safe passenger vessels, fuel efficient inland
waterways, using concrete for building new landing
station pontoons, introducing Bangladesh Class,
Planning for a comprehensive Maritime University,
etc. Overall, a robust national maritime policy could
now be debated and put on place.
4.4 Remarks
The vessel with outriggers is the best way to
against passenger vessels capsizing due to
overcrowding in inland waterways. Just doing some
simple modification and adding two more outriggers
at each side of ship, the stability can be improved.
Shape of outriggers need not be used as main hull
ship. They are easier and faster to install. It is a
convenient and economic way to solve the vessels
capsizing for third world countries and developing
countries like Bangladesh.
ACKNOWLEDGEMENTS
The author is grateful to his ex-student Chen Cheng
for his work [2] in his final year project, part of which
has been produced in this paper. The author is also
grateful to his long time senior friend Ing Dick E.
dArnaud, Senior Consultant, Inland Navigation and
Technology Consultants, Brasil.
For this case against passenger vessels capsizing

due to overcrowding in inland waterways in the third
world countries and developing countries like
Bangladesh, need to find out an economic way. So
being utilized existed vessel is the fundamental of the
concept design. It is easier and faster to build and not
cost too much. It satisfied the requirements of this
case convenient and economic. Ship with
outriggers also provides large deck area and higher
stability satisfying for this concept design.
REFERENCES
[1] Sailing
Monohull
vs.
Mutihull;
http://www.boatsafe.com/nauticalknowhow/022
599d.htm
[2] Chen, C., Concept Design of an Inland
Passenger Vessel against Capsizing due to
Overcrowding, BEng Hons Thesis, School of
Marine Science and Technology (MAST),
Faculty of Science, Agriculture and Engineering
(SAgE), Newcastle University, UK.
The main components of TRIBON that have been

used for this design process was (1) Surface and
Compartment and (2) Calc and Hydro.
The vessel with outriggers has the range of
stability is wider than monohulls; the angle of
52

AERODYNAMIC CHARACTERISTICS OF A COMPOUND WING DURING

GROUND EFFECT
Saeed-Jame, Adi-Maimun, Agoes-Priyanto, Nor-Azwadi
Department of Marine Technology, Faculty of Mechanic Engineering
Universiti Teknologi Malaysia (UTM)Skudai, Johor, Malaysia
E-mail: jsaeed2@live.utm.my
ABSTRACT
The paper numerically investigated the flow characteristics over a compound wing during ground effect. The
compound wing is divided into three parts where one rectangular wing in the middle and two taper reverse
wings with negative dihedral angle in sides. The NACA6409 airfoil was employed as section of wings. Three
dimensional (3D) Computational Fluid Dynamic (CFD) was applied as a computational scheme. The k-
turbulent model was utilized for characterization of flow over wing surface. The computational results of a
rectangular wing with aspect ratio 1.5 and angle of attack 2 with different ground clearance were compared
with experimental data of published work. Next, the principal aerodynamic characters of compound wing and a
rectangular wing were computed for various ground clearance. The numerical results of CFD simulation of
compound wing were compared with the rectangular wing and have an acceptable improvement in lift and drag
ratio, although its stability a little reduced. The major modification of lift to drag ratio of compound wing
occurs at extreme ground effect.
Key words: Wing-in-ground effect, Compound wing, Aerodynamic characteristics, CFD model.
1. INTRODUCTION
Nomenclature
b
Wing Span
c
Chord length
CL
Lift Coefficient
CD
Drag Coefficient
CM
Moment coefficient
D
h
h/c
k
L
L/D
M
S
Sij
U
Drag Force
height of trailing edge above the ground
Ground clearance
Turbulent kinetic energy
Lift force
Lift to drag ratio
Pitching moment at c/4 from the leading
edge
Reference area (= bc)
Mean rate of deformation tensor
Free stream mean velocity
X CP
Center of pressure from the leading edge
angle of attack
Viscosity
Turbulent viscosity
Turbulent energy dissipation rate

air density
The wing-in- ground effect (WIG) crafts are

classified as a middle form between ships and
aircrafts. WIG crafts can fly proximity the any surface
such as ground, sea, snow and ice etc. A high pressure
air (air cushion) is generated from interface between
wing of the WIG craft and the ground. The dropping
of down-wash angle because of the ground effect
guides to an enhancement in lift and decline of induce
drag, with a raise of effective aspect ratio for the
wing. Rising in the lift force and decreasing of
induced drag provides an augmentation on the lift to
drag ratio (L/D) [1]. The type of air cushion is the
principal difference between hovercraft and WIG
craft. A static air cushion holds hovercraft, while the
WIG craft is borne by dynamic air cushion. The small
aspect ratio of wing and high lift to drag ratio are
other differences from a conventional craft. Currently,
the suitable expansion of high power computing and
computational fluid dynamic (CFD) grows the
numerical aerodynamic characteristics of WIG crafts
[2].
There are a lot of concepts of wing-in-ground
effect. First time, the idea of ram wing was put into
operation by Troeng [2]. Practically all WIG craft
employ high pressure ram air for improved lift,
53
adoption of pressure distribution on the lower surface

of wing due to ground effect and increasing of
pitching angle. They showed a thick wake region on
suction surface at small ground clearance and high
angle of attack. Because of merging the boundary
layers of wing and ground plate in extreme ground
effect especially at high angle the drag force has an
enhancement. Ahmed et al. [5] studied the
aerodynamic characteristics of NACA4412 airfoil
section in a low turbulence wind tunnel with moving
ground model at a Reynolds number of 3.0 106 by
changeable the angle of attack from 0 to 10 and
ground clearance (h/c) 0.05-1. They measured the
pressure distribution, velocity and wake region of
flow over the airfoil surface, also lift and drag forces.
They showed a reduction of suction on upper surface
when the airfoil come close to the ground for all pitch
angle. Lift force reduces when decreasing ground
clearance for small angle of attack (<6), although it
augmented for greater angle of attack by improvement
pressure distribution on the pressure side. Further,
they illustrated the drag force is higher in extreme
ground effect due to transformation of the lower side
pressure distribution. Jung et al. [6] performed
extensive tests in the closed-type wind tunnel. Lift
and drag forces and the pitch moment of NACA6409
were measured as main aerodynamic parameters such
as the aspect ratio (AR), angle of attack (), ground
clearance (h/c) and endplate shape were varied. They
demonstrated the ground effect caused a decline in the
tip vortex and the wake following the wing. The lift
enhanced because of the ground effect at short ground
clearance when the endplate is not fitted to wing. Due
to the boundary layer that extends on extreme ground
(h/c< 0.1), the lift force may possibly be a little
underestimated. Also they proved by smoke trace test,
the flow transitory under the pressure side is reserved
by the endplate that causes a reduction on the tip
vortex by the pressure difference between the pressure
and suction sides of the wing. Furthermore, the
influence of endplate on lift force for smaller aspect
ratio of wing is grater than larger aspect ratio wing.
The drag force is reduced by the ground effect when
the wing moves toward the ground. The explanation is
that the induced drag diminishes due to the drop of the
tip vortex at the wing tip. Another finding of Jung et
al. [6] is that when the angle of attack and the aspect
ratio increased, the center of pressure shifted ahead to
the leading edge of the wing. The center of pressure
moved to the leading edge by adding an endplate to
wing. As the ground clearance o decreased, the center
of pressure also moved forward to the leading edge of
wing.
nevertheless the fault of ram wing WIG craft is its

stability. According this concept a number of WIG
craft included a low aspect ratio wing (approximately
square) and a large horizontal tail installed out of
ground effect which supplies the essential stability.
The wing is often built-in endplates with the aim of
increase ground effect. Tandem-Airfoil-Flairboat
(TAF) is defined by the proposal of assembling two
short wings in tandem. Both wings have roughly the
same dimensions with rather small space in between
and no horizontal tail. This arrangement presents good
stability in extreme ground effect, but unstable when
out of ground effect. A special class of ram wing
called as Lippisch idea where the main wing includes
an inverse dihedral wing along the leading edge. This
design embraces more longitudinal stability rather
than a low aspect ratio ram wing. A smaller horizontal
tail is required for longitudinal stability requirement
in low ground clearance and jump modes during
cruise condition. The Lippisch concept uses a greater
aspect ratio of wing rather than ram wing concept
which get near to 3. The lift to drag ratio of Lippisch
craft were around 25. The completed type of Lippisch
craft is Airfish. This WIG craft generally flies on the
proximity to the ground. The operational range
ground clearance of Airfish craft is between 0.1m and
1 m. The design of Airfish is described for utilizing
ground effect but it allows to dynamic jumps about 5
m in some transitory condition. The plan Hoverwing
craft utilizes a plain system of flexible skirts to hold
an air cushion between the twin hulls. This static air
cushion is employed just through takeoff, therefore
assisting the craft to accelerate with smallest power
before shifting to true ground effect mode [2]. The
Hydrofret concept is classified to use both static air
cushion and dynamic ground effect. The concept is
planned in two models. The first is a ram-wing
catamaran balanced by a large aspect ratio forward
wing tail. In the different description a large aspect
ratio rear wing is employed instead the tail wing [2].
Chawla et al. [3] described Wing-in-ground effects
from a wind-tunnel research of a NACA 4415 airfoil
section for wing model with an aspect ratio of 2.33.
They employed moveable flap and detachable end and
center plates. They showed ground effect of wing is
appeared for ground clearance (h/c) up to 1. The
influence of endplate on lift to drag ratio is higher for
lower ground clearance. Ahmed and Sharma [4]
studied on the pressure distribution over the wing
surface at different ground clearances and angles of
attack for measuring the lift, drag and the mean flow
over the surface of the wing, also they depicted the
wake region for mean and fluctuating velocities of
flow. The convergentdivergent passage shape
between the wing and the ground at certain angle of
attack gives a suction effect that causes a local
reduction in lift. The higher lift force mostly related to
The turbulent flow around two-dimensional wing

was numerically investigated for fixed and moving
ground boundary by Chun and Chang [7]. An
incompressible Reynolds Average Navier-Stokes
54
(RANS) equation with finite difference method was

applied for numerical model.
According to their
computational results, the difference in the lift and
moment coefficients simulated by two bottom
conditions is trivial, but the drag coefficient simulated
by the fixed bottom is to some amount smaller than
that by the moving one. Aerodynamic characteristics
of three-dimensional wings in ground effect for Aerolevitation Electric Vehicle (AEV) are numerically
investigated for various ground clearances and wing
spans at the chord-length based Reynolds number of
2 10 6 by Moon et al. [8]. The design of AEV system
is based on small wing span for decreasing costs of
the structure and making of cruising channel. This
system uses the tandem wing concept to satisfy the
requirement lift. They showed the increasing of lift to
drag ratio versus very low ground clearance (h/c<0.1)
is nonlinear. Moreover, the lift to drag ratio for small
span of AVD wings enhances between 10-40%.
Ockfen and Matveev [9] researched numerically on
airflow over NACA4412 airfoil section with favorable
flap pattern that get better aerodynamic characteristics
in extreme ground effect. The Spalart-Allmaras
turbulence model of the Navier-Stokes equations via
Fluent6.3 was used for various Reynolds number and
angle of attack during ground effect. They depicted
with small flap deflection the lift to drag ratio has a
considerable enhancement, although for high flap
angle the pressure drag increasing subsequently lift to
drag ratio would be lesser than the wing without flap.
different ground clearance (h/c= 0.1 ,0.15, 0.2, 0.25,

0.3), angle of attack 2, aspect ratio 1.5 for validation
and 1.25 for present model and velocity of airflow
30.8 m/s (60 knots). Ground clearance (h/c) is defined
of the distance ratio between wing trailing edge and
ground surface (h) to wing chord length (c). The
numerical scheme considered a steady state,
incompressible by means of k- turbulent model of the
Navier-Stokes equations for flow over wing surface.
The CFD models applied Fluent 6.3 software and high
speed computer. The transport equations for the
turbulent kinetic energy (k) and turbulent dissipation
energy () are expressed as follows.
( k )
+ div ( kU ) = div t grad ( k ) + 2 t S ij S ij
t
(1)
( )
2
+ div( U ) = div t grad ( ) + C1 2t Sij Sij C2
k
k
t

(2)
where t is turbulent viscosity.
t = C
k2
(3)
The adaptable constants C, k, , C1 and C2 have

the following values:
C= 0.09 k = 1.00 = 1.3 C1= 1.44 C2 =1.92.
The aerodynamic coefficients and center of
pressure in this numerical study were determined as
follows:
M
L
D
,C =
, CM =
and
CL =
D
2
2
0.5 U 2 Sc
0.5 U S
0.5 U S
The lift to drag ratio of only wing decreases as

other part add to wing [2]. Kirillovikh [10] reported
the lift-to-drag ratio of a wing with aspect ratio 23
and ground clearance 0.2 would be about 3545 that it
is so large. When other part of crafts such as hull and
pylon are included to wing, the reduction of lift-todrag ratio arises consequently lift to drag ratio in this
sample reach to 12-16.
X CP = 0.25 +
The aim of this paper is aerodynamic

characteristics of a compound wing during ground
effect. The compound wing has been completed by
three parts that a rectangular wing in the middle and
two taper reverse wing with negative dihedral angle in
sides. The numerical model used the three
dimensional (3D) CFD using finite volume scheme.
The standard k- turbulent model has been utilized for
turbulent flow around wing. The numerical result of
the present wing is compared with a rectangular wing.
CM
.
C L cos + C D sin
(a)
2. CFD NUMERICAL STUDY

Present numerical study was carried out by a
model of rectangular wing and compound wing with
NACA6409 airfoil section. The principle dimensions
of both wings (Figure 1(a)-1(b)) are shown in Table 1.
These simulation were prepared with respect to
(b)
Figure 1: (a) The rectangular wing, (b) the compound
wing.
55
Table 1: Principle dimension of wings.

Total wing span (b)
Chord length (c)
Middle wing span
Taper ratio
Dihedral angle
pitch angle 2. This figure shows a good agreement

between computational and experimental results. The
trend of both results illustrates a growth of lift
coefficient when ground clearance decreases.
83.4 cm
66.7 c m
41.4 cm
0.8
13
Table 2: Lift coefficient versus ground clearance

with angle of attack 2 and AR = 1.5 for experimental
and numerical result.
The number of mesh for each simulation is about

+
4,000,000-4,500,000. The y values for turbulent is

less than 100. This number of elements has good
enough convergence for aerodynamic characteristics.
The present simulation used symmetry plan as shown
Figure 2 for ram wing and Figure 3 for compound
wing. This is to shorten the simulation time although
the results achieved will be the same.
Ground
clearance
Numerical
0.1
Experimental
0.3741
0.3980
0.15
0.3566
0.3930
0.2
0.3418
0.3682
0.25
0.3389
0.3682
0.3
0.3286
0.3682
0.45
Lift Coefficient (CL)
0.4
Figure 2: The meshing of rectangular wing.
0.35
0.3
0.25
0.2
0.15
Numerical
0.1
Experimental
0.05
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Ground clearance (h/c)
Figure 4: Lift coefficient (CL) versus ground clearance

(h/c) for angle of attack 2 and AR = 1.5.
3.2 Drag Coefficient (CD)

The computational results and experimental data
[6] of drag coefficients were listed in Table 3. The
tendency of both results has good agreement but the
computational results are a slight greater as shown in
Figure 5. The increment of drag coefficient due to
increasing of ground clearance was established by
numerical and experimental scheme.
Figure 3: The meshing of compound wing.
3. VALIDATION OF CFD SIMULATION

The experimental results of Jung et al. [6] are taken
to validate the present CFD simulation. The accuracy
of numerical aerodynamic characters such as lift
coefficient, drag coefficient, lift to drag ratio, moment
coefficient and center of pressure were compared with
experimental results. For validation purpose, the
numerical results of a rectangular wing with NACA
6409 airfoil section were proved.
Table 3: Drag coefficient versus ground clearance

Ground
clearance
0.1
0.15
0.2
0.25
0.3
3.1 Lift Coefficient (CL)

The present numerical results and experimental
data [6] of lift coefficient were summarized in Table
2. Figure 4 depicts a comparison of lift coefficient for
different ground clearance with aspect ratio 1.5 and
56
Numerical
Experimental
0.0384
0.0408
0.0415
0.0425
0.0424
0.0308
0.0308
0.0348
0.0328
0.0348
versus ground clearance is shown in Table 5. Also,

Figure 7 confirms the good agreement between
computational and experimental results of moment
coefficient. According the data in Table 6 the center of
pressure moved to lending edge as the ground
clearance decreased for both methods. Furthermore,
there is acceptable conformity for numerical method
to determine center of pressure as shown in Figure 8.
Drag Coefficient (CD)
0.045
0.04
0.035
0.03
0.025
0.02
0.015
Numerical
0.01
Experimental
0.005
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Table 5. Moment coefficient versus ground clearance

with angle of attack 2 an AR=1.5 for experimental
Ground
clearance
Figure 5: Drag coefficient (CD) versus ground

clearance (h/c) for angle of attack 2 and AR = 1.5.
3.3 Lift to Drag Ratio (L/D)

Lift to drag ratio from numerical simulations and
experimental data [4] are shown in Tables 4 and
Figure 6. Generally, the trend of lift to drag ratio of
the numerical method has an approximate agreement
with experiments. Moreover, both simulations
demonstrate lift to drag ratio increases with smaller
ground clearance.
Ground
clearance
0.1
0.15
0.2
0.25
0.3
Numerical
Experimental
9.743
8.741
8.228
7.969
7.749
12.903
12.742
10.571
11.212
10.571
Moment Coefficient (CM)
Table 4. Lift to drag ratio versus ground clearance

Experimental
0.1
0.0873
0.0700
0.15
0.0869
0.0760
0.2
0.0857
0.0753
0.25
0.0877
0.0748
0.3
0.0862
0.0783
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
Numerical
Experimental
0.05
0.1
0.15
0.2
0.25
0.3
Figure 7: Moment coefficient (CM) versus ground

clearance for angle of attack 2 and AR = 1.5.
14
12
10
L/D
Numerica
l
Table 6: Center of pressure versus ground clearance

8
6
Numerical
Experimental
Ground
clearance
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Ground clearance
Figure 6: Lift to drag ratio (L/D) versus ground

clearance (h/c) for angle of attack 2 and AR = 1.5.
3.4 Moment Coefficient (CM) and Center of

Pressure (Xcp)
The small variation of moment coefficient for
present numerical results and experimental data [6]
57
Numerical
Experimental
0.1
0.4826
0.4242
0.15
0.4928
0.4449
0.2
0.5000
0.4580
0.25
0.5079
0.4545
0.3
0.5114
0.4649
0.35
0.6
0.45
0.5
0.35
Lift Coefficient (CL)
0.4
Xcp/c
0.4
0.3
0.2
Numerical
Experimental
0.1
0.3
0.25
0.2
0.15
RECTANGULAR WING
0.1
COMPOUND WING
0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.05
0.35
0.1
0.15
0.2
0.25
0.3
0.35
Figure.9: Lift coefficient (CL) versus ground clearance

for angle of attack2.
Figure 8: Center of pressure (XCP/c) versus ground

clearance for angle of attack 2 and AR = 1.5.
The drag coefficients of rectangular and compound

wing versus ground clearance with angle of attack 2
and aspect ratio 1.25 are depicted in Figure 10, in
addition the reduction of drag coefficient for
compound wing is calculated by Eq.5 in Table 8. This
reduction is between 9-10.30% which the maximum
reduction occurs in ground clearance 0.1.

The affect of taper reverse wing with negative
dihedral angle of compound wing was exposed in
Tables 7-11 and Figures 9-13. The results of lift
coefficients versus ground clearance with angle of
attack (AOA) 2 and aspect ratio 1.25 of rectangular
wing and compound wing are shown in Figure 9. The
increment of lift coefficient for compound wing is
computed by Eq.4 in Table 7.There is considerable
improvement on lift coefficient in small ground
clearance for compound wing as compare with
rectangular wing. It can see the increment of lift
coefficient is 20.25% at ground clearance 0.1,
However, a small reducing in lift coefficient appeared
at higher ground clearance.
Icrement (%) =
C L (Compound )
C L (Re c tan gular )
Re duction (%) = 1
(4)
Compound
Increment
of CL
%
0.1
0.3186
0.3831
20.25
0.15
0.3091
0.3178
2.80
0.2
0.3017
0.2953
-2.14
0.25
0.2932
0.2834
-3.35
0.3
0.2904
0.2728
Groud
clearance
Rectangular
Reduction
of CD %
0.1
0.0332
0.0297
10.30
0.15
0.0347
0.0314
9.37
0.2
0.0353
0.0321
9.05
0.25
0.0355
0.0322
9.55
0.3
0.0358
0.0326
9.02
Compound
0.04
0.035
Drag Coefficient (CD)
Rectangular
(5)
C D (Re c tan gular )
Table 8: Drag coefficient versus ground clearance

with angle of attack 2 for rectangular and compound
wing.
Table 7: Lift coefficient versus ground clearance with

angle of attack 2 for rectangular and compound wing.
Ground
clearance
C D ( Compound )
0.03
0.025
0.02
0.015
RECTANGULAR WING
0.01
COMPOUND WING
0.005
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
-6.08
Figure 10: Drag coefficient (CD) versus ground

clearance for angle of attack2.
58
pressure of wing is another principal parameter for

stability during take-off and landing conditions. Table
11 and Fig.13 illustrates the variation of center of
pressure of both wings versus ground clearance.
According to presented results the center of pressure
of compound wing is a slight closer to leading edge of
wing.
The lift to drag ratio of rectangular and compound

wing versus ground clearance for angle of attack 2
summarize in Table 9, in addition, the increment of
lift to drag ratio of compound wing related to
rectangular wing is determined by Eq.6. This
increment in small ground clearance is so high where
it is around 34% in ground clearance 0.1 that can be
related to the high efficiency and saving energy of
present wing. The trend of lift to drag ratio versus
ground clearance for both wings is shown in Figure
11.
Increment (%) =
The resulted of positive moment coefficient

decreases angle of attack.
Re duction (%) = 1
L / D( Compound )
L / D(Re c tan gle )
C M (Re c tan gular )
(7)
(6)
Table 10:
Moment coefficient versus ground
clearance with angle of attack 2 for rectangular and
compound wing.
Table 9: Lift to drag ratio versus ground clearance

wing.
Grod
clearane
Rectangular
0.1
9.606
12.877
34.06
0.15
8.920
10.118
13.43
0.2
8.550
9.199
7.59
0.25
8.249
8.814
6.86
0.3
8.109
8.371
3.23
Compound
C M ( Compound )
Groud
clearance
Increment
of L/D %
0.1
0.15
0.2
0.25
0.3
Rectanguar
Compound
0.0819
0.0829
0.0832
0.0823
0.0830
0.0778
0.0716
0.0703
0.0699
0.0688
Reduction
of CM %
4.94
13.63
15.56
15.01
17.15
Moment Coefficient (CM)
0.09
14
12
L/D
10
8
6
RECTANGULAR WING
4
0.08
0.07
0.06
0.05
0.04
0.03
RECTANGULAR WING
0.02
COMBINATION WING
0.01
0
0
COMPOUND WING
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Figure.12: Moment coefficient (CM) versus ground

Table 11: Center of pressure versus ground clearance
wing.
Figure 11: Lift to drag ratio (L/D) versus ground

The variation of moment coefficients for rectangular
and compound wing versus ground clearance with
angle of attack 2 and aspect ratio 1.25 are shown in
Table 10 and Figure 12. A moment coefficient that
causes a decreasing on angle of attack was named
positive moment. The reduction of moment
coefficient for compound wing is calculated by Eq.7
in Table 10. This reduction at extreme ground
clearance (h/c=0.1) is small indicates the stability of
compound wing has a little decline. The center of
Groud
clearance
0.1
0.15
0.2
0.25
0.3
59
Rectanguar
0.5062
0.5174
0.5249
0.5296
0.5349
Compound
0.4527
0.4748
0.4873
0.4960
0.5013
Reduction
of Xcp/c
%
10.57
8.24
7.17
6.35
6.27
[2] Rozhdestvensky, K.-V., Wing-in-ground effect

vehicles, Elsevier Journal of aerospace science,
Vol. 42, pp. 211-283 (2006).
[3] Chawla, M.-D., Edwards, L.-C., and Franke, M.E., Wind-tunnel investigation of
wing-inground effect, Journal of Aircraft, Vol. 27, No.
4, pp. 289 (1990).
[4] Ahmed, M.-R., and Sharma, S.-D., An
investigation on the aerodynamics of a
symmetrical airfoil in ground effect, Elsevier
Journal of Experimental Thermal and Fluid
science, Vol. 29, pp.633-647 (2004).
[5] Ahmed, M.-R., Takasaki, T., and Kohama, Y.,
Aerodynamic of NACA4412 airfoil in ground
effect, AIAA Journal, Vol. 45, No. 1, pp. 37
(2007).
[6] Jung, K.-H., Chun H.-H., and Kim, H.-J.,
Experimenta investigation of wing-in-ground
effect with a NACA 6409 section, JASNAOE
Mar Sci Technol, Vol. 13, pp. 317-327 (2008).
[7] Chun, H.-H., and Chang, C.-H., Turbulence
flow simulation for wings in ground effect with
two ground conditions: fixed and moving
ground, International Journal of maritime
engineering, pp.211-227 (2003).
[8] Moon, Y.-J., Oh, H.-J., and Seo, J.-H.,
Aerodynamic investigation of three-dimensional
wings in ground effect for aero-levitation electric
vehicle, Elsevier SAS Journal of Aerospace
and Technology, Vol. 9, pp. 485-494 (2005).
[9] Ockfen, A.-E., and Matveev, K.-I., Aerodynamic
characteristics of NACA4412 airfoil section with
flap in extreme ground effect, Inter J Nav Archit
Oc Engng, Vol. 1, pp. 1-12, (2009).
[10] Kirillovykh V.-N., Russian ekranoplans,
Proceedings of the international workshop on
twenty-first century flying ships, pp. 71.117
(1995).
0.6
0.5
Xcp/c
0.4
0.3
RECTANGULAR WING
0.2
COMBINATION WING
0.1
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Figure 13: Center of pressure (XCP/c) versus ground

5. CONCLUSION
The target of this research is to investigate the
aerodynamic characteristics of an especial ram wing
concept named compound wing. Based on the
computational results the lift and drag coefficient of
the compound wing has considerable modification as
compare with rectangular wing for small ground
clearance. The taper reverse wing with negative
dihedral angle in the sides of compound creates a
greater decreasing of down-wash velocity due to the
ground effect that leads to a higher augmentation in
lift and reduction of drag, as well as an increase of left
to drag ratio for the wing. The high increment of lift
to drag ratio for present wing in extreme ground effect
(34%) is recognized a good efficiency for WIG craft.
The moment coefficient has a little diminishing in
lower ground clearance, also the center of pressure of
present wing has a small shift forward to leading edge
of wing as contrast with rectangular wing that can
reduce the stability. A horizontal tail out of ground
effect is one alternative to modify stability. For extra
investigate, the aerodynamic characteristics and
stability of present wing with horizontal tail wing will
be numerically researched and compared with
experimental data using UTM wind tunnel.
ACKNOWLEDGEMENTS
The authors would like to thank the Ministry of
Science, Technology, Innovation (MOSTI) Malaysia
for funding this research under vote number 79344,
and UTM due to high performance computing facility.
REFERENCES
[1] Yun, L., Bliault, A., Doo, J., WIG Craft and
Ekranoplan:Ground Effect Craft Technology,
Springer Science+Business Media, LLC (2010).
60

DYNAMIC RESPONSE OF OFFSHORE TRICERATOPS UNDER

ENVIRONMENTAL LOADS
Srinivasan Chandrasekaran1, Madhuri Seeram2, A. K. Jain3 and Gaurav4
1
Corresponding author and Associate Professor, Department of
Ocean Engineering, IIT Madras, India, E-mail: drsekaran@iitm.ac.in
3
Professor, Department of Civil Engineering, IIT Delhi, India,
E-mail: akjain@civil.iitd.ac.in
2
PhD Scholar, Department of Ocean Engineering, IIT Madras, India,
E-mail: madhuri.seeram@gmail.com
4
Research Scholar, Department of Civil
University of Minnesota, Minneapolis, USA,
E-mail: gaura003@umn.edu
Engineering,
ABSTRACT
Offshore Triceratops is relatively new concept in deep water oil explorations in recent times. They exhibit
structural advantages in comparison to other platforms suitable for deep water applications. Structural
integration of deck system, ball joint, Buoyant Leg Structures (BLS) and foundation system make them attractive
for the encountered environmental loads. BLS is a positively buoyant system that serves as a buoyant chamber
and storage chamber as well. Ball joints provided between deck and BLS do not transfer rotation; but transfers
lateral displacements, making them more effective to encounter undesirable yaw motion caused due to
aerodynamic loading. Detailed dynamic analysis of offshore triceratops under wind, wave and current are
presented. Based on the analytical studies carried out using ANSYS AQWA, it is seen that the deck response is
relatively less in comparison to the buoyant legs, making advantageous for operation in moderate sea states.
Key words: Ball Joint, Buoyant Leg Structure, Non-linear Dynamic Analysis, Triceratops.
Experimental investigations and simplified
analyses employed to study the dynamic response of
TLPs show that the analytical methods are valid and
practical [27]. Experimental investigations made to
estimate tether force of deep water TLP model closely
agree with the analytical methods, validating the
employed analytical methods [25]. Researchers also
emphasized the necessity of estimating the wave
forces on TLPs at displaced position to account for
non-linear effects in their dynamic response behavior
[21].
Yoshida et al. [28] discussed linear response
analysis of TLP under regular waves by considering
the flexibility of superstructure in the equations of
motion; response motions, tension variations of
tendons and structural member forces were solved
simultaneously. The applicability of this method was
confirmed by comparison with the test results on two
kinds of small-scale TLP models. A new spectral
description of the longitudinal wind velocity
fluctuations over the ocean for estimating the wind
induced response to TLP was used [1]. Haritos [12]
modeled the response of TLP under wind and wave
loading; the study highlighted the aerodynamic
influence on surge response, in particular. Analytical
studies conducted on TLP model by estimating wave
forces using three-dimensional singularity method
proved to closely agree with that of the results of
experimental investigations carried out. The study
1. INTRODUCTION
Recent advancements in oil and gas exploration
demands more innovative platform designs to
alleviate wave loads in ultra-deep waters. Considering
the fact that oil reserve found in shallow waters are
almost been exploited, future exploration is expected
to be towards ultra-deepwater (1000 to 3000m and
deeper). This necessitates innovative platform
geometries that are capable of alleviating encountered
environmental loads at these greater depths. Tension
leg platforms (TLP), sub-sea systems, semi
submersibles, FPS, FPSOs and Spars are suitable for
greater depths of about 1000m to 3000m. However,
they are expensive due to factors namely: i) large deck
payloads; ii) large hull size, iii) complex station
keeping systems etc. TLP, a most commonly preferred
platform for deep water, is a hybrid design; it is stiff
in vertical direction and compliant in horizontal
direction, making it advantageous to alleviate the
encountered environmental loads. On the other hand,
offshore triceratops has a structural integration of
deck structure, ball joint, BLS and tethering system,
making it more suitable for ultra-deep waters. Ball
joints are special components that transfer
displacements from BLS to deck and vice-versa;
pitch, roll and yaw motions are not transferred.
61
than that of the four-legged (square) TLP.

Chandrasekaran & Jain [5] studied the aerodynamic
behavior of triangular TLP due to low frequency wind
force and random waves. The effect of the offset of
aerodynamic center (AC) and center of gravity (CG)
of the platform on the coupled response of triangular
TLP is discussed. Results shows that the lowfrequency wind alter the response of TLP to a
considerable extent. Studies report that the influence
of hydrodynamic coefficients in TLP response is
nonlinear; they also influence the plan dimension of
TLP and its site location. Chandrasekaran et al. [7]
presented response behavior of triangular TLP under
impact loading and investigated to be safe for the
considered Mathieu stability analyses. Zeng et.al [29]
developed theoretical model for analyzing the
nonlinear behavior of a TLP with finite displacement
in which multi-fold nonlinearities are taken into
account, i.e., finite displacement, coupling of the six
degrees of freedom, instantaneous position,
instantaneous wet surface, free surface effects and
viscous drag force. Chandrasekaran et al. [6]
presented the response behavior of triangular TLP
under regular waves using Stokes nonlinear wave
theory. Results show that the coupled response in
surge and pitch degree of freedom obtained using
Stokes theory is lesser than that obtained using the
Airys theory. Chandrasekaran et al. [8] studied the
response behavior of triangular TLPs using Dynamic
Morison
equation
considering
nonlinearities
associated with vorter shedding effects.
White and Copple [24] introduced an innovative
effective platform named triceratops in ultra deep
water application. They highlighted the inherent
characteristics that facilitate its cost effective
construction and installation while minimizing
potential problems during its service life.
It is seen from the literature that dynamic response
analyse on offshore triceratops are scarce despite the
platform design being innovative. The present study is
focused on the analytical estimate of response of
triceratops under different sea states.
also showed a close agreement of the simulated time

history with respect to the experimental ones [17].
Nordgren [20] analyzed high frequency vibration of
TLP using spectral analysis. Results used to estimate
the fatigue life of the platform was found to be
adequate. However, few researchers conducted linear
analysis in frequency domain to estimate the surge
response under random waves. The nonlinear time
domain and frequency domain approaches used in the
study show that the varying damping ratio does not
influence the mean surge response of TLP [10-11].
Kareem and Zhao [16] analyzed the response of TLP
for random wind loading using Gaussian distribution
and equivalent quadratization method. The higher
order response cumulants were developed based on
Volterra series. A direct integration and Kac-Siegert
technique was used to evaluate the response
cumulants. The results showed a good comparison
with simulation. Jain [14] examined nonlinear
coupled response of offshore TLP to regular wave
forces considering coupling between surge, sway,
heave, roll, pith and yaw DOFs. The nonlinearities
arising from hydro-dynamic drag, change in tether
tension, variable submergence effect was highlighted
in the study. Higher order nonlinear transfer functions
were also alternatively employed by researchers to
estimate the TLP response; the results showed linear
and nonlinear components of the response, explicitly
[13, 2].
Mekha et al. [19] studied the implication of the
tendon modeling on the global response of TLP in
general and on the limits of the tendon forces in
particular. Thiagarajan and Troesch [23] conducted
model tests to examine the heave effects of uniform
current in the presence of disk at TLP columns. The
hydrodynamic interactions among TLP members are
generally included in the motion and structural
response analysis; elastic mode is applied to solve the
radiation problem [9]. Tabeshpour et al. [22] studied
the nonlinear dynamic response of TLP in both time
and frequency domains under random sea wave
loading. The time history of random wave is
generated based on Pierson-Moskowitz spectrum
while hydrodynamic forces are calculated using
modified Morison equation. The power spectral
densities (PSDs) of displacements, velocities and
accelerations are calculated from nonlinear responses.
Statistical linearization techniques were employed to
model the tendon forces more effectively [18]. Yan et
al. [26] presented stress response of a tension leg
platform in extreme environments.
Chandrasekaran & Jain [3-4] compared the
dynamic behaviour of square and triangular TLP
under regular and random waves loading. The results
show that the triangular TLP exhibits a lower response
in the surge and heave degrees of freedom than that of
the four-legged (square) TLP. The triangular TLP
attracts more forces in the pitch degree of freedom
and the response in this degree of freedom is more
2. ANALYSIS OF TRICERATOPS
2.1 Structural Modeling
Triangular TLP and triceratops at 600m water depth
are modeled; for the basis of comparison, buoyancy
and water depth are kept same for both the models.
Details of the models considered in the analysis are
given in Table 1. Triceratops is modeled in ANSYS
AQWA software by using cylinders for BLS structure
and plate for the deck. Meshing of the model is done
using quadrilateral plate elements. Total No. of nodes
and elements are 3006 and 2940 respectively. Total
No. of diffracting nodes and elements are 1635 and
62
1608 respectively. The Plan and elevation of the

model are shown in Figs. 1- 2.
The behavior of Triceratops is studied for different
environmental loads such as waves, wind and current.
The waves are generated using Pierson-Moskowitz

spectra. The waves, wind and current data is given in
Table 2.
Table 1. Details of Triangular TLP and Triceratops.

Description
Water depth
Notation
D
Units
m
TLP
Triceratops
600
600
Steel
Steel
Material
Unit Weight of the
material
Centre to Centre
distance of the legs
kg/m3
7850
7850
Pb
70
70
Diameter/equivalent
dia. of each leg
Freeboard
Draft
17
17
22
22
74.7
32
Length of each leg
54
96.7
Tether Length
Unit weight of
surrounding fluid
Buoyancy of TLP
including pontoons
568
525.3
kN/m3
10.25
10.25
FB
kN
521600
521600
Area of deck
m2
1732.41
1732.41
Self weight of
TLP+Payload
kN
351600
351600
Total Tether force
Tt
kN
170000
170000
rx
35.1
35.1
ry
35.1
35.1
Radius of gyration
about x axis
Radius of gyration
about y axis
Radius of gyration
about z axis
AE/l of the tether
rz
35.1
35.1
kN/m
84000
84000
Area of the tether
m2
0.24
0.22
Diameter of tether
0.55
0.53
Figure 2. Elevation of the model

Figure 1. Plan of offshore triceratops model
63
Table 2 Wave, current and wind data
80
15m, 15s
12m, 15s
70
Hz (m)
Tz sec
(1)* 10
10
(2)* 12
15
(3) *15
15
* Spectrum Number
Current
Data
VC (m/s)
0, 0.5
0, 1
0, 1.5
10m, 10s
Wind
Data
VW (m/s)
35
40
45
Snn (w) m2s/rad
Wave Data
60
50
40
30
20
10
0
0
0.2
0.4
0.6
0.8
Frequency (rad/s)
1.2
Figure 5. Wave Spectra considered for

the present study.
3. ANALYSIS OF RESULTS
The response amplitude operators (RAO) of each
sub-structure of triceratops, in surge and roll degreesof-freedom are shown in Figures 6-7 respectively;
wave height and time period of (Hs,Tz) are taken as
(15m-15s; 12m-15s; and 10m- 10s) in the present
study. Current velocity of 1.5m/s and wind velocity of
45m/s are considered. For the encountered
environmental loads, peak responses in different
degrees-of-freedom are given in Table 3.
Figure 3. Numbering of structure in the
analysis
In the analysis, the response of three BLS structures
& the deck are evaluated separately. The structures
numbering is given in Fig. 3. The three BLS structures
are connected to the foundation system with the
tethers while their connection to the deck is through
universal joints (ball joints). The connection of the
BLS system to the sea bottom is shown in Fig. 4. The
spectra considered for the present study is shown in
Fig. 5.
Figure 6. Surge responses of triceratops
Figure 4. Connection of BLS system to sea

floor
The Pierson-Moskowitz spectrums considered for the
study is shown in Fig. 5.
Figure 7. Roll responses of triceratops
64
Table 3 Peak responses of triceratops for different loads

Description
Surge
Sway
Heave
Roll
Sp1+Wind
1.15
0.19
0.02
0.158
(2)
(1)
(2)
(1,3)
Sp2+Wind
1.15
0.2
0.02
0.165
(2)
(1)
(2)
(1)
Sp3+Wind
1.0
0.19
0.02
0.16
(2)
(2)
(2)
(1)
Sp1+Wind
1.15
0.19
0.02
0.16
+C
(2)
(1,3)
(2)
(1,3)
Sp2+Wind
1.15
0.2
0.022
0.17
+C
(2)
(1)
(2)
(1)
Sp3+Wind
1.0
0.17
0.02
0.16
+C
(2)
(2)
(2)
(1)
Sp: Spectrum
C : Current
(*)
:
Structure #
REFERENCES
Pitch
0.49
(2)
0.49
(2)
0.48
(2)
0.5
(2)
0.5
(2)
0.48
(2)
Yaw
2
(1)
2.2
(1)
2.1
(1)
2.1
(1)
2.2
(1)
2.1
(1)
[1].
[2].
[3].
[4].
4. DISCUSSION OF RESULTS
Surge and roll RAOs, as seen in Figures 6-7 show
that surge response in the deck (marked as structure
#4) is comparable with the surge response of
individual BLS units. This indicates a collective
response of the deck and BLS as an integral unit,
ensuring efficient connectivity between BLS and the
deck. Excessive surge and roll response in BLS 2
shall be attributed to variation in tether tension caused
in leg 2 under the considered sea states. It is seen
from the Table 3 that yaw motion is significant
manifestation of presence of wind forces under the
wave action; however, heave response in the BLS
remains under the permissible values. It is also seen
that the presence of current does not influence
dynamic response of TLP, significantly.
[5].
[6].
[7].
[8].
5. CONCLUSIONS
Results show that the deck exhibits less roll
response under the chosen sea states, highlighting the
advantage of ball joint between the deck and BLS.
However, the surge response is considerate, ensuring
the effective control of roll motion between the
elements namely the deck and BLS, respectively. This
type of behavior is advantageous to upkeep more
facilities on the deck system and for comfortable
operation during moderate sea states.
The paper presented the analytical studies on
triceratops under three different critical sea states,
which is relatively new in the literature. The focus
was to highlight the advantages of the structural
configuration, making it suitable for deep waters.
However the comparison of its behavior with other
deep water platforms is not in the scope of the present
study and hence not presented.
[9].
[10].
[11].
[12].
[13].
65
Ahsan Kareem. M (1985), Wind induces

Response Analysis of Tension Leg Platforms,
Journal of Structural Engineering, Vol. 111,
No.1, pp. 37-55
Bar Avi. P (1999), Nonlinear Dynamic
Response of a Tension Leg Platform, Journal
of Offshore Mechanics and Arctic
Engineering, Vol. 121, pp. 219-226.
Chandrasekaran. S and Jain. A. K (2002),
Dynamic behavior of square and triangular
offshore tension leg platforms under regular
wave loads, Ocean Engineering, Vol. 29, pp.
279-313.
Triangular configuration Tension leg
platform behaviour under random sea wave
loads, Ocean Engineering, Vol. 29, pp. 18951928.
Aerodynamic behavior of Offshore triangular
Tension Leg Platforms, Proceedings of
fourteenth International Offshore and Polar
Engineering Conference, Toulon, France,
May 23-28, pp. 564-569.
Chandrasekaran. S, Jain. A. K, Chandak. N.
R (2007), Response behavior of triangular
tension leg platforms under regular waves
using Stokes nonlinear wave theory, Journal
of waterway, port, coastal and ocean
engineering, ASCE/May/June, pp. 230-237.
Chandrasekaran. S, Jain. A. K, Gupta. A and
Srivastava. A (2007), Response behavior of
triangular tension leg platforms under impact
loading, Ocean Engineering, Vol. 34, pp. 4553.
Chandrasekaran. S , Abhishek Sharma and
Shivam
Srivastava
(2007),
Offshore
triangular TLP behavior using dynamic
Morison equation, Journal of Structural
Engineering, Vol. 34, No. 4, pp. 291-296.
Chuel-Hyun Kim, Chang-HO Lee and JaSam Goo (2007), A dynamic response
analysis of tension leg platforms including
hydrodynamic interaction in regular waves,
Ocean Engineering, Vol. 34, pp. 1680-1689.
Ertas. A, Lee. J-H (1989), Stochastic
Response of Tension Leg Platform to Wave
and Current Forces, Journal of Energy
Resources Technology, Vol. 111, pp. 221230.
Ertas. A and Eskwaro-Osire. S (1991), Effect
of Damping and Wave Parameters on
Offshore Structure under Random Excitation,
Nonlinear Dynamics, Vol. 2, pp 119-136.
Haritos. N (1985), Modelling the response of
Tension Leg Platforms to the effects of wind
using simulated traces, Mathematics and
computers in simulation, Vol. 27, pp. 231240.
Inyeol Paik and Jose M. Roesset (1996), Use
of Quadratic Transfer Functions to Predict
[14].
[15].
[16].
[17].
[18].
[19].
[20].
[21].
[22].
Response of Tension Leg Platforms, Journal

of Engineering Mechanics, Vol. 122, No.9 pp
882-889.
Jain. A. K (1995), Nonlinear coupled
response of offshore TLP to regular waves,
Ocean Engineering, Vol. 24, No. 7, pp. 577592.
Jefferys. E. R. and Patel. M. H (1982)
Dynamic Analysis Models of Tension Leg
Platforms, Journal of energy Resources
Technology, Vol. 104, pp. 217- 223.
Kareem. A and Zhao. J (1994), Analysis of
Non-Gaussian Surge Response of Tension
Leg Platforms under Wind Loads, Journal of
Offshore Mechanics and Arctic Engineering,
Vol. 116, pp. 13-144.
Kobayashi. M, Shimada. K and Fujihira. T
(1987), Study on Dynamic Responses of a
TLP in Waves, Journal of Offshore
Mechanics and Arctic Engineering, Vol.109,
pp.-61-66.
Low. Y. M (2009), Frequency domain
analysis of a tension leg platform with
statistical linearization of the tendon
restoring forces, Marine structures, Vol. 22,
pp. 480-503.
Mekha B.Basim, Philip Johnson. C, Roesset.
M Jose (1996), Implication of Tendon
Modeling on Nonlinear Response of TLP,
Journal of Structural Engineering, Vol. 122,
No. 2, pp. 142-149.
Nordgren. R. P. (1987), Analysis of high
frequency vibration of tension leg platforms,
Journal of offshore mechanics and arctic
engineering, Vol. 109, pp. 119-125.
Spanos P. D and Agarwal V. K (1984),
Response of a Simple Tension Leg Platform
Model to Wave Forces Calculated at
Displaced Position, Journal of Energy
Resources Technology, Vol. 106, pp. 437443.
[23].
[24].
[25].
[26].
[27].
[28].
[29].
66
Tabeshpour. M. R, Golafshani. A. A and Seif.

M. S (2006), Comprehensive study on the
results of tension leg platform responses in
random sea, Journal of Zhejiang University
SCIENCE A, Vol. 7 (8), pp. 1305-1317.
Thiagarajan K. P., Troesch A. W (1998),
Effects of Appendages and small currents on
the hydro dynamic Heave Damping of TLP
Columns, Journal of Offshore Mechanics and
Arctic Engineering, Vol. 120, pp. 37-42.
White. N. Charles, Copple. W. Robert and
Cunyet Capanoglu (2005), Triceratops: An
effective platform for Developning Oil and
Gas fields in deep and ultra deep water,
Proceedings of the fifteenth International
Offshore and Polar Engineering Conference,
Seoul, Korea, June 19-24, pp 133-139.
William C. de Boom, Pinkster Jo. A and
Peter S. G. Tan (1984), Motion and tether
force prediction of a TLP, Journal of
Waterway, Port, Coastal and Ocean
Engineering, Vol.110, No. 4, pp.472-486.
Yan Fa-suo, Zhang Da-gang, Sun Li-Ping
and Dai Yang-shan (2009), Stress verification
of a TLP under extreme wave environment,
journal of Marine science applications, Vol.
8, pp. 132-136.
Yoneya. T and Yoshida. K (1982), The
Dynamics of Tension Leg Platforms in
Waves, Journal of Energy Resources
Technology, Vol. 104, pp. 20-28.
Yoshida. K., Ozaki. M. and Oka. N (1984),
Structural Response Analysis of Tension Leg
Platforms, Journal of Energy Resources, Vol.
106, pp. 10-17.
Zeng Xiao-hui, Shen Xiao-peng and Wu
Ying-xiang (2007), Governing equations and
numerical solutions of tension leg platform
with finite amplitude motion, Journal of
Applied Mathematics and Mechanics, Vol.
28 (1), pp 37-49.

PRELIMINARY FEASIBILITY STUDY AND DEVELOPMENT COASTAL REGION

OF AMBON CITY AS AMBON WATER FRONT CITY IN SUPPORTING
POTENTIALS AND REGIONAL COMPETITIVENESS
Pieter Berhitu1, Wilma Latuny2, Marthin Efruan3
1
Lecturer
Pattimura University, Ambon, Indonesia
Email: pietber @yahoo.com
2
Lecturer
Email:Wilma.latuny@advance-maluku.com
3
Lecturer
Email: efruanmarthin@ymail.com
ABSTRACT
Like most other coastal city in Indonesia, Ambon city have the potential of coastal areas with a long and
beautiful coastline. Today, the coastal area is less attention by government and investor in its development
Coastal area of Ambon is still considered a rear area of the city; have not seen yet as the front porch of the city.
Now, Coastal area of Ambon has grown and developed with various activities to fulfill the lives and livelihoods
of both Maluku and Ambon community. Interest to coastal areas caused by the government's awareness of the
existence of the "era of waterfront", seeing that the coastal has the potency, has a dynamic nature of
development and the opportunity for development. The wish to build and develop the coastal area of Ambon, in
order to maximize waterfront function as; the heart of the city of Ambon, the place services and trading center
and the veins of city movement. According to the Strategic Planning of Ambon city in 2006-2013 want that the
City Planning in related to the City Coastal Development will be directed towards the development of coastal
areas to the Beach City (Coastal City) or the Development of Ambon Water Front City. For this purpose, the
preliminary feasibility study and development of coastal areas of Ambon City are needed to bear the
construction and development of the city in supporting of potential and regional competitiveness. The effort to
formulate the concept of feasibility, planning and development need to be based on design principles, which put
to three main areas, namely: 1). Creation of image or identity of the coastal area, using a variety of
environmental characteristics. 2). Intensity restricted in coastal areas, to control development with
consideration of human values, environmental and development impact. 3). Restricted areas in the region, to
avoid land use conflicts of interest. From the results of this feasibility study as a whole aims to make a statement
about the zoning of coastal areas with a direction of development that is based on the character and typology of
the area and then making the Ambon coastal city development scenarios based on the structure and patterns of
land use, environmental strategic, prospect of economic development, supporting of accessibility and
infrastructure. Therefore, spatial planning, development control guidelines, economic development and
investment opportunities will be done related to potential areas and regional competitiveness. Based on the
results of feasibility analysis conclude that the coastal areas can be developed into Ambon waterfront City.
Key Word; Coastal Area, Ambon City, Water front City
1. INTRODUCTION
conflicts of interest. Having formulated the concept

then held the proposed zoning the waterfront area in
accordance with the direction of development that is
based on the character and typology of the area.
Subsequently made coastal development scenarios
based on Ambon city structure and land use patterns,
the strategic environment, economic development
opportunities, support infrastructure and accessibility.
After the scenario continued with the preparation of
spatial plans, guidelines and development control
Investment program. The purpose of this study is;
1. The creation of images (image) in Ambon as a
waterfront city that provides a visual character,
opinions, experiences, and overall impression of the
city.
2. Awakening access to waterfront areas and
provide assurance of achieving an easy, and
convenient
movement
of
population.
3. The realization of building orientation towards
In Ambon City Strategic Plan 2006 - 2013, wants

the City Planning in conjunction with the City Coastal
Development will be directed towards the
development of coastal areas to the Town Beach
(Coastal City) or the Water Front Development
Ambon City. For this reason the desire to realize the
strategic plan the city is done it is necessary studies
and plans the development of the collar of that goal.
In an effort to formulate a planning concept that it
must be based on design principles, which set three
main areas, namely: (1) Creation of image or identity
of the waterfront area, with the use of various
environmental characteristics of the area, (2) Intensity
in the region restricted the water's edge, to control
development with consideration of human values,
environmental and development impact and (3)
Restricted areas in the region, to avoid land use
67
the water's edge so it does not make the water's edge

as the backyard. 4. Establishment of a spatial planning
concept in Ambon after going through a review
proper of Ambon to Waterfront City. [2, 3]
include all activities in the field of maritime economic

development. Coastal urban population of at least
inhabited by 500 families with income levels ranging
between Rp. 1 million / household / month. Revenue
for the ideal spin in the coastal town because in it
there is a shopping center, education center (fisheries
education), banks (BPR), health facilities, recreation,
sport facilities, entertainment facilities, and others.
With the velocity of money for it, it is estimated the
city will grow and can enhance economic growth
area. In addition, local public savings through banks
or other financial institution in local level, dapat
membuat kota tersebut secara independent terus
karena growing availability of public capital access is
not sulit yang penduduk job Composition coastal
town is very situational but can be used persentase
follows. The arrest of around 40%, fisheries-based
industries around 25%, 10% farming, tourism, 5%,
5% of feed industry, transportation 5%, 5%, port
services and other work 5%. With the composition of
such work, will be able to create a beach town as a
new growth center in the hinterland areas which are
expected to be a triger to attract the surrounding area.
Coastal
town
regional
capability
as an attractor or as a prime mover of growth, because
growth of the fishing industry ranging from the
availability of vessels, fishing gear, ice plants, cold
storage, fish canneries, fishing port, packaging,
shipbuilding and workshop are available at the town
beach. Economic infrastructure as they grow and
develop, because the provision of public facilities
such as water, electricity, telephone and road access.
Economic and public infrastructure required in the
coastal town can be done by the elements involved
(stakeholders). For employers may conduct economic
infrastructure
investment,
while
government
involvement in the form of provision of public
facilities and community involvement in the form of
production [10, 15]
2. LITERATURE REVIEW
2.1 Beach Cities Development
Naturally, the town which has grown into a
modern metropolis and the city always start the
grandeur of the places are remote, then turned into a
place of concentration of people and activity as it is
considered strategic, the strategic places that usually
serve as a place of economic transaction. When a
place has evolved into a place of exchange (economic
transactions), then later it was followed by the
development of settlements in evolutive population.
Such an interest in long-term urban development, the
population will rapidly change to meet the standards
of individual and social needs towards a more rapid
and effective. It also encourages the growth of
modern industrial cities, the availability of supporting
infrastructure
such
as
electricity,
water,
telecommunications and services will grow along with
the social dynamics of society which leads to
maturity.
In economic review, acknowledged that the city
can be a place of growth (growth) because the
transaction of goods and services (goods and services)
are usually clustered in the city. But the population
has exceeded the limit, (Crowded) then the city will
experience growth diseconomy of scale, because its
development started because the negative impact of
uncontrolled urban dominant than the positive impact.
Therefore, by shifting the emergence of small towns
and medium enterprises in coastal areas, it is possible
to create new sources of growth, as the main trigger in
encouraging the development of remote areas. Cities
in coastal areas could allow for growth, provided
there is economic belt as the main buffer that city
growth. Coastal urban development are formulated as
a system of urban and regional development of the
islands, which shows the function of the hierarchy of
the city, the pattern of regional infrastructure,
including transportation, distribution infrastructure,
which refers to the geographical conditions of the
region and exploiting its natural resource potential,
both natural resources that can be recovered as well as
natural resources can not be recovered. Both are
managed sustainably. Management of maritime
resources, to remain sustainable, so in systems
development, taking into account sought wisdom
(wisdom), the local community. This is aligned to be
no gap between the native views of the newcomers
who will inhabit the coastal cities. [1, 3, 4, 16].
3. RESEARCH METHODOLOGY
3.1 Research Type
This descriptive research with observational
methods to get his primary and secondary data and
then reviewed to obtain the conclusions of the
Feasibility Study.
3.2 Study Location

Location of Study in general covers the coastal area
of Ambon city begins from the area in District
Nusaniwe Nusaniwe Village to Village House in
District Three Baguala Ambon Bay. Which consists
of three Districts, 14 villages and 12 villages. Study
sites can be seen in Figure 1
2.2 City Beach Center for Economic Activity

Coastal city is a center of economic activity,
industry, commerce, education, government, or
68
industry. Special development areas inland wake Kea

red / hills dominated for settlement area, while trade
in services activities located along the coast in
primary arterial road. Settlement area is partly
incompatible with the direction of the plan because it
uses the slopes of the hills with a slope of more than
30% and this will be harmful to the environment[2, 3,
10, 12]. Under the direction of urban development,
which is covered in RUTRK, the allocation of
development activities in Ambon City can be
described as follows: (a) Restricted physical
development at the City Center area, except for
services and trading activities and (b) the
development of industrial activities directed to the
Batu Gong is developed in accordance with the
existing potentials where industrial activity is
concentrated.
Development of settlement areas is directed
along the coastal village of Tawiri, Hative Large,
House Three, Poka, Durian Patah, Waiheru, Nania,
Old State, Passo and limiting residential development
in areas that have been densely crowded. [2, 3]
Figure 1. Study Location
4. DATA ANALYSIS
The data have been obtained and then further
analyze the collected include: Analysis of Physical
and non-physical, suitability and feasibility of land,
environmental carrying capacity, environmental
capacity resources, aspects of disaster in the coastal
area (tidal waves, tsunamis, landslides, abrasion) ,
Utilization of coastal space, infrastructure and
utilities, accessibility and transport, population and
socio-cultural, economic and economic development
opportunities, environmental conditions and strategic
opportunities for growth and development of the
region, policies and related regulations.
5.2 Function and Role of Regions Observation

in Ambon City Area Scope
Observation area that extends lengthwise from
Castle Village to Village House Three very important
for overall Ambonsecara City area. Functions
contained throughout this region in accordance with
the direction of development set out in Ambon RUTK
are as follows:
Higher Education Center
Local and Regional Trade
Research and development of marine, agriculture,
fisheries and shipping
Manufacture of large, medium and small
Government district, village and town
Defence and Security (military)
Settlement and Housing
Ports and major transport chain
With the function of this region may play a role and
become an attraction for urban activities such as
commerce, offices, education and housing. On the
other hand there are areas that need to be protected
and preserved its existence.[2,3,10,11,12,13,15].

5.1 Area Studies Are Planning and
Determination
Ambon City Area 359.45 km2 with a total area of
32.4 km2 awake or 8.6% (based on the Appropriation
and Use Plan Land Inventory Book II BPN Province
1996). However, most areas of this city can not be
utilized for the region woke up because one of the
problems is the topography of the area that is not
possible. According to the General Plan Regional
Spatial policy Ambon physical expansion of urban
activity directed to the underdeveloped regions, is
intended to be evenly distributed throughout the
development of urban areas. Based on land suitability
and potential areas suitable for development in
general is towards the East. Direction of physical
development of Ambon City currently can be seen
from the trend of development of the area of Ambon
city awoke [2, 3, 6, 8, 9, 10, 11, 12]. From field
observations and maps of existing land use, it turns
out that the development of the region woke up
quickly lead to the east from the Red Rock area
Galala red Kea, Passo and Pokka follow the road
network along the coast. The development area is
located Kea awoke this anger among other settlement
activities, governance, trade and services, and
6. AS A COASTAL DEVELOPMENT
AMBON CITY BEACH (WATER FRONT
CITY)
6.1 Results and Zoning Ambon City Coastal
Development
In Analysis and Development Study of Ambon City
Beach As may be realized if the development should
maintain the functioning of protected areas, a)
Securing / do not disturb the area of strategic
installations, b) Harnessing the potential of diversity
69
to supportt coastal dev

velopment as: Services andd
trading area.
a
Area settlements, tourism areaa,
residential area, educatioon area, sea traansport and porrt
area, service area and acccessibility. c)) Balancing thhe
developmeent of the moddern sector - trraditional, withh
preserving
the
locaal
culture,
d)
Utilizze
environmeentally-friendlyy
technologiies
in
thhe
settlement of physical development.
d
e) To preparre
evacuationn routes in a state of dannger of naturaal
disasters (ttsunamis, tidall waves) [6, 8, 9, 10, 11, 122,
13, 14, 15]].
Figu
ure 5. Settlemeent Patterns and distribution of
City
y Regions
Figu
ure. 2 Distributtion of Segmennt Planning
F
Figure.
6 Land Use Eksiting
6.2. Zoning of Coastaal Development Potentiall

In the study of cooastal develop
pment planninng
C
Beach ass a substance divided by thhe
Ambon City
Developm
ment zoning based
b
on carry
ying capacity of
coastal sppaces and can be viewed as follows (Figuure
7);
Figuree. 3 Characteriistics of Spatial Use

Segmentss I, II and III
Figure 7. Potenntial Developmeent Area
Figuure. 4 Characteeristics of Spatiial Use

Segmen
nts VI and VIII
70
manage a gradual and ccontinuous so as to supportinng

developm
ment and providde kesejahtraan
n together for thhe
communitty
6.3 Spatiaal Planning of

o Coastal Deevelopment
Under the
t plan the structure
s
of sppace in Ambonn
City Amboon RTRW Yeear 2012, which is a furtheer
translation of Ambon ciity RENSTRA
A years 2006 too
2013 that divides
d
the areea of Ambon city
c be 4 (fourr)
Unit
Developmentt
Areas,
namelyy:
Table 1. Plannning Area delineeation
1. SWP 1,
1 sub-area reffinery to the city
c
center and
service
center
at
City
y
Centerr.
I sub-region with Passo and Hutumurri
2. SWP II,
Service
Passoo.
center
at
3. SWP III, sub-region and
a Laha Threee Houses withh
service centers in the Housee of Threee.
4. SWP IV
V, sub Latuhalaat and surround
ding areas withh
service cen
nters in Amahuusu [2, 3, 15].
Figuree. 8 Structure of space utilizationn planning
Area deelineation
Seegment
Regions Nusaniwe , B
Benteng,
u, dan Waihaoong
Wainitu
Regions Honipopu,, Rijali,

dan Batu Merah
III
Pandan
n
Kasturi
Area,
Galala,, dan Small Haative
IIII
Halongg Area
IV
V
Latta dan
d Lateri Regions
Passo Regions
R
V
VI
Negeri Lama, Nan

nia, dan
Waiherru Regions
V
VII
Regions Hunuth
V
VIII
Regions Tihu, Pok

ka, dan
h Tiga
Rumah
IX
X
S
W
P
II
IIII
7. CONC
CLUSION
REFERE
ENCES
Based on feasibility stuudy developm

ment of coastaal
areas of Ambon
A
city ass the water frront, it can be
concluded as follows: 1).Developmeent of coastaal
areas of Ambon
A
as a water
w
front cityy in accordance
with the su
ubstance of thiis study is feassible with Buillt
Area Devvelopment Trends
T
memp
pertimbangakann
Ambon, Function annd Role in the Regionn
Observatioons Scope of Ambon
A
City Area,
A
Capabilityy
and Suitab
bility and Capaacity of Land, Resources andd
Populationn Culture Social Support, and advancce
planning is the zoning of the coastall area of space
developmeent in accordance with the direction oof
developingg spatial planss. 2). In the framework
f
andd
efforts to manage
m
naturall resources susttainable coastaal
and marinee in the city of Ambon and simultaneouslyy
support thee potential andd competitivenness of the areea
hope the Government
G
thhen in the areaa, the City andd
County shhould work together and bee able to forgge
close ties with both parties
p
diantarra intitusi and
private secctors profesionnalime by form
ming a body oof
work that is able to do the coordinatin
ng, structuringg,
evaluation,, and preparatiion of program
ms and policiees
that are baased on the natiional and regioonal interests inn
professionaalism so that natural resourrces be able too
[1] Anonnimous, Presiddential Decreee No.62 of 20000

on the Coordinnation of National
N
Spatiial
Plan
nning
[2] Anonnimous, Malukku in Figures (2009)
[3] Anonnimous Ambbon City Yearr Strategic Plaan
20066-2012 (2001))
[4] Komaar PD, Beachh Process and Sedimentationn
Schoool of oceanoggrafy, Oregon State
S
Universitty,
New
w Jersey (1976))
[5] Komaar PD, Comp
puter Models of Shore Linne
Channge "CRC, Haardbook of costal process annd
erotiion, inc Boca R
Raton, Ireland (1984)
[6] Pieterr Th Berhitu, M
Model Study of Beach
B
Erosion annd
Sedim
mentation Ambbon Bay, Accredited Scientiffic
Magazine Vol 15 Noo. 3 pp 61-67 (20
003)
[8] Pieteer Th Berhiitu 2004; D
Damage Studiies
Manngrove habitat in the Gulf Coastal
C
Lateri In
Ambo
on, Proceedinggs of the Natiional Seminar on
Theoory and Applicattion of Marine Technology,
T
ISS
SN
1412
2-2332 pp 197 -2201(2004).
[9] Pieteer Th Berhitu F
Forecasting Stud
dy of Wind Wavves
Due Out In Ambon B
Bay, Proceedin
ngs of the Instituute
for Research
R
Unpattii 2245 ISBN No.37 (2005)
71
[10] Pieter Th Berhitu, Analysis of Coastal Land Use and

Spatial Effects on Ambon, Journal of Regional
Planning, (2006)
[11] Pieter Th Berhitu , Shoreline Damage Studies
Amahusu - Eri Outer Bay of Ambon and Spatial
Effects on Ambon, Proceedings of the National
Seminar on Theory and Application of Marine
Technology pp D151-169 (2007).
[12] Pieter Th Berhitu 2008, "Friction Studies of Coastal
Areas To Funtion And The Influence Of Regional
Planalogy Amboiana City", paper presented at the
international seminar MARTEC 2008 at the
University of Indonesia, (2008).
[13] Pieter Th Berhitu, Studies of Coastal Damage To Sub
Nusaniwe Ambon City Spatial Planning, Technology
Journal Vol 5 No 2 Faculty of Engineering Pattimura
University, pp 209-213 (2008)
[14] Pieter Th Berhitu"Analysis of Physical Damage to

Coastal Areas-Liliboy Hatu District of Central
Maluku district Leihitu For
Coastal Spatial
Planning, Technology Journal Vol 5 No October 2,
Faculty of Engineering Pattimura University, (2008)
[15] Pieter Th Berhitu 2009 "Experience Resource
Managament Regional Concept Of Coastal Areas And
Go To Sea Inwroughtly out Parade In Town At
Suporting Area Competitiveness And Potenscy.
Proceedings of the "National Seminar on Theory and
Application of Marine Technology. Faculty of Marine
Technology ITS SurabayaISSN 1412-2332, December
(2009)
[16] Sudharto P. Hadi, Environmental Dimensions of
Development Planning, Gajah Mada University
Press, (2001).
72

MODELING OF GAS TURBINE CO-PROPULSION ENGINE FOR ECOTOURISM

VESSEL TO IMPROVE THE SAILING SPEED
O. Sulaiman1, O.S. Han2, W.B. Wan Nik3, A. H. Saharuddin4, A.S.A.kader5, M.F. Ahmad6
1,2,3,4,6
Department of Maritime Technology, Faculty of Maritime
Studies and Marine Science, University Malaysia Terengganu,
Terengganu, Malaysia
5
Department of Mechanical Engineering,
Technology Malaysia, Skudai, Johor. Malaysia
E-mail: absaman@fkm.edu.my
University
E-mail: o.sulaiman@umt.edu.my
ABSTRACT
Sailing speed is an important factor in choosing marine engines. The uses of gas turbine as co-propulsion
engine for improving sailing speed of ecotourism vessels are to fulfill requirement of SAR operation. Gas
turbine co-propulsion engine has an advantage of high power to weight ratio in comparison to other heat
engines. This paper presents the study on diesel engine, simple cycle gas turbine and regenerative gas turbine
performances. The relation between the thermal efficiency of heat engine and fuel consumption is used to
estimate fuel consumption rate. The design of heat engine can be determined by the specific heat ratio and
pressure ratio of the operation cycle which will give necessary impacts to the thermal efficiency of the heat
engine. Results from the numerical calculation for the implementation of gas turbine will provide the decision
support. The paper also discusses the impact of co-propulsion engine to the ships stability and proper power
rating of gas turbine co-propulsion engine estimated by numerical calculation in order to achieve maximum
sailing speed up to 35 knots.
Key words: Gas turbine, regeneration, sailing speed, thermal efficiency, fuel consumption
1. INTRODUCTION
Comparative to other heat engines, gas turbine will

have the advantages of high power to weight ratio.
The gas turbine provides the same output power as
the diesel engine having more compact design and
smaller in size and weight in comparison with diesel
engine. However, under certain circumstances, the
diesel engine will show higher in fuel efficiency in
comparative to gas turbine [1,2].
The sailing speed of ecotourism vessel can be

improved by several methods. In this research
implementation of gas turbine is proposed as copropulsion system such that the speed of the
ecotourism vessels sailing speed increases up to
35knots. The vessel under study is an important
transport connecting the mainland from Mersing jetty
to Tioman Island. High speed sailing is necessary for
the vessel to carry out the search and rescue operation
in open sea under emergency circumstances. Besides,
improving passenger ferry sailing speed will
overcome the problem of vessels shiftment delays
during peak season in May.
The design of gas turbine will give impacts to the

performances of gas turbine. Design must take into
account on specific heat ratio and the pressure ratio in
order to produce a high performance gas turbine copropulsion engine. These two variables give
significant change of thermal efficiency of gas turbine
co-propulsion engine.
Gas turbine also called a combustion turbine is a

rotary engine that extracts energy from a flow of
combustion gas. In order to adapt the function, the gas
turbine compose of four important components,
which are compressor, combustion chamber, turbines
and exhaust. Energy is added to gas stream by
combustor through ignition of the mixture of
atomized fuel and air. The gaseous streams are then
directed through a nozzle toward a turbine. The hot
gases stream will spin the turbine and empower the
compressor.
2.
MODELLING PROCESS
The thermodynamic properties of each heat engine

were emphasized three types of heat engine were
selected and put into study. Diesel cycle, Bryton cycle
and combine cycle are studied to examine the
properties of individual heat engine.
73
Table 2. M-code for Diesel engine

%M
code for thermodynamic properties of diesel engine
k=1.4;
r=[2:2:24];
rco=2;
a=r.^(k
1);
b=(rco^k)
1;
c=k*(rco
1);
e=b./(a*c);
nD=1
e;
plot(r,nD,'red');
legend('at
k=1.4');
xlabel('compress
ratio,r');
ylabel('Dieselefficiency,nD');
title('thermal efficiency vs comprssion ratio');
Table 3. M-code for simple cycle gas turbine

%M-code for thermodynamic properties of simple gas turbine
k=1.4;
rp=[1:2:24];
x=(k-1)/k;
nB=1-rp.^-xplot(rp,nB,'magenta');
xlabel('pressure
ratio,rp');
ylabel('Thermal
efficiency,nB');
legend('at
k=1.4');
title('Thermal efficiency of gas turbine vs pressure ratio');
Fig 1: thermodynamic properties of co-propulsion

engine
A survey is done by visiting the passenger Fast Ferry
Company located at Mersing jetty, Johor. Data
collections are done on the vessel understudy. These
included ships particular general arrangement,
propulsion engine specification, sailing speeds and
fuel consumption rate [3,4].
Table 4. M-code for regeneration gas turbine
The thermodynamic properties of the following heat

engine were presented in curve to examine the
properties of each heat engine. The plotting tools;
Mathlab is applied for plotting purposes. The
thermodynamic formula needs to translate to the Mcode in order to present a relation curve [5,6]. Table
1-4 show the M-code for propulsion engines.
%M-code for thermodynamic properties of regeneration gas turbine

k=1.4;
r=[1:2:50];
t=0.5;
x=(k-1)/k;
nR=1-t*(r.^x);
plot(r,nR,'black');
xlabel('pressure ratio,r');
ylabel('Thermal efficiency,n');
legend(regen turbine at k=1.4);
Table1: Thermal efficiency of co-propulsion engine

Types of copropulsion
engine
Thermal efficiency formula
title(regen
efficiency);
gas
turbine
thermal
The fuel consumption rate of each heat engine will

then translate from the thermal efficiency using
formula stated below:
Diesel
engine
Simple cycle
Gas turbine
After determining the types of co-propulsion

engine to implement, the power output selection can
be performed by numerical calculation by using the
related formula follow the sequence as shown in
Figure 2.
Regeneratio
n gas turbine
74
services. Besides, the company owned a few cargo

ships modified from the ordinary fisherman boat to
transfer cargoes in between Pulau Tioman to the
mainland to full fill the demands in the Tioman
Island. (Refer to Table 6, and Table 7).
Table 6. Details of the ferry company

1
Company name:
BlueWater Express
Mersing,Johor
Location:
Name
of
Company's owner:
En.Rizam Bin Ali
Types of business:
passenger fast ferry
Routine:
Mersing jetty to Tioman
Figure 2. Arrangement of machinery onboard ship
Distance:
35n miles
Table5: Formula for numarical power calculation
Name of Vessel:
Gemilang 1
Types of vessel:
Fiber single hull vessel

PT.Bintan
Shipping
Bioteknik Tanjung Pinang
Shipyard
Types of power
Simplified
Formula
Effective
horse
power (EHP)
EHP=RV
Effective horse
power required
to tow a hull
without
a
propeller.
thrust
horsepower is
power delivered
by propeller to
the water
Delivered
horsepower
(DHP) is the
power that is
delivered by the
shaft to the
propeller.
Shaft
horse
power is the
power delivered
by engine to the
shaft
after
gearing
and
thrust bearing.
The
power
delivered by the
prime mover at
its connection
flange is called
brake
horsepower.
Thrust
(THP)
horsepower
Delivered
Horsepower
Shaft horsepower
Brake horse power
9
10
Vessel
Manfacturer:
Maximum number
passenger:
100 passengers
Table 7. Ship Hull details

1
Length overall
23.7m
beam
5.20m
draft
2.20m
Hull types:
Single hull
Materials
Detail of lighting
system
Number of pump
required
Vessel
average
sailing speed
DWT(base
on
dimension)
Fiber class
12fluerecent light(45watt),12
others light bulb(24watt)
6 batt pump,1 electrical pump,2
mechanical pump(ramp pump)
6
7
8
9
20knots
271.2tonnes
The vessel under study is named Gemilang 1.

The vessel was constructed by PT. Bintan Shipping
Bioteknik in Tanjung Pinang, Indonesia. The ship hull
has the dimension as shown in the Table. The
machinery used onboard was drafted for the reason of
recommendation on the implementation of copropulsion engine ( See Figure 3 and 4 ).

The data acquired during the survey are presented in
tablelar form. The table 6 below shows the detail of
the passenger ferry company. Bluewater Express ferry
services are a company established in 1999. The core
business offered are the ferry services to the
passengers coming to Pulau Tioman. The company
currently owned 8 fast ferries for the passenger
75
Figure 3. Machinery arrangement in Gemilang 1
Figure
engine room
arrangement plan)
5.
GA-plan
of
Gemilang
(general
Gemilang1 is equipped with 2 propulsion engines

and capable to propel the ship at sailing speed up to
20 knots. In order to fulfill the demand of the
machinery that is necessary for sailing the desired
sailing speed, redundant systems such as water pump
system, lighting system, and the electronics devices
for the navigational purposes are installed. The vessel
is also equipped with a generator power rating up to
50kW. The properties of the propulsion engine are
presented in the Table 8 below.
Table 8: Propulsion engine specification
Figure 4. Mechanism of CI diesel propulsion engine

in Gemilang 1
In predicting the efficiency of gas turbine and
diesel engine, assumption has been made in order to
standardize the condition at which the cycles
performed. In comparing the performance of gas
turbine versus diesel engine; we need to make some
assumption on the working fluid for both of the
system. The air is necessary in carrying out the
combustion process. Fresh air entering the
combustion chamber was considered under the cold
76
air standard assumptions. Where by the specific heat

ratio k, is represented by k=1.4 (specific heat ratio
value under room temperature) (See Figure 6).
Figure 7. Thermal efficiency of each types copropulsion engine

The performance of gas turbine and diesel engine
overlaps at pressure ratio 5. In the middle state of the
curve, the diesel engine has higher thermal efficiency
with the increasing pressure ratio. From the curve
shown, it is observed that the simple cycle gas turbine
engine is less efficient in comparison to diesel engine.
Figure 6. Thermal efficiency of each types co-
The gas turbine operation can be improved by

applying the regeneration cycle. The temperature of
exhaust gas leaving the turbine is higher than the
temperature of the air leaving the compressor. By
leading the heat exhaust gaseous through the heat
recuperates to preheat the compressed air from the
compressor can improve the thermal efficiency of the
gas turbine. Figure 8 shows the thermal efficiency
curve between diesel engine, simple cycle gas turbine
and the regenerative gas turbine at variable
temperature ratio [6,7].
propulsion engine
Besides, assumption is made on that the pressure

ratio and cutoff ratio are similar in term of working
condition. Compression ratio r, is defined as the ratio
of the volume of its combustion chamber; from its
largest capacity to its smallest capacity. It is a
fundamental specification for many common
combustion engines. While the pressure ratio , for
Figure 8 illustrates the regenerative gas turbine with

the minimum temperature ratio between the exhaust
gas and the compressed air shows higher thermal
efficiency in the early stage, the thermal efficiency of
the following gas turbine decrease gradually with the
increasing of the pressure ratio. From the diagram it
is that the gas turbine withregenraton is the ideal
selection for the co-propulsion engine because it
shows high thermal efficiency in low pressure ratio.
Low pressure ratio carries significant information of
low back work ratio and horse power of the following
engine.
gas turbine is defined as ratio of the pressure at the

core engine exhaust and fan discharge pressure
compared to the intake pressure to the gas turbine
engine.
Where:
The study involved the feasibility of implementing a

gas turbine to improve the vessel sailing speed up to
35 knots. The study relates the operation of the gas
turbine and diesel engine with thermal efficiency of
the cycle.
Figure 7 illustrates the relation between thermal
efficiency with pressure ratio for simple cycle of gas
turbine and diesel engine. It shows that at the early
state of the curve, gas turbine show steeper increment
in thermal efficiency with the increasing pressure
ratio.
Figure 8. Thermal efficiency of each type of copropulsion engine
77
3.1 Fuel consumption of co-propulsion

engine
The fuel consumption criteria to consider
during the marine engine selection, the optimum
usage of fuel will ensure profits to the passenger
company. By using the formula as stated below, the
thermal efficiency of co-propulsion engine can be
interrelated.
In this case, we select the diesel fuel as the source for

the heat engine. The diesel fuel having the net heating
value of 130000btu/gallon. Substitute the net heating
value, then the fuel consumption rate can be
represented by the curve plotted by MATLAB as
shown in Figure 9.
Figure 10. Recommendation for distribute the copropulsion engine
3.2
Power calculation
The current existing diesel engines remain as a main

propulsion engine to sail the ship at economic speed.
The minimum power required to propel the vessel is
computed theoretically. The numerical modeling
involve assumption on the numbers of crews, weight
of luggage and cargoes carried to estimate the dead
weight tones of the vessel under studies. The brake
horse power obtained base on theoretical calculation
at variable speed is shown in the Table 9.
Figure 9. Fuel consumption (gallon) of each type of

co-propulsion engine:
From the curve shown in Figure 9 it can be
concluded that the fuel consumption rate versus
estimated ratio for the diesel engine, simple cycle gas
turbine, and regenerative gas turbine shows the
similar trend. The regeneration gas turbine with the
temperature ratio 0.3 showing a moderate fuel spent
over the power production. Hence, the regenerative
gas turbine will be the ideal selection as co-propulsion
engine among the others. Location for the
regeneration gas turbine in engine room [8].
Figure 11. Brake horse power at various speeds

Table 9. Validation on the power calculation
78
From the diagram the power rating of the diesel

engine calculated based on theoretical calculation are
closed to the diesel engines currently applied on that
following vessel. From the survey, we knew that there
were 2 diesel engines with power rating of 700hp
each applied on the vessel to propel the vessel to sail
at optimum speed. On economy aspect, the selection
of higher horsepower propulsion engines is necessary
for the vessel to sail at optimum speed instead of
sailing a vessel with full speed at engine maximum
performances. The speed control can be done by
adjusting on the throttling valve located at fuel pump
attached to diesel engine. Besides, in real
environment, there are some other factors to take into
consideration. Air resistant due to the size of the
superstructure of the vessel may require higher power
propulsion to propel the vessel to move forward.
The result from the numerical calculation on the

power output of co-propulsion engine shows that the
minimum brake horsepower required for the copropulsion engine is 1274.85hp.The recommended
horse power for co-propulsion to implement is
1300hp.That is 10% margin of power excess to suit
the speed of vessel. The regeneration gas turbine is
selected after performing the analysis by plotting
curve. The exhaust gas released by regeneration gas
turbine was retracted and used to reheat the
compressed gas exiting from the compressor.
4. CONCLUSION
The paper proposed to improve ecotourism vessel
sailing speed by implementing a gas turbine as copropulsion engine. The study of feasibility of
implementing a gas turbine as co-propulsion engine
relates the performances of the gas turbine to the
thermal efficiency and fuel consumption. The
objective of this research is to improve the speed of
vessel up to 40 knots with minimum fuel
consumption.
For the co-propulsion engine, the output power

becomes the terminology chosen in selecting the
marine engine. Referring to the curve shown, the
resistances of the vessel differ at variable speed.
Hence, numerical calculation on the power at various
speeds is necessary in order to ensure the vessel can
sail at desired speed (See Figure 12).
REFERENCES
[1]
Boyce, M.P.,Cogeneration and combined Cycle

Power Plant, Chapter 1, ASME
Press, NY, 2001
[2]
Cogeneration and Combined Cycle Power Plant,

Chapter 1, Gas turbine (third edition)
Engineering Handbook,Boyce M.P Meherwan
P.Boyce,Phd. And Pe
[3]
Enchancement of performances of gas turbine

engines by inlet air cooling and
Cogeneration
system,
Yousef
S.Najjar,
Mechanical Engineering department (Thermal),
King Abdulaziz University
[4]
Farmer, R., Design 60% Net Efficiency in

Frame 7/9H Steam Cooled CCGTGas turbine
World,May-June 1995
Gas turbine (third edition) Engineering
Handbook, Meherwan P.Boyce, Phd. And Pe
Marine Engineering, Roy L. Harrigton, Asistant
manager, Engineering Technical
Department, New port News shipbuilding
Figure 12. Types of resistances at variable speed

Figure13 shows the types of power at various speed.
The minimum power required to propel a vessel to
move forward is the effective horse power. The brake
horsepower is the highest power and will encounter
power loss in each transition state from the engine to
the shaft following with the propeller.
[5]
[6]
and Dry Dock Company

[7]
Thermodynamic optimiation of the reheat

chemically recuparated gas turbine,Roberto
Carapellucci,Adriano Milazzo,Department of
Engineering University of LAquila,Italy
[8]
Thermodynamic,an engineering approach six

edition,Yunus A . cengel,university of
nevada,Reno,Michael A.Boles,University of
north Carolins state www.wikipedia.com/gas
turbine
Figure 13. Types of power at variable speed
79

80

QUALITATIVE METHOD FOR ANTIFOULING LONG LIFE PAINT FOR

MARINE FACILITIES OR SYSTEM
O. Sulaiman1, C.M.Sam 2, W.B. Wan Nik4, A.S.A.kader5, A. H. Saharuddin6
1,2,3,4,6

5
Department of Mechanical Engineerng, University Technology
Malaysia, Skudai, Johor. Malaysia
ABSTRACT
Ship Hull fouling reduces the fuel efficiency and speed of affected ships, increase in frequency of ship dry
dockings, reduces propeller efficiency and accelerated corrosion rate. Antifouling paints are used to coat the
underwater area of ships to prevent organisms such as algae and molluscs attaching themselves to the hull of
the ship. As a result, antifouling paints which are alternatives to TBT systems such as controlled depletion
systems (CDPs), tin-free self-polishing copolymers (tin-free SPCs) and foul release systems were developed in
marine industry. However, most of these paints cannot last for long because misapplication. Paint is not a
finished product until it has been applied and dried on an appropriate substrate at the designed performance
film thickness. High performance paint systems are especially sensitive to misapplication and knowledge of the
paint characteristics. Also recommended film thickness is vital to obtain optimum results to improve paints
performance and reduce maintenance cost. Therefore, proper application is critical to the performance of the
paint system. This paper presents the result of study made on the problems of fouling on ship hull structures and
deduced a qualitative model for ship paints application in order to prolong the life span of antifouling paint.
Key words: antifouling, application, hull fouling, paint, qualitative
of affected ships, consequently increases their

operating costs due to the increase in frequency of
ship dry docking. It also reduces propeller efficiency
[2,3] and accelerates corrosion [1].
1. INTRODUCTION
Performance of ships depends on their speed and
power generated. Likewise, economic and efficiency
of ship operation is achieved at optimum speed,
power and according to frequency of dry docking of
ship. Hull fouling poses a lot of hindrances to design
power and efficiency of ship. The fouling material
can consist of either living organisms (biofouling) or a
non-living substance (inorganic or organic).
The new IMO convention defines antifouling

systems as a coating, paint, surface treatment,
surface or device that is used on a ship to control or
prevent attachment of unwanted organism. Biocidal
anti-fouling paints have been applied to the bottoms
of ships for decades. The paints slowly leach into the
water, killing anything attached to the ship hull, but
leachates have been found to accumulate in harbors
and the sea.
Marine fouling is a perennial problem for vessels,

ports and anything kept in the sea for a period of time.
The sea is teeming with the tiny larvae of marine
organisms that swim around until they find
somewhere to settle and grow. Smooth surfaces are
particularly attractive to many of these creatures, and
are quickly encrusted. This slow down ships in
seaways, block pipes and speed up corrosion. This
study focuses on the problems of ship hull fouling
which hull fouling reduces fuel efficiency and speed
Among all the different solutions proposed

throughout the history of navigation, tributyltin (TBT)
paints have been one of the most effective deterrents
to hull fouling organisms, but studies have linked
TBT accumulations to deformations in oysters and
sex changes in whelks. As a result, restrictions on the
81
This study seeks to examine the related issues of

antifouling (AF) paints, ship paints application and
aims to deduce fouling prevention systems and
enhance the performance of antifouling paints. This
includes the study of biology of the fouling process,
historical development of AF paints and also the
proper way for ship paints application.
usage of TBT in vessels were imposed. The

International Maritime Organization (IMO) adopted
the Antifouling Systems (AFS) Convention in 2001.
The Convention called for a global ban on the
application of TBT-based antifouling paints by 1st
January 2003 and the prohibition of the presence of
such paints on the surface of vessels by 1st January
2008 (IMO, 2005). National bans on the use of the
TBT paint will result in an increased hull fouling,
unless environmentally friendly replacement paints
are accepted by the shipping industry.
2. QUALITATATIVE APPROACH
The model design is for the ship paint application
procedur produced through interview carried out with
the ship yard personal in MMHE and M-Set. Data are
collected from Painting and Blasting Department,
reviewed about ship painting process and interview
with the Classification Society and Paint Maker in
order to get the further information. This is to make
sure the procedure is compliance with the standard
and IMO requirement. Data is analysed by
considering the whole ship painting process and how
the ship painting procedure is carried out according to
the standard. And the analysis leads to deduce a
qualitative model for ship paint application procedure
[2].
The paint industry has been urged to develop

environmental friendly TBT-free products able to
replace the TBT-based paint that yield the same
economic benefits and cause less harmful effects on
the environment. There are controlled depletion
systems (CDPs) and tin-free self-polishing
copolymers (tin-free SPCs). CDPs are upgrade of
traditional soluble matrix technology by means of
modern reinforcing resins. The reaction mechanisms
are assumed to be equivalent to those of conventional
resin-based AF paints. The tin-free SPCs are designed
for the same reaction mechanisms with sea water as
tributyltin self-polishing (TBT-SPC) paints.
Paint is not a finished product until it has been
applied and dried on an appropriate substrate at the
designed performance film thickness. When the paint
is applied to the exterior layer to a ship hull, it is
subject to a variety of parameters that can degrade the
paint and reduce its useful life-span. These parameters
need to be taken into consideration during ship paints
application. Thus, ship paints application procedures
are very important in order to enhance the
performance or quality of antifouling paints. The
parameters which need to be considered during
application of the paints to the ship hull include
surface preparation, paint application, paint materials,
curing time, environmental conditions, locations,
personal quality, inspections and others.
3. RESULT AND DISCUSION

The model is produced with the intention of giving
a guideline for all level of personnel on the standard
of workmanship in the ship repair division; especially,
the blasting and painting parts in order to satisfy ship
owners and classification societies. This model
indicates the elements accuracy to be kept in the
process of blasting and painting repairs or
modifications and the finished quality obtained.
The model is developed from several references,
historical data and case study related to antifouling
paints. The model can serve as a guideline of the
standard of workmanship for painting process that
mitigates fouling of ships hull. . The quality of the
end product relies on the whole ship paint application
process. Thus, we must always keep in mind that
Quality is built in the process, not in the inspection
(See Figure 1)
The performance of any paint coating depends on

the correct and thorough preparation of the surface
prior to coating. The most expensive and
technologically advanced coating system will fail if
the surface preparation is incorrect or incomplete.
Additionally, methods of applying the paints are by
brush, roller, conventional (air) spray, conventional
(pressure pot) spray and airless spray. Although the
application methods are very important, the
application technique or skills of personnel also play a
vital role. When applying marine paints, the most
important factors to consider are the condition of the
substrate, the surface temperature, and the
atmospheric conditions at the time of painting.
Appropriate ship paint materials can effectively
prevent attachment or accumulation of fouling on the
ship hull bottom. Furthermore, inspection by the
coating inspector is necessary to make sure the
coating is properly applied.
From the existing ship paint application above, the

gaps to improve the existing procedure is deduced.
The improved ship paint application flow chart is
showed in section 3.2. This flow chart can make the
paint application becomes more efficient (See Figure
1).
82
3.3.
UNDERWATER
HULL
CLEANING ASSESSMENT
PRE-
The step for the inspection process is to conduct an

underwater assessment of the fouling growth that has
occurred since the last inspection and evaluate the
coating condition. This will be completed before any
hull cleaning is performed. Normally, ship hull can be
divided into 6 quadrants as showed in Figure 3. The
six quadrants are: I - starboard forward, II - starboard
aft, III - port aft, IV - port forward, V - starboard
waterline, and VI - port waterline.
Figure 1. Existing flowchart for ship paint

application
3.1.
Improved flow chart for ship paint
application
Figure 3: Hull quadrants (IRTA, 2009)
3.3.1.
Fouling Assessment
Fouling growth on each boat hull will be evaluated

on a 0 5 scale. 0 represents the optimal condition
and 5 the worst condition. Table 1 determines the
numeric ratings and provides a description of what
type of fouling growth is associated with each rating.
The paint makers inspector will record the fouling
rating for each quadrant and provide any additional
observations or comments, such as noting the type of
fouling present on the hull surface.
Table 1. Fouling rating scale
Rating
0
1
2
Figure 2. Improved flowchart for ship paint

application
3.2.
VESSEL COMES INTO DOCK
Since it is a class requirement, for each vessel needs

to come into dock to undergo the bottom survey
inspection of its underwater area every two and half
years. Vessel will be put in the dry dock upon arrival
in the shipyard. The vessel will be moored into the
dock and when it had successfully sat on the keel
blocks, the water in the dock will be pumped dry[4].
Fouling Growth
No silting, biofilm or fouling growth present.
Light silting or biofilm. Little to no discoloration; Paint
surface still clearly visible beneath.
Heavy biofilm; Light to moderate silting as indicated by
discoloration (a solid, discernible, physical layer);
Painted surface may be slightly obscured.
Low to medium levels of fouling present; Dark algae
impregnation; Hard growth may be present (tubeworms,
barnacles, bryozoans, etc.); Painted surface definitely
obscured.
Medium to high levels of fouling present; Hard growth
present, such as tubeworms, barnacles, bryozoans, etc.;
Macrofoulers may include mature forms that may be
densely grouped; Paint surface no longer visible beneath
fouling in areas.
High levels of fouling present; Lengthy, soft algae and
hard, tube worms and possibly barnacles impregnating
the coatings; Macrofoulers may be densely grouped;
Coral** growth can be seen to extend out from the hull;
Paint surface no longer visible beneath fouling.
*0 is best condition; 5 is worst condition; ** Coral is

the local term used for limestone tubes of worms that
grow on the coatings surface.
83
3.3.2.
Coating condition assessment
Hull cleaning standard by fresh water
Coating condition for the entire hull need to be

evaluated based on Table 2 which identifies the rating
scale of coating condition. The colour of undercoat
also need to be recorded when the coating was applied
to the ship hull. Ratings of 1-3 represent antifouling
painted surface appearance associated with normal
physical wear due to underwater cleaning action or
hydrodynamic effects. Ratings 4 and 5 indicate either
excessive cleaning actions or blistering due to internal
failure of the paint system.
Surface preparation by using fresh water can be

divided into 4 levels. Table 3 is the levels or
categories for fresh water surface preparation:
Table 3. Categories for fresh water surface
preparation
Fresh Water
Washing/Pressure
Low Pressure Water
Washing
Pressure:
Less than 68 bar (1000
psi)
High Pressure Water
Washing
Pressure:
Between 68-680 bar
(1000-10000psi)
Table 2. Coating condition rating scale

Coating
condition
rating
1
2
3.4.
Coating Description
Antifouling paint intact, new or slick finish. May have a
mottled pattern of light and dark portions of the original
paint colour.
Shine is gone or surface lightly etched. No physical
failures.
Physical failure on up to 20% of hull. Coating may be
missing from slightly curved or flat areas to expose
underlying coating. Coating has visible swirl marks
within the outermost layer, not extending into any
underlying layers of paint.
Physical failure of coating on 20-50% of bottom.
Coating missing from slightly curved or flat areas to
expose underlying coating. Coating missing from intact
blisters or blisters which have ruptured to expose
underlying coating layer(s). Visible swirl marks expose
dunderlying coating layer.
Physical failure of coating on over 50% of bottom.
Coating missing from intact blisters or blisters which
have ruptured to expose the underlying coating layer(s).
Visible swirl marks exposed underlying coating layer.
SURFACE PREPARATION
3.4.1.
It can remove surface salts, dust and loose

surface debris.
For 68-204 bar (1000-3000 psi)

It can remove salts, dirt, loose coatings and
leached layer of antifouling coatings.
For 204-680 bar (3000-10000psi)
It can perform selective removal of coatings
and intact coatings.
High Pressure HydroBlasting (WaterJetting)

Pressure:
Between 680-1700 bar
(10000-25000 psi)
It can remove all existing old paint or heavy

rust. It scales to WJ 3 (Water jetting standards
NACE 5/ SSPC-SP 12) to a uniform matt
finish with at least two thirds of the surface
being free of all visible residues (except mill
scale) and the remaining one-third containing
only randomly dispersed stains of previously
existing rust, coatings and foreign matter.
Ultra High Pressure

Hydro-Blasting
(Water Jetting)
Pressure:
Above
1700
bar
(25000
psi),
but
normally 2000-2800
bar (30000-40000psi)
It can remove all existing old paint or heavy

rust. It scales to WJ 2 hydro jetting standard
of uniform matt finish with at least 95% of
the surface area being free of all previously
existing visible residues and the 5%
containing only randomly dispersed stains of
rust, coating and foreign matter.
3.5.
Good surface preparation is one of the most

important process of the entire coating procedures, as
great percentage of coating failures are usually
associated with poor surface preparation. All paint
systems will fail prematurely if the surface
preparation is not done according to standard
procedures requirement. If contaminants such as loose
rusts, oil, grease, dirt, salts, chemicals, dusts, etc. are
not removed completely from the surface intended for
coating, the paint adhesiveness as well as
cohesiveness and its quality would be affected.
Osmotic blistering would also occur resulting in
premature failure of the coating in service. There is no
paint system that would give optimum performance
result over a poorly prepared steel surface.
Cleaning Quality
SALT TEST
The purpose of carrying out the salt test is to

prevent coating failure due to effects of salt elements
on the surface before coating. In order to prevent the
defect, salt test is carried out to measure the level of
salt and to make sure that salt content is at minimum
level. Normally, salt test is carried out by using
Bresle kit sampler. Figure 4 showed the flow chart
of salt test measurement by Bresle kit sample.
i. Remove protective backing and foam centre from Bresle Patch and apply
to the hull surface, press firmly around the perimeter of the patch to ensure
a complete seal.
ii. Syringe with 3 ml of deionised water is inserted into the patch through
spongy foam perimeter and inject 1.5 ml of deionised water into
patch and do not remove syringe, reposition needle and evacuate any air in
the patch.
Hull cleaning
There are various methods available for cleaning

and preparing steel surfaces prior to painting. The
choice and methods of surface preparation would
depend on the location where the intended area of the
vessel is required and the availability of equipment to
be used. Hull cleaning includes hard scrap and fresh
water washing. Hard scraping shall be carried out to
remove slimes, weeds, shells, barnacles, etc. Besides
that, approved detergents shall be used to remove any
oil or grease present on the hull.
iii. Once air has been removed, inject remaining 1.5 ml of water and removed
the syringe from the patch. Surface of patch is then rubbed gently for 10 to
15 seconds to allow water to dissolve surface contaminants.
iv. Syringe is then inserted into the patch and extract the solution from patch.
Salt content will be measured by salt-meter. Acceptable salt reading should
not exceed 30 p.p.m.
Figure 4: Flow chart of salt test measurement by
Bresle kit sample.
84
and larger size in order to achieve a higher blast

profile on the steel substrate.
3.6.
CONDITION OF SHIP HULL (SIDE
SHELL
AREA)
PRIOR
TO
GRIT
BLASTING
Besides that, it is important that the correct blast

profile is achieved before the substrate is coated. Paint
manufacturers should specify the blast profile for each
coating, in terms of the pattern required for that paint.
The instrument to measure the blast profile is called
Blast Profile gauge and the reading is in micron. In
general, thicker coatings will require a profile with a
greater peak to trough measurement than a thin
coating.
Next, the Owners representative will mark out the

corroded area for blasting based on rust grade and
shall discuss it with the Yards Painting executive.
There are four types of rust condition using Swedish
Standard which listed in Table 4. When all parties had
agreed on the total blast area and the blasting grade.
Table 4. Rust condition using Swedish Standard
Steel surface largely covered with
adhering mill scale with little, if any
rust.
3.7.1.
The most commonly referred standards are Steel

Structure Painting Council (SSPC), National
Association of Corrosion Engineers (NACE) and
Swedish Standards or International Standards
Organization (ISO) (1988). Each standard is divided
into four standards of cleanliness, broadly described
as; brush off, commercial, near white metal and white
metal. Whilst each standard may be differ slightly in
requirements and terminology. The Table 5 indicates
the grades for steel surfaces using blast cleaning.
Rust Grade A
Steel surface has begun to rust and
from which mill scale has begun to
flake.
Rust Grade B
Steel surface on which the mill scale
has rusted away or from which it can
be scraped, but with slight pitting
visible under normal vision.
Table 5. Preparation grades for steel surfaces using

blast cleaning
Rust Grade C
Steel surface on which the mill scale
has rusted away and on which pitting
is visible under normal vision.
Rust Grade D
Shipyard shall draw up a work schedule based on

the agreed areas and instruct the blasting contractor to
proceed with the blasting works. The blasting time of
inspection is usually divided into two sessions, once
before noon and another late in the evening. This is to
allow sufficient time for the blasters to produce a
larger blast area so that when the paint is mixed and
applied, there will not be much wastage for the
coverage.
3.7.
Blast cleaning standard
Commercial
SSPC
Brush
Off
SP 7
NACE
No. 4
SWEDISH
Sa. 1
3.7.2.
Near-White
Metal
SP 10
White
Metal
SP 5
No. 3
No. 2
No. 1
Sa. 2
Sa. 2
Sa. 3
SP 6
Blast profile
The correct blast profile is very important prior to

painting. If the blast profile is produced too high, an
inadequate coating coverage will result over any high
and sharp peaks and this could lead to premature
coating breakdown. However, grit blasting can also
result in an insufficient surface profile and may
simply re-distribute contamination over the steel
surface trapping contaminants under the surface.
GRIT BLASTING
Grit blasting is the commonly used method for

preparing a surface for the application of paint. When
properly carried out, grit blasting can remove old
paint, rust, salts, fouling, etc., and provides a good
mechanical key (blast profile) for the new coating.
3.8.
PAINT APPLICATION
The paint application is to provide a film which

give protection or decoration of ship hull being
painted. The success of any application and
subsequent performance depends on some variables
such as surface preparation, film thickness of the paint
system, methods of application and conditions during
application.
Copper grit is one of the blast media widely used

for blasting in shipyard and is obtained as cooper slag
waste from melting the copper metal at a very high
temperature. It is a by-product and is often referred to
as hard coarse-grained silicieous sandstone. This is
the base for grit and can be found or prepared in
different sizes for different types of blasting known as
grit blasting. It usually comes in sizes ranging from
830cc (meshes) and 1030cc but most shipyard prefers
the former over the latter because of its coarseness
3.8.1.
Film thickness measurement
The wet film thickness (WFT) of the coating is

measured and can be converted to a dry film thickness
(DFT) following the paint makers guidelines for that
product. The wet film thickness measurement can be
determined by how much coating should be applied to
85
reach the specified DFT. The dry film or wet film

ratio is based on percentage of solids by volume of the
coating being used. The basic formula to measure the
WFT is:
WFT =
3.8.2.
spraying, but there is some risk of painters inhaling

spray droplets. Antifouling compositions offer special
problems because of the poisons they contain, this
applies both to the older copper poison types and
more particularly to the organometallic poisons. Thus,
suitable protective equipment must be used.
DFT ( m) 100
VolumeSolid (%)
Commonly, painting must be inspected regularly to

ensure that specifications regarding surface
preparation, wet and dry film thickness, drying times,
mixing of two-pack materials, overcoating intervals,
quality of workmanship and others are met.
Ship side coating thickness range
Different ship side area has different Dry Film

Thickness (DFT). Table 6 shows coating condition
thickness scale.
Table 6. Coating condition rating scale
Locations/Areas
200 400 (epoxy)
Boot Top
150 300 (epoxy)

75 100 (tie coat)
150 200 (antifouling)
Flat Bottom
1. High
pressure fresh
water
wash (3000
5000 psi)
2. Grit blast Sa
2.0 or Sa 2.5
The proper ambient temperature for steel hull

painting process should be 3C above dew point.
Most paints can tolerate high humidity but
condensation must not form on the surface being
painted. During the painting process for the ship hull,
the relative humidity must be below 85%.
Furthermore, paint should not be applied during fog,
mist or raining. Generally, under these conditions, it is
difficult to maintain the steel temperature above the
dew point. Besides that, condensation is forbidden
during hull painting process.
200 300 (epoxy)

75 100 (tie coat)
175 300 (epoxy)
75 100 (tie coat)
Propeller
375 400
Rudder
500
Condition during application
There are some factors which must be considered

during paints application. The major factors are
condition of substrate, temperature, relative humidity,
weather conditions and condensation.
DFT range, m
Topside
Vertical Side
3.8.3.
Surface
Preparation
3.8.4.
Method of paint application
Airless spray is now almost a universal method for

ship side paint application. This is by far the most
important and efficient method for the application of
heavy-duty marine coatings, which allows the rapid
application of large volumes of paint as well as the
application of high build coatings without thinning.
Airless spray method can reduce the overspray and
bounce back problems. Moreover, it follows that ships
paints must be formulated and manufactured to be
suitable for application by airless spraying.
3.9.
ANTIFOULING PAINTS FOR THE
SHIP HULL AREA
Generally, there are two basic mechanisms
employed in coatings to prevent fouling settlement
which are toxic antifouling and foul release coating.
Toxic antifouling means that prevention of fouling by
a surface coating requires the maintenance, in the
water in contact with the coating, of a concentration
of toxin that is lethal to all of the target organisms.
Meanwhile foul release coating has a surface with
very low surface energy which reduces the efficiency
of the attachment process dramatically, i.e. a nonstick surface is presented to the organism.
Airless spray is a technique of spray application

which does not rely on the mixing of paint with air to
provide atomization. Atomization is achieved by
forcing the paint through a special and precise
constructed nozzle or tip by hydraulic pressure. The
choice of tip determines the film thickness applied per
pass of the spray gun and should be selected in
accordance with the coating manufacturers
guidelines. The speed of each pass and volume solids
of the paint will influence film thickness. Airless
spray equipment normally operates at fluid line
pressure up to 5000 psi (352 kg/cm2) and care should
be taken periodically.
3.9.1.
Typical ship paints properties
Ship paints properties are very important for us to

understand because this is important on choosing the
correct and proper types of ship paints.
3.9.2.
Coating compatibility
Coating compatibility is important when the

maintenance or repair work is carried out, to ensure
that the repair coat will adhere to the original paint,
otherwise failures will occur between the individual
layers (inter-coat adhesion failure). Incompatibility
between coating types, such as epoxy anticorrosive
coatings with some types of antifouling paints, can be
overcome by the use of a tie coat, which has good
Modern products are most commonly applied by

airless spray. One airless spray gun is capable of
spraying between 50 and 80 litres of paint per hour,
i.e. covering 150 400 m2/hr at the required film
thickness. Moreover, airless spray application
produces less overspray than conventional air-assisted
86
Furthermore, Convention and Best Management

Practices is to prevent marine pollution by the
removal of ship antifouling coatings. Thus,
management for AFS waste collection is very
important throughout the process. The adoption of
management practices for the application and removal
of antifouling systems can reduce the release of
biocides into the natural environment. The aspects
include choice of antifouling system, and collection,
treatment, and disposal of spent coatings which have
an impact on the release of biocides into the
environment. If not managed properly, it may result in
high concentrations of biocides in the marine
sediments in areas close to where application and
removal activities are conducted [1, 4].
adhesion to both paint types and is therefore applied

onto the anti-corrosive layer before the antifouling
layer is applied. Thus, paint compatibility is a factor
which must be taken into consideration.
3.9.3. Performance of antifouling paints

determined by hull roughness
Ships underwater hull is painted to protect the
substrate and also prevent undue roughness. The most
significant cause of hull roughness is fouling.
Normally, paint fouling control technology can be
characterized into 4 types: Controlled Depletion
Polymer (CDP), TBT Free Self Polishing Copolymer
(SPC), Hybrid TBT Free Self Polishing Technology
and Foul Release Technology. Each type of paint
fouling control technology has different Average Hull
Roughness (AHR) value. Table 7 shows the AHR
value for these 4 types of AF paints.
3.11.
Each model has their own standards in order to

make sure the standard is controlled and complied
with the rules and requirements. To ensure the model
is controlled efficiently, quality assurance plays an
important role. The purpose of a quality assurance
system is to prevent problems from occurring, detect
them when they do, identify the cause, remedy the
cause and prevent recurrence. Quality Assurance
mechanism in this model is to ensure that accuracy
and precision throughout a procedure. The parties
involved in this procedure include Yards Painting
executive, Shipowners representative, suppliers,
Paint Manufacturer and Surveyors. The responsibility
of Yards Painting executive is to use checklists and
inspection records to ensure that the standards are
followed. Besides that, they will conduct audit by QA
department on a monthly basis. If sign of any
incompliance is found, yard person such as Project
Manager will issue Quality Assurance Note (QAN) or
Non Conformance Report (NCR) towards suppliers.
QAN is only for light or small incompliant, but NCR
is for heavy incompliant.
Table 7. Average Hull Roughness (AHR) value for

antifouling paints
Types of paint
fouling control
technology
Controlled Depletion
Polymer (CDP)
TBT Free Self
Polishing Copolymer
(SPC)
Hybrid TBT Free Self
Polishing Technology
Foul Release
Technology
Quality Assurance
Average Hull Roughness

(AHR)
40 microns/year
20 microns/year
30 microns/year
5 microns/year
The paint fouling control which has smallest AHR

value has lower percentage increase in power needed
or fuel used. Thus, foul release technology can save
more power and fuel used.
3.10. ANTIFOULING SYSTEMS
REGULATIONS AND CONVENTION
Generally, antifouling system has their
regulation to control the harmful antifouling systems
on ships. This is very important for us to understand
in order to enable the ship become compliant. In this
research, I include Antifouling System (AFS)
requirement and the Convention and Best
Management Practices on marine pollution by
removal of antifouling coatings from ships. From the
AFS requirement, there is a prohibition on the
application or re-application of organotin compounds
which acts as biocides in antifouling systems. When
existing vessels replaced the antifouling after 1
January 2003, they complied with this requirement or
provided sealer to avoid a non-compliant antifouling
to avoid leaching. All vessels after 1 January 2008
shall either not bear such compounds on their hulls or
external parts or surfaces; or apply coating that forms
a barrier (sealers) to such compounds such as leaching
from the underlying non-compliant antifouling
systems.
Furthermore, Paint Manufacturers inspector must

have a widely experience and good judgement in
order to make sure the paint job was completed as
specified. They need to take concern on many aspects
throughout the painting process. Adhesion test or
dolly test which is commonly known among the
paint inspectors, would have to be carried out when
the external hull of the ship is being grit blasted and
applied fresh coatings. This is to ensure the paint
adhesion onto the substrate hull can withstand a pullout pressure of not less than 300 psi. The higher the
pressure of the pull-out test, the stronger the
adhesion of the paint onto the steel substrate is. All
parties must take concern on their responsibilities to
ensure the quality obtained is in compliance. Every
daily log, tests and inspections work must be recorded
for future evaluations of the painting. There are no
any by-pass steps that can be skipped. Throughout the
process above, the quality of the paint is assured [2,5].
87
the model results are complied with the standard

requirements.
4. CONCLUSION
Fouling is unwanted accumulation material on solid
surface. There is either living organism (biofouling)
or non living substance. Antifouling paints are used to
prevent the biofouling. There are two types of
antifouling paints which are toxic and non-toxic
alternatives to TBT systems. Nowadays, there is a
trend to use the foul release technology which is also
known as non-toxic alternatives to TBT systems such
as silicon-type foul release AF paint. This product is
expensive and requires longer working period to
accomplish but the long term benefits can be seen
from the smooth and faster speed of the vessel
reaching its destination and its cost effective saving
which is believed to be about 40%. According to
LNG Carrier Owners Manual, foul release coatings
are proving themselves to be the ideal solution for
LNG hull and propeller fouling control. Not only can
they keep hulls and propellers smooth and free of
macro-fouling for extended service period of up to 60
months, but in addition, since they do not use biocides
to control fouling, they can be an integral part of an
LNG environmental management plan.
5. RECOMMENDATION
There are many aspects in this research which could
be investigated in the future. Some suggestion and
recommendations on future study are as follow:
i.
ii.
iii.
The practical way to measure hull fouling is

to use a professional diver to not only
measure but to survey or inspect the general
condition of the hull as a whole and record
with video camera or CCTV. This is because
hull fouling varies along the hull.
There are many ship paints application
methods discussed in this paper Therefore, it
is suggested that investigation for each
methods in much more details.
The performance of AF paints can be
determined by many factors. Thus, it is
suggested that investigation for performance
or quality of AF paints to be done in much
more detail.
REFERENCES
High performance paint systems are especially

sensitive to misapplication and knowledge of the
application characteristics and recommended film
thickness is vital to obtain optimum results. For
optimum service life, the surface must be completely
free of all contaminants that might impair
performance and should be treated as such to assure
good and permanent adhesion of the paint system.
The quality of surface preparation has a direct relation
with the lifetime of a system. Nowadays, the paint
application method commonly used airless spray. The
degree of skill of the personnel can affect the
performance of paints.
Throughout the whole ship painting process, it is
necessary to inspect the work as it progresses if there
is to be any reasonable assurance that a paint job was
completed as specified. There are many failure cases
due to poor workmanship occurred after the work has
been completed and has been paid for. Beside that,
quality assurance is part of quality management
focusing on increasing the ability to fulfill
requirements of the process.
As a conclusion,
88
[1]
Edyvean, R.G.J. 1982. Fouling and Corrosion

by Microalgae, PhD Thesis, University of
Newcastle-upon-Tyne, Newcastle-upon-Tyne.
[2]
IMO. 2005. Antifouling Systems. International

Convention on the Control of Harmful Antifouling
Systems on Ships. International
Maritime Organization. London.
[3]
IRTA. 2009. Safer Alternatives to Copper

Antifouling Paints. San Diego.
[4]
Kan, S., Shiba, H., Tsuchida, K. and Yokoo, K.

1958, Effect of fouling of a ships hull and
propeller upon
propulsive
performance.
International Shipbuilding Progress, Vol. 5,
pp. 15-34.
[5]
Mosaad, M.A. 1986. Marine Propeller

Roughness
Penalties,
PHD
Thesis.
Department
of
Marine
Technology.
University of Newcastle-upon-Tyne.

HATCH COVER ANALYSIS OF CAPESIZE BULK CARRIERS

AND SUGGESTIONS FOR ALTERNATE MATERIALS
K. Kunal1, C. Kannan2 & S. Surendran3
1 Research scholar
Department of Ocean
Engineering
IIT Madras, India
kunalsky.kumar9@gmail.com
2 Project Officer
National Institute of Ocean
Technology
Chennai, India
Kannan021@gmail.com
3 Professor
Department of Ocean
Engineering
IIT Madras, India
ssur@iitm.ac.in
ABSTRACT
Hatch cover failure leads to flooding of the forward cargo compartment and occasionally results in fatal
casualty. The foremost hatch cover and the next one within 25% length of the vessel are prone to the impact by
shipping of green water. The loss of hundreds of bulk carriers with the precious lives is a nightmare to both
designers and insurers. The tail of Rayleigh distribution showing the extreme rare wave heights is a matter needed
great attraction of designers and analysts. Though rare, the extreme wave heights are responsible for the failure of
bulk carrier ships. The tricky point is that the more we increase the strength of the hatch cover the more is its self
weight. The position of vertical centre of gravity of vessel goes up and that is not desirable for commercial ships in
many aspects. Deficiency in the scantling of bulk carrier covers is brought out in this paper. Suggestions to use
better material having higher strength to weight ratio is made. The use of composite reduces maintenance costs
and improves structural and operational performance.
Keywords: Hatch cover, shipping green water, wave breaking, composite laminate, flooding, delamination.
bulk carriers can induce extremely large impact

pressures and can cause capsize and loss of the ship in
no time. As per Rosenthal [7] a highest wave of about
28 meters was found in the South Atlantic The hatch
cover of a capsize bulk carrier ships are analyzed here
for both UR S 21 and UR S21+40% extra loads. Is the
extra load is only 40%?. More loads are also to be
considered along with a strong hatch cover system
especially within 25% of the length of the ship from the
most forward part. Can we manage the hatch cover
problems using ordinary steels? Is there any alternate
material possible? Such critical queries are to be
addressed by the researchers working in the field.
1. INTRODUCTION
The pathetic situation in many marine accidents is
that the vessel broke up sinking immediately before the
Master could send a distress massage. A recent
observation on the seas reveals a fact that there is
possibility of waves with crest to trough of 20 to 30
meters [1]. Such waves of non-linear nature encounter
the moving ship and the mass of the water above the
main deck is broken and collapsed onto the deck and
hatch cover of the vessel. During the head-on encounter
with the incoming waves the ship may be heaving,
pitching or undergoing a coupled motion. The collapsed
wave on top of the hatch meets the moving ship with
high relative velocity and giving more inertial force
onto the deck. The dynamic force of wave impacts
should also be considered in the structural analysis of
the deck, hatch covers, hatch coamings and other
vulnerable areas. Current design criteria generally
consider only 11m wave height as per IACS
recommendation [3]. The authors also suggest that the
IACS recommendation 34 should be modified to
consider more possible height of waves for safe marine
operations. Wave impacts in the closed hatch covers of
A typical hatch cover of a capsize bulk carrier is

considered for this study. The hatch cover is designed
in accordance with IACS rule URS21. A 40% higher
load would mean approximately 10-15% additional to
URS21 weight. It is better to reduce the weight for the
parts of hull above the load water line. Hatch coamings
and covers made of steel are at levels above the main
deck and challenging the righting arm stability. The
worse is the case when their own masses are increased
to protect the ships from unforeseen wave conditions.
89
With a view to reducing structural weights, composites

are being tried now a day.
2. HATCH COVERS LOCATION AND

DETAILS
Hatch covers are designed as per classification rules.
But when it comes to ship strength, there are many
challenges. The strength and weight of hatch covers are
mutually related. Weight directly affects the freight
capacity and hence returns for the ship operator. The
hatch system design has to outlive the hydrodynamic
loads Sea wave profiles converted into time domain will
predict a wave even 2.4 times the significant wave
height. This is reported as per Rayleigh distribution of
wave heights [1].
Table 1 shows the particulars of the vessel
considered. Figure 1 shows the plan view of the hatch
cover. The size of Hatch cover No. 1 is the foremost one
and17 m long and 15 m width roughly. The size of the
second hatch cover (Hatch cover no.2) is 20m long and
15 m width. Hatch cover No. 1 has a steel weight of 270
kg/m2 and that of no. 2 has a steel weight of 220 kg/m2.
Hatch cover no. 1 is stiffened for a load of 6- 6.5 T/m2
and the hatch cover No. 2 is stiffened for a vertical load
of 4 - 4.5 T/m2.
Figure. 1 Plan of Hatch Cover of Capesize

bulk carrier
Table 1. Main Particulars of Capsize Bulk Carrier

Item
Length
Between
Perpendiculars
Breadth moulded
Draft
Depth moulded
Displacement
Dimension
271.0 m
45.0 m
18.15 m
24.6 m
189197.0
Tonnes
In this study three different sections are taken from

the hatch cover and analyzed under three different load
conditions (2.8, 3.5 & 4.5 T/m2).Thereafter, a 40% extra
load condition was also considered. First a section
containing longitudinal stiffener was considered and the
analysis for 2.8 T/m2 was done and then repeated
analysis for a load of 3.5 and 4.5 T/m2 respectively.
Thereafter, the load was increased to 40% and the
weight of the hatch cover up to roughly 15% to respond
to this load. Then another sections containing side
girder and centre girder also considered for the analysis.
The sectional details of the girders and ordinary
stiffener are as shown in Figure 2.
Figure. 2 Sectional details of girders and stiffeners.

Study of hatch covers is very important as deck
flexing could spring this stiffened panels followed by
water entry, rapid flooding and capsizing. The analysis
is started with a load corresponding to 2.8T/m2. This is
90
knl
for a wave crest of height roughly 2.8m above the hatch

cover. As per Figure 3, modeling and analysis using
ANSYS is done and the results are discussed, maximum
equivalent stress of 57.229 MPa is occurred at the node
number 4338. The maximum centre deflection is seen to
be 0.599961 mm, which is well within the design value
provided by any codes for the end conditions assumed
and the same for this is fixed supports. The loading is
considered as equivalent static one due to a breaking
wave producing a head equal to the height of the wave
crest above the hatch cover. Similar approach is adopted
by other analysts. Higher load cases were analyzed and
purposely avoided here in view of optimization of
space.
= non-linear factor on relative motion for

point 2
= 1.025 tons/m3
g
= acceleration due to gravity, 9.81m/sec2
h
= height from base line to hatch cover.
T
= draft at the hatch cover
relu6 = linear relative motion in head sea at
selected point in hatch cover1 at centre line.
Substituting suitable values in Equation (1) the design
pressure is calculated approximately as 5 T/m2 for world
wide and 6.3 T/m2 for North Atlantic Sea.
4. APPLICATION OF COMPOSITES
The feasibility of application of composites in
vulnerable parts of hull was established by Hackman
Sandel and Kockums AB [2]. Although, much weight
saving is possible for hatch covers having good strength
to counter the load due to shipping green water or wave
breaking, transverse bulkhead to withstand grain load
will be more massive and bulky. This issue may be dealt
with using a composite having better properties. For the
same load the flexural rigidity of composite materials
will be a few times (multiplication by a factor) that of
the steel. Deflection for the same load will be more for
the case of composites. This cannot be accepted for
longer marine vessels. As the elastic modulus is only a
fraction of steel the section modulus has to compensate
for the total flexural rigidity or EI value. However hatch
cover can be made with higher section modulus as it is
just a cover. Honey combs and sand witch models
discussed by Hackman Sandel and Kokums AB [2] are
feasible. Scott and Somella [5] discussed in detail the
feasibility of glass reinforced plastic in cargo ship.
Costly fibers can be used to increase the strength to
weight ratio, depending on the situation. With proper
fiber orientation we can get desired strength and
different lamina can be stacked together to give
sufficient strength. The factors such as the choice of
fiber and the resin, fiber volume fraction, fiber
orientation, ply stacking sequence and number of plies
and very important design drives to the composite
structure.
Figure 3 Analysis of hatch cover section with

longitudinal stiffener subjected to 2.8T/m2.
3. DESIGN LOADS
When the moving ship encounters head-on waves,
there is relative upward motion of the ship with respect
to the wave. The rising forward part of hull in pitch or
coupled pitch-heave motions give rise to more
accelerations and the hatch cover and deck meet the
falling water mass, from the broken wave in an extreme,
cause immense load on the hatch cover and deck.
Nestegard and Krokstad [4] studied on water impact on
deck. Search of available literature shows that the
pressure on top of the hatch cover may be calculated as
follows.
Thickness of ply depends on the fiber volume fraction

and we can get accurate value from design chart.
In this study three different load conditions [2.8,
3.5&4.5T/m2] are considered. This hatch cover consists
of four different laminate (high strength carbon-epoxy,
S-glass epoxy, Kevlar epoxy and high strength graphite
epoxy).
Pd = kdyn. .g. (knl.relu6-(h-T)) [kN/m2] ------ (1)

Where:
= dynamic factor relative to static pressure
kdyn
height.(Taken as 1.4)
Bottom layer is S-glass epoxy named as material 3.

Next one is E-glass epoxy, Kevlar-epoxy and upper
layer is high strength carbon epoxy.
91
Figure 4. Finite element modeling of Composite

plate with boundary Conditions.
Figure 6.Normal stress in fiber direction material 2.
To reduce the delamination area in a composite plate
stitching through the thickness is also required [6].
Figure 5. Orientation of fiber

Stiffener is also replaced with graphite-epoxy
composite. For 4.5 T/m2 load a composite plate with
stiffener is modeled for folding type hatch cover having
dimensions 2775 and 8474 mm. For each
layer maximum stress and strain calculated. Thickness
of each ply is 9.5 mm. Fiber orientations are 30 , 60 , 60
and 28 in order from bottom to top for plate and 0 &
90 for stiffeners.
Figure 7. Shear stress in material 2.
From Figures 6, 7, 8, 9, 10 & 11 it is clear that all the

stresses developed are less than the permissible stresses,
for the composite materials .
92
Figure 8. Normal stress in fiber direction

in material 1.
Figure 10. Shear stress in material 1
Figure 11. Deflection at middle of plate
5. CONCLUSION
Figure 9. Normal stress in perpendicular to fiber
direction in material 1.
From the above analysis we can conclude for almost

all cases the equivalent stress are well below the
ultimate strength of the material that is high strength
low alloy steel. The end connections assumed for centre
and side girders are simply supported and in actual
scenario it may not be true. It may not be fixed too.
93
There will be some condition in between the fixed

and simply supported ends.
REFERENCES
[1]
Craig B. Smith, Extreme Waves, Joseph Henry

Press, Washington D C(2006)
[2]
Hakman Sandell and Kockums AB,. Case study

WP3f; composite materials in a troll bulk cargo
vessel, LASS, Lightweight Construction
Applications at Sea(2009).
[3]
IACS Unified Requirements S 21 , Vol 1, (1997)
[4]
Nestegard, A., and Krokstad, J. R ( JIPDEEPER); Deepwater analysis tools, Offshore

Technology Conference, OTC 10811, Houston,
May (1999).
[5]
Robert J. Scott and John H. Somella, Feasibility

study of glass reinforced plastic cargo ship, Ship
structure committee, SSC-224, (1971).
[6]
Robert M. Jones, Mechanics of Composite

Materials. Taylor & Francis New York, London.
(1999).
ACKNOWLEDGEMENTS
[7]
Rosenthal W, Lehners, Danker H , Detection of

extreme single wave and wave statistics. (2003)
The authors would like to thank Ministry of Shipping,

New Delhi for their financial support. The authors
would also like to thank Dr. Dilsha Rajapan of NIOT,
Chennai and Prof. S K Bhattacharyya, Head
(Department of Ocean Engineering) and Prof. R.
Sundaravadivelu of IIT Madras
[8]
Tetsuya Yao, Atsushi Magaino, Toshiro

Koiwa, Shugo Sato, Collapse strength of hatch
cover of bulk carrier subjected to lateral pressure
load, Marine structures, pp 687-709(2003)
The reduction in the topside weight would provide

increased payload and better sea-keeping of the ship.
Composite laminate is proposed for the same. A
composite material, which is free of corrosion, is
desirable for marine structure. The results of the
detailed analysis will be discussed in the main paper.
SCRIMP and resin transfer moulding processes can be
used for high quality composite production and
consideration of low safety factors for fabrication.
Here feasibility of using of composite materials is
proved with folding type hatch cover analysis made up
with hybrid laminate. After comparing the average
density of laminate with steel it is found that composite
structure weight is 0.56 of corresponding steel hatch
cover weight. It is hoped that Bangladesh and other
countries dealing with marine transportation will
appreciate the ideas raised in this paper.
94

A STUDY ON NUMERICAL ANALYSIS OF UNSTEADY FLOW OVER TWO

DIMENSIONAL HYDROFOILS
N. Mostafa1, M. M. Karim2, and M. M. A. Sarker3
1
Graduate School of Environment and
Information Science, Yokohama National
University, Yokohama, Japan.
E-mail: mostafa_nur@yahoo.com
2
Department of Naval Architecture and
Marine Engineering, BUET, Dhaka,
Bangladesh
E-mail: mmkarim@name.buet.ac.bd
3
Department of Mathematics, BUET,
Dhaka, Bangladesh
E-mail: masarker@math.buet.ac.bd
ABSTRACT
The time dependent characteristics performance of cavitating flow around CAV-2003 hydrofoil has been
simulated using pressure-based finite volume method. A bubble dynamics cavitation model is used to investigate
the unsteady behavior of cavitating flow and describe the generation and evaporation of vapor phase. For
choosing the turbulence model and mesh size a non cavitating study is conducted. The cavitating study presents
an unsteady behavior of the partial cavity attached to the foil at different time steps in the case of =0.8.
Moreover, this study is focused on cavitation inception, the shape and general behavior of sheet cavitation, lift
and drag forces for different cavitation numbers.
Key words: Cavitation, CAV2003 hydrofoil, finite volume method, k- turbulence model, unsteady flow.
cavitation process is by the so called barotropic state
law that links the density of vapor-liquid mixture to
the local static pressure.
A cavitation model, based on bubble dynamics
equation [10] is used for computation of cavitating
flows. The non-cavitating operation as well as
influence of mesh and turbulence model, mainly by
comparing the values of lift and drag are studied in
the previous paper of the author [5]. Now two
cavitating conditions are separately analyzed, i.e.,
=0.8 where an unsteady partial cavitating behavior is
obtained and =0.4 where a supercavitating flow is
observed. Cavitating flow at different cavitation
numbers is analyzed and finally an unsteady partial
cavitating behavior at =0.8 is observed at different
time steps.
1. INTRODUCTION
Cavitation in hydraulic machines causes different
problems like vibration, increase in hydrodynamic
drag, pressure pulsation, and change in flow
kinematics, noise and erosion of solid surface. Most
of these problems are related to transient behavior of
cavitation structure. Cavitation erosion is strongly
related to unsteady fluctuations of the cavitation zone.
Hence a study of unsteady cavitation behavior is
essential for a good prediction of the problem. To
investigate cavitation phenomena and validate
numerical procedures, a number of investigations
were performed in the past [1, 4, 8, 9]. In the last
decade various methods for numerical simulation of
cavitating flow were developed. Most of the studies
treat the two phase flow as a single vapor-liquid phase
mixture flow. The evaporation and condensation can
be modeled with different source terms that are
usually derived from the Rayleigh-Plesset bubble
dynamics equation.
Recently different authors proposed to consider a
transport equation model for the void ratio, with
vaporization/condensation source terms to control the
mass transfer between two phases [10]. This method
has the advantage that it can take into account the
time influence on the mass transfer phenomena
through empirical laws for the source term. It also
avoids using quantities like bubble number density
and initial bubble diameter. The other way to model
The numerical model uses an implicit finite volume
method associated with multiphase and cavitation
model. For numerical simulation of cavitating flow, a
bubble dynamics cavitation model is used to describe
the cavity formation. The RNG k- turbulence model
with enhanced wall treatment is used as a turbulence
model. The Reynolds number (Re=5.9 105) based on
chord length is used .The corresponding y+ is 5-15. A
second order central scheme is used for discretization
for space except for the convective terms. The
convective term in the momentum equation is
95
discretized by the QUICK scheme for non cavitating

flow and second order implicit scheme is used for
cavitating problem. Pressure based solver SIMPLE is
used as the velocity pressurecoupling algorithm
C e and C c are empirical constants, and k is

the local kinetic energy, is the surface tension, f v
is vapor mass fraction and f g is mass fraction of
where
noncondensable (dissolved) gases. Values of
C c are 0.02 and 0.01 respectively.
3. MULTIPHASE MODEL
A single fluid (mixture model) approach is used.
The mass and momentum conservation equations
together with the transport equation and the equation
of the turbulence model from the set of equations
from which fluid density (which is the function of the
vapor mass fraction f v ) is computed. The
The section of the hydrofoil is presented in Figure 1

which shows a schematic view of the CAV2003
hydrofoil geometry. The hydrofoil is placed at an
(mixture density-vapor mass fraction)
relation is [2]:
5. GEOMETRY AND COMPUTATIONAL

DOMAIN
fv
1 fv
angle of attack 7 . The equation of the upper surface

of the symmetric foil geometry is provided as
2
The volume fraction of the vapor phase
( v ) is
x
a4
c
where a 0 = 0.11858 , a1 = 0.02972 ,
related to the mass fraction of the vapor phase with:
m
v
v = fv
(2)
(3)
The momentum conservation equation for the mixture
coordinate along the chord line. The flow field around

the hydrofoil is modeled in two dimensions. The flow
from left to right with the hydrofoil of chord length
c = 0.1 m submersed in an incompressible fluid is
considered. The hydrofoil is located at the middle of a
channel of length 10c and height 4c. Figure 1 shows
the total 2D computational domain and boundary
conditions. The inlet boundary condition is specified
velocity inlet with a constant velocity profile which is
6 m/s. Upper and lower boundaries are slip walls, i.e.,
symmetry boundary condition. The outlet uses a
constant pressure boundary condition. The foil itself is
( m vrm ) + .( m vrm vrm ) = p +

(4)
is t
r
r T
r r
. m v m + v m + mg + F
[ (
)]
And the transport equation for the vapor is:
( m f v ) + .( m vrm f v ) = Re Rc
t
(5)
4. CAVITATING MODEL
evaporation) and vapor condensation, respectively.

The source terms can be expressed as [7].
Re = C e
l v
2 pv p
(1 f v f g ),
3 l
p < pv
when
Rc = Ce
when
l v
p > pv
Symmetry
Velocity Inlet
The working fluid is assumed mixture of liquid,

liquid vapor and noncondensible gas. Source terms
Re and Rc define as vapor generation (liquid
(8)
a 2 = 0.00593 , a3 = 0.07272 ,
a 4 = 0.002207
y = y c and x = x c is the dimensionless
The mass conservation equation for the mixture is:
( m ) + .( m vrm ) = 0
t
y
x
x
x
x
= a0
+ a1 + a 2 + a3 +
c
c
c
c
c
(1)
4c
Pressure Outlet
m fv
C e and
10c
(6)
a no-slip wall, i.e., u = 0, v = 0 at the foil surface.

Figure 1. Schematic diagram of the flow field around
CAV2003 hydrofoil with boundary condition.
2 pv p
fv ,
3 l
(7)
96
of Kawamura et al. [6]. There exists a good agreement

but some difference in magnitude may be due to the
k- turbulence model used by Kawamura et. al. [6].
The difference in pressure distribution on the face side
is found very small. Similar comparison is shown in
Figure 3 for =0.8. The time history of the lift and
drag coefficients computed by mixture model at =
0.8 and = 0.4 are shown in Figure 4 respectively.
The characteristics of the curve of lift and drag
coefficients are almost similar. The contours of
pressure coefficient and vapor volume fraction for
cavitation numbers =0.8 and 0.4 are shown in Figure
5. These contours show the expansion of cavity and
their sizes for different cavitation numbers. At =0.8
the half of the hydrofoil is covered with vapor and at
=0.4 the back surface is covered with vapor and it
decreases with the increase in cavitation number and
cavity length become shorter.
Figure 6 and Figure 7 show the instantaneous field
of pressure distribution and vapor volume fraction
respectively computed for five time levels t=0.47s,
0.49s, 0.52s, 0.55s and 0.57s in the case of =0.8.
These figures clearly show the creation and collapsing
of cavity over the hydrofoil surface.
Figure 2. The overall view of grid lines in mesh.
6. CAVITATING ANALYSIS
This section presents results computed for the
typical cavitation numbers =0.8 and =0.4. For
simulation the convergence criterion is determined by
observing the evaluation of different flow parameters
(velocity magnitude at inlet, static pressure behind the
hydrofoil) in the computational domain. For
computation, each value of residual is taken as 10-4.
Time step size has a great influence on simulation of
cavitating flow. Different time step values are tested,
eventually the time step for unsteady computation is
set to 510-5 and approximately 30 iterations per time
step are needed to obtain a converged solution. To
predict the behavior of the cavitating flow for the
values of cavitation number = 0.8 and = 0.4, we
first present comparisons of the computed timeaveraged lift and drag coefficient for cavitating flow
with Pouffary et al. [7], Courtier -Delgosha et al. [1],
Kawamura et al, [6] and Yoshinori et al. [11]. Table 1
shows that the lift coefficient and the drag coefficient
are in good agreement with published results.
Pressure coefficient Cp
-0 .4
P re s e n t
k a w a m u ra ,T . e t a l.[6 ]
0 .4
0 .8
Table 1. Comparison of time-averaged lift and drag

coefficient at cavitation numbers = 0.8 and = 0.4
0 .0
0 .2
0 .4
0 .6
0 .8
1 .0
x /c
=0.4
CL
CD
CL
CD
Present
0.44
0.077
0.214
0.076
Pouffary
0.456
0.0783
0.291
0.086
CourtierDelgosha
0.450
0.0700
0.200
0.065
Kawamura
0.399
0.047
0.187
0.063
Yoshinori
0.417
0.0638
0.160
0.056
-1.0
Pressure coefficient Cp
=0.8
0 .0
Present
Kawamura et.al[6]
-0.5
0.0
0.5
1.0
The time average values of lift and drag

coefficients calculated by present method for the
cavitation number = 0.8 are very close to the
numerical result of Pouffary et al. [7]. However, the
results show little discrepancy at cavitation number
= 0.4. This discrepancy may be attributed due to the
fact that different researchers used different
turbulence models. The comparisons of the pressure
distribution on the foil surface for =0.4 is shown in
Figure 3. It compares the present result with the result
0.0
0.2
0.4
x/c
0.6
0.8
1.0
Figure 3. Comparison of the pressure coefficient on

the foil surface at =0.4 and=0.8 respectively.
97
1.0
CL
Lift Coefficient CL
0.8
0.6
0.4
0.2
0.0
-0.2
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Time(s)
Drag Cooefficient CD
0 .1 6
CD
0 .1 2
0 .0 8
0 .0 4
0 .0 0
0 .0
0 .1
0 .2
0 .3
0 .4
0 .5
0 .6
T im e (s)
0.6
Lift Coefficient CL
0.5
CL
0.4
0.3
0.2
0.1
0 .0
0.1
0.2
0.3
0.4
0.5
Figure 5. Contour of pressure coeffcient and

contour of vapor volume fraction at =0.8 and
=0.4
0.6
T im e(s)
It is observed that the length of sheet cavity grows

gradually until the cavity trailing edge almost reaches
the end of a foil, and then reverse flow emerges near
the foil trailing edge. The reverse flow propagates
towards the leading edge along the back surface of
edge foil causing collapse of the sheet cavity. Table 2
shows the summary of the cavitation parameters
Drag Cofficient CD
0 .1 1
0 .1 0
CD
0 .0 9
0 .0 8
where l max , t max and lt
0 .0 7
0 .0
0 .1
0 .2
0 .3
0 .4
0 .5
max
are maximum cavity
length, maximum cavity thickness and the position of

maximum cavity thickness respectively. As the
cavitation number decreases, the maximum cavity
length and maximum cavity thickness increase. On
the other hand the position of maximum cavity
thickness is almost constant at 75% of the maximum
cavity length except for = 0.4. In the case of = 0.4,
a super cavitating flow fully develops and the
cavitation also appears on the pressure side near the
0 .6
T im e (s )
Figure 4. Time history of lift coefficient and drag

coefficient at =0.8 and =0.4 respectively.
98
trailing edge. It is also seen that the time-averaged lift

coefficient decreases and the time-averaged drag
coefficient increases as the cavitation number
decreases. However, after =0.9, it decreases
slightly.
Figure 8 shows the variation of maximum cavity

length and maximum cavity thickness with the
cavitation number. The idea of change pattern can be
obtained from this figure.
t=0.47s
t=0.47s
t=0.49s
t=0.49s
t=0.52s
t=0.52s
t=0.55s
t=0.55s
t=0.57s
t=0.57s
Figure 7. Countour of the volume fraction at different

times in the case of =0.8
Figure 6. countour of the presure coefficient at

different times in the case of =0.8
99
thickness show good correlation with cavitation

numbers.
Table 2. Comparisons of time-averaged lift and
drag
coefficient at cavitation numbers = 0.8
and = 0.4
t
l
l
C
C
max
max
t max
3.5
0.667
0.024
1.5
1.2
1.1
1.0
0.9
0.8
0.4
0.098
0.16
0.21
0.25
0.45
0.49
1.00
0.0211
0.0329
0.0465
0.047
0.0772
0.0784
0.28
0.76
0.79
0.73
0.73
0.78
0.71
0.66
0.582
0.57
0.566
0.560
0.51
0.44
0.214
0.0378
0.0425
0.0446
0.0476
0.0783
0.077
0.0763
Maximum cavity length
1.2
Mximum cavity length

Maximum cavity thicness
0.4
0.9
0.3
0.6
0.2
0.3
0.1
0.4
0.6
0.8
1.0
1.2
1.4
Maximum cavity thickness
0.5
1.5
0.0
REFERENCES
0.0
Cavitation number
Figure 8. Plot for maximum cavity length and

maximum cavity thickness versus cavitation number
7. CONCLUTION
Two-dimensional finite volume method has been
applied to simulate in compressible flow around
CAV2003 hydrofoil. For cavitation number =0.8, an
unsteady partial cavitating behavior and for cavitation
number =0.4, unsteady supercavitating behavior is
simulated. In general, the results are very promising.
The instantaneous pressure distribution and vapor
volume fraction computed are shown at different time
steps. Therefore, it is clearly understood that the
length of the sheet cavity grows gradually towards the
trailing edge and reverse flow emerge near the foil
trailing edge and propagates towards the leading edge
and causing the creation and collapsing of cavity over
the hydrofoil surface. Finally, analysis is done for
different cavitation numbers and the computed
maximum cavity length and maximum cavity
[1] Coutier-Delgosha, O. and Jacques A. A.,

Numerical reduction of cavitating flow on a
twodimensional symmetrical hydrofoil with a
single fluid model, Fifth International
Symposium on Cavitation (Cav2003), Osaka,
Japan, November 1-4, (2003).
[2] Dular, M., Bacher, R., Stoffel, B. and irok, B.
Experimental
evaluation
of
numerical
simulation of cavitating flow around hydrofoil,
European Journal of Mechanics B/Fluids 24,
522-538, (2005).
[3] Frobenius, M., Schilling, R., Bachert, R. and
Stoffel, B., Three-dimensional, unsteady
cavitation effects on a single hydrofoil and in a
radial pump measurements and numerical
simulations, Part two: Numerical simulation,
in: Proceedings of the Fifth International
Symposium on Cavitation, Osaka, Japan, (2003).
[4] Kubota, A., Hiroharu, A. and Yamaguchi, H., A
new modeling of cavitating flows, a numerical
study of unsteady cavitation on a hydrofoil
section, J. Fluid Mech. 240, 5996, (1992).
[5] Karim, M. M., Mostafa, N. and Sarker, M. M. A.,
Numerical prediction of unsteady cavitating
flow around hydrofoil, Proceeding of 16th
Mathematical Conference of Bangladesh
Mathematical Society, 17-19 December, (2009).
[6] Kawamura, T. and Sakuda, M., Comparison of
bubble and sheet cavitation models for
simulation of cavitation flow over a hydrofoil,
Fifth International Symposium on Cavitation
(Cav2003), Osaka, Japan, November 1-4, (2003).
[7] Pouffary, B., Fortes-Patela, R. and Reboud, J.L.,
Numerical simulation of cavitating flow around
a 2D hydrofoil: A barotropic approach, Fifth
International
Symposium
on
Cavitation
[8] Stutz, B. and Reboud, J. L, Measurements
within unsteady cavitation, Exp. Fluids 29,
545552, (2002).
[9] Schnerr, G. H. and Sauer, J., Physical and
numerical modeling of unsteady cavitation
dynamics, in: 4th International Conference on
Multiphase Flow, ICMF-2001, New Orleans,
USA, (2001).
[10] Singhal, A. K., Li, H., Atahavale, M. M. and
Jiang, Y., Mathematical basis and validation of
the full cavitation model, J. Fluids Eng. 124,
617624, (2002).
[11] Yoshinori S., Ichiro N. and Tosshiaki I.,
Numerical analysis of unsteady vaporous
cavitating flow around a hydrofoil, Fifth
International
Symposium
on
Cavitation
100

STATIC AND DYNAMIC ANALYSIS OF SEMI-SUBMERSIBLE TYPE

FLOATERS FOR OFFSHORE WIND TURBINE
A.C.Mayilvahanan1 and R.Panneer Selvam2
Research student
Department of Ocean Engineering
Indian Institute of Technology Madras
m.vahanan@gmail.com
Associate professor
Department of Ocean Engineering
Indian Institute of Technology Madras
pselvam@iitm.ac.in
ABSTRACT
The wind potential in offshore site faster and steadier than on land and makes lesser wear on the turbine
components besides generating more electricity per turbine. Wind energy is a reliable source of sustainable
power generation and has been an active area of research globally to economically harness the energy for
human use. Such a consistent source of wind energy pushed the engineers to install wind turbines near the
coasts. As more farms came on the shore, the coastal community perceived as an eyesore and nuisance which
raised the demand of installing the wind turbine further offshore. In shallow water, fixed structures like
tripods, jackets, and truss-type towers, monopiles and gravity base are functionally and economically feasible
for depths up to 100 m. In deep waters, a floating substructure can be more economical for offshore wind
turbine. The floating structures can also be deployed in shallow water depth regions for its inherit
advantages like construction and installation, mobility, maintenance etc.Different floating options like TLPs,
SPAR, Semi-submersibles, barges besides new floating structural configurations are being actively analyzed.
In this study a semi-submersible type floater of different aspect ratios is investigated for its performance
under wave and wind loading.
Keywords: offshore wind energy; floating offshore wind turbines; static stability; hydrodynamic analysis;
semi-submersible floater, spectrum, response amplitude operator.
1. INTRODUCTION
As the offshore wind posses less turbulence, low wind
shear, vast area availability and high wind speed
within a shorter distance from shore, offshore wind
energy has been the focus of many of the engineers
and researchers these days. The support structure for
the offshore wind turbine can be fixed or floating.
The floating structures can also be deployed in
shallow water depth regions for its inherent
advantages like construction and installation, mobility
and maintenance. As the interest worldwide is in the
development of offshore wind farms, different
floating option like TLPs, SPAR, Semi-submersibles,
barges besides new floating structural configurations
are being actively analyzed. The performance of the
semi-submersible floater under wave and wind
loading for supporting a 5-MW offshore wind turbine
is undertaken.
2. BACKROUND
Several types of floating supporting structures for
offshore wind turbine have been investigated by
different researchers around the world. Bulder et al.
(2002) presented the technical and economical
feasibility of floating wind energy system in the

depth range of 50m. Based on the cost analysis he
concluded that the Tri- floater concept was
technically and economically feasible option.
Musial et al. (2004) addressed the different types of
floating platforms for offshore wind turbine. They
made the cost comparison analysis for Tri-floater
and TLP. Based on the comparison study, it was
concluded that cost of TLP is lesser than the Trifloater concept. Lee (2005) examined two floater
concepts namely, a three legged tension-leg
platform and a four legged taut-moored system, for
floating wind turbine systems of 1.5MW capacity
for their performance in wind and wave. Tempel
(2006) focused on the design basis of supporting
structure for 2.0MW offshore wind turbine. A
monopile supporting structure was the focus of his
study, the methodology of aerodynamic load
calculation on wind turbine and total system
response analysis in the frequency domain method
were presented. Wayman (2006) studied the three
different types of floating supporting structures;
shallow drafted barge (cylinder), TLP (surface) and
TLP (submerged). It was concluded that the barge
101
has lowest cost of construction among the others,

but TLP submerged supporting structure exhibited
least response in all conditions.Tracy (2007) carried
out a parametric study on concrete ballasted
cylinder with different combinations of draft and
mooring systems. The aim of the work is to study
the performance of the semi-submersible floaters
with different aspect ratios to support a 5 MW
NREL offshore wind turbine in 100m water depth
for different sea-states in Indian coastal waters
under wind and wave loading. The schematic view
of the semi-submersible floater with wind turbine is
shown in Figure 1.
4. STATIC STABILITY OF FLOATERS

The hydrostatic calculation is initially carried out to
determine the optimal size and shape of the semisubmersible floater that will provide sufficient
stability in unmoored operating conditions. The
parameters that have been considered for the static
stability analysis are adequate restoring in pitch
motion to limit pitch angle to 10 degrees beyond
which the wind turbine loses substantial efficiency,
fixed metacentric height (GM) of 1.0m. The system
should be stable within the standard threshold value
of heel angle and also must maintain an acceptable
steady-state heel angle (less than 10 degree) in
maximum static wind loading conditions (Wayman
and Sclavounos, 2007 and Tracy, 2007).
The static stability analysis was carried out for six
different semi-submersible floaters, with a square
deck and fixed GM of 1.0 m, of varying D/d ratio (i.e.
ratio of the diameter of pontoon to the diameter of
column) and h/H ratio (i.e. ratio of the height of
column to height of pontoon). The D/d ratio is 1.0 for
h/H ratio equal to 1.0 and D/d ratio is 1.5 for H/h ratio
ranging from 2.0 to 5.0. The semi-submersible is
designed with restoring moment greater than that of
the design restoring moment given by Eq. (2). The
detailed comparisons of hydrostatic and mass
properties of the semi-submersible type floater are
given in Table 1. The semi-submersible floater mass
has been varied from 2285.32 t to 2405.52 t and the
ballast from 3360 t to 4800 t for the different models
considered in this study.
The static wind thrust is calculated based on the 1-D
blade momentum theory, the disk is considered
friction less and there is no rotational velocity
component in the wake. The force in the stream wise
direction resulting from the pressure drop over the
rotor is the thrust, FThrust and is given by Eq.1.
FThrust = 2 CT a(1 a )Vo 2 A
Figure 1. Schematic view of the semi-submersible

type floater
3. DETAILS OF THE NREL 5 MW WIND TURBINE
MODEL
The wind turbine used in this study is the National

Renewable Energy Laboratory (NREL), USA, 5-MW
offshore baseline wind turbine model. This model
does not correspond to an operating turbine, but it is a
realistic representation of a three-bladed upwind 5MW wind turbine; its properties are drawn and
extrapolated from operating machines and conceptual
studies. NREL suggested a 5 MW wind turbine or
higher for offshore (Jonkman, 2009). It is variable
speed, upwind rotor orientation model with a rotor of
126 m diameter at a hub height of 90 m and mass of
the turbine is 697.46 t.
(1)
where Vo is the wind speed, A is rotor area; is

density of air; CT is the coefficient of thrust; a is
Axial inflow factor and is taken as 0.333 (Freris,
1990). The value of a is considered for the condition
at which the turbine generates more (max) power. The
design restoring moment
(k55, Design) for the semisubmersible floater in pitch motion is given by Eq.2
(Freris, 1990, Manwell, 2002).
k 4 4 , D esig n = F B Z B + g I T M S g Z S
(2)
where FB is the buoyant force; Z B is the centre of

buoyancy; IT is transverse moment of inertia of the
floater; M S is the total mass and center of the gravity
of the system; Z S is the center of the gravity of
system.
102
Table 1 Hydrostatic and mass properties of semisubmersible floaters
Details of semisubmersible floaters
Model Model
1
2
Model
3
Breadth to length ratio

(B/L)
Pontoon diameter to
column diameter (D/d)
ratio
1.5
1.5
Column height to
pontoon height (h/H)
i
Pontoon diameter, D
(m)
Free board (m)
6.5
6.5
6.5
Operating draft (m)
17.0
17.5
18.0
Displacement (t)
7052
6890
6605
5. AERODYNAMIC LOAD ON ROTOR BLADES
The blade element momentum theory (BEM), based

on momentum theory and blade element theory, is
used to find the forces on the rotor blades. Momentum
theory finds the forces at the blade based on the
conservation of linear and angular momentum. Blade
element theory determines forces at a section of the
blade based on the function of blade geometry.
9.8
15.3
15.2
Pontoon height, H (m)
10.3
7.3
5.2
Column diameter,d (m)
9.8
10.2
10.1
Column height, h (m)
10.3
7.3
10.5
Free board (m)
6.5
6.5
6.5
FD = 0.5CD ( ) airVrel2 br
(3)
Operating draft (m)
20.7
14.5
15.7
FL = 0.5CL ( ) airVrel2 br
(4)
Displacement (t)
6343
7905
7359
Details of semisubmersible floaters
Model Model
Model
Breadth to length ratio

(B/L)
Pontoon diameter to
column diameter (D/d)
ratio
1.5
1.5
1.5
Pontoon diameter,D(m)
14.9
14.8
14.6
Pontoon height, H (m)
4.3
3.5
3.5
Column diameter,d (m)
9.9
9.9
9.7
Column height, h (m)
12.8
14.0
15.0
Column height to
pontoon height (h/H)
The combination of these two methods is used to

analyse the blade elements and is called as strip
theory or blade element momentum (BEM) theory. In
this calculation the aerodynamic interactions between
the strips are ignored (Freris, 1990, Manwell, 2002).
The drag (FD) and lift forces (FL) for each section of
the blade are given by Eq.3 and Eq.4.
where FL is the aerodynamic lift force; FD is the

aerodynamic drag force; CL is the aerodynamic lift
coefficient; CD is the aerodynamic drag coefficient b
is the airfoil cord length ; is the angle of attack ;
Vrel is the relative velocity and r is the radial length
of blade sections. The axial force on the rotor axis is
combination of lift and drag forces as given in Eq.5.
Fx = FL cos + FD sin
(5)
where FL is the aerodynamic lift force; FD is the

aerodynamic drag force; is the angle of inflow.
6. WAVE ENVIRONMENT IN INDIAN COASTAL

WATERS
The wave parameters, frequency of the spectral peak

(p) and significant wave height (Hs) of the Indian
Ocean environment (east and west directions) are
referred from the literature published (Sannasiraj,
2007). Three types of sea states are considered for the
analysis namely moderate, rough and very rough
based on the magnitudes of Hs as 1.67m (sea state-4),
3.22m (sea state-5) and 5.30m (sea state-6). The
corresponding peak frequency associated with these
sea states are 0.914 rad/s, 0.688rad/s and 0.60rad/s
respectively. In the present study, P-M spectrum has
103
been used for the sea state representation. The

spectral density for fully developed seas represented
by P-M spectrum is given by (Chakrabarti, 1987):
S ( ) = g 2 5 exp[1.25(
S ( ) is
where
5 H s 0
2
the
4
) ]
0
wave
(6)
spectral
ordinate;
, HS is significant wave height, 0 is the

16 g 2
peak frequency and g is the acceleration due to
gravity. For wind environment, power-law is used to
represent the mean wind speed variation.There are
many mathematical wind spectrum models are
available to represent the turbulence. In offshore
context, the effect of lower frequency components of
longitudinal velocity fluctuations is important
(Manwell, 2002). In our present study Harris wind
spectrum is considered for the analysis. The
mathematical form of the Harris wind spectrum is
given by (simiu and scanlan, 1996 and Manwell,
2002).
nS ( f )
=4
u 2
(2 + 2 )
5
6
1800 f
U 10
Figure 2 Multisurf model of semi-submersible floater
(7)
where S ( f ) is the spectral ordinate at frequency f ;
u2 is the shear velocity or frictional velocity of flow

field.
7. HYDRODYNAMIC ANALYSIS USING WAMIT
WAMIT (Wave Analysis MIT) uses threedimensional boundary integral equation method
(BIEM), to solve the linearized hydrodynamic
radiation and diffraction problems for the interaction
of surface waves with stationary (zero forward speed)
floating structures in the frequency domain. The
hydrodynamic analysis of the floaters are carried out
using WAMIT for three different wave directions
namely 0, 45 and 90. The water depth considered
is 100m and the analysis is carried for wave periods
ranging from 2 s to 32 s. It is necessary to introduce
external damping as suggested by Chakrabarti et al.
(2007) in the diffraction analysis to get reasonable
estimates of RAOs in the regions of interest. A
reasonable damping ratio of 4% in heave and 5% in
pitch were considered in the WAMIT analysis for this
floater. The panelized view of the semi-submersible
floater is shown in Figure 2. The comparison of the
RAO for 0 wave heading angle is shown in Figure 3.
104
Figure 3 RAO comparison of semi-submersible with

different aspect ratios for 0 wave heading angle
8. LOAD CONDITIONS FOR THE RESPONSE

ANALYSIS
The correlation between inflow wind field and the sea

state is not considered in this study. The wave and
wind events are considered as independent events for
the analysis.
8.1 Case (I) Response Analysis of Floater with Wave
Only
This condition may exist in case of pre-installation

stage of the offshore wind turbine. The maximum
significant height of surge response is 2.91m, the
maximum significant height of heave response is
3.63m and occurred for sea state-6 of 0 wave
heading. The significant height of pitch response is
less than 1 for all the sea states. The response statics
of the semi-submersible floater listed in Table 2 and
the response spectrum obtained are shown in Figure 4
for case I for a wave heading angle of 0.
Table 2 Response statistics for Case I
Sea states
Sea state4
00
heading
Sea state5
00
heading
Sea state6
00
heading
Modes
Surge(m)
Heave(m)
Pitch(deg)
Surge(m)
Heave(m)
Pitch(deg)
Standard
deviation
Hrms
Hs
0.108
0.306
0.432
0.037
0.104
0.147
0.008
0.023
0.032
0.369
1.044
1.476
0.312
0.883
1.249
0.019
0.054
0.076
Surge(m)
0.728
2.058
2.910
Heave(m)
0.906
2.563
3.624
0.031
0.086
0.122
Pitch(deg)
Figure 4 Response spectrums of semi-submersible

floater for different sea states - Case I
8.2 Case (II) Response Analysis of Floater with Wind
and Wave
The maximum significant height of surge response is

5.72 m for sea state-6 of 0 wave heading angle. The
maximum significant height of heave response is
3.63m and occurred for sea state-6 of 0 and 45
wave heading angles. The maximum significant
height of pitch response is 2.2 occurred for all the
sea states as the system response is wind dominated
one. The wind load is constant for all the sea states
since the operating wind speed is considered for the
determination of aerodynamic forces and moments.
But the wave load will vary according to the sea state
and the angle of attack. The response statics of the
semi-submersible floater listed in Table 3 and the
response spectrum obtained are shown in Figure 5 for
case II for a wave heading angle of 0
Table 3 Response statistics for Case II
Sea
states
Sea
state- 4
00
heading
Sea
state- 5
00
heading
Sea
state- 6
00
heading
105
Modes
Standard
deviatio
n
Hrms
Hs
Surge(m)
1.234
3.491
4.937
Heave(m)
0.037
0.104
0.147
Pitch(deg
0.435
1.230
1.739
Surge(m)
1.284
3.631
5.135
Heave(m)
0.312
0.883
1.249
Pitch(deg
0.435
1.231
1.741
Surge(m)
1.429
4.041
5.715
Heave(m)
0.906
2.563
3.624
Pitch(deg
0.436
1.233
1.743
response statistics are compared for three different sea

states.
REFERENCES
Bulder, B.H., A. Henderson, Huijsmans and J.M.
Peeringa (2002) Floating offshore wind turbines for
Shallow waters, ISOPE Conference
Chakrabarti, S. K., (1987), Hydrodynamics of Offshore
Structures, Computational Mechanics Publications,
Southampton, UK
Jonkman J M, Butterfield S, Musial W, and Scott G
(2009), Definition of a 5-MW Reference Wind Turbine
for Offshore System Development, Technical report,
NREL Laboratory, USA
Jonkman, J. M. And M. L. Buhl, (2007) Loads
Analysis of a Floating offshore wind turbine Using Fully
Coupled Simulation, Wind Power Conference
Jonkman, J. M. (2007), Dynamics Modeling and Loads
Analysis of an offshore Floating wind turbine, National
Renewable Energy Laboratory Technical Report
Lee, K. H. (2005) Responses of Floating wind turbine to
wind wave Excitation, M.S Thesis, Department of Ocean
Engineering, Massachusetts Institute of Technology
Manwell J.F (2002), Wind energy explained, John
Wiley & Sons.USA
Sannasiraj S A (2007), Coastal problems and mitigation
measures- Including the effects of Tsunami, Indo-Japan
workshop, India
Tempel, J.V. (2006) Design of supporting structures for
offshore wind turbine, Ph.D.Thesis, Delft University,
Netherlands
Tracy, C. (2007) Parametric Design of Floating wind
turbine, M.S. Thesis, Department of Ocean Engineering,
Massachusetts Institute of Technology
Figure 5 Response spectrums of semi-submersible

floater for different sea states- case II
Wayman, E.N. And P.D. Sclavounos (2007) Coupled

Dynamic Modeling of Floating wind turbine Systems
offshore Technology Conference.
9. CONCLUSION
Static stability, dynamic and response analyses are

carried out to find out the behavior of semisubmersible floater in three different seas states. All
these floaters were designed with transverse
metacentric (GM) height equal to 1.0. Hydrodynamic
analysis was carried out using WAMIT, for different
semi-submersible type floaters with D/d aspect ratio
ranging from 1.0 to 1.5 and h/H aspect ratio ranging
from 1.0 to 5.0. The semi-submersible floater with
D/d=1.5 and h/H=1.0 yielded minimum RAO in
heave and semi-submersible floaters with D/d=1.0
and h/H=1.0 and D/d=1.5 and h/H=5.0
yielded
minimum RAO in pitch for the range of wave period
excitations between 5 sec and 20 sec. The response
analysis is carried out in frequency domain and
106

EXPERIMENTAL INVESTIGATION OF SLAMMING LOADS ON A

WEDGE
Tanvir Mehedi Sayeed1, Heather Peng2 and Brian Veitch3
1
Faculty of Engineering and Applied
Science, Memorial University of
Newfoundland, St. Johns, NL, Canada.
E-mail: tms120@mun.ca
2
Science, Memorial University of
E-mail: hpeng@mun.ca
3
Science,
Memorial
University
of
E-mail: bveitch@mun.ca
ABSTRACT
This paper reports an initial series of wedge drop test results performed at Ocean Engineering Research Center
(OERC) to investigate the slamming loads acting on a wedge of 100 deadrise. The aim is to get accurate twodimensional impact loads to incorporate into the simulation of planing hull motion. Parameters varied in these
drop tests were the mass of the wedge and the drop height. These factors were found to have negligible influence
in predicting the maximum pressure coefficient. The analytical prediction method developed by Chuang [3] is
found to be an accurate tool for determining maximum slamming pressures. In the future, more experiments
could be performed varying the deadrise of the wedge to verify Chuangs [3] prediction method. Then this
method could be implemented in the numerical simulation of planing hull motions in waves.
Key words: Slamming, Hydrodynamic impact, Simulation, Planing hull.
1. INTRODUCTION
The study of planing craft is closely related to the
fundamental study of two-dimensional wedge impact
problem. Von Karman [11] was one of the pioneer
researchers in this field. He reduced 3D problem to
2D and simplified the cross section of floats of seaplanes to a wedge. He developed a theoretical model
based on the momentum theorem and the water-added
mass. His works were applied to the maximal
pressure estimation on the floats of hydroplanes
during sea landings. A similar study of twodimensional water impact on solid bodies was
conducted by Von Herbert Wagner [12]. Instead of
considering a wedge, Wagner reduced the problem to
dropping a plate on the water surface, considering that
the virtual plate width varies over time. There is good
agreement between the Wagner and the Von Karman
formulae in the particular case of wedge entering
water. Water Rise or Splash-up was not considered by
Von Karman but Wagner took this into account by
assuming some approximation. Payne [7] claimed that
refinements added to Wagners theory actually caused
more offsets and the original Von Karmans theory is
superior. He presented a model to calculate maximum
pressure away from keel which is an improvement on
Von Karmans theory. One of the first real drop tests
with wedge-shaped models were conducted by
Chuang [2]. The tests were performed with one rigid
flat bottom model and five rigid wedges with deadrise
angles of 1, 3, 6, 10 and 15 degrees respectively. The
pressures were measured at the keel and away from
the keel. The data from test results was used to
provide a set of charts or empirical relations for

estimating the maximum impact pressure due to rigidbody slamming of the wedges. It was concluded that
the effect of trapped air needs to be taken into account
for wedge angles between 0 and 3. S.L. Chuang [3]
& Stavovy [9] developed a prediction method for
determining slamming pressures of a high speed
vessel in waves. This method is based on the Wagner
wedge impact theory, the Chuang cone impact theory
and NSRDC drop tests of wedges and cones. Then he
conducted some slamming tests of three dimensional
models in calm water and in waves. The experimental
results matched quite accurately with predicted
results. Zhao et al. [15] presented non-linear boundary
element method to solve water entry problem and
verified the results with experimental drop tests of a
wedge and a bow flare section. Engle & Lewis [4]
conducted experimental drop tests and made a
comparison between experimental results and several
numerical methods relating to the maximum waterimpact pressure of a symmetrical wedge for different
initial impact velocities. Wu et al. [14] analyzed 2-D
wedge free fall motion based on velocity potential
theory. They compared the similarity solution and
time domain solution with experimental drop test
results. Breder [1] performed the drop tests to
examine pressure loads on a rigid structure. He
conducted the two-dimensional wedge drop tests with
controlled vertical velocity, while earlier experiments
involved free fall water entry problem. Peseux et al
[8] used finite element method to solve highly nonlinear hydrodynamic impact problem. A series of drop
tests were conducted on rigid and deformable cone-
107
shaped structures. They validated their numerical

results with experimental data. Yettou et al. [14]
presented the results of experimental investigations of
the pressure distribution on a free-falling wedge
varying parameters such as drop height, the deadrise
angle and the mass of the wedge. A thorough
experimental investigation of constant velocity water
entry problem has been performed recently by
Tveitnes et al. [10]. They have provided some useful
information relating to wetting factor, flow
momentum drag and added mass based on those
experimental data analysis.
been specifically designed to achieve rigidity and

stiffness on impact and also to ensure that there is no
ingress of water on the inside of the wedge. The
wedge had a rectangular top on which attachments
were fitted to vary the mass of the wedge.
2. THE EXPERIMENT
The experiments were performed in the deep tank
of the Ocean Engineering Research Center (OERC) at
Memorial University of Newfoundland. Only vertical
drop tests were conducted varying the mass of the
wedge and the drop height with the 100 deadrise
model (Figure 2). The tests were conducted in calm
water and the wind induced loading was negligible.
The detail of the experimental set-up is given in
Mandeep et al. [5]. Still the instrumentation and data
acquisition system is briefly described here for better
understanding.
2.1 Description of the experimental set-up

The frame (Figure 1) used in the experiment was
constructed using T-Slotted aluminum extrusions and
was attached to the deep tank.
Figure 2. Design of wedge of dead rise angle 10

(Ref. [5]).
2.2 Instrumentation
A potentiometer cable extension transducer Celesco
(PT5MA-150-S47-DN-500) with a range of 150
inches has been used along with two accelerometers
(CTC Model AC140-2A) range 50g to measure the
instantaneous vertical position and accelerations. The
velocity is calculated by a numerical differentiation of
the position signal.
Four Piezoelectric pressure transducers (Kistler
Model 211B4) were used to measure the pressure on
the wedge surface. Their range is 0-200 psi and each
of them has diameter of 5.5 mm. They were arranged
along the median of transducer attachment on one side
of the wedge. Among the four pressure transducers,
three of them were close to apex and one was at the
corner end of the side.
Two rectangular electromagnets (BRE-4080-110)
of size 4" wide x 8" long x 2.5" high each
manufactured by Bunting Magnetics Co. have been
used so as to achieve remote automatic release of the
trolley and wedge. Magnets have a rating of 1000 lbs
for lifting application and are powered by 110 volts
DC power supply (BPS1-0150-110).
2.3 Release Mechanism
Figure 1. Front view of experimental frame with

wedge attached to it (Ref. [5]).
The wedge apex was aligned perpendicular to the
longitudinal axis of the tank. The wedge was attached
to a trolley made of similar aluminum extrusions. This
trolley slides on the guide rails fitted to the frame. The
idea behind guide rails is to achieve high vertical drop
speeds and high impact load bearing capacity. The
linear motion guide rails have been custom designed
from Macron Dynamics Inc. for these experiments
[5].
The 10 deadrise model (Figure 3) was made from
0.5 inch thick clear acrylic sheets. The wedge has
The Electro-magnets were fitted on the top of frame

on underside of the cross-bar (Figure 4) and was
electrically controlled to trigger the release of the
trolley.
2.4 Data acquisition & Sampling Frequency

Data acquisition was carried out using a
combination of 8-channel (Differential) data
acquisition system DaqBook/2001 (16 bit 200 kHz)
Series and an 8-channel expansion card DBK85 (both
manufactured by IOtech Inc.) to get 16-differential
channels. Preliminary experimental runs have been
carried out to verify measurement repeatability and
accuracy of the sampling frequency. To check the
validity of the chosen sampling frequency, the well-
108
known NyquistShannon theorem has been applied.

The best sampling frequency was found 5 kHz for the
drop tests that have been carried out.
3.5
2.5
Impact zone
In air before
impact
Velocity (m/s)
1.5
0.5
-0.5
0.8
1.2
1.4
1.6
1.8
Time (s)
Figure 5. Wedge velocity as a function of time (No

extra mass, drop height =60 cm).
3. EXPERIMENTAL DATA ANALYSIS

The 100 degree deadrise model was dropped from
drop heights of 40 cm and 60 cm with additional
masses of 20 kg and 40 kg. For each configuration,
the trends of the results are similar, thats why only
few cases represented here.
Figure 4 shows a typical case of the wedge
displacement curve recorded by the position cable
transducer. The wedge apex hits the water after a free
fall distance of 60 cm with a maximum velocity 3.43
m/s which is much clear in Figure 5.
200
PT-1
180
PT-2
PT-3
160
PT-4
140
Pressure (kPa)
Figure 3. Trolley release mechanism (Ref. [5]).
Figure 7 shows the spatial pressure distribution at

specific times. Each solid line corresponds to a
discrete approximation of the spatial pressure
distribution at the time of the peak pressure. At
t=2.0168s, only one point is shown by circle which
corresponds to the spatial pressure distribution at the
time of the peak pressure on transducer no.1. In this
case the water has reached only the first transducer
and the pressure on all the other transducers is very
small or close to zero which are not shown in the
figure for the sake of clarity. The dotted line joins the
peak pressures of all transducers.
120
100
80
60
1
40
20
0.8
Impact zone
-20
2.01
2.02
2.03
D isplacem ent (m )
2.04
2.05
2.06
2.07
2.08
Time (s)
0.6
In air before impact
Figure 6. Recorded pressure by four pressure

transducers during impact (extra mass=20 kg, drop
height =40 cm).
0.4
0.2
200
t=2.0168 s
t=2.0204 s
t=2.0242 s
t=2.0524 s
Peak pressure
180
160
-0.2
1.2
1.4
1.6
1.8
2.2
2.4
2.6
2.8
140
Pressure (kPa)
Time (s)
Figure 4. Displacement raw data as a function of time

(No extra mass, drop height = 60 cm).
120
100
80
60
A typical case of the wedge velocity with time

identifying the free fall and impact zone is presented
in Figure 5. This curve is obtained by filtering the
position raw data and using a numerical
differentiation. It is evident that the zone of interest
where the maximum kinetic energy is lost lasts less
than 50 ms following the impact.
40
20
0
100
200
300
400
500
600
700
Position of transducers (mm)
Figure 7. Pressure distribution on the face of the

wedge at different times during impact (extra
mass=20 kg, drop height =40 cm).
109
To describe the influence of the parameters on

slamming pressure, two non-dimensional quantities
have been used:
Drop height=40 cm
90
h
v(t )dt
(1)
and pressure coefficient, C p =
p pa
(2)
1
(v(t ))2
2
velocity of the wedge, pa is the atmospheric pressure
which is assumed to zero and is the mass density of

water.
Figure 8 shows the influence of the mass of the
0
wedge on pressure coefficient for the 10 deadrise

model. Three different masses were considered and
they were dropped from the same height of 40 cm.
The results are plotted for the time when the peak
pressure is located at pressure transducer no.3 for all
cases. It is evident that the mass of the wedge has
negligible influence in predicting maximum pressure
coefficient at the maximum dimensionless entry
depth, which was also reported by Yettou et al. [14].
The value of the maximum pressure coefficient is
found in the order of 80 and is approximately
constant, which also matches with the experimental
results of Zhao et al. [15] and analytical results of Mei
et al. [6].
Figure 9 shows the influence of initial drop heights
0
on pressure coefficient for the 10 deadrise model.

Two drop heights were considered for the model with
an extra mass of 40 kg. The results are also plotted for
the time when the peak pressure is located at pressure
transducer no.3 for all cases. Same conclusion can be
drawn that the maximum pressure coefficient is
independent of drop height which was also observed
by Yettou et al. [15]. The magnitude of the maximum
pressure coefficient is also in the order of 80 as was
found by the experimental results of Zhao et al. [15]
and analytical results of Mei et al. [6].
100
No extra mass
Extra mass=20 kg
Max. Cp
Extra mass=40 kg
80
Pressure coefficient,C
70
60
50
70
60
50
40
30
10
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
Entry depth,
Figure 9. Effect of drop height on pressure coefficient

as a function of dimensionless entry depth with extra
mass=40 kg
The maximum entry depth for the above cases
corresponds to pressure transducer no. 3 which was
the last transducer in contact with water at that instant.
This maximum entry depth is also found remaining
constant in the order of 1.5 as was reported by Yettou
et al. [14].
4. COMPARISON WITH CHUANGS [3]

PREDICTION METHOD
S.L. Chuang [3] developed a prediction method for
determining slamming pressures of a high speed
vessel in waves. This method is based on the Wagner
wedge impact theory, the Chuang cone impact theory
and NSRDC drop tests of wedges and cones.
According to this method, the pressure acting normal
to the hull bottom in the slamming area may be
separated into two components [10]:
1. The impact pressure pi , due to the normal
2.
component to wave surface of the relative

velocity between the impact surface and the
wave.
The planing pressure p p , due to the
tangential component to wave surface of the

relative velocity between the impact surface
and the wave.
The planing pressure is usually small and
insignificant compared with the impact pressure. The
total pressure due to normal velocity component of
the vehicle both normal and tangent to the wave
surface is therefore
pt = pi + p p
(3)
In this paper, we only summarized the simplified
case of wedge impact pressure in calm water. To
estimate the maximum impact pressure, the pressure
velocity relation is written as,
40
30
20
10
Max. Cp
20
where h is the vertical height of a given point on the

wetted surface measured from the apex, v(t ) is the
90
Drop height=60 cm
80
Pressure coefficient,C p
The entry depth, =
100
0.2
0.4
0.6
0.8
1.2
1.4
pi = kVn2
(4)
where k is a non-dimensional coefficient, is the
mass density of water and Vn is the relative normal
Max
1.6
Entry depth,
Figure 8. Effect of extra mass on pressure coefficient

as a function of dimensionless entry depth at drop
height =40 cm.
velocity of impact body to wave surface.
110
The relative normal velocity, Vn is determined on
180
the hypothesis that only the velocity component of the

moving body normal to the impact surface and the
velocity component of the wave normal to its surface
generate the impact pressure [10]. For the case of
calm water impact, Vn becomes
140
Pressure (kPa)
120
(5)
Vn = Vv cos 2
where Vv is the vertical impact velocity and is the
100
20
-20
2.05
2.1
2.15
0.0233896 + 0.0013578 0.00003132

3
180
268.49
166.51
150.67
262.26
249.41
186.53
171.61
268.64
265.29
PT-2
PT-3
PT-4
140
Pressure (kPa)
120
100
80
60
40
20
-20
2.02
2.04
2.06
2.08
2.1
2.12
2.14
2.16
2.18
2.2
Time (s)
Figure 11. Comparison of recorded pressure with

Chuangs [3] prediction method (extra mass=20 kg,
drop height 40cm).
Chuang (1973)
PT-1
PT-2
PT-3
PT-4
250
Pressure (kPa)
200
Table 1: Comparison of maximum pressure with

Chuangs [3] prediction method
280.64
Chuang (1973)
PT-1
160
For all the cases of drop tests, pressures have been

calculated using this method. It has been found that in
each case, this method can predict the maximum
pressure quite accurate for practical use which is
summarized in Table 1.
Figure 10 through Figure 12 show the comparison
of recorded pressure with Chuangs [3] prediction
method for three cases. It is evident that Chuangs [3]
method can predict the maximum pressure almost
exactly, provided that the vertical impact velocity is
accurate. The main reason of the discrepancy of the
results is due to the dynamic noise. Since the velocity
is obtained by filtering the raw signal of displacement
data and then differentiating them, the vertical impact
velocity as an input was not perfectly accurate. This
also caused a little bit time delay in predicting the
maximum pressure.
Maximum pressure
[kPa]
(Chuangs [3] method)
2.25

Chuangs [3] prediction method (No extra mass, drop
height =40 cm).
(7)
where is the impact angle which is equal to the
deadrise angle in the present case.
Maximum
pressure [kPa]
(Experimental
result)
151.84
2.2
Time (s)
k1 = 2.1820894 0.9451815 + 0.2037541 2
No extra mass, 40 cm
drop height
No extra mass, 60 cm
drop height
20 kg extra mass, 40
cm drop height
cm drop height
cm drop height
cm drop height
60
(6)
The best approximate values of k1 is expressed by the

following equation obtained through the method of
curve fitting [10]. For 2.2 < 11 deg:
Configration
80
40
deadrise angle.
The non-dimensional coefficient, k is determined as
follows,
k = k1 cos 4
Chuang (1973)
PT-1
PT-2
PT-3
PT-4
160
150
100
50
150.49
2.45
2.5
2.55
2.6
2.65
Time (s)

Chuangs [3] prediction method (extra mass=40 kg,
drop height =60 cm).
5. CONCLUSIONS
An initial series of free fall drop tests have been
performed with the 100 deadrise wedge varying the
111
drop heights and the mass of the wedge. For each

configuration, the maximum peak pressure was found
in either pressure transducer no. 2 or 3, which
signifies that the peak pressure tends to increase from
keel towards the chine. There was a big gap between
pressure transducer no.3 and 4, which should be
covered with more pressure transducers in the next
experiments to depict the more accurate and complete
spatial pressure distribution. The maximum pressure
coefficient for this 100 model is found approximately
constant and in the order of 80 and does not depend
on drop heights and mass of the wedge. Chuangs [3]
prediction method has been found to predict
maximum slamming loads quite accurately for each
case, though dynamic noise caused some
discrepancies.
Follow up experiments would be performed
varying the deadrise of the wedge and changing the
deadrise in the same section to account for more
actual ship hulls. Some oblique drop tests also need to
be performed to get the insight of the slamming
phenomenon more accurately. Finally model tests
need to be carried out with planing hull in waves to
further verify Chuangs [3] method. Then this method
would be incorporated in the simulation of planing
hull motion.
ACKNOWLEDGEMENTS
The authors would like to thank Luke Rae, Trevor
and Daniel Sutow for their help in experimental set-up
and data analysis. The financial support by Natural
Sciences and Engineering Research Council (NSERC)
and Atlantic Innovation Fund (AIF) is also greatly
appreciated.
REFERENCES
[1] Breder, J., Experimental Testing of Slamming
Pressure on a Rigid Marine Panel, Masters
Thesis, Naval Systems KTH Aeronautical and
Vehicle Engineering, Stockholm, Sweden
(2005).
[2] Chuang, S.L., Experiments on Slamming of
Wedge-Shaped Bodies, Journal of Ship
Research, Vol. 11, No. 3, pp. 190-198 (1967).
[3] Chuang, S.L., Slamming Tests of Threedimensional Models in Calm water and Waves,
NSRDC report 4095 (1973).
[4] Engle, A., Lewis, R. (2003): A Comparison of

Hydrodynamic Impacts Prediction Methods with
Two-dimensional Drop Test Data Marine
Structures 16, pp. 175182.
[5] Mandeep, S.P., Qui, W., Veitch, B.,
Experimental Design and Plan for Determining
Slamming Loads, Proceedings of 8th Canadian
Marine Hydromechanics and Structures
Conference, St. Johns, NL, Canada (2007).
[6] Mei, X., Lui, Y., Yue, D.K.P., On the Water
Impact of General Two-dimensional Sections,
Applied Ocean Research 21, pp. 1-15 (1999).
[7] Payne, P. R., The Vertical Impact of a Wedge on
a Fluid, Ocean Engineering, Vol. 8, No. 4, pp.
421-436 (1981).
[8] Peseux, B., Gornet, L., Donguy, B.,
Hydrodynamic
Impact:
Numerical
and
Experimental Investigations Journal of Fluids
and Structures 21, pp. 277-303(2003).
[9] Stavovy, A.B., Chuang, S.L., Analytical
Determination of Slamming Pressures for HighSpeed Vehicles in Waves, Journal of Ship
Research, Vol. 20, No. 4, pp. 190-198 (1976).
[10] Tveitnes, T., Fairlie-Clarke, A.C., Varyani, K.,
An Experimental Investigation into the
Constant Velocity Water Entry of Wedge-shaped
Sections, Ocean Engineering 35, pp. 14631478 (2008).
[11] Von Karman, T., The Impact on Seaplane Floats
during Landing, Technical Notes, National
Advisory Committee for Aeronautics (1929).
[12] Wagner, H., Phenomena Associated with
Impacts and Sliding on Liquid Surfaces,
translation of ber sto und gleitvorgnge an
der oberflche von flssigkeiten, Zeitschrift Fr
Angewandte Mathematik Und Mechanik, 1932
NACA
Library,
Langley,
Aeronautical
Laboratory (1936).
[13] Wu, G.X., Sun, H., He, Y.S., Numerical
Simulation and Experimental Study of Water
Entry of a Wedge in Free Fall Motion, Journal
of Fluids and Structures 19, pp. 277-289 (2004).
[14] Yettou, El-M., Desrochers, A., Champoux, Y.,
Experimental Study on the Water Impact of a
Symmetrical Wedge Fluid Dynamics Research
38, pp. 4766 (2005).
[15] Zhao, R., Faltinsen, O.M., Aarsnes, J., Water
Entry of Arbitrary Two-dimensional Sections
with and without Flow Separation Proceedings
of 21st Symposium on Naval Hydrodynamics,
Trondheim, Norway (1996).
112
NUMERICAL INVESTIGATION OF BOW SLAMMING ON SHIPS

WITH LARGE FLARE
Md. Mashiur Rahaman1, Kun Zheng2, Hiromichi Akimoto 3
1
Dept. of Environmental & Ocean Eng.
Graduate School of Engineering
The University of Tokyo
7-3-1 Hongo, Bunkyo-ku, Tokyo
113-8656, Japan
rahaman@triton.naoe.t.u-tokyo.ac.jp
2
Department of Systems Innovation
113-8656, Japan
tei@triton.naoe.t.u-tokyo.ac.jp
3
Department of Systems Innovation
113-8656,Japan
akimoto@sys.t.u-tokyo.ac.jp
ABSTRACT
Prediction of slamming loads due to wave impact is important not only for structural design but also for
safe maneuvering of ship. In the present study, computational fluid dynamics simulation technique is used for
predicting the slamming loads on ships with large bow flare advancing in waves. Finite volume method
(FVM) is used for discretization of time-dependent Reynolds averaged Navier-Stokes (RaNS) equation.
Overlapping grids technique is employed for simulating waves, ships interaction with waves and resultant
motion of ships. The free surface is captured by density-function method. The numerical results of two
container ship models SR108 and KCS with Fn=0.33, /LPP=0.133and high amplitude regular head sea and
120 deg. Oblique waves are considered. Simulations are conducted for 3 degrees of freedom (Heave, Pitch
and Roll). First, motions data are validated with experimental results and finally slamming phenomena is
analyzed by visualization technique.
Keywords: Slamming, Bow flare, Container ships, Visualization technique.
1. INTRODUCTION
Twenty first century is the age of containerization
for maritime transportation. Now-a-days almost 90%
of sea transport is by container. Therefore, the
number of container ships is increasing rapidly. For
larger accommodation of containers and; safe and
convenient cargo handling for containers, the recent
tendency has been to widen the bow flare angle. This
raises the flare slamming pressure excessively and
causes structural damage. Yamamoto et al. [1]
reported a serious structural damage due to bow flare
slamming.
Flare slamming describes dynamic wave impact on
the bow side shell structure above the design
waterline and during water entry, the bow structure is
subject to high pressure loads. It is therefore
necessary to evaluate the slamming loads in a
practical and sufficiently accurate manner.
Computational Fluid Dynamics (CFD) is one of the
possible method.
During design stage, slamming pressures are
usually obtained by using empirical formulae given
by the classification societies. However, there is
sizeable difference in the quantitative levels of
slamming pressure obtained by these formulae as
mentioned by Boitsov and Koudrin [2]. Therefore,
the necessity for direct calculation methods is
increasing for novel design.
Slamming has challenged many researchers since

von Karmans work [3]. He idealized the impact as
2D wedge entry problem on calm water surface to
estimate the water impact load on a seaplane during
landing with small deadrise angle. Zhao and
Faltinsen [4,5] used Boundary Element Method
(BEM) for predicting slamming loads on 2D wedge
and flared ship sections for splash effects and
variations in impact velocity. Their numerical
results for ship section compared favorably with the
experiments but over predicted for case of wedge
sections. They also showed that 3D flow effects are
significant and the numerical results were sensitive
to the length of the induced jet part. Sun and
Faltinsen [6] also used BEM for 2D slamming of
bow-flare ship with roll angle. They pointed out that
large roll angle causes very high localized pressure
in the flare area. Hermundstad and Moan [7,8] used
non-linear strip theory for predicting slamming
loads on ship hulls and validate the procedure for a
120m car carrier and 290m cruise vessel in bow and
bow quartering regular and irregular waves of
different heights.
To obtain slam loads on 2D sections, Arai et al.
[9,10] used finite difference method for discretization
of the Euler equations and volume of fluid (VOF)
method for free surface evolution. He obtained
favorable trends but the pressure did not co-relate
well with drop tests results probably due to
113
negligence of impact velocity variations in the

numerical procedure. Muzaferija et al. [11] and
Sames et al. [12] used finite volume method (FVM)
for discretization of the Navier-Stokes equations and
high resolution interface capturing (HRIC) scheme
for free surface to predict the slamming loads on a
2D sections by considering varying re-entry
velocities. Muzaferija et al. [11] obtained slam loads
on wedge section considering three-dimensional
effects and numerical results matched reasonably
well with the experimental results of Zhao et al. [5].
Sames et al. [12] mentioned that prescribed vertical
velocity histories significantly affected the
determination of realistic pressure levels. Reddy et al.
[13] simulated slam loads on 2D wedge section using
CFD techniques. They concluded that impact
velocity variations, domain size and three
dimensionality of flow are significant on the
numerical procedure for slamming loads prediction.
Ogawa et al. [14] conducted model tests of a postpanamax container carrier to examine the relationship
between ship motions and the water impact pressure
on the bow flare in heave and pitch free conditions.
They reported that the large magnitude of the impact
pressure is rare or not frequently occurred in the long
crested wave and the flare angle has an effect on it.
Most recently Kapsenberg and Thornhill [15] used
modified version of classical momentum theory
(monty) to capture the effects of wave steepness,
wave direction and CFD application ANSYS CFX to
predict slam loads on the bow section of a 173 m
ferry. They conclude that when the incoming wave
was well matched to the wave profile data, CFD
results are in good agreement with experiment.
Comprehensive review on slamming has also been
reported in [16,17].
2. NUMERICAL METHODS
WISDAM-X, developed at Miyata & Akimoto
Laboratory, The University of Tokyo utilizes
overlapping grid systems for obtaining ships
interaction with waves and resultant ship motions.
The total solution domain has two parts; inner
solution domain (O-H type grid) near the vicinity of
the hull and outer domain (rectangular grid) located
several ship lengths away from hull surface. The
numerical modeling of WISDAM-X is mentioned in
Table 1. Details are mentioned in Orihara and Miyata
[18].
The inner solution domain provides high resolution
around free surface and hull whereas the outer
solution domain extends to the outer boundary,
located several ships length away from the hull
surface. In half ship simulation, symmetry boundary
conditions are used at the center plane.
Table 1. Numerical modeling

Governing
RaNS equation,
equation
continuity equations
Pressure solution
Marker-and-Cell (MAC)
algorithm
method
Spatial
3rd order upstream
discretization
(advection term)
2nd order central (others)
Time
Explicit Euler method
discretization
Variables
Staggered mesh
arrangement
Turbulence model
Baldwin-Lomax model
Dynamic SGS model
Free
surface
Marker density function
treatment
(MDF) method
The size of computational grids for the present
analysis is mentioned in Table 2.
Table 2. Size of computational grids
Ship
types
Inner
Outer
and
hull
conditions
Half
1312171
SR108
Full
13121141
1463141
Half
1302170
KCS
Full
13021139
3. SIMULATION TOPOLOGY
3.1 Ship geometry
In real seas, among different types of slamming,
the occurrence of bow flare slamming is significant
and violent in case of container ships due to its
geometrical shape. Therefore, two different types of
container ship model SR108 (designed by National
Maritime Research Institute (NMRI), Japan
formerly Ship Research Institute, Japan) and KCS
(KRISO Container Ship; developed by Korean Ship
Research Institute now KORDI) are chosen for the
present analysis. The main principal particulars and
body plan of these two container ships are shown in
Table 2 and Figure 1 respectively.
Table 2. Principal particulars of ships
Particulars
Length
between
perpendiculars (m)
Length of waterline (m)
Breadth, moulded (m)
Depth (m)
Draught (m)
Displacement
Volume
(m3)
Block coefficient
SR108
175.00
KCS
230.00
178.20
25.40
15.40
9.50
24742
232.50
32.20
19.00
10.80
52030
0.5716
0.6505
The geometrical shape differences between the

two ships are mentioned in Table 3.
114
5. RESULTS AND DISCUSSIONS

Figures 2 and 3 represent the time history
comparison of numerical results with experimental
data [19] for heave and pitch motions respectively in
case of KCS. In the figures, the amplitude of heave
motions matches well with experiment but phase
difference occurs with increase of time whereas the
pitch agrees well with experiment.
KCS
Figure 2. Time history of heave motion
SR108
Figure 1. Body plan of container ship models
Table 3. Geometrical shape difference
Item
Bulbus bow
Flare angle
Stern shape
SR108
Small Bulb
High
Round
KCS
Present
Moderate
Transom
3.2 Simulation conditions

The designed Froudes number for KCS and
SR108 is 0.26 and 0.275 respectively. In present
numerical investigation, one extreme case is
considered. Therefore, simulations are conducted
with Froudes number, Fn= 0.33 and wave length
ratio, /LPP=1.33 in regular head waves and 120
degree oblique wave with heave and pitch (HP) free,
and heave, pitch and roll (HPR) free conditions . The
wave amplitude ratio is 1.12% of LPP for KCS and
1.21% for SR108.
Figure 3. Time history of pitch motion
Figure 4 shows the time history comparison of

total drag coefficient in head sea and oblique wave
conditions with heave and pitch (HP) free; and
oblique wave with heave, pitch and roll (HPR) free
conditions for KCS. The total drag co-efficient, CT is
calculated from the total drag, RT.
CT =
1 2
v S
2
where S is the wetted surface, v is speed of the ship
4. ANALYSIS METHOD
Due to unavailability of experimental results for
slamming, present analysis is made in the following
procedure.
KCSs motion data is validated with

experimental data
SR108 is simulated with same conditions of
KCS
Flow field data in the region of bow flare
section is visualized
Finally, slamming phenomena is analyzed
based on visualization technique
RT
and
is the density of water.
It is seen from Figure 4 that inclusion of roll

motion reduces the total drag coefficient. The peak
drag co-efficient for head wave with HP, oblique
wave with HP and oblique wave with HPR is 0.0181,
0.0116 and 0.0115 respectively. So, from practical
point of view roll motion must be considered in the
analysis of slamming.
115
Figure 4. Comparison of time history of total drag coefficient
Figure 5. Velocity vector on KCS in head wave with heave and pitch motions for initial instance, wave
crest and wave trough at FP of the ship
116
Slamming in the bow flare section is a local

phenomenon which has the global effects on the
ships hull. Present analysis is carried out by
visualization of velocity vector relative to the hull
with pressure color bar on the bow section with ship
motions in waves. However, due to unsteady

characteristics, three time instances in a single
steady wave period are considered like initial
instance, wave crest and wave trough at FP of the
ship.
Figure 6. Velocity vector on SR108 in head wave with heave and pitch motions for initial instance, wave
Figures 5 and 6 show the time sequence
visualization of velocity vector relative to the hull
with ship motions in the bow flare section on KCS
and SR108 in head wave with heave and pitch free
conditions respectively. In the figures, at the initial
instance of wave hit, for SR108, there is no velocity
in the flare section.
Both in KCS and SR108, the velocity vector is

inclined with flow directions but the angle of
inclination is higher in case of KCS.
In Figure 5, when the ship is on the wave crest,

separation of velocity vector occurs close to the
design water line in case of KCS. Again when the
ship is on the wave trough, the magnitude of
velocity is higher at some particular points in the
upper part of the bow flare section for KCS.
Figures 7 and 8 show comparison of the time

history of heave and pitch motions for KCS in
head wave and oblique wave. In the figures,
around 1.5 deg. pitch up and 0.1 meter heave
difference occur in case of head wave. Also there
is a difference in encounter frequencies.
There will be difference in the total drag, ships

motion in the two different kinds of ship models but
the differences that are identified from Figure 5 are
because of different bow shapes and flare angle.
117
Figure 7. Time history of heave motion
Figure 9. Velocity vector on KCS in oblique wave with heave and pitch motions for initial instance, wave
Figure 9 shows the time sequence visualization
of velocity vector relative to the hull with ship
motions in the bow flare section on KCS in oblique
wave with heave and pitch free conditions. No
separation of velocity vector near the design water
line like head wave in Figure 5 occur in this
condition when the ship is on the wave crest.
However, when the ship is on the wave trough, the

velocity vector before the flare section is going
upward towards bow region and after that it is
nearly parallel to the flow direction. It is also
noticed that the inclination of velocity vector for
oblique wave is low when comparing with head
wave in Figure 5.
118
Figures 10 and 11 show the comparison of time

history of heave and pitch motions for KCS in
oblique wave with HP; and oblique wave with HPR
free conditions. In the figures, there are almost no
differences in the amplitude of motions and
encounter frequencies.
Figure 10. Time history of heave motions
Figure 12 shows the time sequence visualization

of velocity vector relative to the hull with ship
motions in the bow flare section on KCS in oblique
wave with heave, pitch and roll free conditions. No
significant differences occur in the velocity vector
for KCS on oblique wave with HP (Figure 9) and
oblique wave with HPR (Figure 12).
Figure 12. Velocity vector on KCS in oblique wave with heave, pitch and roll motions for initial instance,
wave crest and wave trough at FP of the ship
119
In present numerical simulation, pressure is nondimensionalized by density and free stream velocity. The
non-dimensional pressure values of the cases for the
present analysis are given in Table 4.
Table 4. Non-dimensional pressure value
Model
KCS
SR108
Condition
Oblique wave with HP
Oblique wave with HPR
Head wave with HP
Head wave with HP
Value
0.00088
0.00094
0.00158
0.00211
It is seen from Table 4 that increasing the degrees of

freedom reduces the pressure and in head wave with
heave and pitch free conditions, SR108 has the higher
pressure than KCS.
6. CONCLUSIONS
A computational fluid dynamics technique, called
WISDAM-X and visualization technique has been
presented for investigating the slamming phenomena
on the bow region of two container ship models. The
degree of accuracy of ship motions with experimental
results is satisfactory and the visualization technique
describes which kinds of ships motions would be
included and specific region of bow to be considered
for analysis of ship slamming. In future, detailed
analysis of flow particulars and prediction of loads
due to slamming in the specific bow region that is
indentified by the present study will be carried out.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Takuya
Ohmori and Mr. Hiroyuki Saito of IHI Corporation,
Japan for their valuable comments and allow using
the mesh of KCS.
REFERENCES
[1] Yamamoto, Y., Iida, K., Fukasawa, T.,
Murakami, T., Arai M., and Ando, A.,
Structural damage analysis of a fast ship due
to bow flare slamming, International
Shipbuilding Progress, Vol. 32, pp. 124-136
(1985).
[2] Boitsov, G.V., Koudrin, M.A., Slamming
sstrength requirements for ship bow structures,
Journal of Marine Technology, Vol.38,No.2,
pp.102-105(April 2001)
[3] Karman, von, The impact of seaplane floats
during landing, NACA Technical Report 321.
[4] Zhao, R., and Faltinsen, O., Water entry of
two dimensional bodies, Journal of Fluid
Mechanics, Vol. 246, pp. 593-612 (1993).
[5] Zhao, R., Faltinsen, O. and Aarsnes,
J.V.,Water entry of arbitrary two dimensional
sections with and without flow separation,
Proceeding of Twenty first Symposium on
Naval Hydrodynamics , pp. 118-138 (1996).
[6] Sun, H., and Faltinsen, O., Water entry of a

bow-flare ship section with roll angle, Journal
of Marine Science and Technology, Vol. 14, pp.
69-79 (2009).
[7] Hermundstad, O.A. and Moan, T. , Efficient
calculation of slamming pressures on ships in
irregular seas, Journal of Marine Science and
Technology, Vol. 12, pp. 60-182 (2007)
[8] Hermundstad, O.A. and Moan, T. , Numerical
and experimental analysis of bow flare
slamming on a Ro-Ro vessel in regular oblique
waves, Journal of Marine Science and
Technology, Vol. 10, pp. 105-122 (2005)
[9] Arai, M. and Matsunaga, K., A numerical
study of water entry of two-dimensional shipshaped bodies, PRADS, Vol. 75, pp. 1-8.
[10] Arai, M., Cheng, L.Y. and Inoue, Y., A
computing method for the analysis of water
impact of arbitrary shaped bodies, Journal of
the Society of Naval Architects of Japan, Vol.
176, pp.233-240(1994)
[11] Muzaferija, S., Peric, M., Sames, P.C. and
Schellin, T.E., A two-fluid Navier-Stokes
solver to simulate water entry, Proceedings of
22nd Symposium on Naval Hydrodynamics,
pp.638-651(2000).
[12] Sames, P.C., Schellin, T.E., Muzaferija, S. and
Peric, M., Application of a two-fluid finite
volume method to ship slamming, Journal of
Offshore Mechanics and Arcitic Engineering
(OMAE), Vol. 121, No.1, pp. 1-7(1999)
[13] Reddy, D.N., Scanlon, T., Chengi, K.,
prediction of slam loads on wedge section
using computational fluid dynamics technique,
Proceedings of 24th Symposium on Naval
Hydrodynamics, pp. 37-49 (2002)
[14] Ogawa, Y., Matsunami R., and Arai, M., The
effect of a bow flare shape on the water impact
pressure, International Journal of Offshore
and Polar Engineering, Vol. 16, No. 2, pp.
112-117(2006)
[15] Kapsenberg, G.K. and Thornhill E.T. , A
practical approach to ship slamming in waves,
Proceedings of 28th Symposium on Naval
Hydrodynamics, California, 12-17 September
(2010)
[16] SNAME, Notes on ship slamming, Society of
Naval Architects and Marine Engineers,
Technical and Research Bulletin, pp. 230(1993).
[17] Faltinsen, O., Landrini, M. and Greco, M.,
Slamming in marine applications, Journal of
Engineering Mathematics, Vol. 48, pp. 187217(2004)
[18] Orihara, H. and Miyata, H., Evaluation of
added resistance in regular incident waves by
computational
fluid
dynamics
motion
simulation using an overlapping grid system,
120
Journal of Marine Science and Technology,

Vol. 8, pp.47-60(2003)
[19] http://www.gothenburg2010.org/Instructions_K
CS/Case_2.4/Case_2-4.htm
[20] Orihara, H. and Miyata, H., Evaluation of
added resistance in regular incident waves by
computational
fluid
dynamics
motion
simulation using an overlapping grid system,
Journal of Marine Science and Technology,
Vol. 8, pp.47-60(2003)
121

DRAG ANALYSIS OF DIFFERENT SHIP MODELS USING

COMPUTATIONAL FLUID DYNAMICS TOOLS
Salina Aktar 1, Goutam Kumar Saha2 and Md. Abdul Alim 3
1
Department of Natural Science
Stamford University
51, Siddeswari, Dhaka, Bangladesh
E-mail: salina854@gmail.com
2
Marine Engineering
Bangladesh University of Engineering and
Technology , Dhaka-1000, Bangladesh
Email: goutamkumar@name.buet.ac.bd
3
Department of Mathematics
Technology , Dhaka-1000, Bangladesh
Email: maalim@math.buet.ac.bd
ABSTRACT
Drag analysis based on CFD (computational Fluid Dynamics) simulation has become a decisive factor in the
development of new, economically efficient and environment friendly ship hull forms. In this research work
three-dimensional Finite Volume Method has been applied to determine the drag coefficient. The numerical
solutions of the governing equations have been obtained using commercial CFD software package FLUENT
6.3. Two conventional models namely Wigely and Series 60 are simulated to compute drag coefficient at
different Froude number in case of steady turbulent. Two turbulence models, namely, Standard k-, and Shear
stress transport (SST) k- are used to analyze turbulent flow. Velocity vectors as well as contour of pressure
distribution have also been displayed graphically. The computed results show good agreement with the
experimental measurements/numerical results obtained by other researchers.
Keywords: Drag co-efficient, Viscous drag co-efficient, Wave drag coefficient, Froude number, Total Pressure ,
Velocity vectors.
1. INTRODUCTION
One of the most active fields of ship
hydrodynamics research today is the development of
methods for computing the drag coefficient of the
steady viscous flow with free surface around a ship
hull. In these days with the development of new
numerical tools, the advances in computer technology
and the increase capability of data processing,
Computational Fluid Dynamic (CFD) has made
remarkable progress and allowed good results to be
obtained. The interest and demand of the industry to
implement new methods is one of the most important
reasons that influence the development of CFD. In
ship hydrodynamics, drag is also being named as
resistance.
It is vital to define the hydrodynamic performance
of the hull, to calculate the engine power, capable to
overcome the hydrodynamic resistance produced by
the interaction of the hull with the flow. CFD allows
ship designers to create a computer-generated model
of a ship and then test the ship at various speeds in a
simulated environment. The result from the CFD
simulations is necessary to understand the
complicated flow characteristics for an optimal hull
design, which includes a low drag and high
propulsive efficiency.
This allows designers to determine if the total

resistance of the ship is at an acceptable level from a
financial standpoint as well as a physical standpoint.
The financial perspective relates to the cost of the
engine and the fuel that the engine consumes in order
to meet the ships mission requirements.
There have been some experimental works on drag
characteristics around the ship models available for
validation of CFD. Sangseon [1] and Sakamoto et al
[2] made an extensive review on various types
resistance for the Wigley parabolic hull based on
ITTC 1957 Model-Ship Correlation Line and using
RANS(Reynolds-Averaged-Navier-Stokes)
simulation. Fonfach et al [3] used Series 60 model on
drag calculation by CFD code. Gray[4] investigated
frictional resistance using another type of ship model.
Previously one research work by Banawan et al [5]
and one conference by Ozdemir et al[6] were
presented for the computational analysis flow around
ships. Computation of drag coefficient using turbulent
model on RANS simulation were described by
Repetto [7] and Senocak et al[8].
The computed results has been compared with
other CFD method named Boundary Element Method
(BEM) from Saha[9]. The two models most used in
CFD to solve the turbulent phenomena, are the
Standard k- model and the Shear Stress
123
Transport(SST) k- model have been studied in

[10,12].
Turbulent Model Equations:

a) Standard k- model
The standard k model is a semi-empirical model
based on model transport equations for the turbulence
kinetic energy k and its dissipation rate The model
transport equation for k is derived from the exact
equation, while the model transport equation for
was obtained using physical reasoning and bears little
resemblance to its mathematically exact counterpart.
2. MATHEMATICAL MODEL
Total drag coefficient is normally broken down into
a Froude number dependent component-wave drag
co-efficient (residuary drag co-efficient) and a
Reynolds number dependent component-viscous
drag co-efficient (frictional drag co-efficient).
In the derivation of the k model, it was assumed

that the flow is fully turbulent, and the effects of
molecular
viscosity
are
negligible.
The
standard k model is therefore valid only for fully
turbulent flows.
The bracketed names give an alternative breakdown:

Total drag coefficient (Cd)
= Wave drag co-efficient (Cw) +
Viscous drag co-efficient (Cv)
Transport equations for the Standard k- model are

given by:
= Residuary drag co-efficient (Cr) +

Frictional drag co-efficient(CF)
The turbulent flow around hull including the free
surface is computed using the Reynolds-averaged
NavierStokes (RANS) equations for a threedimensional steady, incompressible and viscous
turbulent flow. Both air and water are considered as a
single fluid with variable properties. Two turbulent
models namely the Standard k- and Shear Stress
Transport k- are used in this category.
2.1 Governing Ship Flow Equations
u i
ui
xi
2

C 1 G k C 3 G b C 2
x j x j
k
k
In these equations
Gk represents the generation of
turbulent kinetic energy due to mean velocity

gradients, Gb is the generation of turbulent kinetic
energy
The coordinate system (x, y, z) for calculating the

viscous drag and the wave making drag is defined to
represent the flow patterns around hull form as
positive x in the opposite flow direction, positive y in
port side and positive z upward where the origin at
the aft perpendicular of the hull form, as shown in
Fig.1.
t
G kG b Y M
xi x j k x j
kj
due
to
buoyancy, Ym
represents
the
contribution of the fluctuating dilation in

compressible turbulence to the overall dissipation.
Here C1 , C2 and C3 are set equal to 1.44, 1.92
and 0.09, respectively. k =1.0 and
=1.3
are the
turbulent Prandtl numbers for k and .

b) The Shear Stress Transport (SST) k- Model
The SST k- turbulence model is a conglomeration of
the robust and accurate formulation of the Standard k model in the near-wall region, with the Standard k-
in the far field. The SST k- is more accurate and
reliable for a wider class of flows than the Standard k, including adverse pressure gradient flows.
Figure 1: Co-ordinate system
2.1.1Governing Equations of Fluent Model

Continuity equation:
u i
0
x i
(1)
Transport equations for the SST k- model are given

by:
k
k ku i k Gk Y kS k
t
xi
x j x j
Momentum transport equation:
u
1 p
ui i
xi xi
xi
u uj
i ui uj
xj xi xj
u u j 2
ui u j t i k i j
x j xi 3
(2)
(3)

ui GY DS
t
xi
xj x j
In these equations, Gk represents the generation of
turbulence kinetic energy due to mean velocity
124
gradients, G represents the generation of , k and
3. COMPUTATIONAL METHOD
represent the effective diffusivity of k and ,

respectively, Yk and Y represent the dissipation of k
3.1 Computational Method for Viscous Drag
and due to turbulence, D represents the crossdiffusion term, S k and S are user-defined source
terms. The constants applied in the high Reynolds
number form of the SST k- turbulence model are
equal to:
k .11.176,.1 2.0, k ,21.0,.2 1.168, 1 0.31,

i,1 0.075, i,2 0.0828, k 0.41.
2.1.2 Governing Viscous Drag Equations
Typically the friction drag coefficient is predicted
using the ITTC57 ship model correlation line or
some similar formulation.
Frictional drag coefficient CF
Cf
0.5 V 2 S
is co-efficient of friction and Cf
where C f
0.075
log10 Rn 2
3.2 Computational Method for Wave Making Drag
The computational domain triad is subdivided into

two numerical grids; the numerical grid on the ship
hull surface and the numerical grid on the water
surface belong to the ship hull. The flow is
superimposed from point sources located near to the
patch centers and the boundary conditions are
satisfied in the average over surface "patches".
2.1.3 Governing Equations of Wave Making

Drag
2 2 2
0
x2 y2 z2
(4)
During the iterations the non-linear free surface

boundary condition is applied. The numerical grid on
the water surface is updated with respect to the
calculated wave heights and the calculated waterline.
A convergence has been achieved, if the changes of
the calculated deformation of the free water surface
and the calculated wave making resistance are less
than a certain criteria for each one.
Water-surface condition:
1

2
2 g
Ship hull surface condition:
n. 0
(5)
(6)
Free surface condition: The kinematic and dynamic

boundary conditions on the free surface:
(7)
x x y y z 0 at z =
1
g . U 2 0 at z =
2
(8)
Pressure on the hull surface by Bernoullis equation:
p p
1
U 2 . gz
2
(9)
Hydrodynamic force in the x-direction:
Rw
p p n dS
To solve the governing equations, the fluid domain is

subdivided into a finite number of cells and these
equations are changed into algebraic form via
discretisation process. Finite volume method is used
for the discretisation. The convective terms are
discretised using Second Order Upwind scheme and
the diffusion terms by central- difference scheme. The
coupling between the pressure and velocity fields was
achieved using SIMPLE (Semi-Implicit Methods for
Pressure- Linked Equation) algorithm.
To solve the governing equations, the fluid
domain is subdivided into a finite number of cells and
these equations are changed into algebraic form via
discretisation process. Finite volume method is used
for the discretisation. The convective terms are
discretised using Second Order Upwind scheme and
the diffusion terms by central- difference scheme. The
coupling between the pressure and velocity fields was
achieved using SIMPLE (Semi-Implicit Methods for
Pressure- Linked Equation) algorithm.
(10)
S S
4.
The ship model used for this study is Wigley

parabolic and Series 60, which are standard for shiphydrodynamics research, and are chosen because
these are used by ITTC research program. Two types
of hull modes used for the experimental and
computational test are, given bellow, and the
longitudinal profiles of the 3D model are shown in
Figure 2 and 3. The principal particulars of Wigley
and Series 60 hull are described in Table 1.
S Mean wetted surface, S Fluctuating part

Wave making drag Cw
COMPUTATIONAL DOMAIN AND

BOUNDARY CONDITION
Rw
0.5 V 2 S
125
The water condition was modeled as in the

experimental test, which means fresh water at 100
Celsius , density = 1000 kg/m3, dynamic viscosity =
0.0008 kg/m-s, thermal conductivity = 0.677 w/m-k.
4. GRID
GENERATION
SIMULATION CRITERIA
AND
In this study, 3D unstructured tetrahedral grids are

constructed around the Wigley parabolic and Series
60 hull. Unstructured grid of 224189 mixed cells with
49483 nodes is constructed on the surface of Wigley
model which are shown in Figure 4(a) and 4(b).
Figure 5(a) and 5(b) shows the same pattern of grid of
1018918 mixed cells with 194443 nodes for the
model Series 60.
Figure 3: Wigley hull model
Figure 3: Hull of Series 60 model
(a)
(b)
Figure. 4. Schematic diagram of the flow field around hull
with boundary condition
Figure 5: (a) Unstructured grid flow domain around and

(b) Enlarged view of grid around the hull of wigley
model
Table 1: Principal particulars of Wigley and Series 60 model

Dimension
Length
Between
Perpendicular [LPP]
Wigley
1.00[m]
Series60
1.00[m]
Breadth [B]
0.10[m]
0.133[m]
Draft [T]
0.0625[m]
0.053[m]
Block Coefficient
[CB]
0.44
Wetted surface area
0.135[m2]
0.60
0.168[m2]
The boundary condition was employed to simulate the

condition on the towing tank. At the inlet a uniform
flow is given. The free surface elevation was fixed at
the inlet and at the outlet a hydrostatic pressure outlet
boundary condition was used downstream; the
hydrostatic pressure at the outlet was calculated
assuming an undisturbed free surface. Smooth walls
with a free-slip condition were assumed for the top,
floor and the sidewall. Smooth walls with a non-slip
condition (u, v, w = 0) were assumed in the entire
hull.
It is ensured that the computation domain and the

number of grids are sufficient enough to calculate the
drag coefficient on the hull accurately. In external
flow simulations using Standard k- and SST k- the
computational grid should be in such a way that
sufficient number of grid points are within the
turbulent sub-layer of the ensuing boundary layer.
Unstructured tetrahedral grid is chosen because it is
easily adjustable to complex geometry.
Plotting the flowing parameters against the
number of iteration assessed convergence: Residuals
for mass, momentum and turbulence (target criteria =
1E-3) and Drag forces (X directions). The maximum
number of iterations was equal to 500. However, if
the convergence criteria are reached for all residuals,
the simulation was stopped before reaching 500
iterations.
126
8
7
6
5
Cd10-2
4
Cv10-3
Cw10-2
3
2
1
0
(a)
0.1
0.2
0.3
0.4
0.5
Fn
(a)
100
90
(b)
80
Figure 6.(a) Unstructured grid flow domain around

and (b) Enlarged view of grid around the hull of
series 60 model
70
60
Cd10-2
50
Cv10-3
40
Cw10-2
30
The computation of viscous (Cv), wave (Cw) and

total (Cd) drag coefficient by standard k- (SKE) both
for model wigley and series 60 at various Froude
number has been showed in Table 2. In Figure 8(a)
and (b) we see that with the increasing values of Fn, at
first Cw and Cd increases significantly and then
decreases whereas Cv decreases frequently. Different
Froude number [0.173, 2.05, 269, 0.355, and 0.476]
has been taken because wave making drag at these
Froude number are high that are hump positions.
To compare and validate the numerical results, use
have been made with another numerical named
BEM[9] and experimental result [3] for Series 60 hull
and also computed (SST) with another numerical
result named BEM[9] for Wigley hull .
20
10
0
0
0.1
0.2
0.3
0.4
0.5
Fn
(b)
Figure 8. Fn vs various drag coefficient of (a) wigley hull
(b) series 60 hull.
Table 4: Comparison of computed Cd by SKE, SST and
BEM method for Wigley hull
Fn
Table 2:Computed value of Cd, CV and CW of Wigley

hull
SKE
SST
-2
BEM
-2
5.210-2
0.173
3.910
0.205
4.610-2
6.010-2
3.910-2
6.110
Fn
Cd10-2
Cv10-3
Cw10-2
0.269
6.510-2
5.810-2
8.210-2
0.173
3.9
1.30
3.7
0.355
7.210-2
5.610-2
3.510-2
0.205
4.6
0.87
4.5
0.476
6.910-2
5.410-2
19.110-2
0.269
6.5
0.47
6.4
0.355
7.2
0.25
7.1
0.476
6.9
0.13
6.7
Table 5: Comparison of computed Cd with BEM method

and Experimental result for series 60 hull
Fn
Cd10
0.173
12.6
-2
Cv10
8.25
-3
Exp.
Fn
SKE
BEM
0.173
12.610-2
12.110-2
10.0710-2
0.205
30.810-2
22.510-2
44.210-2
11.7
0.269
90.710-2
39.110-2
102.0610-2
64.9710-2
Table 3: Computed value of Cd, CV and CW of Series 60

hull
Cw10
-2
0.205
30.8
5.65
30.2
0.355
87.610-2
32.610-2
0.269
90.75
3.08
90.4
0.476
85.110-2
60610-2
0.355
75.6
5.97
75.0
0.476
85.1
8.72
84.2
127
From Table 4 it is seen that the computed values by

SKE and SST agrees well with another computed
result by BEM [9] for Wigley hull. The difference
between the SKE and SST by 32% where as it is 30%
for BEM method. As can be seen that results obtained
from SKE are in good agreement with experimental
data. From Table 5 for the Series 60 hull it is also
seen that difference between the computed (SKE) and
BEM method is 27% whereas it is 2.51%from the
experimental [3] result. From both of the Table 2 and
3 the computed value has a nice agreement with the
BEM and experimental result.
Drag convergence history is shown in Figure 9.
Contours of total pressure around Wigley and Series
60 are shown in Figure 10(a) and 10(b) respectively.
Figure 11(a) and 11(b) show the plot of velocity
vectors around the hull. The path lines around the
hulls are shown in Figures 12(a) and 12(b).
(a)
(b)
Figure 10: Plot of total pressure of (a) Wigley and
(b) Series 60 hull
(a)
Figure 9: Drag convergence history
(b)
7. CONCLUSION
In this research work we have obtained drag coefficient of two models: Wigley and Series60 by using
CFD code FLUENT. We also compared numerical
results computed by SKE and SST models of
FLUENT and compared with BEM and experimental
values. From comparative values it is seen that total drag
coefficient for
FLUENT model, BEM and experimental
shows a good agreement in the range of Froude
number from 0.2 to 0.3.
Whenever the range
exceeds 0.3 there occur a large variation among the
computed, BEM and experimental result. This
variation can be minimized by decreasing the the
mesh size, refining mesh etc.
Based on the results of a CFD simulation, a ship
designer can choose optimum speed with minimum
power annd then proceed to a model test for
experimental result.
Figure 11: Plot of velocity vectors around (a) Wigley and

(b) Series 60 hull
(a)
8. ACKNOWLEDGMENTS
The authors thank MARTEC 2010 conference
organizer for giving us the opportunity to present the
research work.
(b)
Figure 11: Plot of path lines around (a) Wigley and (b)
Series 60 hull
128
9. REFERENCES
[1] Sangseon, J., Study of Total and Viscous
Resistance for The Wigley Parabolic Ship Form,
Iowa Institute of Hydraulic Research, IIHR
Report No.261. (1983).
[2] Sakamoto,N.,
Wilson,R.V
and
Stern,F.,
Reynolds-Averaged Navier-Stokes Simulations
for High-Speed Wigley Hull in Deep and
Shallow Water Journal of Ship Research, Vol.
51, No. 3, pp. 1872003,September (2007)
[3] Fonfach, J. M. A. and Soares,. C. G, Improving
The Resistance Of A Series 60 Vessel with a cfd
code, European Conference on Computational
Fluid Dynamics,ECCOMAS CFD,J. C. F.
Pereira and A. Sequeira (Eds),Lisbon, Portugal,
1417 June (2010).
[4] Gray, A. W., Preliminary Study on The Use of
Computational Fluid Dynamics to Determine
The Frictional Resistance of a Trimaran Ship,
M.S thesis, Department of Mechanical and
Aerospace
Engineering,
West
Virginia
University, Morgantown, USA, (2007).
[5] Banawan, A. A. and Ahmed, Y. M., Use of
Computational Fluid Dynamics for The
Calculation of Ship Rsistance, and Its Variation
with
The
Ship
Hull
Form
Parameters, Alexandria
Engineering
Journal, Vol. 45, No.1, (2006)
[6] Ozdemir, Y. H., Bayraktar, S. and Yilmaz, T.,
Computational Investigation of A Hull, 2nd
International Conference on Marine Research
and Transportation, Ischia, Naples, Italy, 28th30th June, (2007).
[7] Repetto, R.A., Computation of Turbulent FreeSurface Flows Around Ships and Floating
Bodies, PhD. Thesis, Technical University HamburgHarburg, Argentina (2001).
[8] Senocak,. I and Iaccarino,G, Progress towards

RANS simulation of free-surface flow around
modern ships,Center for Turbulence Research
Annual Research Briefs (2005)
[9] Saha,G.K.,Numerical optimization of ship hull
forms from viewpoint of wave making
resistance based on panel method , Ph.D
Thesis,
Yokohama
National
University,
Yokohama, Japan[2004]
[10] Mulvany, N. J., Chen, L., Tu, J. Y. and
Anderson, B., Steady-State Evaluation of TwoEquation.RANS (Reynolds-averaged Navier
Stokes) Turbulence Models for High-Reynolds
Number Hydrodynamic Flow Simulations,.
Maritime Platforms Division, Defence Size and
Technology Organization, Deparment of
Defence, Australian Government (2004)
[11] Versteeg, H.K. and Malalasekera, W., An
Introduction to Computational Fluid Dynamics,
Longman Scientific & Technical, England
(1995).
[12] Hoffmann, K. A. and Chiang, S.T.,
Computational Fluid Dynamics (Third Edition),
Vol. 2, Wichita KS: Engineering Educational
System, (1998).
[13] http://www.cfdonline.com/Wiki/Turbulence_
modeling
129

DYNAMIC RESPONSE BEHAVIOUR OF MULTi -LEGGED ARTICULATED

TOWER WITH & WITHOUT TMD
Srinivasan Chandrasekaran1, Bhaskar K2, Lino Harilal3 and Brijith R4
1
Department of Ocean Engineering,

Indian Institute Of Technology, Madras, India.
E-mail: drsekaran@iitm.ac.in
Tel: (Off) 91-044-22574821,
Fax: +91-044-22574802,
Mobile: +91-0-9444583179
Department of Civil Engineering

National Institue of Technology, Tiruchirappalli, India.
Email: kbaskar@nitt.edu
Department of Ocean Engineering,

Indian Institute Of Technology, Madras, India.
Email: linoharilal@gmail.com
Mobile: +91-0-9884606772
Department of Civil Engineering

National Institute of Technology, Calicut, India
Email: brijithrajan@gmail.com
ABSTRACT
Articulated tower platform is one of the compliant structures that are economically attractive especially as
loading and mooring terminal to deep waters. These platforms are lighter when compared to the conventional
fixed platforms. An articulated tower is a linear structure, flexibly connected to the sea bed through a
universal joint and held vertically by the buoyancy force acting on it. The part of the tower emerging from the
water supports the super structure designed to suit the particular application e.g. a tanker to be loaded etc.
As the connection to the sea bed is through the articulation the structure is free to oscillate in any direction
and does not transfer any bending moment to the base. Design methodologies of these towers ensures reduced
motion characteristics with less deck acceleration while loads at the articulated joint are kept to minimum;
this is required to establish sufficient stability under working conditions. In this paper the dynamic response
characteristics like bending stress variations and the displacement of the Multi-Legged Articulated
Tower(MLAT) are quantified through experimental investigations done by the author. A Tuned Mass Damper
[TMD] is a small secondary mass-spring-damper system attached to the bottom of the deck plate. Its natural
frequency is tuned near to the natural frequency of the MLAT vibration mode that is to be controlled. The
TMD inertia forces produced by this motion are approximately anti-phase to the dynamic wind forces driving
the MLAT. MLAT motion and hence the stresses are thus greatly reduced with the wave forces primarily
driving the TMD instead of the MLAT. The energy of this motion is dissipated by the internal damping
mechanism associated with the TMD. The MLAT itself is modelled as a single degree of freedom mass-springdamper system. The mass of this system is selected to give the same kinetic energy (at the TMD attachment
point) as the tower in the vibration mode under consideration. This mass, together with the natural frequency
of the vibration mode, define the effective spring constant. The damping ratio for this system is assumed closer
to 1% of the critical. The TMD mass is selected on the basis of maximum allowable TMD motion, maximum
allowable MLAT deflection and off-tuned performance.
The equation of motion of MLAT has been formulated based on the model. The various dynamic responses of
the structure have also been studied. The Bending Moment along the height of the model and the RAO are
plotted from the results obtained from the experiment.
Key words: Multi-Legged Articulated Tower, Tuned Mass Damper, Universal Joints,bending moment.
.
1. INTRODUCTION
The demand for oil and gas has brought the
offshore drilling and production of hydrocarbon
deposits to greater water depths. New concepts of
structural systems have been developed which are
suitable for deep water structures. Articulated
towers belong to the class of compliant towers

which have been found quite suitable for deep water
applications. Similar to a reed which bends but
does not break the articulated structures withstand
with suppleness the combined effects of the waves,
wind and currents.
131
There has been an increase in the use of

mobile offshore systems for the storage and loading of
oil into attendant tankers, particularly for fields that
have a limited production capability, or are too remote
from refining or terminal facilities to warrant the
laying of a pipe line. A typical mobile loading and
storage system is the articulated buoyant loading
tower which may have either a single universal joint
or a greater number of joints in the intermediate level
which can be used at very deep water depths. The
articulated tower with universal joints in the
intermediate level is called a multi-hinged articulated
tower.
The extension of the concept of the singleleg articulated tower led to the development of a new
type of platform with several columns which are
parallel to one another. The columns are connected by
universal joints, both to the deck and to the
foundation. The use of universal joint ensures that the
columns always remain parallel to one another and the
deck remains in horizontal position. There is no
rotation about the vertical axis of the columns. This
type of platform is called a multi-leg articulated tower
which has three or more columns. The advantage of
this system is that the pay loads and deck areas are
comparable with the conventional production
platforms in moderate water depths.
The single-leg articulated tower has been
used successfully for single-point mooring, control
tower and flare structure. This also can be used as a
production platform for marginal fields, as an early
production platform, or as a processing unit in deep
waters.
As the connection to the seabed is through
the articulation, the structure is free to oscillate in any
direction and does not transfer bending moment to the
base. Since the articulated tower is a compliant
structure and it freely oscillates along with the waves,
the wave force on the structure is much less than that
of a fixed structure. The dynamic amplification factor
is low compared to the other fixed structures since its
natural frequency is much less than the frequency of
the wave. Different types of articulated towers are
shown in Fig. 1.
Fig. 1 Types of Articulated towers

A Tuned Mass Damper [TMD] is a small secondary
mass-spring-damper system attached to the bottom of
the deck plate. Its natural frequency is tuned near to
the natural frequency of the MLAT vibration mode
that is to be controlled. When the MLAT begins to

oscillate, it excites the TMD into motion. The TMD
inertia forces produced by this motion are
approximately anti-phase to the dynamic wind forces
driving the MLAT. MLAT motion and hence the
stresses are thus greatly reduced with the wave forces
primarily driving the TMD instead of the MLAT. The
energy of this motion is dissipated by the internal
damping mechanism associated with the TMD.
The MLAT itself is modelled as a single
degree of freedom mass-spring-damper system. The
mass of this system is selected to give the same
kinetic energy (at the TMD attachment point) as the
tower in the vibration mode under consideration. This
mass, together with the natural frequency of the
vibration mode, define the effective spring constant.
The damping ratio for this system is assumed closer to
1% of the critical. The TMD mass is selected on the
basis of maximum allowable TMD motion, maximum
allowable MLAT deflection and off-tuned
performance. Brief literature review is as follows:
Concepts of single legged articulated towers, well
known in the literature [1-2] are extended to multilegged articulated platforms; in the latter, more legs
are connected parallel. Universal joints connecting
shaft to foundation and deck necessitates legs of the
tower to remain parallel and to keep the deck in a
horizontal position under environmental loads.
Buoyancy chamber, ballast, guy wires, axial piles and
tendons, used as restoring forces reduce the structural
stiffness that is required to constrain the structure's
motion. Variation of the outer diameter of buoyancy
chamber influences structures response more than
that of its length and position [3]. Structural responses
of single and double hinged articulated towers under
combination of wind and waves showed more
response under random wave condition in comparison
to the regular one [4-5]; researchers simulated
different wave and wind combinations using Monte
Carlo simulation technique.
The Articulated tower studied under
combined action of wind, waves and current also
showed increased response in comparison to those
under waves only [6]. A single leg inclined mooring
(SLIM) tower is developed as a concept for an
articulated structure to moor tankers with a single
hawser in shallow water. Instability of an upright
buoyant tower, addressing it as a special purpose
single point mooring (SPM) is discussed [7]. Results
showed that SLIM motions become near chaotic at
higher wave heights. Analysis methodologies of
slender articulated towers generally focus on their
nonlinear response characteristics; these algorithms
retain a relative motion quadratic drag component to
model the wave-structure interaction [8]. Studies
reported on single legged articulated towers under
deterministic and random waves included nonlinear
132
terms due to geometry and fluid-structure interaction,

arising from large angle rotations and drag and inertia
components, respectively [9]. It is shown that the
tower response depends on wave frequency and
amplitude; for most of the frequencies, tower exhibits
quasi-periodic behavior and for certain frequencies it
showed chaotic behavior. Nonlinear dynamic
response of multi hinged articulated tower carried out
using Lagrange equation approach showed
satisfactory results when applied to towers at deepwaters. Analytical studies conducted on tankermoored articulated tower showed that the tanker
dominates the response of the system and does not
respond to wave loads as rapidly when the tanker is
fully loaded [10]; the study considered operation
condition and environmental excitations as well.
Dynamic responses of multi-legged articulated towers
are dominated by drag forces and use of simplified
Morisons equation approach that neglects diffraction
effects is used in the current study. Experimental
studies on articulated towers are limited to few
researchers who demonstrated the influence of added
mass on the tower response [11]; however, the study
was limited to unidirectional regular waves.
Fujino and Abe [12] investigated the
characteristics of a multiple tuned mass damper
(MTMD) whose tuning frequency bandwidth was
distributed around the natural frequency of the
structure subjected to control. Wind tunnel model
tests on Sydney Tower were performed by Vickery
and Davenport [13], and on Citicorp Center, New
York City by lsyumov et al [14]. Those are two
notable earlier studies in which a model of a TMD
was incorporated in the building model to study the
effect of a TMD in reducing wind-induced responses.
Tsai and Lin [15] numerically developed plots to
obtain optimum damper parameters for harmonic
excitations.
Diameter of the tower

Thickness of the tower
Leg spacing
= 0.1 m
= 0.05 m
= 0.3 m
4.2 Parameters considered for the study

4.2.1 Regular Wave
Wave Height
Wave period
Wave angle
: 0.06 m, 0.09 m, 0.12 m and 0.15 m

: 1.3 s, 1.5 s, 1.7 s and 1.9 s
: 0o
4.2.2. Mass distribution

Sand ballast
Water ballast
3. OBJECTIVE
To have a detailed study on the experimental

investigation being done on the dynamic
response of MLAT with and without TMD
To formulate the equation of motion of
MLAT with and without TMD
4. EXPERIMENTAL MODEL
Materials for the model were selected such that the
elastic limit of material was not exceeded. The fourlegged articulated tower model was fabricated to a
scale of 1:200 using Perspex tubes of 100mm
diameter. Universal joints are used to connect the
Perspex leg to the base which acts as a hinge at the
base and oriented along the direction of propagation
of wave.
4.1 Model details (ref. Fig (2, 3 and 4))

Model height
Fig. 2 Articulated Tower
= 1.2 m
133
Fig 3(i) Mass distribution of the tower - water filled

columns
Fig 4(ii) Model with instrumentation - model fitted

with accelerometers and strain gauges
4.3 RESULTS AND CONCLUSIONS.
Heightofthemodel(m)
The bending moment along the height of the model

without TMD, is more when compared to the results
obtained from the model with TMD. The model is
tested with TMDs of mass 100gm, 150gm, 200gm
and 250gm. The results showed that as the mass of the
TMDs are increased the bending moment is
decreased. However in certain cases the bending
moment are higher for lower masses like 150gm.
Fig 3(ii) Mass distribution of the tower - sand filled

columns
BMdiagramfor
6cm,1.3secwave
withvaryingTMD
NOTMD
2
0
0
10
100gm
TMD
BM(Nm)
Fig 5.BM diagram
themodel(m)
BMdiagramfor
9cm,1.7secwave
withvaryingTMD
Fig 4(i) Model with instrumentation - universal joint
2
0
NOTMD
0
20
40
BM(Nm)
100GM
TMD
Fig 5.BM diagram

Inferences drawn from the study bring detailed insight
to the dynamic response behavior of the multi-legged
articulated towers under regular and random waves.
134
RAOOFSANDBALLASTAND9CM
WAVEHEIGHT
RAOofWaterBallastconditionand
constantwaveheight12cm
20
0
2
0
0.5
0.6
0.7
0.8
AOofBM
100gm
TMD
150gm
TMD
5
0
0.6
0.7
0.8
0.5
0.7
100gmTM
D
0.9
RAOBM
5
0
0.5
0.6
0.7
0.8
100gm
Tmd
5
0
0.5
0.6
0.7
Ff
0.8
0.7
0.75
100gm
TMD
0.8
10
0
0.5
0.6
0.7
0.8
100gm
TMD
10
0
0.5
0.6
0.7
0.8
100gm
TMD
5
0
0.5
0.6
0.7
0.8
100gm
TMD
The RAO graphs of the sand ballast

conditions are almost similar; starting with a
value 2.5 to 3 and decreasing to a value of
1.5 to 2. Along the frequency, 150 gm TMD
is showing higher value than the other cases
of TMD except in 15cm case. And the least
value by 250 gm TMD.
In the water ballast case for all wave height
the graphs start from a value of 3 to 4 and
end of in a RAO value of 2 to 2.5. The
graphs are not having a shape similar to the
sand ballast case as there is sloshing within
the legs.
Whereas in the no ballast condition it can be
seen that the RAO is very high value
compared to the sand and water ballast
condition. We can also see that the RAO is
least for the 250 gm TMD.
150gm
TMD
100gm
TMD
RAOofNoBallastconditionand
150gm
TMD
0.8
RAOofNoBallastconditionand
ofBM
RAOofBM
5
0
0.75
f
RAOofSandBallastconditionand
RAOofBM
0.65
15GMTMD
0.9
AOofBM
RAOofBM
5
0
0.7
RAOofNoballastconditionand
WAVEHEIGHT
0.5
0.7
5
0
100GM
TMD
RAOofBM
RAOofBM
WAVEHEIGHT
0.5
0.65
200gm
TMD
RAOofBM
RAOofBM
100gm
TMD
150gm
TMD
Based on the studies on experiments being conducted,

following conclusions are drawn:
135
Bending moment along the tower and deck

acceleration increases with the wave height.
By providing buoyancy chambers at the
appropriate locations, this sensitivity to the
increase in response can be controlled.
Bending moment and the deck acceleration
decreases with the increase in wave
frequency.
Bending moment of the structure decreases
with increase in the mass of TMD. However
in the case of 150gm TMD we see a
fluctuation. This may be due to resonance.
References
[1] Nagamani, K. and Ganapathy, C., 1996,
Finite element analysis of nonlinear
dynamic response of articulated towers,
Computers & Structures, Vol. 59(2), pp.
213-223.
[2] Sellers, L.L. and Niedzwecki, J. M., 1992,
Response characteristics of multi-articulated
offshore towers, Ocean Engg., Vol. 19(1),
pp. 1-20.
[3] Mas Murtedjo, Eko B. Djatmiko and Hendri
Sudianto, 2005, Influence of buoyancy
parameters on the dynamic behavior of
articulated tower, J. Mechanical, Vol. 19,
pp. 32-47.
[4] Mohd. Moonis Zaheer and Nazrul Islam,
2008, Fluctuating wind induced response of
double hinged articulated platform,
Proceedings of the 27th International
Conference on Offshore Mechanics and
Arctic Engineering, OMAE2008-57723,
Estoril, Portugal.
[5] Baskar, K and Nagamani, K., 1996,
Dynamic response of two hinged articulated
tower, Proceedings of the International
conference in Ocean Engineering, COE 96,
IIT Madras, India.
[6] Datta, T. K. and Jain, A. K., 1990,
Response of articulated tower platforms to
random wind and wave forces, Computers
& Structures, Vol. 34(1), 137-M.
[7] Subrata Chakrabarti, 2008, Instability
analysis of offshore towers in waves,
Engineering Structures, Vol. 30, 3193-3203.
136
[8] Bar-Avi, P. and Benaroya, H., 1996, NonLinear dynamics of an Articulated tower in
the ocean, Journal of Sound and Vibration,
Vol. 190(1), pp. 77-103.
[9] Oded Gottlieb, Solomon C.S, Yim and
Robert T. Hudspeth, 1992, Analysis of
nonlinear response of an articulated tower,
International journal of offshore and polar
engineering, Vol. 2(1), pp. 61-65.
[10]Bruce Joseph Gernon and Jack Y. K. Lou,
1987, Dynamic response of a
tanker
moored to an articulated loading platform,
Ocean Engg, Vol. 14(6), pp. 429-512.
[11]Nagamani, K. and Ganapathy, C., 2000, The
dynamic response of a three-leg articulated
tower, Ocean Engg, Vol. 27, pp. 1455
1471.
[12]Fujino, Y. and Abe, M. 'Dynamic
characterization of multiple tuned mass
dampers', Proceedings of 1st International
Conference on Motion and Vibration
Control, Yokohama, September 1992, pp.
176-181
[13]Vickery, B. J. and Davenport, A. G. "An
investigation of the behaviour in wind of the
proposed Centrepoint Tower, in Sydney,
Australia', Engineering Science Report
BLWT-1-70, University of Western Ontario,
Canada, 1971
[14]Isyumov, N., Holmes, J. D., Surry, D. and
Davenport, A. G. 'A study of wind effects for
the First National City Corporation project
New York, USA', Boundary Layer Wind
Tunnel Laboratory Special Study Report,
BLWT-SS1-75, University of Western
Ontario, Canada, 1975
[15]Tsai, H. C. and Lin, G. C. 'Optimum tuned
mass dampers for minimizing steady-state
response of support excited and damped
systems', Earthq. Engng Struct. Dynam.
1993, 22, 957-973

INVESTIGATION OF THE RESPONSE OF A BOXED SHAPED FREE FLOATING

BUOY IN REGULAR WATER WAVES
Muammer Din Arif1, M. Rafiqul Islam2, A. K. M. Sadrul Islam3, Enaiyat Ghani Ovy4
3
Professor
Department of Mechanical & Chemical
Engineering
Department Head
Civil Engineering Department
Islamic University of Technology (IUT)
Board Bazar, Gazipur, Dhaka: 1704, Bangladesh
Email: sadrul@iut-dhaka.edu
Industrial Lubricants Engineer

ExxonMobil, Bangladesh
M.Sc. (Mechanical Engineering)
Department of Mechanical & Chemical Engineering
139, Rd: 1 (West), DOHS Baridhara, Dhaka, Bangladesh
Email: marif@mtu.edu
Lecturer
Department of Mechanical & Chemical
Engineering
Board Bazar, Gazipur, Dhaka: 1704, Bangladesh
Email: enaiyat_ovy@yahoo.com
Professor
Department of Naval Architecture and Marine Engineering
Bangladesh University of Engineering & Technology (BUET)
Dhaka: 1000, Bangladesh
Email: rafiqis@name.buet.ac.bd
ABSTRACT
The density, intermittency, and predictability of wave power make it the ideal renewable energy source for
zero pollution electricity generation. A numerical analysis was carried out using Visual Studio FORTRAN 5.0,
employing 3-D Sink Source Method, on a box shaped free floating buoy, in linear water waves. The motion
responses (6 degrees of freedom) of the buoy were determined with the aid of a Boundary Element wire mesh,
for given water depth (100m), wave incident angle (180), and wave period (3 to 26 s). The main objective is to
use the results to aid in the subsequent design of an Oscillating Water Column (OWC) Wave Energy Absorber.
The FORTRAN algorithm calculated wave loads and buoy motions for each individual wave period input by the
user. The program first determined the velocity potentials of the incident waves. This was subsequently used to
compute the pressures and thus, forces and moments exerted by the incident waves, on the wetted surface of the
buoy. From the calculated parameters, the various motions of the buoy i.e.: heave, surge, sway, roll, pitch, and
yaw were found. Finally, the desired wave period(s) for optimal buoy motions were determined. It was found
that the magnitudes of the motions of interest (heave, pitch, and surge) were the highest when the wave
frequencies were in proximity of the buoys natural frequencies. The buoy is a similitude of the full scale wave
energy absorber. Thus, these results will be used in subsequent research, by the authors, involving the evaluation
of the optimal performance of a Single Free-Floating Backward Bent Duct Buoy, housing an Oscillating Water
Column and turbine--to generate electricity.
Keywords: Wave Energy Absorber, Single Body Free Floating Buoy, Backward Bent Duct Buoy, Oscillating
Water Column, Renewable Energy, Boundary Element Method, 3-D Sink Source Method, Numerical Model.
1. INTRODUCTION
Sea and ocean wave energy has been gaining
popularity as a viable and clean renewable energy
source, namely for electricity generation. Duckers [1]
stated that the potential world wave energy resource
was similar in magnitude to that globally available
from hydro-energy (2000 GW). He also discussed that
waves are typically 2 to 3 times more powerful in
deep offshore areas compared to coastal regions. The
authors, accordingly, have modeled and determined
the motion responses of a free floating buoy in waters
of 100 m depth. A large amount of research and
model-testing has been done with, offshore and

coastal, wave energy convertors since the early 1970s.
The research topics have been manifold: (1)
characterization of wave energy resources, (2)
theoretical and numerical analysis of the
hydrodynamics of wave energy absorption, (3) design
and construction of varied energy convertors, (4) the
development of power take-off mechanisms (PTOs)
such as: air and water turbines, power hydraulics, and
linear electrical generators etc. Pioneers in this field
include: McCormick [7], Shaw [8], and Masuda [6].
Masuda [6] was the first to develop a working
137
navigation buoy housing an Oscillating Water Column

(OWC) which ran an air turbine [2], as shown in
Figure 1.
of simplification, the authors assumed a moored free

floating buoy, which is buoyant in normal sea state
and excited only by regular sinusoidal propagating
waves.
Figure 1. Schematics of a Backward Bent Duct

Buoy (BBDB) Wave Energy Convertor
These buoys have been commercialized in Japan
since 1965. The seminal work of Slater [9] also
helped boost the worldwide interest in wave energy
convertors. Presently, research and testing of wave
energy convertors are underway in UK, Norway,
Japan, India, Lisbon, Portugal, Greece, Denmark,
Sweden, Canada, USA etc. From the standpoint of
available technology and economy, wave energy
convertors are most profitable for power generation in
nations where other methods of electric power
generation are costly or non-existent [1].
In the current research, the authors investigated the
motion responses of a box-shaped buoy, in linear
water waves, considering 6 degrees of freedom
(DOF); namely: heave, surge, sway, pitch, roll, and
yaw. Three of these degrees of freedom, heave, pitch,
and surge, were considered important because these
three motions are directly responsible for producing
the oscillations in the water column of a wave energy
convertor [1]. The complex hydrodynamic
mathematical relations were solved and approximated
using Finite Boundary Element and 3-D Sink Source
methods. Both these methods have been perennially
used in the design of ships and large offshore
structures.
2. METHOD
2.1 Discussion
A free floating buoy in normal sea states has 6 DOF
as shown in Figure 2. Heave, pitch, and surge are the
most important motion modes for OWC generators
housed in free floating buoys. Thus, the origin of
these motions and the buoys resonance behavior in
these modes were analyzed in detail.
The incident waves impart forces and moments on
the buoy which manifest as these 6 DOF. The induced
motion is mainly linearly-excited motion [3].
High frequency motions are mostly caused by nonlinear waves and transient effects such as ringing
and springing. Similar non-linear effects, current,
and wind forces also cause slow drift motions of the
buoy. To counter such effects a mooring system is
utilized as shown in Figure 1. Therefore, for the sake
Figure 2. 6 DOF of Buoy

Hong et al. [4] did extensive numerical analysis on
the motions of and drift forces on Backward Bent
Duct Buoy (BBDB) OWC wave energy absorbers.
Their model consisted of buoyancy module connected
to a box housing the OWC. The OWC housing is
submerged and connects to the sea water through a
duct. In their research they utilized a numerical tank
to approximate the pressure drop in the air chamber
above the OWC, using linear wave theory.
Falcao [2] modeled a rectangular box-shaped buoy
containing a PTO. He varied the linear damping
coefficient of the PTO in order to determine the
energy conversion efficiency of the buoy under
normal sea states.
Faltinsen [3] used 2-D panel elements to discretize
the wetted surface of a large rectangular box-shaped
body in order to numerically evaluate sea loads on
offshore structures.
All three of the above models: Hong et al. [4],
Falcao [2], and Faltinsen [3], were used as guidelines
and for validation of the results obtained by the
authors.
2.1.1 Heave, Pitch, and Surge Motions
The typical BBDB housing an OWC has three
types of resonance behaviors: (1) the resonance
behavior of the OWC, (2) the coupled resonance
motion, and (3) the uncoupled resonance motion for
each DOF. The highest amplitudes of motions are
attained when the wave period nearly equals one of
the buoys aforementioned natural period types. This
fact is very important in the effective design of the
buoy, since, greater response motion amplitude means
greater extraction of wave energy for power
generation. The numerical buoy model used by the
authors lacks an OWC and thus, the first type of
resonance behavior was absent.
138
The uncoupled and undamped natural periods for

the three motions (heave, pitch, and surge) can be
determined from the following general formula (refer
to Table 1 for explanation of the variables):
Tni = 2 {(Mii + Aii)/Cii}1/2
(1)
resonance surge response motion was seen in the

results of the simulation.
2.1.2 Approach of the Computer Simulation
The hollow box-shaped buoys wetted surface was
discretized into 2 dimensional rectangular elements as
shown in Figure 3.
The natural uncoupled heave period for most free

floating structures is in the range of 4 to 16 s [3]. The
dominating exciting mechanism around this period is
always linear. The restoring force for a buoy is due to
the change in the buoyancy forces and is related to the
water plane or wetted surface area. The uncoupled
natural heave period of the buoy (Tn3 = 4.2 s) was
evaluated using a modified version of equation (1) as
defined below:
Tn3 = 2 {(M + A33)/(gAw)}1/2
(2)
Figure 3. Outline of the Simulation Approach
Table 1. Magnitudes and Nomenclatures of the

Variables Used
Variable
Magnitude
Unit
Nomenclature
21.45
Length of buoy
16.5
Width of the buoy
Aw
353.925
Water plane area (L*B)
1,085,412
kg
Mass of the buoy
A33
0.8M
kg
Heave added mass
A55
N.m
Pitch added moment
A11
kg
GML
Surge added mass

Longitudinal metacentric
height
r55
Pitch radius of gyration
Center of rotation of the buoy
Waterline of the buoy
Next, the velocity potential of the incident

sinusoidal water waves, on infinite (100m) water
depth, was determined. The dynamic pressure on the
wetted surface was then found. Finally the wave loads
and motion responses were evaluated.
The uncoupled natural period in pitch of the buoy is

estimated from the following modified version of
equation (1). The order of magnitude of Tn5 (4.1 s) is
similar to Tn3, the uncoupled natural heave period.
Tn5 = 2 [{M(r55)2 + A55}/{g(GML)}]1/2
(3)
The uncoupled natural period for surge motion of

the buoy was also estimated by the equation below:
Tn1 = 2 {(M + A11)/C11}1/2
(4)
For a moored structure, like the buoy, the

magnitudes of the uncoupled natural periods of surge,
sway, and yaw are in minutes [3] and, therefore, too
long compared to wave periods occurring under
normal sea states. Thus, no noticeable uncoupled
2.2 3-D Sink-Source Method

Three dimensional Sink-Source technique was used
to analyze the linear wave induced loads and resultant
buoy motions. The technique has been extensively
used in wave loading calculations for large volume
structures like ships and offshore rigs [5].
In this method, differential equations representing
a boundary value problem, were solved numerically
using the Boundary Element Method. The concept of
3 dimensional sources over the discretized wetted
surface was employed. The sources were considered
as harmonically oscillating variables with zero mean
speeds that generate 3 dimensional radiation wave
fields. The source density, velocity potential, and fluid
pressure were considered to be constant for each
individual element.
2.3 Boundary Element Method
The determination of wave loads on and motion
responses of the buoy was achieved using a wellestablished numerical computation technique known
as Boundary Element Method. This method
discretized the wetted surface continuum of the buoy
into 2 dimensional quadrilateral elements. Each
element was defined by a unique number (1 to 260)
and four nodes (1 to 300). The nodes were shared
between the elements to preserve continuity and
transmit loads and motions. 2-D rectangular elements
worked well as the buoy was assumed to be hollow as
represented in Figure 4.
139
natural resonance period of a moored buoy is in

minutes. Thus, no resonance motion was seen in
normal sea states (wave periods from 7 to 26 s).
Figure 4. Discretization of the Buoy

The discretization resulted in a set of linear
equations for the unknown source densities of the 6
DOF. The no-fluid-flow condition, through the wetted
surface, was satisfied at the geometric center of each
quadrilateral panel. Thus, the velocity, pressure, and
all other subsequent calculations were performed at
the geometric center of each element.
2.4 Computations
For the numerical computations of wave loads and
buoy motions, the computer code, used for the
calculations, was initially developed for wave force
analysis on a box-shaped body. It was subsequently
modified by Islam [5] for hydrodynamic interactions
and drift force calculations of multiple floating
structures.
The input to the program was from two sources.
Initially the program received parameters from an
input file. Once the program was compiled, built, and
executed; the command prompt received user defined
inputs for the incident wave periods. The outputs:
wave loads, response motions, and phase angles were
then generated.
Figure 5. Surge Motion (X) of Buoy

A similar graphical result (chain-dot line with linear
damping C = 2.5) was obtained by Falcao [2]. The
discrepancy between the two results, Falcaos and the
authors, resulted due to the fact that Falcao [2]
modeled a rectangular buoy with a simple PTO
whereas the authors did not include any PTO. Wave
periods in the range of 7s to 26s were selected to be
representative of normal sea states in North Atlantic
(3.3 to 23.7 s) and North Pacific (5.1 to 22.5 s);
Faltinsen [3].
The heave excitation force Fz for wave angular
frequencies () from 2.02 rads/s to 0.00001 rads/s
was analyzed. This range for corresponds to wave
periods from 3s to 628400s (T ). The results
obtained by the authors are represented in Figure 6
(dashed line). The results were compared with those
obtained by Hong et al. [4] (solid line).
3. RESULTS AND VALIDATIONS

The incident wave angle was taken to be 180 and
fixed mooring conditions imposed on the buoy. It was
assumed that the buoy was moored to the bottom of
the sea bed by a cable of negligible weight. The
mooring line prohibited any drift motion of the buoy.
This condition resulted in making some of the wave
loads and motion magnitudes zero or insignificant.
Thus, the force Fy (sway), moments: Mx (roll) and Mz
(yaw), and motions: Y (sway), Ya (yaw), and Ra (roll)
were insignificant. The noticeable parameters
forces: Fx (surge) and Fz (heave), moment My (pitch),
and motions: X (surge), Z (heave), and Pa (pitch)
were non-zero as shown in Figures 5 to 9.
From Figure 5 it was observed that the surge
motion of the buoy increased as the wave periods
increased. The surge magnitude increased as the wave
period was increased from 7 to 26 seconds. However,
the increase in surge magnitude was steady and no
sharp rises or peaks were observed. As already
mentioned in the Discussion section, the uncoupled
Figure 6. Heave Force (Fz) on Buoy

Hong et al. [4] obtained a resonance peak at 0.65
rads/s. This was due solely to the OWC and the
140
fluctuations smoothened with increasing equivalent

linear damping . No such peak was therefore observed
in the dashed curve as the authors did not model the
OWC in the buoy.
The heave motion amplitude Z was analyzed and
compared with the results obtained by Hong et al. [4]
as represented in Figure 7.
motion was also slightly shifted, to higher frequency,

from = 0.85 rads/s (Figure 8, solid line and long
dashes) to = 1.1 rads/s (Figure 8, small dashes).
Figure 8. Pitch Motion (Pa) of Buoy
Figure 7. Heave Motion (Z) of Buoy

In Figure 7 the long dashes and solid line are results
obtained for heave motion by Hong et al. [4]. They
obtained different curves for two different linear
damping conditions (long dashes for = 0 and solid
line for = 300). The authors results are represented
by the small dashes in Figure 7. Again an OWC
resonance peak was observed at = 0.56 rads/s as
shown in Figure 7 (long dashes). This was not seen in
the authors results as no OWC was modeled. The two
other peaks, at = 0.85 and 1.1 rads/s, in Figure 7
(long dashes and solid line), were due to the
resonance of coupled surge-heave-pitch and
uncoupled heave motions of BBDB-1 respectively.
The authors obtained a similar coupled surge-heavepitch resonance peak at a much higher frequency, =
1.1 rads/s (Figure 7, small dashes). The authors
resonance peaks amplitude, due to uncoupled heave
motion, was much smaller at 0.053 compared to Hong
et al. [4] in Figure 7. It also occurred at a higher
frequency, = 1.904 rads/s compared to 1.1 rads/s in
Hong et al. [4] model. The signs of another potential
resonance peak, of coupled surge-heave-pitch motion,
was detected around = 2.3 rads/s, in the authors
results. However, this frequency was too high for
normal sea state waves and hence, ignored.
The pitch motion Pa results were plotted as shown
in Figure 8 (small dashes) and compared with the
results obtained by Hong et al. [4] (Figure 8, solid line
and long dashes).
In this case a similar OWC resonance peak was
obtained at 0.56 rads/s (solid line), which was nonexistent in the authors results (small dashes). The
resonance peak due to coupled surge-heave-pitch
A second resonance peak, due to uncoupled pitch

motion, was observed near = 2 rads/s as shown in
Figure 8 (small dashes). This agreed with the findings
of Faltinsen [3] that the resonance frequency of heave
and pitch uncoupled motions are similar in magnitude.
A third resonance peak was observed near = 1.5
rads/s as represented in Figure 8 by small dashes. This
will be investigated further in subsequent research by
the authors.
The pitch moment My was plotted as shown in
Figure 9 by small dashes and evaluated against the
results obtained by Hong et al. [4], also shown in
Figure 9 by solid line ( = 300) and long dashes ( =
0).
Figure 9. Pitch Moment (My) of Buoy

Hong et al. [4] obtained OWC resonance peak near
= 0.65 rads/s (Figure 9, long dashes) which was, as
expected, non-existent in the authors graph (Figure 9,
small dashes). The authors, however, obtained the
141
forecasted coupled surge-heave-pitch resonance peak

near = 1.1 rads/s (Figure 9, small dashes) and
uncoupled pitch resonance peak near = 1.964 rads/s
(Figure 9, small dashes). The third and fourth peaks,
near = 1.6 and 1.8 rads/s (Figure 9, small dashes),
require further investigation.
4. INFERENCES
The results of the numerical model and simulation
were also compared against the previous seminal
work of Islam [5] for further verification. The
discrepancies between the results obtained by Hong et
al. [4], Falcao [2] and the authors were attributed to
different approximations and assumptions made in
modelling. These discrepancies and the differences in
modelling were addressed in the previous section.
The BBDB-1, modelled by Hong et al. [4]
incorporated an open and submerged duct which was
not present in the authors buoy. The BBDB-1 also
contained an OWC while the authors buoy was
considered to enclose a vacuum inside. The authors,
unlike Hong et al. [4], also considered a moored buoy.
These differences of geometry, inertia and relative
movements of the OWC with respect to BBDB-1
inner walls account for the disparity between the two
sets of results.
The rectangular buoy, modelled by Falcao [2],
contained a simple linear damping PTO. Such a PTO
was absent in the authors model. Falcao [2], unlike
the authors, did not consider mooring conditions. As a
result, the disparity between the two sets of results
was apparent in the graphical comparison.
The authors also modelled a moored buoy in order
to make the ensuing calculations more amenable. The
mooring acted against various motion modes of the
buoy. These in-congruencies need to be studied more
extensively prior to designing an efficient buoyant
wave energy absorber.
5. CONCLUSIONS
As mentioned earlier, extensive research has been
done on the behavior and motion of free floating
buoys under varying conditions and assumptions.
Special interests of the researchers have always been
focused on the resonance response motion of such
buoys. The results of the authors simulation, adjusted
for discrepancies caused by different initial
assumptions, agreed with such established research
observations regarding resonance motions
So, this computational model could be used to
accurately determine the wave loads on and motion
responses of a buoy of known dimensions, especially
near the conditions of interest, i.e. resonance motion.
in predicting parameters of interest near the elements

edges or at the buoys sharp boundary corners. It was
noted that the fluid flow actually separated at the
sharp corners from the buoys boundary surface.
Consequently, the boundary element model has
singularities at these regions. The handicap could be
mitigated through the use of smaller dimension panel
elements near the aforementioned corners.
However, the model and simulation results were
accurate enough that, the model could be used in the
optimal design of a Backward Bent Duct Buoy
(BBDB) with an Oscillating Water Column (OWC);
for the conversion of wave energy into electricity. The
authors are of the opinion that their work will further
the research in such renewable energy sectors.
ACKNOWLEDGEMENTS
The authors are grateful to the Faculty and Staff of
Technology (BUET) and Islamic University of
Technology (IUT) for lending access to the
institutions computing facilities.
REFERENCES
[1] Duckers, L. J., Wave energy: crests and troughs,
Renewable Energy, Vol. 5, Part II, pp. 1444-1452
(1994).
[2] Falcao, AF de O., Wave energy utilization: a
review of the technologies, Renewable and
Sustainable Energy Reviews, Vol. 14, pp. 899-918
(2010).
[3] Faltinsen, O. M., Sea Loads on Ships and Offshore
Structures, U.K.: Cambridge University Press (1998).
[4] Hong, D. C., Hong, S. Y., and Hong, S. W.,
Numerical study on the reverse drift force of floating
BBDB wave energy absorbers, Ocean Engineering,
Vol. 14, pp. 1257-1294 (2004).
[5] Islam, MR., A study on motions and second order
drift forces on multi-body floating system in waves,
Doctoral thesis, Yokohama National University, Japan
(2001).
[6] Masuda, Y., Wave-activated generator,
International Colloq Exposition Oceans, France:
Bordeaux, (1971).
[7] McCormick, M. E., Ocean Wave Energy
Conversion, New York: Wiley (1981).
[8] Shaw, R., Wave energy: a design challenge,
Chichester: Ellis Horwood (1982).
[9] Slater, SH., World progress in wave energy,
International J. Ambient Energy, Vol. 10, pp. 3-24
(1989).
Admittedly, the model does have some inherent

limitations, i.e. no wind loads, drift forces, or nonlinear wave excitations were considered. The velocity
potentials were evaluated at the panel elements
geometric centers. Hence, the model is less accurate
142

NUMERICAL MODELLING IN ASSESSING IMPACT OF BANK PROTECTIVE

WORKS AT UPSTREAM ON SILTATION IMPROVEMENT AT DOWNSTREAM
Masum Ur Rahman1 and Sarwat Jahan2
1
Junior Specialist, Institute of Water Modelling, House 496, Road
32, New DOHS, Mohakhali, Dhaka, Bangladesh,
E-mail: mur@iwmbd.org
2
Senior Specialist, Institute of Water Modelling, House 496,
Road 32, New DOHS, Mohakhali, Dhaka, Bangladesh,
E-mail: srj@iwmbd.org
ABSTRACT
Bank protective works, in general, lead depending of the river, near the structure as well as within the
mid section. Such deepening is mainly due to constriction provided by the bank protective works along the
bank. In response to changes in water and sediment discharges at upstream, the river morphology at
downstream adjusts in order to establish dynamic equilibrium. The time scale for such adjustment varies
depending on the existing hydro-morphological condition of the river. This paper is an attempt to reveal
investigation of the physical process involved for such response of the Jamuna River near Hurasagar outfall.
Upstream of Hurasagar is exhibiting bank erosion and on the contrary, right bank channel immediate
downstream of the outfall is experiencing siltation. Relentless siltation ultimately causes hindrances to free
movement of the water based transport system hampering overall navigability within this area. Applying
numerical modelling technique, investigation is carried out to envisage the improvement of siltation problem
at downstream adopting a bank protection works at upstream. Reflection of upstream bank protection works
on river morphology are analyzed with the application of advanced two-dimensional mathematical modeling
tool, MIKE21C developed by DHI Water and Environment, Denmark.
Key words: Siltation, Navigation, Numerical modelling, Two-dimensional modelling, bank erosion, bank
protective works, erosion, deposition.
1. INTRODUCTION
Bangladesh is blessed with numerous rivers,
which are characterized by unpredictability in their
behavior. Every year many of the rivers pose
different shape, follow different path showing their
swinging nature. Building up of sediments, i.e.
sedimentation, along the river bed may cause
instability in channel morphology and this
instability includes impedance of flow and
increasing intensity of inundation. Additionally,
reduction of conveyance area causes inadequacy of
navigability hampering easy transportation through
waterway. Sedimentation not only affects human life
but also fish, plants, and other wildlife. Rivers
provide humans with many valuable uses. They
provide drinking water, water for crops, a source of
power, transportation, fish and wildlife, and
recreation. The disruption that sedimentation causes
prevents us from utilizing these resources to their
full potential. Although sedimentation is a natural
occurrence, it can be exacerbated by mankind
through activities such as those involved with
construction.
For the last few years, sequential deposition at

the bed level of the right anabranch of the Jamuna at
Koitola had been observed. Such relentless
deposition from year to year poses threat to
navigation around this area and subsequently, the
conveyance capacity of this channel is reducing
which in turn increases possibility of spilling over
the flood plain. On the contrary, bank erosion at
upstream of the Hurasagar near Kaijuri bazaar to
Binotia Bazar poses serious threat and in the last
few years, the rate is several hundred meter per year.
Observing the propagation rate, changing critical
channel pattern and its adverse consequences,
Jamuna Meghna River Erosion Mitigation Program
(JMREMP) formulated a study for investigation.
Necessity of the River bank protective works at
upstream, its required extent, and its impact at
further downstream, i.e. in improving the siltation at
downstream of Hurasagar outfall, has been
investigated by Institute of Water Modelling (IWM)
under the study framework. The recommendation
comes from the study are to protect the erosion
affected area from Kaijuri Bazar to Binotia Bazar,
which in addition to protecting bank, likely to
improve the downstream siltation problem.
143
This paper is an attempt to present the role of

bank protection works at upstream in improving
siltation at downstream. Reflection of upstream
bank protection works on river morphology has
been analyzed with the application of advanced twodimensional (2-D) mathematical modeling tool,
MIKE21C, developed by DHI Water and
Environment, Denmark.
alluvial chars, permanent and moving, are the

characteristics of the river. The formation and
removal/migration of the chars during monsoon
controls the erosion of bed and banks. The
characteristics of the Jamuna River are summarized
in Table 1 (Source: RSP FAP24, 1996).
Table 1: Summary characteristics of the
Jamuna River
Description
Length in Bangladesh
Average bankful width
Discharge
2. STUDY AREA
The study area covers 45 km long reach of the
Jamuna River near the outfall of the Hurasagar. The
upstream boundary of the study area is at 12 km
away from the Bangabandhu Bridge and the
downstream boundary is at 11.8 km away from the
Jamuna-Ganges-Padma confluence. Though the area
of interest is around the outfall of the Hurasagar but
45 km long coverage has been considered since the
morphology of any local area is governed by the
upstream morphology. Study area is shown in
Figure 1.
Makar
Khal
Zamirtola
Elasin
Bangabandhu
Bridge
Binotia
Bazar
Hurasaga
r Outfall
Koitola
Protection
works upto 7
km
100 to 400 m/year

30-40 m/year
Basic issues, which are responsible for dynamic

changes of the Jamuna, are variability in
hydrological factors, sediment load/morphological
variability, soil characteristics of the bank and bed,
and man-made structures. Past studies (BRTS, FAP,
and CEGIS) indicate that the Jamuna is in the stage
of dynamic equilibrium but any abrupt hydraulic
condition might change the whole system even
within one season. Whenever there is any abrupt
change like a major flood (1998 or 1988 flood) or
construction of any massive man-made structure
within its regime takes place, or earthquake, the
Jamuna again tries to adjust itself but changing its
domain. Such adaptation for the Jamuna might take
from several years to centuries and ultimately result
in changing its properties like shifting of bank line,
channel alignment, bed topography, etc.
Dhaleswary
Kaijuri
Bazar
Average discharge
Bankful discharge
Dominant discharge
Braiding index
Water surface slope
Grain diameter (d50)
Average yearly sediment transport
Average yearly coarse suspended
sediment transport
Maximum erosion rate
Minimum erosion rate
Parameter
240 km
11 km
1950 ~ 102,535
m/s
20,400 m/s
48,000 m/s
38,000 m/s
46
8.5 6.0 cm/km
0.2 0.14 mm
600 Million tons
200 Million tons
Ghior
Khal
Ghior
Bazar
PIRDP
A
3. APPROACH AND METHODOLOGY

Manikgan
j
Figure 1: Study area of the Jamuna

River under investigation
2.1 Characteristics of the Jamuna River
The Jamuna River is one of the world's large
rivers, ranking in the top three in terms of both
sediment and water discharge. The river is in fact a
multi-channel one, i.e. channels of many different
sizes, from hundreds of meters to kilometers wide,
and of different patterns including braiding,
meandering and anastomosing pattern. Numerous
In order to analyze the baseline hydrodynamic

and morphological conditions in the vicinity of the
PIRDP area and investigate the erosion deposition
mechanism of the Jamuna River around PIRDP area
for the proper maintenance of the navigational route,
a two-dimensional (2-D) morphological model has
been developed using MIKE21C modelling tool.
Prior to the mathematical modelling exercise,
data analysis have been done extensively to
understand the hydraulic process related to the
erosion-deposition at Kaijuri bazaar and Koitola,
respectively. The outputs from such analysis have
not only formed the basis of understanding but also
144
enhanced the confidence to interpret the model

generated outputs.
The core of the mathematical morphological
modelling study is the application of the MIKE21C
modelling system, which is an advanced twodimensional mathematical modelling technology for
simulation of unsteady flow, sediment transport, and
morphology. A morphological model is a combined
hydrodynamic and sediment transport model. The
hydrodynamic flow field is updated continuously
according to changes in bed bathymetry. The
solution of hydrodynamics is solved at a certain
time step prior to solution of the sediment transport
equations. Subsequently, a new bed level is
computed and the hydrodynamic model proceeds
with the next time step. Other sub-models such as
bank erosion, bank line update, alluvial bed
resistance, bed forms, and graded sediment are also
included in the simulation.
The equation solved
hydrodynamic model is:
in
the
curvilinear
p
H p q q
+
+
+
= 0 (1)
t
s n R s Rn
where,
Eb
z
S
h
,,
model
Bank erosion in m/s

local bed level
near bank sediment transport
local water depth
calibration coefficients specified in the
4. MODEL DEVELOPMENT AND

SIMULATION
The 2-D mathematical morphological model has
been developed and calibrated based on field survey
and secondary data. The computational grid is
developed with surveyed banklines as well as
banklines digitized from recent satellite image of the
study area. The bathymetry of the model is
developed with data corresponding to September
2007 and May 2008.
Computational grid of the two-dimensional
model and initial bathymetry for the study area
covering 50km stretch of the Jamuna River has been
shown in Figure 2. Grid of the model has 134310
computational grid cells where the grid resolution is
363 along the river length and 370 across the river
width.
where,
s, n
Coordinates in the curvilinear coordinate
system
p, q
Mass fluxes in the s- and n-direction,
respectively
H
Water level
Rs, Rn Radius of curvature of s- and n-line,
respectively
The transport of bed material and the bed level
change are computed from the following sediment
continuity equation:
(1 n ). z + S x +
t
where,
Sx
Sy
n
z
t
(x,y)
Se
S y
= S e
y
(2)
total sediment transport in x-direction

total sediment transport in y-direction
porosity
bed level
time
Cartesian coordinate system
lateral sediment supply from bank erosion
Figure 2: Curvilinear Grid and initial

bathymetry
The sediment transport model can include bank

erosion in the continuity equation:
Eb = .
z
S
+ . +
t
h
(3)
The developed model has been calibrated for

2007 monsoon. Figure 3 shows the observed and
simulated water levels at Mohangonj, Bera, Pabna
for 2007 hydrological year and acceptable
agreement between observed and simulated water
level data is revealed from the figure.
145
material into the river flow. Overall siltation

situation at downstream of Hurasagar may get
improved if these sources of sediment can be
blocked. Sediment coming from upstream cannot be
obstructed, so recommendation can be made at this
point to stop the bending at upstream since this
bending is causing bank erosion, which is ultimately
slumped and mixed with the flow.
12
Observed
Simulated
Water Level (m,PWD)
11
10
6
7-Jun
27-Jun
17-Jul
6-Aug
26-Aug
15-Sep
5-Oct
25-Oct
Date
Figure 3: Comparison of observed and simulated

water level at Mohangonj, Bera, Pabna
The verification of the morphological model has
been done using observed sediment transport data.
Figure 4 shows the simulated and observed (FAP24)
sediment transport at Sirajganj. The simulated
sediment transport indicates satisfactory agreement
with the observed data. Since the simulated
sediment transport, using Van Rijn sediment
predictor, match well with the observed transport
data, it can be considered that the Van Rijn sediment
transport predictor generates acceptable sediment
flow for the present case.
120000
Simulated
Sediment Discharge (kg/s)
100000
Observed_Sirajganj
80000
60000
40000
Simulation results of the base model and the

model with bank protection works have been
analyzed to compare between the prevailing and
with protection work for assessing the improvement.
Influence of the proposed bank protection works
from sediment distribution aspect has been assessed
with the determination of the sediment passing
through different channels. It has been observed
from the model results that the model with bank
protective works generates less amount of sediment
if compared with without bank protective works.
Base model generates 79005 million kg of total
load including the sediment coming from further
upstream throughout one monsoon whereas in case
of revetment option, it is reduced by 546 million kg,
Figure 5. Since the bank erosion could not take
place due to presence of revetment, i.e. bank erosion
is not allowed in the model within the protected
stretch at Kaijuri, so no additional bank eroded
material is added with the sediment load, i.e. the
flow only carries the sediment which was coming
from upstream. On the contrary, in case of base
condition, sediment load coming from upstream is
increased in volume with addition of bank eroded
material at Kaijuri. Since less sediment is being
carried, obviously less siltation is generated in the
right most anabranch near the outfall of Hurasagar.
20000
0
0
10000
20000
30000
40000
50000
60000
70000
80000
Discharge(m3/s)
Figure 4: Comparison of observed and simulated

sediment discharge of Jamuna River at
downstream of Hurasagor outfall and Sirajgonj.
In order to assess the change in morphology due
to implementation of the bank protection works at
upstream, apart from the base model comprising the
present field condition, another model with a bank
protective work superimposed on the base model
has been simulated thereafter for 2007 flood event.

Siltation is directly related to the sediment load
coming from upstream and intrusion of bank eroded
146
Base Condition
With Revetment
Condition
Figure 5: Simulated sediment load for base

condition and with revetment scenario through
different channels for 2007 flood event
Since the model indicated less amount of

sediment through the channels at mouth of
Hurasagar, so it is expected that the bed would
experience less siltation, i.e. bed scour would take
place. Reflection of the sediment on the channel bed
has been determined with extracted cross-sections
for base model and model with bank protection
scenario, Figure 6. Though bed scour in with
revetment condition is not significant, it is
important that scouring has initiated. At both
Location A and Location B, shown in figure, bed
erosion occurs at the thalweg position.
REFERENCES
Bed level(m)
Cross section at B
-3
-6
Base
9km Revetment
-9
1500
2500
Sediment generated due to bank erosion at

Kaijuri Bazar aggravates the siltation condition at
Koitola. However, this is not solely responsible for
siltation at Koitola rather one of the potential
causes. Morphological model has been deployed to
simulate the river behaviour and thus to establish the
theorem on contribution of the bank eroded
sediment on the morphology at downstream. Prior to
simulations, base model was calibrated and the
results of the calibration were found within the
acceptable range. IWM recommended bank
protection works at upstream of Hurasagar outfall,
i.e. from Kaijuri Bazar to Hurasagar outfall, under
the JMREMP formulated study. Adopting the
recommendation, both the base model and model
with recommended bank protection works has been
simulated for the flood event of 2007. Model result
shows effectiveness of recommended upstream bank
protection works at Kaijuri Bazar to improve the
siltation at downstream of Hurasagar outfall near
Koitola. Although the response is very minor, it
shows initiation of the process and this is how a part
of the sediment can be reduced that is responsible
for the siltation at downstream.
With the aid of mathematical modelling, it is
possible to assess the impact on river morphology
for any human intervention. Modelling application
for prediction of the impact in the river engineering
sector can guide the planners and policy makers for
planning and thus to strengthen the development
works.
Cross section at A
500
6. CONCLUSION
3500
4500
Distance
Figure 6: Bed profiles for base and

with revetment condition at Koitola
for 2007 flood event
5500
[1] Center for Environment and geographic

Information Services (CEGIS), Long-term
Bank Erosion Processes of the Jamuna River,
Jamuna-Meghna River Erosion Mitigation
Project, (2007).
[2] DHI (2009), Mike 21, Non-Cohesive
Sediment Transport Module, User manual and
Reference
manual,
DHI
Water
and
Environment, Hrsholm, Denmark.
[3] Flood Action Plan (FAP) 24, Morphological
Characteristics, Final Report; Bangladesh
Water Development Board, (1996).
[4] Halcrow, River Training Studies of the
Brahmaputra River, Master Plan Report, Main
Report, (1993).
[5] Rahman M.U., Jahan S., Kamal M.M.,
Response of Climate change on the
morphological behavior of the major river
system of Bangladesh, Proceeding of
International MikebyDHI2010 Conference,
Denmark, (2010).
147

THERMOMECHANICALLY TREATED ADVANCED STEELS FOR STRUCTURAL

APPLICATIONS
M.A. Islam
Md. Aminul Islam
Materials and Metallurgical Engineering Department
BUET, Dhaka-1000, Bangladesh
E-mail: aminulislam@mme.buet.ac.bd
ABSTRACT
It is true that earthquake and other accidents, especially accidents of passenger vehicles and associated
damages are very unpredictable. However, degree of damages to human lives and properties due to various
accidents can be minimized by using proper materials as well as design for buildings and auto bodies. Large
volumes of steels are being used every day for civil constructions and various auto bodies. For the safety of
residents of buildings and passengers of vehicles good quality steel undoubtedly can play a very important
role. Considering the safety of various structures and people associated with them, thermomechanically
treated advanced steels ushered a new horizon in the field of construction of buildings, bridges, auto bodies,
etc. In this paper, various advanced steels developed for structural applications will be discussed with their
merits and demerits.
Keywords: Thermomechanical treatment, Advanced structural steels, DP steel, TRIP steel, UFG steel.
1. INTRODUCTION
Throughout the whole world, environmental
pollution is really a very big issue. It is important to
mention that heavy industrialization is significantly
responsible for the overall worldwide environmental
pollution. Till now, as engineering material, steel is
used in the largest volume. As a result, steel industries
are under tremendous pressure from the
environmentalists to reduce the steel production.
Considering the environmental pollution, auto makers
are also trying to reduce the vehicle weight for lower
fuel burning related air pollution. At the same time,
structural design engineers are looking for high
strength and earthquake resistant steel bars for
reinforcing concrete to reduce the overall weight of
the structures. However, it is really not so straight
forward job, because any possible reduction in steel
consumption (i.e. overall reduction in structural
weight) must be adopted without compromising the
safety of the users of buildings and auto bodies. In
this situation, application of high strength steel
rods/bars/sheets might provide necessary solution.
Through the addition of various alloying elements, the
strength of steel can be increased. But the key
problem concerning this steel strengthening route is
that it reduces the ductility of the steel as well as the

bendability and/or stretch formability (1-4). In order
to avoid problem related to inferior ductility,
materials scientists have developed various advanced
structural steels by proper microstructures of the steel
through precise control of various thermomechanical
treatment parameters. Thermomechanically treated
steels might be plain carbon as well as alloy steels.
The latter group of steels (alloy steels) are designed to
achieve further higher strength along with other
property requirements depending on service
conditions [5-6].
The aim of this paper is to present various types
of advanced structural steels developed so far and
their benefits if they are used as structural steels as
substitutes of conventional plain carbon or alloy
steels.

It is to be noted that there are various technological
options to increase the strength of structural plain
carbon steels such as by increasing carbon content,
addition other alloying elements, decreasing
ferrite/pearlite grain size, changing ferrite/pearlite
149
Figure 1. Microstructure of TMT steel.
1000
800
Stress MPa
structure to high strength bainitic structure by water

spray/sudden cooling, proper cleaning of molten
steels, thermo-mechanical treatment and so on [7].
Here in this paper, initiative will be taken to discuss
techniques to increase the strength of steels by various
thermomechanical treatments with their possible
advantages. There are various thermomechanical
treatment techniques such as quenching and
tempering (QT, note: in Bangladesh some steel
companies are marketing this type of product as TMT
steel), severe refining the grain size of steels (ultrafine
grained; UFG steel), controlling the soft and hard
phases of the steels (e.g. dual phase; DP steel),
retaining and controlling the proportion of
unstable/metastable phases, i.e. phases with stress
induced transformation behaviour (transformation
induced plasticity; TRIP steel), etc.
Average
Outer
L
600
Core
Transition Zone
400
200
0.0
20 0
5.0
10.0
15.0
Figure 2. Stress-strain diagram of various phases

in QT or TMT steel.
2.1 QT or TMT Steel

This group of structural steel might be plain carbon
or alloy type. In general, it is bar type product used
for reinforcement of concrete. However, it can also be
used for load bearing machine components or other
structural applications. In the case of conventional
bars, the hot rolled products are cooled naturally in
the air. So, from surface to core, the microstructures
remain fully ferrite-pearlite. But for QT/TMT steel
bars, they are forced to cool suddenly by passing them
through water chamber just after final rolling. So, the
outer layer of the bar gets quenched with subsequent
formation of tempered martensite at the outer layer.
The structures of the core remain as normal ferritepearlite (Fig. 1) and there also exits a transition zone
in between the outer tempered martensite and ferritepearlite core. Tempered martensite is a very strong
phase, which provides high strength and relatively
soft ferrite-pearlite structure at core provides
necessary ductility of the steel. The stress-strain
diagrams of individual phases of QT steel as well as
the combined one are presented in Fig. 2.
2.2 Ultra Fine Grain Steel

Grain refinement has been always one of the most
important subjects in microstructure control of
structural metallic materials and also an important
field of research for physical metallurgists. The
minimum grain size that might be achieved through
conventional method is around 10 micron. Through
controlled thermomechanically processed rolling, this
value can be up to 5 micron [8]. However, it has
become recently possible to fabricate the steel with
ultrafine grained structures of which mean grain size
is smaller than 1 micron. This is possible in carbon
steel through heavy one-pass deformation at low
temperature around 500oC followed by rapid cooling
of sheets or wire to achieve fine grain through
controlled recrystallization [8]. It has been well
accepted that decrease in ferrite-pearlite grain size
causes increase in yield strength of the steel product
as shown in Fig. 3 and a typical ultrafine grained
ferrite-pearlite structure is shown in Fig. 4.
Tempered Martensite
Yield Strength, MPa
450
Ferrite/Pearlite Structure
375
300
225
225
Transition Structure
110
40
16
10
4.5
Ferrite Grain Size, m

150
Figure 3. Effect of ferrite grain size on tensile yield

strength of plain carbon steel.
alloying element. Because of high strength and

ductility combination, the earthquake resistance of
this steel also very good. However, so far the
production cost of this steel is very high because of
expensive heat treatment schedule along with precise
control system. As a result, these steels are used in
some advanced structural applications, e.g. in auto
bodies.
2.4 Transformation Induced Plasticity or

TRIP Steel
Figure 4. A typical ultrafine grained ferritepearlite steel.
The main limitation of this strengthening method is
the gradual brittleness in the steel and that drastic
decrease in thickness during rolling needs a very
powerful rolling mill.
2.3 Dual Phase (DP) Steel

This is, in general, carbon steel with controlled
amount of manganese and silicon content. During the
production of this steel, the stock material is hot
rolled with a precise rate at every specific temperature
so that initially some ferrite grains may come out
from austenite. Gradually the proportion of austenite
becomes lower with relatively higher carbon content.
This high carbon austenite then transforms to
martensite/bainite due fast cooling after final
deformation. Ferrite is a relatively soft phase, which
ensures higher ductility. However, martensite/bainite
is very strong which retards the deformation of soft
ferrite under loading condition. So, at a certain stage,
the resultant deformation becomes slower and thus
the steel provides reasonably high strength. The
microstructure of this steel is shown in Fig. 5.
In this steel, similar to DP steel, ferrite matrix is

strengthened by hard and strong martensite/bainite
structures. However, retained austenite which is kept
untransformed during the controlled processing helps
to increase the ductility of this steel (9-11). So, TRIP
steel is also termed as high strength multiphase steel.
Because of retained austenite transformation during
loading, TRIP steel shows higher ductility than that of
DP steel. Because of higher strength, at the same time
higher ductility, TRIP steel provides excellent
earthquake resistance and resistance to crush as it has
the capability to absorb energy under severe loading
condition [12-14]. The microstructure of a typical
TRIP steel is shown in Fig. 6 and the stress-strain
diagrams of DP and TRIP steels are presented in Fig.
7. The unique advantage of this steel is that, during
the severe shock of earthquake retained austenite
absorbs energy to transform to bainite, which
ultimately reduces the severity of earthquake related
damages.
Figure 6. Microstructure of transformation

induced plasticity (TRIP) steel.
Figure 5. A typical ultrafine grained ferritepearlite steel

From this steel, high strength along with moderate
elongation is achieved without adding any expensive
The earthquake resistance of any structural steel

depends on its energy release rate under cyclic
loading. The energy release rate curves of
conventional and TRIP steels are shown in Fig. 8.
From this figure it is clear that TRIP steel has
significantly higher energy release rate compared to
that of its rival conventional steel, because TRIP steel
shows a very high strength along with a very high
percentage of elongation.
151
3. CONCLUSIONS
Nominal Stress, MPa
1000
DPDP
Steel
Steel
800
TRIPSteel
Steel
TRIP
600
400
200
0.0
10.0
20.0
30.0
40.0
% Elongation
Figure 7. Nominal stress-strain diagrams of dual
phase and transformation induced plasticity steels.
In order to produce TRIP steel, both chemical

compositions, degree of deformation at different
levels and corresponding temperatures and final
cooling rate must be controlled properly. So, similar
to DP steel, the production cost of this steel is also
very high. Because of production cost, these steels are
not commercially produced in many countries and
that their applications are also limited in
manufacturing various auto bodies.
4. REFERENCES
[1] Takashi, M., Kawano, O., Hayashida, T.,
Okamoto, R. and Taniguchi, H. High Strength
Hot-rolled Steel Sheet for Automobiles,
Nippon Steel Technical Report No. 88, pp.1-12
(2003).
[2] Beynon, D., Jones, T.B. and Fourlaris, G., Effect
of High Strain Rate Deformation on
Microstructure of Trip Steels Tested Under
Dynamic Tensile Conditions, Materials Science
and Technology, Vol.21, No.1, pp.103-112
(2005).
Very Poor Energy

Release
Conventional
Steel
Very
Energy
Tensile strengths of structural steels can be

increased by various ways. However, till now,
thermomechanical treatment method is found to be
the most efficient method of strengthening of steels
without compromising the good combination of
mechanical properties. The main benefits of the
thermomechanically treated steel bars are due to
relatively lower volume of steel consumption
compared to any conventional steel in any structural
application and their outstanding earthquake
resistance. Another benefit of this method is that no
expensive alloying element is required to increase the
strength of the steel. If it is possible to reduce the
production cost of thermomechanically treated steel
and make this steel more popular, then there is no
doubt that construction cost of buildings will be
reduced to some extent. At the same time, safety of
the building will also increase significantly, which is
very crucial for a society with a heavy population like
that in Bangladesh.
[3] Tomota, Y., Tokuda, H, Adachi, Y., Wakita, M.,

Minakawa, N., Moriai, A. and Morii, Y.,
Tensile Behavior of TRIP-aided Multiphase
Steel Studied by In-situ Neutron Diffraction,
Vol.52, pp.5737-5745 (2004).
Good
Release
[4] Hulka, K., The Role of Niobium in Cold Rolled

TRIP steel, J. Materials Science Forum,
Vol.473, pp.91-102 (2005).
TRIP Steel
[5]
Panigrahi, B.K., Srikanth, S. and Sahoo, G.,

Effect of Alloying Elements on Tensile
Properties, Microstructure and Corrosion
Resistance of Reinforcing Bar Steel, J. Mat.
Engg. Perform., ASM International, Vol.18,
pp.1102-1108 (2009).
[6]
Manoharan, R., Jayabalan, P. and Palanisamy,

K., Experimental Study on Corrosion
Figure 8. Energy Release behaviour of conventional and

TRIP structural Steels.
152
Resistance of TMT Bar in Concrete, ICCBT,

pp.239-250 (2008).
[7]
Islam, M.A., Various Routes for Production of

High Strength Steels-A Review, Proceedings
of International Conference of SPPM2010, held
in Dhaka on 24-26 February, No.E19 (2010).
[8]
On-line Quenched and Self-Tempered High

Strength Steel, Nippon Steel Technical Report,
pp. 38-40 (2010).
[9] Islam, M.A., Chain, S. and Tomota, Y., Tensile

and Plane Bending Fatigue Properties of Two
TRIP Steels at Room Temperature in the Air- A
Comparative Study Published in the Journal of
Materials Engineering and Performance, ASM
International, Vol. 16, No. 2, pp.248-253
(2007).
[12] Panigrahi, G.K., Srikanth, S. and Sahoo, G.,

Effect Alloying Element on Tensile Properties,
Microstructure and Corrosion Resistance of
Reinforcing Steel Bar, J. Mat. Engg. Perf.,
ASM International, Vol.18, pp.1102-1108
(2009).
[13] Jha, G., Sinha, A.K., Bandyopadhyay, N and
Mohanty, O.N., Seismic Resistant Reinforcing
Bars, Journal of Practical Failure Analysis,
Vol.5, pp.53-56 (2001).
[14] Tsuji, N., Ultrafine Grained Steels Managing
Both High Strength and Ductility, Journal of
Physics Conference, Serial 165012010, pp.1-6
(2009).
[10] Islam, M.A., Structure and Properties of

Thermomechanically Treated TRIP Steel,
BSME-ASME Conference, Held in Dhaka,
pp.245-251 (2008).
[11] Islam, M.A. and Tomota, Y., Plane Bending
Fatigue Behaviour of Interstitial Free Steel at
Room Temperature International Journal of
Materials Research (formerly Metallkunde,
German), Vol. 97, No. 11, pp.1559-1565
(2006).
153

COMPUTATIONAL DRAG ANALYSIS OVER A CAR BODY

M. M. Islam1 and M. Mamun2
1
Department of Mechanical Engineering, Bangladesh University of
Engineering and Technology, Dhaka, Bangladesh.
E-mail: engg_monir09@yahoo.com
2
Department of Mechanical Engineering, Bangladesh
University of Engineering and Technology, Dhaka, Bangladesh.
E-mail: mdmamun@me.buet.ac.bd
ABSTRACT
In the recent times, CFD simulations, with the advent of computer architectures with superfast processing
capabilities are rapidly emerging as an attractive alternative to conventional wind tunnel tests which are either
too restrictive or expensive, for aerodynamic styling of a car. In vehicle body development, reduction of drag is
essential for improving fuel consumption thus protects the global environment and driving performance, and if
an aerodynamically refined body is also aesthetically attractive, it will contribute much to increase the vehicles
appeal to potential customers. This paper outlines the process taken to optimize the geometry of a vehicle.
Vertices and edges were imported into Gambit and a computational domain created. An unstructured triangular
mesh was then applied. The CFD program Fluent was used to iterate toward a converged solution with the goal
of obtaining a better flow around the car and drag force. The results are analyzed and only the drag force is
compared with a recognized journal to validate the results. These practices were detailed in hopes that further
research would use the ground work laid out in this paper to redesign existing vehicles in order to improve
handling and increase fuel efficiency.
Key words: CFD, car, drag, velocity vector, pressure contour, Gambit, Fluent.
1. INTRODUCTION
The rapidly increasing fuel prices and the
regulation of green house gasses to control global
warming have given tremendous pressure on the
design engineers to enhance the current designs of the
automobile using minimal changes in the shapes. To
full fill the above requirements, design engineers have
been using the concepts of aerodynamics to enhance
the efficiency of automobiles [2, 4]. A lot of
emphasize is laid on the aerodynamics in car design
as an aerodynamically well designed car spends the
least power in overcoming the drag exerted by air and
hence exhibits higher performance- cruises faster and
longer, that too on less fuel (Figure 1).
Apart
from
improved
fuel
economy,
aerodynamically superior car offers better stability
and handling at highway speeds and also
minimization of harmful interactions with other
vehicles on the roadway. In optimization of car
aerodynamics, more precisely the reduction of
associated drag coefficient (CD), which is mainly
influenced by the exterior profile of car, has been one
of the major issues of the automotive research centers
all around the world. Average CD values have
improved impressively over the time, from 0.7 for old
boxy designs of car to merely 0.3 for the recent more
streamlined ones [1].
Figure 1. Fuel economy with reduction in drag

coefficient (CD)
The Figure 2 shows the description of the fuel energy
used in a modern vehicle at urban driving and
highway driving. The shape of the vehicle uses about
3 % of fuel to overcome the resistance in urban
driving, while it takes 11% of fuel for the highway
driving. This considerable high value of fuel usage in
highway driving attracts several design engineers to
enhance the aerodynamics of the vehicle using
minimal design changes [2].
The effect of drag on the moving vehicle is
proportional to the square of velocity, so with increase
155
in velocity (at approximately 50 km/h), aerodynamic

drag becomes one of the most prominent factors
contributing to the total drag experienced by the
vehicle [5].
imported into Gambit and a computational domain

created. An unstructured triangular mesh was then
applied. The CFD program Fluent was used to iterate
toward a converged solution with the goal of
obtaining a better flow around the car and drag force.
The results are analyzed and only the drag force is
compared with a recognized journal to validate the
results. These practices were detailed in hopes that
further research would use the ground work laid out in
this paper to redesign existing vehicles in order to
improve handling and increase fuel efficiency.
2. PROCEDURE
Figure 2. Typical Fuel Energy usage at urban and

highway driving
Aerodynamic drag (AD), which compares the drag
force, at any speed, with the force it would take to
stop all the air in front of the car influences fuel
consumption of a car, especially at higher speeds
(Figure 3) and hence is considered a crucial factor in
judging its performance. An aerodynamically well
designed car spends the least power in overcoming the
drag and hence yields higher performance - cruises
faster and longer that too on less fuel [1].
The first step of CFD is to design a computational

model. This general process is done using a program
called Gambit (Version 2.3.16). In this program,
geometry of the object is created in AutoCAD then it
imported in Gambit. After the modeling is done, a
grid is applied. This grid, also called a mesh,
represents the computational domain. All the faces
were meshed using a Tri element and an interval size
of 0.003. This creates an unstructured grid of
triangular mesh elements (Figure 4). Unstructured
meshes are more applicable, especially to irregular
faces. After the meshing was finished, the outer
domain for the problem was created. A real brick
(shaped volume) was created around the car with
dimensions of 10m x 5m x 5m (width*depth*height
or x*y*z) with the coordinate system centered within
the brick, then it was moved 2.5 meters in both the y
and z directions in order to obtain proper positioning.
Then volume mesh was done using a Tet/hybrid
element and TGrid type, the interval size was left at
the default value of 1 (Figure 5).
Figure 4. Meshed faces of the car
Figure 3. Increasing influence of drag co efficient

(CD) at higher speeds
This paper outlines the process taken to optimize
the geometry of a vehicle. Vertices and edges were
156
Figure 5. Side view of car and mesh
After the meshing was complete, the boundary

types must be set. Boundary-type specifications
define the physical and operational characteristics of
the model at those topological entries that represent
model boundaries. All of the boundaries can be
found in Table 1. Finally, the mesh was exported to
Fluent for analysis.
Fluent (Version 6.2.16) is used to solve the
exported problem. Fluent applies fluid flow physics in
order to get solutions. The choice of the turbulent
model depends on the required level of accuracy,
available computational resources, and the required
turnaround time. For the problem analyzed in this
paper, standard k - turbulent model is selected. The
k - model is one of the most common turbulent
models. It is a semi - empirical, two equation model,
that means, it includes two extra transport equations
to represent the turbulent properties of the flow. The
first transported variable is turbulent kinetic energy k.
The second transported variable is the turbulent
dissipation . It is the variable that determines the
scale of the turbulence, whereas the first variable k
determines the energy in the turbulence. The model
transport equation for k is derived from the exact
equation, while the model transport equation for was
obtained using physical reasoning and bears little
resemblance to its mathematically exact counterpart.
The Fluent software performs discretization over the
computational domain and allows the user to generate
figures and plots as a means of visually comparing the
results and data.
Table 1. Boundary types
Boundary Conditions
Face
Boundary Type Zone Name
In front of the
Velocity Inlet
Inlet
car
Side opposite
Symmetry
Side
the car
Above the car
Symmetry
Top
Below the car
Wall
Ground
Two faces
Symmetry
Symmetry
created at
Plane
the car center
plane
Behind the car
Pressure Outlet
Outlet
Faces of the car
Wall
Car
the hood. Both the front and rear end of the roof are
negative pressure areas. Particularly, air slows down
when it is approaches the front of the car and results
that more air molecules are accumulated into a
smaller space. Once the air stagnates in front of the
car, it seeks a lower pressure area, such as the sides,
top and bottom of car. As the air flows over the car
hood, pressure is decreasing, but when reaches the
front windshield it briefly increasing. When the higher
pressure air in front of the windshield travels over the
windshield, it accelerates, causing the decreasing of
pressure. This lower pressure literally produce a lift force on the car roof as the air passes over it.
Figure 6. Pressure contour around the car body

Figure 7 shows that air velocity is decreasing as it
is approaching the front of the car. Air velocity then
increases away from the car front. It can be seen that
there are small separation sectors in the areas above
the radiator, near the bumper, and above the front and
rear windshields. And some eddy flow is found in the
wake of the body. In the area near to the symmetric
section, air flow is parallel to ground while it is apart
from the roof, and slightly deflects to both sides. Air
between floor and the ground goes upwards while
leaving from the body. The main direction of the wake
is upward. In the areas near to the sides of the body,
air from the bottom is whirling and a large eddy is
found. Air from the roof of either side whirls softly, it
forms rather small eddies, which are not obvious.

Figure 6 shows the pressure contour around the
car, it can be easily found that there are positive
pressure areas at the front of the body, especially in
the area before the radiator and the boundary area
between hood and the front windshield. However
negative pressure area is also found at the front end of
157
Figure 7. Velocity vector around the car body
Simulation is done for different air velocity

and the value of drag force is obtained for each
velocity. Variation of drag force with air velocities is
very much in confirmation with theoretical
understanding. For validation of the result, only
variation of drag force with air velocities are
compared with M. Desai [1] experiments result. He
did the experiments in two methods on a
geometrically similar, reduced scale (1:15), clay
model, differing from actual car (my model) only in
size and simulating dynamically similar flow
situations. He placed the model in a wind tunnel test
section blows air at different velocities. He measures
the variation in static pressure at different nodes by
traversing Pitot tube in horizontal and vertical
direction, thereby covering the entire flow field on
upstream as well as downstream sides of the car and
calculates in first method. In second method, he used
pressure tapings along the centerline of the car over
its profile.
some lower than experimental value. But it is able to

overcome in higher speed.
4. CONCLUSION
This paper presents the process taken to optimize
the geometry of a vehicle. The geometry was
imported into Gambit, a computational domain
created and unstructured triangular mesh was then
applied. The CFD program Fluent was used to obtain
the flow characteristics around the car. The results are
analyzed and only the drag force is compared with a
recognized journal to validate the results. Comparison
of the numerically derived results with that of the
experiments obtaining from well recognized journal
shows a good correlation. However further
investigations are suggested in order to reduce the
differences in the results at different conditions. These
practices were detailed in hopes that further research
would use the ground work laid out in this paper to
redesign existing vehicles in order to improve
handling and increase fuel efficiency. It is also
recommended that the methods used in this paper can
be applied to other vehicles of interest and used to
create better vehicles. Applying this study to different
geometry of vehicles to reduce the drag optimum
geometry can be designated, which is the ultimate
objective.
REFERENCES
Figure 8. Variation of drag force with air velocity

(Computational /Experimental from journal)
Figure 8 shows the comparison of drag force
between the numerical result and the two
experimental results. It appears a good agreement
among the results; however in lower speed range the
numerical drag force is a little bit lower than the
experimental results. But in higher speed range a good
agreement displays. Results differ due to the size of
the model, especially in lower velocity. Size, however,
plays a major factor in the aerodynamics of the
vehicle. For same velocity, Reynolds number will be
less if the size of the car is small. The boundary layer
thickness and turbulence transition as well as the flow
separation points are dependent on the Reynolds
number, therefore determining the most dominant
source of drag. The boundary layer creates viscous
drag through shear stress onto the surface, and
separation creates pressure drag. On a full scale car,
the pressure drag is stronger than viscous due to the
high Reynolds number and complex shape [3]. So for
the vastly different Reynolds numbers, Pressure
separation drag is the dominant source of drag on
these model cars and therefore the numerical value is
[1] M. Desai, S A Channiwala and H J Nagarsheth,

A comparative assessment of two experimental
methods
for
aerodynamic
performance
evaluation of a car, Journal of scientific &
Industrial research, Vol. 67, pp 518 522 (July
2008)
[2] P. N. Krishnani, CFD study of drag reduction of
a generic sport utility vehicle, Masters thesis,
Mechanical engineering department, California
State University, Sacramento (fall 2009).
[3] R. T. Kell, Aerodynamic Analysis of F1 in
Schools Challenge Car Models, Initial Thesis
Report, ACME, UNSW@ADFA, Canberra,
Australia (2009).
[4] Sugiono, W. M. Hong, O. Ilias, Developing the
database for vehicle body shape design with the
less of aerodynamic resistance and vibration
ECCM 2010 , IV European Conference on
Computational Mechanics, Palais des Congrs,
Paris, France (May 16-21, 2010).
[5] S. N. Singh, L. Rai, A. Bhatnagar, Effect of
moving surface on the aerodynamic drag of road
vehicles Proceeding of IMechE. Vol. 219 Part
D: J. Automobile Engineering, pp 127- 134
(August 2004).
158

AN EXPERIMENTAL ANALYSIS OF
THE INTERNAL FLOW EFFECTS ON MARINE RISERS
Motohiko Murai and Marcio Yamamoto
Yokohama National University
Graduate School of Environment and Information Sciences
79-7 Tokiwadai, Hodagaya-ku, Yokohama, Japan
Email: m-murai@ynu.ac.jp, yamamoto@ocean.jks.ynu.ac.jp
ABSTRACT
Nowadays, tubular structures are widely deployed offshore. The Offshore Industry names as riser the pipe
that conveys hydrocarbon or other work fluid (e.g. drilling mud) between the wellhead (or other subsea
equipment) on the sea bed, up to the platform at the surface. Up until now, most of risers researches have
been focused on the hydrodynamics loads and the boundary conditions, i.e., the motion of the supporting
vessel at the top end, and in same risers configuration the interaction between pipe and soil. This work is
focused on the effects of the pipes internal flow on its mechanical behavior. The investigation has been
carried out mainly by the way of experimental analysis, but some discussion about numerical analysis is also
included.
Keywords: Marine Riser, Internal Flow, Dynamic Behavior.
1. INTRODUCTION
The Exploration & Production of offshore
hydrocarbon fields have increased tremendously in
the last 30 years. The new Exploratory Frontiers
have advanced into deeper and deeper waters. In
order to develop such new discoveries, one essential
subsea equipment is the Riser. The Offshore
Industry names as riser any pipe that connects a
subsea wellhead or other equipment located at the
sea floor to the platform at the surface.
The riser can be classified into drilling riser and
production riser. The drilling riser is a rigid pipe
usually made of steel. During the first two drilling
phases, the drilling is carried out riser less. After
that, the Blow-Out Preventer (BOP) and the drilling
riser are used. The BOP is lowered down attached to
the drilling risers lower end (Figure 1a). The
drilling risers sections are connected each other
through bolted flanges joints. Then the BOP is
installed on the wellhead, such installation is called
re-entry operation [10].
After the installation of the BOP and drilling riser,
the drilling is resumed. The drillstring is lowered
down into the drilling riser and BOP, and the
drilling riser conveys the drilling mud with rock
cuttings from the wellhead up to the platform.
Following, the next risers classification is the
production riser. The production risers convey fluid
in either upward and downward directions. The
most common flow direction within the production
riser is upward when a multiphase fluid (a mixture
of water, oil and gas) flows from sea bed up to the

platform. But the production riser also can be used
for the downward flow as, for example, to inject
water into the reservoir, to pump the gas used in the
artificial elevation method called gas-lift, or even
to export gas or oil to the shore.
The production risers are also classified into rigid
riser and flexible riser. The rigid riser is also a
common pipe, usually, made of steel. Such riser is
mostly deployed in the vertical position. Such
vertical configuration restricts its application to
fixed platforms and complaint platforms (e.g., the
Tension Leg Platform, or TLP, as shown by Figure
1b), which have a reduced heave motion.
The flexible production risers are composed by
several concentric layers of helical metallic armors
and polymeric material. The metallic armors give
mechanical strength to the pipe and the polymer
layers are responsible for the pipes sealing and
insulation. Such flexible risers are widely deployed
for hydrocarbon production or water injection with
conventionally moored floating platforms, e.g.,
Floating Production Storage and Offloading (FPSO) and
semi-submersibles, as shown by Figure 1c.
Up until now, several researches have been
carried out concerning the effects of the
hydrodynamics loads, motion of the platform and
contact with soil on the risers mechanical behavior.
But only a few researches have concerned about the
effect of the internal flow on the riser dynamics and
this work addresses to this gap.
159
TLP
DRILLING RISER
FLEXIBLE RISER
SSHR
Waves
Current
Drilling
Riser
Tendons
Buoys
Rigid
Riser
BOP
Wellhead
(a)
Buoy
Flexible
Riser
(b)
(c)
Flexible
Riser
Rigid
Riser
(d)
Figure 1. Example of risers: drilling riser during the installation (a), rigid riser (b), flexible riser (c), and a hybrid riser
configuration called Self-Standing Hybrid Riser (d).
This work intends to investigate the internal flow

effects in the risers mechanical behavior by the way
of experimental analysis. Discussion about the
effect of internal flow on the risers dynamic
equation for numerical simulation is also included.
2. EXPERIMENT
Different experiments had been carried out in the
Yokohama National University (YNU) and the
Deep Sea Basin located in the National Maritime
Research Institute (Tokyo, Japan) in order to
investigate the effect of the internal flow on the
risers mechanical behavior. The first experiments
were carried out in the YNU; in such experiment, a
silicon made tubing of 0.6 m of length was used as
model and fresh water was injected into the pipe by
a positive displacement pump. Further, these
experiments were ran dry to separate the effects of
internal flow and the effect of the external
hydrodynamics loads. The experimental results
agreed with the results of other researches [4] that
were obtained analytically. Namely, the natural
frequency has the trend to decrease when the
internal flow rate increases, the effect on the
damping is reduced, and the risers axial tension has
significant effect on the risers mechanical behavior.
The other experiment was carried out at the

NMRIs Deep Sea Basin. This time, a silicon made
tubing of 10 m of length was used as model; and the
experiments were carried out in the water. Again, a
positive displacement pump injected water into the
model, and a harmonic motion was applied to the
top of the pipe. The results shows that pipes
response shape changed when the internal flow rate
increased. Even the pipes Vortex-Induced
Vibration had a different response for different flow
rate.
2.1 Experiment at the YNU

The model used in this experiment is a silicon
made tubing, which the main properties are feature
in Table 1.
Figure 2a shows the schematic of the experimenttal apparatus. The pipes model is clamped at both
ends, but the top end is connected to slider that is
driven by a stepping motor and generates the forced
oscillation on the top. The model has also 4
measurement stations that were named A, B, C and
D from the top. The station A corresponds to the
sliders position.
Table 1. Main properties of the model for the
experiment at the YNU.
Specific gravity
1.18
Modulus of elasticity (E)
4.5 MPa
Length
600 mm
Internal diameter (ID)
6.35 mm
External diameter (OD)
9.35 mm
Second Moment of Area (I)
295.35 mm4
Then the pipes vibration was filmed using a highspeed camera (300 fps). After, a software post
processes the recorded movie tracking and
calculating the time series for each one of the
measurement stations.
Figure 2. Schematic (a) and a picture (b) of the experiment
carried out at the YNU.
2.1.1 Data Analysis

Then we calculated the Fourier Coefficients [3],
viz.
160
a
b
t dt
f t cos
(1a)
f t sin
t dt
(1b)
N
N
where a and b are thhe Fourier coeefficients; N is

ycles that shalll be analyzed;
the numberr of signals cy
f t is the time series; is the signaals frequency
that shall be
b analyzed; an
nd t is time.
show such
h trajectories. The time stepp between eachh
point is 0.01 s; and only a half cycle of thee
me points aree
oscillatoryy motion is plotted. Som
highlighteed only to maake clearer thhe pipe shape..
Further, thhe arrows indiccate the stationns trajectory.
Using thhe Fourier coeffficients, we caan calculate a

function th
hat fits to the tim
me series, nam
mely,
g t
A cos t
(2)
In the above equatiion, A is thhe amplitude

defined by
y Eq.(3a); is the signals freequency ; t is
time; iss the phase deffined by Eq.(3bb).
A
ataan2
(3a)
Figure 4. Staation Cs time seeries.
(3b)
where A and are thhe amplitude and

a the phase
of Eq.(2); and
a a and b are the Fourieer coefficients
defined by
b Eq.(1); thhe function atan2
a
is an
extension of the trigonom
metric function atan that is
available in
i most of com
mputational math
m
libraries.
The response of atan2 is defined inn the interval
, , instead atan is defined
d
in
, .
2.1.2 Ressults
Figure 3 shows the time series of the
t Station A
with input amplitude and
d frequency off about 4 mm
r
respectiv
vely. In the pllot, the black
and 14.5 rad/s,
line is the experimental data and the red
r line is the
fitted curvee calculated ussing equation (22).
Figure 5. Stations trajjectories for the 1st harmonic.
3 harmonic.
Figuree 6. Stations trajeectories for the 3rd
Fig
gure 3. Station As
A input (=14.5 rad/s).
Next, Figure 4 shows the

t Station Css response for
the previouus input (Fig. 8).
8 We can see that the input
with a freequency of 14
4.5 rad/s geneerated a high
frequency response of thee pipe.
The speectral analysiss revealed th
hat the highfrequency response has a frequency of 43.72 rad/s,
mes the top oscillation frequen
ncy.
about 3 tim
Then th
he trajectory function,
fu
viz. equation (2),
was calculated for each
h measuremen
nt station for
d/s. Figure 5 and
a
Figure 6
14.5 rad/s and 43.5 rad/
Figure 5 shows the piipes response due to the topp

n, namely, 14.55 rad/s. Figure 6 shows the thee
oscillation
response for
f 43.5 rad/s. The
T pipes ressponse, shownn
by the Fig
gure 6, seems tto be the 1st modal
m
shape forr
a fixed beam.
b
In fact,, this high freequency (43.66
rad/s) is close
c
to the natuural frequency
y measured in a
previous work [11]; bbut, in the preevious workss
experimenntal apparatuus, the pipe was in thee
horizontall position.
Figure 7 shows the innternal flow efffect on the am-for the four stattions, analyzedd
plitude, eqquation (3a), fo
frequencyy is 43.5 raad/s, and to
op oscillationn
frequencyy is 14.5 rad/ss. We can obbserve that thee
161
internal fllow with 7900 mL/min hass the highest

amplitude.
The Deeep-Sea Basin is equipped with

w 10 pairs off
underwateer high-resoluttion cameras. Each
E
pair filmss
a portion
n of the basinn. A cluster of computerss
processes the images from
fr
those cam
meras, in real-time, traccking and calcuulating the tim
me-series in 3D
D
of each measurement
m
staation.
Figure 7. Amplitude
A
of eacch Station for thee third harmonicc
motionn (43.5 rad/s)
2.2 Exp
periment at the
t Deep-Seea Basin
Table 2 features the main
m
propertiess of the 10 m
long silicoon made tubin
ng that was ussed as model
during the experiment at the Deep-Sea Basin.
B
Table 2. Main propertiies of the modell used in the
experimennt at the NMRI.
Modulus of
o elasticity (E))
Length (L))
1..1 MPa
10 m
External diiameter (OD)
42
4 mm
Internal diaameter (ID)
32
3 mm
Second Mooment of Area (I)
101.1E-9 m4
Figure 8 shows the scaale of the modeel used during

ong its length,
this experiment. The pipe also has, alo
several meeasurement stattions.
Figure 9. Experimental appparatus of the experiment

e
at thee
NMRI.
Figure 9 is a scheematic of the experimentall

apparatus. A positive displacement pump injectss
fresh watter from an auuxiliary tank innto the model;;
and an oscillator
o
generrates a top osscillation. Thee
model is clamped at tthe top end and
a
free in alll
degrees of
o freedom at tthe lower end. In addition, a
weight off 4.6 kg was atttached at this loower end.
2.2.1 Reesults
Figure 10 shows thhe risers respponse in timee
f
the top ooscillation. Suuch data wass
domain for
obtained using
u
the visuaal measuremennt system. Thee
In-Line grraph corresponnds to the pipees response att
same direection of the topp oscillation annd Transversall
is the perpendicular
p
response; succh transversall
response is probably caused by the vorticess
formation abouut the Vortex-shedding. Further info
V
can be
b found in Sarrpkaya [7].
Induced Vibration
Figure 10. Riseer's dynamic behhavior.
Figuree 8. Model with 10 m length used during

experiment at the NMRI.
For lacck of space, only the results for twoo

different flow rates wiill be presented. Figure 111
shows thhe amplitude and phase, as shown byy
equation (3), for eachh one of the measurementt
stations. Such
S
amplitudees and phase were
w
calculatedd
for the signal frequuencies of 0.235
0
Hz (topp
oscillation
n frequency), 0.470 Hz, and 0.705 Hz. Ann
162
d
dz
internal flow of 0.675 L/s is injected at the top of the

model. The measurement stations were numbered
from #1 (at the top) up to #29 (at the bottom).
EI
d xR
dz
T
We can observe, in Figure 11, that the main pipes

response is at the top oscillation frequency, viz.,
0.235 Hz. In such response, there is node around the
node #25 and the top and bottom are vibrating in
opposite phases.
R A
2 I U
A P
AI
xR
t z
AI PI
A
FZR
1
CD D |u
2
d xR
dz
d xR
AI I
mxR
dz
I A I U
UC
xR | u
UC
xR
(4)
where EI is the beam stiffness; xR is the riser

deflection and the dot above means the its time
derivative; z is the coordinate along the unbent
beams axis; I is the density of the external
(internal) fluid; U is the internal fluid velocity; T is
the axial tension in newtons; A AI is the external
(internal) cross sectional area; P PI is the external
(internal) pressure acting on the risers wall; R is the
specific weight of the risers material; FZR is the
resulting forces per unit length in the z-direction;
I is the specific weight of the external (internal)
fluid; m is the mass per unit length (including the
internal fluid mass and hydrodynamic added mass);
CD is the drag coefficient; D is the risers external
diameter; u is the velocity of the external particle due
to surface waves; and UC is the velocity of the
external fluid particle due to current.
Figure 11. In-line results for internal flow rate of 0.675 L/s
(top oscillation 0.235 Hz).
Next, Figure 12 shows the results for a different

internal flow rate, i.e., 1.013 L/s; the top oscillation
frequency is kept the same as the results in Figure 11.
Equation (4) includes the effects of an inviscid

flow represented by the terms containing the internal
flow velocityU. According to Wu & Lou [9], the
internal flow introduces an additional Coriolis force
R
represented by the term 2 I U
. Despite this
term has a time derivative, it does not dissipate
energy from the system. Thus such term is neglected
in the numerical analysis of this work.
3.2 Numerical Model
Figure 12. In-line results for internal flow rate of 1.013 L/s
(top oscillation 0.235 Hz).
We can observe that changing the internal flow rate,

the pipes response changed, especially at 0.235 Hz.
The node of such response is near to station #16 and
the phase between top and bottom is about 90.
A numerical simulation program has been
developed in order to better understand the effect of
the internal flow on the risers mechanical behavior.
3.1 Risers Mechanical Behavior

A riser can be modeled as a slender tensioned beam
[5] with transverse loads acting due to effects of the
hydrostatic and hydrodynamic pressures, namely,
In the numerical simulation, instead to solve the

equation (4) with continuous distributed properties
and infinite number of degrees of freedom, the riser
is discretized in a number of elements; it is assumed
that each element has 2 nodes (one node at each
elements end) with 3 degrees of freedom (2
translational DOFs and 1 rotational DOF), summing
6 DOFs per element. Then the risers dynamic
properties are lumped into such nodes. Thus a
discrete approximate dynamic equation is obtained,
viz.
M
(5).
In the above equation, xS is the vector containing

the displacement of each node in time domain, and
the dot above represents its time derivative; M is the
global consistent mass matrix; B is the global
structural damping matrix (the viscous damping is
not included); K is the global stiffness matrix,
because the riser has an axial tension, the stiffness
163
matrix is composed by the Elastic Stiffness and

Geometric Stiffness Matrix; fS is the vector of
external loads that are concentrated on the nodes,
such forces are composed by the hydrodynamics
forces and the forces caused by the internal flow.
For further details about consistent mass matrix,
structural damping matrix, elastic stiffness matrix,
geometric stiffness matrix, and coordinate
transformation, see Clough & Penzien [1] and Paz
& Leigh [6].
P 2,
This research has investigated the effect of

internal flow on the dynamics of marine risers. By
the way of experimental analysis, we could observe
that when the internal flow rate changes, the pipes
responses also changes. In Figure 7, we can observe
that variation of the pipes vibration amplitude due
to the internal flow. Further, Figure 11 and Figure
12 show the internal flow effect on the pipes amplitude and phase.
Finally, further investigation is necessary to verify
the effect of the internal fluid viscosity on the pipes
mechanical behavior. Further, the development of
numerical simulation program can allow us to model and understand better the effects generated by the
internal flow.
+d
Fx
P1 , U
4. CONCLUSION
U2
Fz
+d
kinetic energy change; and PPF is the frictional

pressure drop.
ACKNOWLEDGMENTS
Figure 13. An infinitesimal riser element.
3.2.1 Forces generated by the Internal Flow

The right term of equation (5) is usually
composed by the external hydrodynamics forces
acting on the riser. In this work, the forces generated
by the internal flow are also included as proposed
by Seyed & Patel [8]. Such forces are calculated by
the balance between the inlet pressure and the outlet
pressure of a piece of pipe (Figure 13) and the
variation of the internal fluids linear momentum as
shown by equation (6).
f
AI P sin
P sin
AI P cos
U sin
[1]
[2]
(6a)
P cos
AI I U cos
REFERENCES
[3]
AI I U sin
f
The authors would like to acknowledge the Japan

Society for the Promotion of Science (JSPS) for its
financial support during this research. The second author
also extends the acknowledgements to the Ministry of
Education, Culture, Sports, Science & Technology of
Japan for the granted scholarship.
U cos
[4]
g dL
(6b)
In the above equation, f and f are the resultant

force exerted by the internal flown on the pipe
element in x-direction and y-direction, respectively;
AI is the internal cross sectional area of the pipe; P
( P ) is the pressure at the lower (upper) end as
shown by Figure 13; I is the density of the internal
fluid; ( ) is the angle between vertical and
lower (upper) end (Figure 13); U ( U ) is the
internal fluid velocity at the lower end (upper) end.
[5]
[6]
[7]
The risers internal pressure drop can be estimated

summing three different terms [2], viz.
P
PPE
PKE
PPF
(7)
where P is the pressure drop along the pipe; PPE

is the pressure drop due to the gravitys potential
energy change; PKE is the pressure drop due to the
[8]
164
Clough, R. W., Penzien, J., Dynamics of

Structures, 2nd Ed., McGraw-Hill (1993).
Economides, M. J., Hill, A. D., EhligEconomides, C., Petroleum Production
Systems, Prentice Hall (1994).
Hamming, R. W., Numerical Methods for
Scientists and Engineers, 2nd Ed., Dover
(1986).
Olunloyo, V. O. S., Oyediran, A. A., Osheku,
Dynamic response interaction of vibrating
off-shore pipeline and moving seabed with
varying geological and geo-mechanical
properties, Proc. of the 23rd
OMAE,
OMAE2004-51353 (2004).
Patel, M. H., Witz, J. A., Compliant Offshore
Structures, Butterworth-Heinemann (1991).
Paz, M., Leigh, W., Structural Dynamics:
Theo-ry and Computation, 5th Ed., Springer
(2004).
Sarpkaya, T., A critical review of the intrinsic
nature of vortex-induced vibration, Journal of
Fluids and Structures, Vol.19, No. 4, pp. 389447 (2004).
Seyed, F. B., Patel, M. H., Mathematics of
flexible risers including pressure and internal
flow effects, Marine Structures, Vol.5, No. 23, pp. 121-150 (1992).
[9]
Wu, M. C., Lou, J. Y. K., Effects of rigidity

and internal flow on marine risers, Applied
Oceand Research, Vol. 13, No. 5, pp. 235-244
(1991).
[10] Yamamoto, M., Morooka, C.K., Ueno, S.,
Dynamic behavior of a semi-submersible
platform coupled with drilling riser during re-
entry operation in ultra-deep water, Proc. of

the 26th OMAE, OMAE2007-29211 (2007).
[11] Yamamoto, M., Murai, M., Maeda, K., Uto, S.,
An experimental study of the interaction
between pipe structure and internal flow,
Proc. of the 28th OMAE, OMAE2009-79312
(2009).
165
WIND SAIL ANALYSIS USING COMPUTATIONAL FLUID DYNAMICS

SIMULATION
A. Sulisetyono 1, A. Nasirudin2, and FA Wibowo3
1
Department of Naval Architecture And
Shipbuilding Engineering, Institut
Teknologi Sepuluh Nopember, Surabaya,
sulisea@na.its.ac.id
2
Shipbuilding Engineering, Institut Teknologi
Sepuluh Nopember, Surabaya,
anasirudin@na.its.ac.id
3
Shipbuilding Engineering, Institut
Teknologi Sepuluh Nopember,
Surabaya,
ABSTRACT
This paper described the behavior of rigid and untwisted sail model in the fluid dynamic analysis. The two
and three dimensional wind sail was modeled numerically using CFD technique to determine the influence of
the main sail design variables such as draft, camber and angle of attack due to the forces coefficients of sail.
The 10-sail models were developed, and each sail model was simulated at the variation of the angle of attack
such as 15o, 20o and 25o respectively. For certain case, the numerical results were validated by the results of
wind tunnel model testing in term of drag and lift coefficient. The simulation showed the design of sail models,
which generated maximal efficiency and the most optimal thrust force, were the sail models with camber value
20% and 45% of draft.
Keywords: sail design, computational fluid dynamics, wind tunnel testing
1. INTRODUCTION
One of the shipbuilding technologies developed is
a ship propulsion system by using renewable energy
sources. This is motivated by the increasingly prices
of oil which affects the operational cost of vessel,
and the impact of burning fuel by ship to the
pollution in the marine environment. The energy
crisis in the world can be anticipated by the
reducing oil consumption in shipping industry, and
begin to switch to the renewable energy sources
such as sunlight and wind.
Indonesia has developed the utilization of
renewable energy as a driving force (propulsion) on
the ship KLM. Maruta Jaya which was a 900
DWT general cargo vessels using sail technology as
the prime propulsion, and a DC motor as a means of
driving aids. Ship "Maruta Jaya" was the concept of
energy-saving ships that was collaboration project
between the Governments of Indonesia and
Germany in the 1980s (Wiriadidjaja, 1997).
Sail design is an important part in the generating
of ship speed optimally. The study was conducted
by Kartika, (1996) shows the sail performance of
900 DWT ships Maruta Jaya was not optimized
yet in generating the sail trust. The development of
sail design in order to improve the performance of
sail is still necessary.
Commonly, the thrust of sail was influenced by

the natural conditions such as, velocity, density of
air; operator skill such as position of the sail against
the wind direction/angle of attack); and the
geometry of the sail such as, wide sail shape,
camber and draft.
This paper is proposed to analyze the sail models
that was developed based on the dimensions of the
main sail KLM Maruta Jaya. Numerical analysis
was conducted by the numerical approach of CFD
(Computational Fluid Dynamics) as well as the
wind tunnel testing. Furthermore, this paper
described the effect of variable shape of the sail
(camber and draft) and the angle of attack against
the thrust.
The variations of draft and camber of sail were
determined based on the research performed by
(Collie, SJ, 2006) and (Nasirudin, A. 2002). There
were 10 types of sail modeled in which have the
variation value of camber and draft. The angle of
attack was determined in the range of angle 10o-20o
follow Marchaj (1982).
The simulation results were obtained by CFD
techniques acceptable while those were validated by
the results of sail model testing in wind tunnel.
167
2. COMPUTATIONAL AND
EXPERIMENTAL METHOD
2.1
Computational Setup
The cross-flow problems of the 2D sail was

solved by using ANSYS software. In this numerical
simulation of the flow was assumed turbulent,
steady, adiabatic and incompressible. Reynolds
Number of the air as the working fluid was 4.33 106,
with the fluid density () = 1.225 kg/m3 and
viscosity () = 1.79 x 10-5 kg/m sec. The flow was
assumed turbulent when the Reynold Number value
of 106<Re<107 (SJ Collie, 2006), and the turbulence
model standard k- was selected with the pressure
correction was 0,001.
of rigid untwisted sail is about 0.2 m. The variation

of camber on the sail was set at 0%, 5%, 10% and
20% respectively have a draft of 45%, 50% and
55%. Fluid boundary condition that was used in
modeling as has been illustrated in Figure 2 while
the length = 15*c = 184.605 m and width = 5*c =
61.535 m.
The sail size was modeled in the CFD geometry

with certain value of chord (c), the length of the
peak curvature (Ymax), and lies on a flat surface (Xd).
The Sail was a rectangular with dimensions of l (p)
= 32 m, c (chord) = 12.31 m, t (thickness) = 0.2 m,
A (area) = 393.92m2, AR (aspect ratio) = 2.6.
Figure 2 Modeling of sail in 2-D
Figure 1. Cross-section of sail (S. J. Collie, 2006)

Table 1 Geometry of sail in 2-D
Geometry
No.
Model
Camber
and Draft
(%)
cb0, dr0
cb5, dr45
0.6155
5.5395
cb5, dr50
0.6155
6.155
Ymax
(m)
Xd
(m)
cb5, dr55
0.6155
6.7705
cb10, dr45
1.231
5.5395
cb10, dr50
1.231
6.155
cb10, dr55
1.231
6.7705
cb20, dr45
2.462
5.5395
cb20, dr50
2.462
6.155
cb20, dr55
2.462
6.7705
The 2D sail model with geometry in Table 1 had

three variations of angle of attack such as 15o, 20o and
25o; and the number variations of the sail were 30
models. The boundary conditions of sail model were
determined while speed or pressure hold on the
boundaries of sail geometry. At the inflow and side
boundaries were obtained with the fluid velocity, Vx =
5.14 m/s and Vy = 0 m/s. At the outflow boundary
was determined with the value of pressure = 0 Pa, and
the velocity value on the surface of sail was 0 m /s.
Grid independence tests was carried out to obtain
the number of elements efficiently in producing a
good accuracy of output data with time computing a
shortest as possible. Tests were conducted with a
variety of elements including 5860, 13,300, 16,770
and 23,110 elements. The grid independence test was
performed on the model A with a camber of 5%, draft
of 45% and 15o angle of attack.
The simulation results in terms of the pressure (p)
and shear force () at each node on the sail surface
was performed for all models of the sail. Based on the
distribution of pressure and shear forces on each node,
it can be calculated the amount of aerodynamic forces
on the sail. Pressure distribution was illustrated
through the relationship graph of pressure coefficient
(Cp) and the ratio of long-distance nodes chord (x /c).
Pressure coefficient itself can be obtained from
equation (1).
Figure 1 shown the cross section of the sail,

where the value Ymax and Xd was determined in
accordance with camber and draft variation of the
model, and those can be seen in Table 1. The thick
(1)
168
where: CP = pressure coefficient, P = pressure

contours of the sail (on the node), P = pressure on
the free flow (free stream pressure), 1.361 N/m2
(obtained at the side), = density of air, 1.225 kg/m3,
v = air velocity, 5.14 m/s, q = dynamic pressure-free
flow 16.182 N/m2.
2.2 Experimental Setup
Tests on-sail model was carried out in laboratory
wind tunnel. Wind tunnel as shown in Figure 3 is a
place of testing specimens having rectangular section
with a length of 150 cm, height of 66 cm, 66 cm wide
and maximum speed of 21 m /s and the motor power
of 5.5 KW.
3. RESULTS
The result of the pressure distribution of 2D sail
model was obtained by using CFD technique and
wind tunnel test. Both results of pressure distribution
were compared.
3.1 Pressure Distribution by CFD

The pressure distributions were resulted from the
CFD simulation for all models in Table 1, and its can
be classified into four typical models since each other
have a different pressure distribution significantly.
First, the model of flat plate O, second A, B, C with a
camber of 5%, third in model D, E, F with a camber
of 10%, and fourth models G, H, I with a camber of
20%. Those four typical models were analyzed for
each group represented by one model. Figure 4 and 5
show the pressure distribution of model O and model
A due to angle of attack 15o, 20o, and 25o respectively
Pressure Distribution Model O
Lower15derajat
Upper15derajat
Lower20derajat
Upper20derajat
Lower25derajat
Upper25derajat
2.5
2
Cp
1.5
1
0.5 0
0.5
1.5
0
0.5
Figure 3 Open circuit, Subsonic Wind tunnel (ITS)
x/c
The test section on Wind tunnel was about

2465.31 cm2. Scale model was adapted follow the test
section in order to avoid blocking of the flow uniform
in wind tunnel due to the size of the model. The scale
of model can be defined, 1:100. In this testing the
model O and model A was selected in which model O
with a camber of 0% and the model A with a camber
of 5% and 45% draft. The model has a span or height,
660 mm, length of chord (c) = 123.1 mm, thickness =
20 mm, pressure taps = 26 holes and the static
pressure tap mounted on the center position p (l) of
the test model.
Wall pressure tap serves the measurement of fluid
flow static pressure that was mounted on the surface
of 2D sail model. Pivot static tube was used to
measure the stagnation pressure and static pressure of
the fluid flow behind the object. The manometer was
used to display the pressure value measured by a wall
pressure tap and pivot static tube. With Froude
similitude equation, the air flow rate was defined
about Vm = 1.54 m /s or the value that comes closest
to that number. Fan rotation frequency was setup to
produce the wind speed desired on the sail model.
Results of measurement by each pressure tap on the
sail model were used to calculate the pressure
distribution on the surface of the model.
1.5
Cp
Figure 4 Pressure distribution of model O

3.5
3
2.5
2
1.5
1
0.5 0
0
0.5
1
1.5
Pressure Distributi on Model A

Lower15derajat
Upper15derajat
Lower20derajat
Upper20derajat
Lower25derajat
Upper25derajat
0.5
1.5
x/c
Figure 5 Pressure distribution of model A

In general, the sail model has pressure distribution
close to airfoil; because the principle work of sail has
similarity with airfoil. A typical airfoil and the sail has
a high transition point which was negative (negative
pressure) on the upper part and the stagnation point
169
(positive pressure) on the lower part. Both points are

important in generating lift force of the sail, and the
lift contributes greatly generate thrust on the sail.
O), typical 2 (model A), typical 3 (model D) and

typical 4 (model G) respectively.
Figure 4 through 9 show that the difference angle

of attack 150, 200 and 250 affect the distribution
pressure on the contour model for typical 1 (model
Pressure Contour, = 15o
Figure 6 Pressure distribution of typical 1 (model O)
Figure 7 Pressure distribution of typical 2 (model A)
Figure 8 Pressure distribution of typical 3 (model D)
Figure 9 Pressure distribution of typical 4 (model G)
170
3.2 Pressure Distribution by Experiment

Figure 10 and 11 show the pressure distribution of
2D sail model from wind tunnel tests for model O
with camber 0% and model A with camber 5%, draft
45% respectively.
In the figures, while the angle of attack was
getting bigger, the transition point getting smaller and
opposite the stagnation point become bigger. On the
upper and lower parts of both models, the bubble
separation does not occur, and only on model A the
trailing edge separation evident with decreasing
pressure from the reattachment.
17. Comparison of CFD simulation and testing

models in wind tunnel were expressed with error.
Percentage error can be defined by the following
equation:
100%
With the CFT in each method, the difference between

the both results were obtained in Table 2. The error
was defined about 2.28% to 7.87%. This shows that
the both results has a small difference under 10%.
Pressure distribution Model O, camber 0%

3
Lower15derajat
Upper15derajat
Lower20derajat
Upper20derajat
Lower25derajat
2.5
2
1.5
Cp
Tabel 2 Error of the force resultant coefficient

CFT WT
1
0.5 0.00
0.50
1.00
1.50
0
x/c
0.5
1
CFT CFD
Error (%)
Mod.
Mod.
Mod.
Mod.
Mod.
Mod.
15
1.01
1.25
0.99
1.35
2.39
7.87
20
1.07
1.26
1.05
1.33
2.28
5.42
25
1.20
1.31
1.12
1.34
6.80
2.24
1.5
However, the percentage of error was expected to
reach smaller values when the pressure tap on the 2D
model wind tunnel more or adjusted by the number of
nodes on the surface of the CFD model. In addition,
errors can also be contributed by the precision of
measurement on the manometer.
Figure 10 Pressure of model O, (camber 0%)
Pressure distribution of Model A, Camber 5%

Draft 45%
3
2.5
2
1.5
3
2
1
0.5 0.00
1.5
0.50
1.00
1.50
0
0.5
LowerWT
UpperWT
LowerCFD
UpperCFD
2.5
Cp
Cp
PressuredistributionmodelOCamber0%, =15o
Lower15derajat
Upper15derajat
Lower20derajat
Upper20derajat
Lower25derajat
Upper25derajat
1
0.5 0.00
0.50
1.00
1.50
x/c
0.5
x/c
1
Figure 9 Pressure of model O, (camber 0%, = 150)
1.5
Figure 11 Pressure of model A, (camber 5%)

Relatively small difference both methods of
pressure distribution is shown in Figure 12 through
171
PressuredistributionModelACamber5%
Draft45%, =20o
PressuredistributionmodelO Camber 0%, =

20o
LowerWT
UpperWT
LowerCFD
UpperCFD
2.5
2
1
Cp
Cp
1.5
0.5 0.00
0.50
1.00
1.50
0
0.5
x/c
1
1.5
LowerWT
UpperWT
LowerCFD
UpperCFD
0.50
1.00
1.50
x/c
Figure 16 Pressure of model A, (camber 5%, = 200)
PressureDistributionModelACamber0%,
= 25o
2.5
LowerWT
UpperWT
2
LowerCFD
UpperCFD
1.5
2.5
PressuredistributionModelACamber5%
Draft45%, =25o
LowerWT
UpperWT
LowerCFD
UpperCFD
2
1.5
0.5 0.00
0.50
1.00
1.50
Cp
Cp
3
2.5
2
1.5
1
0.5 0.00
0
0.5
1
1.5
0.5 0.00
0.50
1.00
1.50
0
x/c
0.5
0.5
1.5
1.5
x/c
3.3 Trust of Sail
Cp
PressuredistributionModelAcamber5%,
draft45%, =150
3.5
3
2.5
2
1.5
1
0.5 0.00
0
0.5
1
1.5
LowerWT
UpperWT
LowerCFD
UpperCFD
0.50
1.00
1.50
x/c
The thrust of sail can be expressed in terms of trust

coefficient, where the sail which has the largest thrust
coefficient means that the screen has the largest thrust as
well. The magnitude of thrust coefficient can be expressed
in terms of driving force, CR and heeling force, CH. Figure
18 show the three variations of wind angle direction due to
ship (the apparent wind angle, ) such as 30o, 60o and 90o
with the value of CR and CH.
Figure 9 shows that the model G (camber 20% Draft
45%) have the most optimum thrust on all three variations
of the apparent wind angle. The polar diagram shows the
value of CR and CH coincides on models which have the
same camber design. And the camber design has a big
influence on the thrust generated than the draft design.
172
largest percentage of 7.87% error. This error occurs

caused by several factors, the biggest factor error
impact are the process of calculating the force on the
wind tunnel tests are less accurate, and scales that are
used less worthy style.
1.6
= 30o
=
1.4
60o
1.2
CD
1
0.8
=
25o
20o
15o
90o
REFERENCES
0.6
0.4
0.2
0
1.5
Model0
Model4
Model8
CR
Model1
Model5
Model9
0.5
Model2
Model6
0
Model3
Model7
Figure 18 Polar diagram of CR and CH
4. CONCLUSIONS
The analysis of 10 (ten) sail design were done by
using approach of CFD simulation and testing models
in wind tunnel. The variation percentage of the draft
on a sail design that were 45%, 50% and 55% cause a
slight decrease in the lift force, but increase in the
drag force. While the percentage of camber on the
model increase with 5%, 10% and 20%, the lift force
also increase significantly. The simulations show the
design of sail models which generated maximal
efficiency and the most optimal thrust force, were the
sail models with camber value 20% and draft 45%.
The comparison results of CFD and wind tunnel
testing showed the two-dimensional analysis has the
[1] Bao, F. and Dallmann, U.Ch, Some phisycal

aspects of separation bubble on a rounded
backward facing step, Science Direct, 2003.
[2] Collie, S.J., Application of computational Fluid
Dynamics to Two Dimensional Downwind Sail
Flows, PhD Thesis, Department of Mechanical
Eng. and Eng. Science, University of Auckland
New Zealand (2006).
[3] Kartika, C.A., Analisis Teknis Perencanaan
Layar sebagai Penggerak Utama pada KLM
Maruta Jaya 900 DWT, BSc. final project,
,Jurusan Teknik Sistem Perkapalan, ITS
Surabaya (1996).
[4] Marchaj C. A., Sailing Theory and Practice,
New York, United State. (1982)
[5] Miyata, H., and Lee, Y.W., Application of CFD
Simulation to Design of Sails, Jour. of Marine
Science and Technology, 4:163-172. (1999)
[6] Nasirudin,A., Desain Layar Simulasi CFD,
BSc final project, Teknik Perkapalan, ITS
Surabaya. (2002)
[7] Versteeg, H.K., and Malalasekera, W., An
Introduction to Computational Fluid Dynamics,
Longman Scientific& Technical, England (1995)
[8] Wiriadidjaja, S., Pengembangan Kapal Layar
Modern untuk Menunjang Perdagangan antar
Pulau di Indonesia, ITS Surabaya. (1997).
173

CORROSION BEHAVIOUR OF ALUMINUM ALLOY IN SEAWATER

W. B. Wan Nik1, O. Sulaiman2, A. Fadhli3, R. Rosliza4
1,2
Faculty of Maritime Studies and
Marine Science
University Malaysia Terengganu,
21030 Kuala Terengganu, Terengganu,
MALAYSIA
Tel: 6019-9146575; Fax: 609-6683193
Email: o.sulaiman@umt.edu.my
3
Faculty of Science and Technology
University Malaysia Terengganu,
21030 Kuala Terengganu, Terengganu,
MALAYSIA
Tel: 6019-9146575; Fax: 609-6683193
Email: fadhli@umt.edu.my:
4
TATi University College,
Teluk Kalung, 24000, Kemaman,
Terengganu, MALAYSIA
Tel: 6019-9162516; Fax: 609-6633123
Email: r.rosliza@yahoo.com
ABSTRACT
Many different types of destructive attack can occur to marine structures, ships and other equipments used
in sea water service. Year upon year the cost of marine corrosion has increased until it is estimated today at 4 %
of the Gross National Product. Aluminium alloys are important materials and commonly used in marine
applications. In the current research we are simulating a ship cargo container, which carries corrosive aqueous
material. In this study different types of aluminum were subjected to aqueous corrosion in salt spray chamber
and normal seawater container. Natural products (such as starch) were used as corrosion inhibitors for various
metals and alloys in aggressive marine environment. In order to study the corrosion behavior of aluminum alloy
AA6063 in seawater, the electrochemical behavior of aluminum alloys was investigated with different inhibitor
concentrations, using electrochemical potentiodynamic reactivation (EPR), potentiodynamic polarization and
electrochemical spectroscopy (EIS). The morphology study was carried out to observe the development of thin
film on the specimen and this study was performed using Scanning Electron Microscopy (SEM). The values of
activation energy confirm this result. EIS data showed that the mechanism of corrosion depends on inhibitor
concentration.
Keywords: Corrosion, Marine Environment, Aluminum alloy.
1. INTRODUCTION
Many ship superstructures and liquid cargo
containers are made of aluminum due to its low
weight and acceptable strength. The most popular
aluminium alloys for use in corrosive environments
such as seawater are the 5xxx and 6xxx series alloys,
which demonstrate adequate strength and excellent
corrosion resistance. These series are highly suitable
in various marine structures, machinery and port
buildings. Alloy 6063 is a heat treatable aluminiummagnesium-silicon alloy. It provides good
combination of extrudability and mechanical
properties. It also responds well to polishing chemical
brightening, anodizing and dyeing.
Ship cargo containers are exposed to sea
environment. Not only the outside of the cargo
containers are exposed to corrosive environment, the
internal compartments are also exposed to corrosive
aqueous media.
Marine corrosion is a perennial problem not only

to moving vessel but also ports and anything kept in
the sea or exposed to sea environment for a period of
time. Corrosion is fundamental process which plays
an important role in economics and safety. Apparently,
corrosion cannot be avoided, but its severity can be
reduced to a lower magnitude.
The term 'aqueous corrosion' describes the
majority of the most troublesome problems
encountered when metal material is in contact with
sea water. Various methods have been employed to
reduce corrosion. Several techniques and methods
have been developed to combat corrosion efficiency
are continually being sought after, as a result of
exorbitant amount spent on corrosion annually (Abd
El Rahim et al. [1], Bethencourt et al. [5]). The use of
inhibitor for the control of corrosion of shore and
offshore metal and alloys which are in contact with
aggressive marine environment is an accepted
practice.
175
Benzoate
compounds
offer
interesting
possibilities for corrosion inhibition and are of
particular interest because of their safe use and high
solubility in water (Rosliza and Wan Nik [10]).
Corrosion inhibitors seem to be attractive
because of their low cost and easy handling,
compared to other preventive methods. Normally
inhibitors are chemicals that interact with a metallic
surface, or the environment this surface is exposed,
giving the surface a certain level of protection.
Inhibitors often work by adsorbing themselves on the
metallic surface by forming a film and reducing
corrosion.
1.1 Corrosion Inhibitors

Inhibitors have always been considered to be the
first line of defense against corrosion. Several
corrosion inhibitors are available today. The objective
of this study was to evaluate the effectiveness of
sodium benzoate as an inhibitor to slow down or
prevent corrosion. This project involves the use of
gravimetric
measurements,
potentiodynamic
polarization
and
electrochemical
impedance
spectroscopy (EIS) to evaluate inhibitive action of
sodium benzoate on corrosion behavior of AA6063
aluminium alloy in seawater. The electrochemical
measurements showed that the presence of sodium
benzoate as an inhibitor significantly decrease the
weight loss, corrosion current densities (icorr),
corrosion rates and double layer capacitance (Cdl),
whilst increasing the polarization resistance (Rp).
Sodium benzoate converts to benzoic acid when
used in acidic mixtures. Benzoic acid has good antimicrobial features, but does not dissolve well in water,
whereas sodium benzoate dissolves very well in
water. It is also used as a corrosion inhibitor in
automotive anti-freeze products.
1.2 UMT Research Vessel and Research

Motivation
UMT Research Vessel was built to fulfil the
universitys research and teaching needs. The vessel is
being employed to support Institute of Oceonography
(INOS) research activities. The latest task is surveying
South China Sea for submarine routes. The vessel also
serves for students activities of Maritime Technology
Department, Nautical Science and Transportation
Department, Marine Science Department, and
Maritime Management Department.
The vessel is 35m long and made of aluminum as
shown in Figure 1. It is powered by 634kw engine and
its gross tonnage is 291. It has several small
compartments to serve as ballast tank and to keep
marine samples. The compartments are also made of
aluminum (Si (0.20%), Fe (0.35%), Cu (0.10%), Mn
(0.10%), Mg (0.45%), Cr (0.1
0%), Zn (0.10%),
Ti (0.10%) and Al (reminder) by weight).
Figure 1. 35 m long UMT Research Vessel.

Beside this RV, UMT has nine other boats or
small vessels. Aboard these vessels, there are quite
expensive equipments. Since they are heavily used for
teaching and research, they are always exposed to sea
environment. Significant corrosion was observed in
sampling tanks which normally contains sea products
and palm oils. Figure 2 shows one of the sampling
ballast tanks which was heavily corroded. This
phenomenon attracts the researchers to study marine
corrosion.
Figure 2. Actual corrosion on ballast tank.
2. METHODOLOGY
2. 1 Materials
In this laboratory scale study, the material
employed was AA6063 (25 x 25 x 1.5 mm coupons).
The actual seawater was used as the test solution.
The inhibitor used was sodium benzoate, NaBz
(1g/dm3 concentration).
Several studies have been carried out on
corrosion inhibitor; many of them concluded that
corrosion was uniform with sodium benzoate.
Benzoic acid and its sodium salt (sodium benzoate)
are the most common, safe, food preservatives and
antimicrobial agents. Benzoic acid and sodium
176
benzoate are classified in the United States as

Generally Recognized as Safe (GRAS) and their use
in food is permitted up to a maximum level of 0.1%.
Commercial availability and low cost of benzoic acid
and sodium benzoate make them attractive candidates
to be incorporated into coatings as environmentally
benign alternatives (Al-Juhni and Newby [2]).
2.2 Methods
Sophisticated corrosion inhibitor test methods,
typically designed to reproduce the most extreme
conditions in a system, have been employed to
improve inhibitor capabilities. Research of uniform
corrosion is the most widespread application of
electrochemical measurements both in the laboratory
and industry. The widespread use of these
electrochemical methods comes with limitation
(Yagan et al. [14]).
3. RESULTS
Figure 3 illustrates the weight loss of AA6063
alloys in seawater with static condition with and
without NaBz as inhibitors. Figure 4 presents the
corrosion rate versus the immersion time of the
aluminium alloy samples in seawater with and
without inhibitor. Figure 5 represents the
potentiodynamic polarization curves of AA6063 in
seawater in the absence and the presence of NaBz.
2.3 Electrochemical Impedance Spectroscopy

(EIS)
Circuit model or circuit description code and
initial circuit parameters were assumed and input by
the operator based on the shape of the EIS spectrum.
The program then fits the best frequency response of
the given EIS spectrum, to obtain fitting parameters.
The quality of the fitting was judged by how well the
fitting curve overlaps the original spectrum. The fitted
EIS data was used to obtain a set of parameters which
can be correlated with the coating condition and the
corrosion of the steel substrate.
Figure 3. Weight losses versus immersion time of

AA6063.
2.4. Electrochemical Analysis Process

The preparation of working electrode before
measurements is the same as described before
(Rosliza and Wan Nik [11]). Electrochemical
potentiodynamic reactivation (EPR) is probably the
most commonly used polarization testing method for
measuring corrosion resistance and is used for a wide
variety of functions. Potentiodynamic polarization is a
technique where the potential of the electrode is
varied at a selected rate by application of a current
through the electrolyte. This technique uses a
potentiodynamic scan over a range of potentials from
passive to active (called reactivation) to measure the
degree of sensitization of stainless steels.
All
electrochemical
measurements
were
accomplished with Autolab frequency response
analyzer (FRA) coupled to an Autolab potentiostat
connected to a computer. The potentiodynamic
current-potential curves were recorded by changing
the electrode potential automatically from -250 mV to
+250 mV with the scanning rate of 5 mVs-1.
Impedance measurements were conducted over a
frequency range of 104 Hz to 10-1 Hz. The data were
presented as Nyquist plots.
Figure 4. Corrosion rate versus immersion time of

AA6063.
177
4. DISCUSSION
Figure 2 shows the severity of one of our vessel
ballast tank. It was the result of almost 20 years of
seawater corrosion. This real life corrosion is very
complicated to be simulated in laboratory
environment.
4.1 Weight Loss Measurements
Figure 5. Potentiodynamic polarization curves for

AA6063.
The values for the corrosion potentials and

corrosion current densities were estimated from the
intersection of the anodic and cathodic Tafel lines.
The corresponding corrosion potentials (Ecorr),
corrosion current density (icorr), anodic Tafel slopes
(ba) & cathodic Tafel slopes (bc) are listed in Table 1.
Table 2 displays the results of the values of the IE (%)
from the weight loss and electrochemical tests.
Table 1. The electrochemical parameters: Ecorr, icorr,

ba, bc and Rp of AA6063 in seawater.
Results from Figure 3 show that the weight

losses of AA6063 for all samples increase with the
immersion time. The corrosion rate tends to be
increased by decreasing temperature, pH, and flow
velocity and by increasing dissolved oxygen
(Hollingsworth and Hunsicker [7]). Values of salinity
and pH measured during the immersion test are stable
(in the range of nature seawater). Baorong et al. [4]
reported that there was no obvious effect when the
seawater salinity was varied from 30% to10%.
Clear differences can be noted between the
samples exposed to seawater containing NaBz as
compared to that immersed in the absence of the
inhibitor. The test without inhibitor addition had the
higher weight loss. There was a reduction in weight
loss; hence reduced the corrosion rate with the
introduction of NaBz as an inhibitor into the seawater
for both cases in static condition.
The obtained results from Figure 4 demonstrate
that the test without inhibitor addition in static
condition had the higher corrosion rate. The results
indicated that introduction of NaBz obviously
minimize the rate of corrosion and abridged
aluminium dissolution in seawater. Therefore this
inhibitor can be considered as efficient inhibitor for
AA6063 in seawater.
4.2. Potentiodynamic Polarization Studies
Table 2. Values of IE (%) for AA6063 samples in

seawater.
The changes observed in the polarization curves

after addition of the inhibitor are usually used as
criteria to classify inhibitors as cathodic, anodic or
mixed (Bethencourt et al. [5]). From Figure 5, it can
be seen that the anodic and cathodic current densities
obtained in the presence of inhibitor are lower than as
compared to that of in the absence of inhibitor. The
corrosion potential (Ecorr) values in the presence of
inhibitor are shifted to negative direction and leftward
displacement in the cathodic branch of the curves.
According to Trabanelli et al. [13], these are the
typical features of cathodic inhibitors, this being in
agreement with the results obtained for other
aluminium alloys.
EIS research has proved that EIS is a powerful
and accurate method for measuring corrosion rates. In
order to access the charge transfer resistance and
polarization resistance that is proportional to the
178
corrosion rate at the monitored interface, EIS results

is interpreted by model of the interface. Advantage of
EIS techniques is the possibility of using very small
amplitude signals without disturbing the properties
being measured. In EIS measurement, a small
amplitude signal, usually a voltage between 5 to 50
mV is applied to a specimen over frequency range of
0.001 Hz to 100,000 Hz.
The EIS instrument records the real resistance and
imaginary capacitance components of the impedance
response of the system. Equivalent circuits that
describe electrochemical of a freely corroding
interface at kinetic equilibrium are (Al-Juhni and
Newby [2], Yagan et al. [14]):
Relaxation time constant with extended

diffusion.
Impedance of pitting processes.
RC representation of an electrochemical
interface.
The results display in Table 1 show that the
values of corrosion current density (icorr) for all
samples increase with the immersion time. The
change in Ecorr is assumed to be related to the growth
of a passive layer at the surface of electrode
(Rosalbino et al. [9]). The rate of cathodic reaction
controls the rate of the corrosion process (as the
cathodic over voltage is much greater than the anodic
one). The corrosion current decreases obviously after
sodium benzoate is added in test solution. It is clear
that the presence of inhibitor causes markedly
decrease in the corrosion rate. The corrosion
resistance of the investigated material is remarkably
higher in the benzoate containing than benzoate free
solution (Bethencourt, et al. [5]).
4.3 Electrochemical Impedance Spectroscopy
impedance. It includes a solution resistance, Rs in

series to a parallel combination of resistor, Rct,
representing the charge transfer (corrosion) resistance
and a double layer capacitor, Cdl, representing the
electrode capacitance (Badawy et al. [3]).
After 180 days immersion time, the charge
transfer resistance, Rct for the aluminium alloy
decreased to the likely formation (or thickening) of a
corrosion product layer on the electrode. The Rct
values were higher for AA6063 after addition of
inhibitor at all the immersion time, indicating that
NaBz is inhibiting the corrosion rate. It can be
concluded that NaBz is an effective inhibitor for the
aluminium alloy.
4.4. Inhibition Efficiency and Adsorption

Isotherm
The inhibitor such as benzoate group adsorbs on
aluminium alloy. As a result a thin film is formed on
the aluminium alloy to retard the corrosion. Thus, in
this case, sodium benzoate worked as filming
inhibitor to control the corrosion rate. Instead of
reacting with or removing an active corrosive species,
filming inhibitors function by strong adsorption and
decrease the attack by creating a barrier between the
metal and their environment (Al-Juhni and Newby
[2]).
The values of the IE (%) from the weight loss
and electrochemical tests were calculated using the
same equations as previously report (Rosliza and Wan
Nik, [12]). The values of IE (%) for the samples at
different immersion time are given in Table 2 and the
trends were similar for all three different types of
measurements.
5. CONCLUSION
The shape of the obtained Nyquist plots shows

single capacitive semicircles, showing that the
corrosion process was mainly charge-transfer
controlled (El-Etre, [6] and Trabanelli et al. [13]). The
general shape of the curves is very similar for all the
samples and this is maintained throughout the whole
test period, indicating that almost no change in the
corrosion mechanism occurred either due to the
immersion time or to the inhibitor addition (Reis et al.
[8]). Inspection of the data reveals that the impedance
spectra consists of a large capacitive loop at high
frequencies corresponding to the relaxation processes
in the oxide film covering the electrode surface (Abd
El Rahim et al. [1]). In all cases, single capacitive
loops are observed which sizes decrease with the
exposure time.
The equivalent circuit fitting for these
experimental data was a Randles circuit. The Randles
equivalent circuit is one of the simplest and most
common circuit models of electrochemical
The corrosion studies of AA6063 aluminium

alloy have been carried out at room temperature using
seawater. The results obtained lead to the conclusion
that NaBz is an effective corrosion inhibitor of
AA6063 in seawater. The potentiodynamic
polarization curves suggested a cathodic character for
the inhibition process in seawater. EIS measurements
clarified that the corrosion process was mainly
charge-transfer controlled and no change in the
corrosion mechanism occurred either due to the
immersion time or the inhibitor addition to seawater.
ACKNOWLEDGEMENTS
The authors greatly acknowledge the assistance
of Marine Corrosion students, namely Kamales and
Fakhratul Ridzuan and lab assistance, Mr. Afandi, Mr.
Yusof and Mr. Arif for their indirect contribution in
179
completing the experiment. The authors also would

like to thank the Ministry of Higher Education of
Malaysia for financially supporting this research
under Fundamental Research Fund (budget code
FRGS59210).
[7]
[8]
REFERENCES
[1] Abd El Rahim, SS., Hassan, HH., Amin, MA.,
The corrosion inhibition study of sodium
dodecyl sulphonate to aluminum and its alloy in
1.0 M HCl solution. Mater Chem Phys, Vol. 78,
pp. 337348, Elsevier (2003).
[2] Al-Juhni, AA., Newby, BZ., Incorpora-tion of
benzoic acid and sodium benzoate into silicone
coatings and subsequent leaching of the
compound from the incorporated coatings.,
Prog Org Coatings, Vol. 56, pp. 135145,
Elsevier (2006)
[3] Badawy, WA., Al-Kharafi, FM., El-Azab, AS.,
Electrochemical behaviour and corrosion
inhibition of Al, Al-6061 and Al-Cu in neutral
aqueous solutions, Corros Sci, Vol. 41, pp.
709727, Elsevier (1999)
[4] Baorong, H., Jinglei, Z., Yanxu, L., Fangying, Y.,
Study on effect of seawater saliniyu on
electrochemical performance of Al anodes,
Department of Marine Corrosion, Institute of
Oceanology,
Chinese
Academy
of
Sciences, China (2001).
[5] Bethencourt, M., Botana, FJ., Cauqui, MA.,
Marcos, M., Rodriguez, MA., Protection
against corrosion in marine environments of
AA5083 Al-Mg alloy by lanthanide chlorides,
Alloys Compounds, Vol. 250, pp. 455460,
Elsevier (1997).
[6] El-Etre, AY., Inhibition of aluminum corrosion
[9]
[10]
[11]
[12]
[13]
[14]
180
using Opuntia extract, Corros Sci, Vol. 45, pp.

24852495, Elsevier (2003).
Hollingsworth, EH., Hunsicker, HY., Corrosion
of aluminum and aluminum alloys, ASM Metals
Handbook, Vol. 13, pp. 582609 (1987).
Reis, FM., de Melo, HG., Costa, I., EIS
investigation on Al 5052 alloy surface
preparation for self-assembling monolayer,
Electrochimica Acta, Vol. 51, pp. 17801788,
Elsevier (2006).
Rosalbino, F., Angelini, E., De Negri, S.,
Saccone, A., Delfino, S., Electrochemical
behavior assessment of novel Mg-rich Mg-AlRE alloys (RE=Ce, Er), Intermetallics, Vol. 14,
pp. 14871492, Elsevier (2006).
Rosliza, R., Wan Nik, WB., The effect of
inhibitor on the corrosion of aluminum alloys in
acidic solutions, Mater Chem Phys, Vol. 107,
pp. 281288, Elsevier (2008a).
Rosliza, R., Wan Nik, WB., Electrochemical
properties and corrosion inhibition of AA6061 in
tropical seawater, Colloids Surf A, Vol. 312, pp.
185189, Elsevier (2008b)
Rosliza, R., Wan Nik, WB., Improvement of
corrosion resistance of AA6061 alloy by tapioca
starch in seawter, Current Applied Physics,
Vol. 10, pp. 221229, Elsevier(2010).
Trabanelli, G., Monticelli, C., Grassi, V.,
Frignani, A., Electrochemical study on
inhibitors of rebar corrosion in carbonated
concrete, Cement Concrete Research, Vol. 35,
pp. 18041813, Elsevier (2005).
Yagan, A., Pekmez, NO., Yildiz, A., Corrosion
inhibition by poly (N-ethylaniline) coatings of
mild steel in aqueous acidic solutions, Prog
Org Coatings, Vol. 57, pp. 314318, Elsevier
(2006).

HYDRODYNAMIC CALCULATION AND EFFECTS OF DEADRISE ANGLE ON

THE PLANING CRAFTS
Hassan Ghassemi 1, Saman Kermani1 and Abdollah Ardeshir2
1
Dept. of Marine Tech.,
Amirkabir University of Technology,
Tehran, IRAN
E-mail: gasemi@aut.ac.ir
2
Dept. of Civil Eng. Technology,
Amirkabir University of Technology,
Tehran, IRAN
Email: ardshir@aut.ac.ir
ABSTRACT
Hydromechanics calculations are one of the important steps in planing crafts design. In this paper, it is
described some parts of prepared software which is used with different modules for hydromechanics calculations
(hydrostatic curves, GZ-), hydrodynamic calculations (pressure distribution, lift and drag forces, wake) and
drawing 3D of planing crafts in AutoCAD. The method is based on the numerical boundary element method
(BEM) and developed Savitskys formulae. The results show that this software has high capability. Methods of
calculations in this paper are boundary element for pressure distribution, free surface wave and Savitskys
method for drag force.
Key words: planing crafts, pressure and lift and drag forces, free surface wave
1. INTRODUCTION
A vessel is planing when the length Froude number
be greater than 1.0~1.2. Hydrodynamic prediction of
the planing is one of the important aspects. Many
parameters may affect on the lift and resistance of the
planing hull [1]. Deadrise angle and hull form are
very important to generate the lift. Multi-chine hull
form may help the craft that to be stable. Because the
streamline is parallel to the longitudinal direction and
causes that the hull to be straight on the surface of the
water [2-3].
Up to date, some researchers carried out on the
hydrodynamics problem of the planing craft. Most of
them are experimental work. Savitsky (1964, 2007)
did very extensive regression method and provided
many practical formulae for determining the drag of
the craft. His method is still useful for calculations of
the drag of prismatic hull form. From the numerical
view point, some people carried out like Savander et
al, Wang et al and Ghassemi et. al. [5-6-7].
In this paper, BEM and practical modified
Savitskys formulae employed to calculate the
hydrodynamic lift, drag, pressure distribution and free
surface pattern. The effect of the deadrise and center
of gravity are presented.
hydrodynamic calculations for planing craft and

drawing 3D plan of vessels in Auto CAD.
One of the most important points in this software is
receiving infinity table of offset with unequal spaces
for sections and water plans. Mean while the table of
offset can have dimension or non-dimension. Also
user can select wanted unit (English or Metric) in
each modules for informations.
2.1 Hydrostatic calculations

This part of software can calculate 20 hydrostatic
parameters including: SWP ,V ,M ,IXX ,IYY ,XC ,CB ,CM
,CW ,CVP ,CHP ,XB ,ZB ,BMT ,BML ,GML ,GMT MUT
,TPC ,GZ and Bonjean curves. User can include the
chine in calculations and also to determine the shape
of vessels bottom (flat, concave, convex). The
method of integration in this part is Gauss (10 points)
and for comparing curves used spline method.
2. SOFTWARE
This software is including of different modules
such as hydrostatic calculations for merchant vessels,
hydrostatic
calculations
for
planing
craft,
hydrodynamic calculations for merchant vessels,
181
1/ 2
(u )du = R (u ) + .... + R
1
1 / 2
R1 = R10 = 0.03333567215
R2 = R9 = 0.07472567458
R3 = R8 = 0.1095431813
R4 = R7 = 0.1346333597
R5 = R6 = 0.1477621124
u10 = u1 = 0.4869532643
u9 = u2 = 0.4325316833
u8 = u3 = 0.3397047841
u7 = u4 =0.2166976971
u6 = u5 =0.0744371695
10
(u10 )
(1)
2.2. Hydrodynamic calculations

For calculations of this part, Savitskys method is
used and the results are shown from Fn = 2.0 to user
definition. These results are including trim changes,
wet length of keel, wet length of chine, length of
stagnation line, wet height of transom, spray area,
total resistance, effective power and relation of total
resistance to weight.
3. Extracting results and Analysis

A prismatic planing craft which its dimensions have
been shown in the table (1) is selected for our
calculations. In this paper first author has used
boundary element method and second author used
developed Savitskys method and Bowles method for
spray area and spray resistance [2].
Figures (1), (2) show body plan and 3D plan of
vessel. Figures (3), (4) show BMT, GMT in different
water lines and GZ curve in different conditions of
center of gravity. These calculations show that vessel
at Fn= 0 has suitable stability. Figure (5) shows
pressure and lift forces which calculated by boundary
element. Pressure distribution in stem area is more
than other area of vessel and pressure distribution in
stern area is near the atmosphere pressure because
separation phenomena is occurred in transom. Figure
(6) shows current of wake and free surface of sea at

two speeds: 10, 20 m/s. Created wave and current
profile in free surface at behind of vessel show that
waves are transverse and convex.
Table (3) shows hydrodynamic calculations of 5
types of planing craft HSC5, HSC4, HSC3, HSC2 and
HSC1 from software. In figure (7) 3 crafts HSC1,
HSC2, HSC3 which are different in dead rise angle
have been compared. By increasing dead rise angle
the vessels draft and total resistance are increased but
slamming is decreased. Averagely total resistance in
HSC2 about 5.5% and in HSC3 about 11%, effective
power and trim in HSC2 about 6% and in HSC3 about
12%, spray area in HSC2 about 10% and in HSC3
about 20% are decreased.
In figure (8) 3 crafts HSC1, HSC4, HSC5 which
are different in longitudinal center of flotation have
been compared. If LCF is near to stem the trim
decreases and wet surfaces will be increased.
Averagely total resistance in HSC4 about 5% is
decreased and there is no change in HSC5, trim in
HSC4 about 35% and in HSC5 about 57% is
decreased, effective power in HSC4 about 2% and in
HSC5 about 14%, spray area in HSC4 about 27% and
in HSC5 about 73% are increased.
Table 1 Characteristics of main vessel
HSC1
L(m)
17
B(m)
4
T(m)
0.8
LCB(m)
3.7
(D)
22
(ton)
25
(D)
18
14
22
22
(ton)
25
25
25
25
Table 2. Characteristics of secondary vessels
HSC2
HSC3
HSC4
HSC5
L(m)
17
17
17
17
B(m)
4
4
4
4
T(m)
0.8
0.8
0.8
0.8
LCB(m)
3.7
3.7
5.7
7.7
Fig. 1 Body Plan
Fig. 2 Three dimensional view of planing craft
182
KG=0.5
KG=1
KG=1.75
KG=2
KG=1.5
BM T , GMT
1.5
GZ
1
0.5
0
0
20
40
60
80
6
5
4
BMT
GMT
2
1
0
100
-0.5
Heel Angle (Phi)
Fig. 4 GZ curves
Aft
0
0.1
0.2
0.3
0.4
0.5
0.6
Fig. 3 BMT, GMT in different drafts
Bow
x/L
WL
0.7
Stern
0.8
0.9
0
0
-0.04
0.002
-C p
CLd, CLs
-0.02
-0.06
-0.08
Bow
x/L
-0.002
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.004
0.006
-0.1
Fn =3.35
0.008
Volume Froude No=3.35

-0.12
Strip-1 (adjacent to Keel)
Strip-4
Strip-9 (near to chinel)
Buoyant Coef. (CLs)

Lift Coef. (CLd)
0.01
Fig. 5 Pressure distribution and lift force on bottom (Fn = 3.35)
0.140551
0.123635
0.106718
0.0898017
0.0728853
0.0559689
0.0390525
0.0221361
0.00521964
-0.0116968
-0.0286132
-0.0455296
-0.062446
-0.0793624
-0.0962788
0.5
0
0.5
1
1.5
2
2.5
Speed: 20 m/s
3
3.5
-1
-0.5
0.5
Speed: 10 m/s
Fig. 6 Simulation of created wake and wave in sea water
Table 3. Results of hydrodynamic calculations for 5 type of vessels from software
183
V(kn)
20.817
31.226
41.635
52.044
62.452
FnV
2
3
4
5
6
Trim
(D)
13.925
8.7712
5.9976
4.4472
3.4987
EHP
(hp)
918.388
969.682
1118.77
1493.2
2166.22
Spray
Area (m2)
4.3170442
4.889846
6.1620329
7.7859871
9.5873042
RT (N)
63948
45013
38950
41589
50278
V(kn)
20.817
31.226
41.635
52.044
62.452
FnV
2
3
4
5
6
Trim
(D)
13.357
8.2744
5.5693
4.0676
3.1538
EHP
(hp)
881.933
921.146
1051.98
1399.63
2035.28
Spray
Area (m2)
4.2459188
4.4695831
5.4523301
6.8398895
8.4548514
RT (N)
61409
42760
36625
38983
47239
V(kn)
20.817
31.226
41.635
52.044
62.452
FnV
2
3
4
5
6
Trim
(D)
12.816
7.8096
5.1751
3.7235
2.8454
EHP
(hp)
846.933
874.963
990.094
1308.87
1901.6
Spray
Area (m2)
4.392318
4.1665248
4.7885636
5.8724748
7.2281124
RT (N)
58972
40616
34470
36455
44136
V(kn)
20.817
31.226
41.635
52.044
62.452
FnV
2
3
4
5
6
Trim
(D)
6.6694
5.8649
4.4295
3.4223
2.7496
EHP
(hp)
499.269
786.419
1131.77
1733.01
2711.78
Spray
Area (m2)
5.7254576
6.2627038
7.8119899
9.7792784
11.955095
RT (N)
34764
36506
39403
48268
62941
V(kn)
20.817
31.226
41.635
52.044
62.452
FnV
2
3
4
5
6
Trim
(D)
2.9741
3.6476
3.2635
2.7003
2.2436
EHP
(hp)
319.045
710.124
1233.33
2052.55
3323.23
Spray
Area (m2)
11.114744
9.2379074
10.208465
12.159307
14.490149
RT (N)
22215
32964
42939
57168
77133
184
Project
Name
HSC1
Project
Name
HSC2
Project
Name
HSC3
Project
Name
HSC4
Project
Name
HSC5
HSC1
Trim (D)
15
HSC2
HSC3
10
5
0
2
2.5
3.5
4.5
5.5
Fn V
HSC1
EHP (hp)
2500
HSC2
HSC3
2000
1500
1000
500
0
2
2.5
3.5
4.5
5.5
Fn V
HSC1
Rt (N)
80000
HSC2
HSC3
60000
40000
20000
0
2
2.5
3.5
4.5
5.5
Spray Area (m^2)
Fn V
HSC1
12
10
8
6
4
2
0
2
2.5
3.5
4.5
HSC2
HSC3
5.5
Fn V
Fig. 7 Comparison of trim, effective power, total resistance, spray area for HSC1, HSC2, HSC3
185
HSC1
Trim (D)
15
HSC4
HSC5
10
5
0
2
2.5
3.5
4.5
5.5
Fn V
HSC1
EHP (hp)
4000
HSC4
HSC5
3000
2000
1000
0
2
2.5
3.5
4.5
5.5
Fn V
HSC1
100000
HSC4
HSC5
Rt (N)
80000
60000
40000
20000
0
2
2.5
3.5
4.5
5.5
Spray Area (m^2)
Fn V
HSC1
20
HSC4
HSC5
15
10
5
0
2
2.5
3.5
4.5
5.5
Fn V
Fig. 8 Comparison of trim, effective power, total resistance, spray area for HSC1, HSC4, HSC5
186
4. Conclusion
References:
In this paper, hydromechanics calculations for

prismatic planing craft have been done by both
boundary element method and Savitskys method. It is
concluded that:
1. Prepared software has different modules for
hydrostatic and hydrodynamic calculations of planing
craft and merchant vessels. In part of hydrostatic
calculations, the Gauss 100 points for integrations is
used, in part of hydrodynamic calculations, Savitskys
method is used and the effects of dead rise angle and
center of gravity has been calculated. Also the lift and
drag forces in different conditions are calculated.
2. Using boundary element method, the pressure
distribution, lift, drag, wake and free surface wave
calculated. The results show that boundary element
numerical method is powerful tool for calculation the
planing crafts.
187
[1]. Faltinsen, O. M. , 2005. Hydrodynamics of

high-speed marine vehicles, Cambrige University
Press.
[2]. Savitsky, D. Hydrodynamic Design of Planing
Hulls. Marine Technology, Vol. 1 , 1, 1964, 71-95.
[3]. Savitsky, D. , Delorme, M. F. and Datla, R. ,
2007. Inclusion of whisker spray drag in
performance prediction method for highspeed
planing hulls, Mar. Technol. , 44(1) : 35 ~ 56.
[4]. Clement, E. P. and Blount, D. L., 1963.
Resistance tests of systematic series of planing hull
forms, Trans. SNAME, 71, 491 ~ 579
[5]. Savender, B. R. , Scorpio, S. M. and Taylor, R.
K. , 2002. Steady Hydrodynamic of planing
surface, J. Ship Res. ,46(4) : 248 ~ 279.
[6]. Wang, D. P. and Rispin, P. , 1971. Threedimensional planing at high Froude number, J. Ship
Res. , 15(3) : 221 ~ 230.
[7]. Ghassemi, H. and Ghiasi, M. , 2008. A
combined method for the hydrodynamic analysis
characteristics of the planing crafts, Ocean Eng . ,
3(3-54): 310 ~ 322.

MAX-PLUS APPROACH TO SCHEDULING PROBLEMS OF BLOCK ASSEMBLY

LINES
Hiroyuki Kajiwara1, Yasuhiro Hitoi2, and Khairul Hassan3
1
Department of Marine Systems Engineering,
Research Institute of Engineering,
Kyushu University, Japan
E-mail: kajiwara@nams.kyushu-u.ac.jp
2
Oshima Shipbuilding Co. Ltd.,
Japan.
E-mail: y-hitoi@ma.osy.co.jp
3
Department of Maritime Engineering,
Graduate School of Engineering,
Kyushu University , Japan
E-mail: hassan@system.nams.kyushu-u.ac.jp
ABSTRACT
The paper shows pull-type scheduling plans of block assembly lines in a shipyard based on max-plus
approach. The shipyard is assumed to have assembly lines for large-scale/medium-scale blocks as conveyer
lines and also have stockyards for storing blocks. The dynamics of assembly lines and stockyards can be
mathematically modelled in the form of linear system representations using max-plus algebra. This makes it
possible to obtain pull-type scheduling plans by solving model predictive equations. Compared with
conventional scheduling plans, max-plus scheduling plans indicate that the capacity overflow in stockyards
can be avoided and also the workplace limitation on conveyer lines can be observed strictly under the
consideration on assembly order of blocks.
Key words: shipyard scheduling, max-plus algebra, block assembly, conveyer line, stockyard.
1. INTRODUCTION
In recent years, the global flow of goods into
BRIC (Brazil, Russia, India, and China) countries
has been accelerated. There are strong demands
from ship owners on bulk carriers to transport
unpackaged bulk cargo such as cereals, coal, ore,
and cement. In Japan, several shipbuilding
companies are trying to improve the productivity of
bulk carriers. In particular, Oshima Shipbuilding is
a shipbuilding company that specializes in the
fabrication of bulk carriers, especially dry cargo
carriers. An air photo of the shipyard is shown in
Photo.1. Although the company has been delivering
annually around 30 ships, the more productivity is
expected. For the purpose, the more elaborate or
systematic scheduling approach is required, that is,
the assembly start times of several hundred blocks
in each production cycle must be determined to
meet the erection dates just in time and not to cause
the overflow in stockyards. It is not so easy to solve
such a scheduling problem using commercial
software packages.
how to change the push type to the pull type. In

particular, in order to solve the actual problem, it is
crucial to develop functions of conveyer model and
stock model with functions of serial and parallel
connections using Max-Plus Toolbox on
MATLABTM [4] or VBA in EXCEL.
The paper follows up the same research direction
to demonstrate that max-plus scheduling plans of 28
ships produced in 2009 to 2010 are effectively
obtained. In the chapter 2, some prerequisites are
given. Pull-type scheduling results for assembly
lines for large-scale and medium-scale blocks are
shown in chapter 3 and chapter 4 respectively.
In our previous papers[1],[2], we have shown that

shipbuilding lines consisting of block assembly
conveyers and block stockyards can be
mathematically
modelled
as
state-space
representations using max-plus algebra[3], which
are scheduled based on model predictive control
theory. The scheduling problem is concerned with
Photo. 1: Oshima Shipbuilding in Japan
189
2.2 Convveyer Model [2]
2. PRER
REQUISITES
S
2.1 Max-P
Plus Algebra
a Based Lineear System
Represen
ntation [1]
A convveyer line is useful to reallize CONWIP

P
(constant works in proccess), in which
h workplacess
are movedd with a constaant pitch . Figg.3 shows suchh
a convey
yer mechanissm, which iss successfullyy
modelled as a linear ssystem represeentation in thee
followingg.
Figure 1. Production System

m
All Connsider a produ
uction system for the -th
product ass shown in Fig
g.1, where
is the start
time,
is the processs time,
i the arrival
is
time of paarts/ materials,, and
is th
he stop time.
Then, the following
f
relation holds:
1 ,
max
Figuree 3. Schematicc Diagram of Conveyer

C
Line
For the k-th product, let n be the nuumber of partss
which aree fabricated in the conveyer line consistingg
of m worrk-places. Assuume that the last
l
parts forr
the (k-1))-th product occupy the beginning
workplacees, and also that the last r paarts for the k-thh
product occupy
o
the beeginning r woorkplaces. Wee
consider the
t following ffour cases.
The firsst equation sh

hows that the start time is
determinedd as the last sttop time or thee arrival time
of parts, whichever
w
is lateer. The second
d equation is a
trivial relaation between the start time and the stop
time. Repplacing the max
m operation and the plus
operation in
i (1) by andd
respectiveely brings
1
Letting
1
1 ,
and omittinng , (2) is wrritten as
1
,
.
This is called as a liinear system representation

r
m
algeb
bra. The first eq
quation is the
based on max-plus
state equattion and the seecond equationn is the output
equation. In general, , ,
are vectors of
appropriatee dimensions and , , aree matrices of
compatiblee dimensions.
The papper[2] derives the results sho

own in Fig.4, 5
and 6 corrresponding to C
Case 1, 3 and 4 respectively.
Figure 4. Max-plus Moodel of Conveyyer Line (Case1)
Figure 2. Productionn System of 2-P

Port Type
Considerr the system with
w 2 kinds of
o inputs and
outputs as shown in Fig.2. In addition
n to
,
as defined in the above,
is the avaailable time of
a machine,,
is the relleased time of a machine.
1
Figure 5. Max-plus Moodel of Conveyyer Line (Case33)
1
.
It is a keey to distinguissh the productioon cycle from

the machinne utilization cycle in orderr to take into
account of resource coonstraints on capacities or
workplacess.
Figure 6. Max-plus Moddel of Conveyeer Line(Case4))
190
another pair of Ship C/ Ship D are constructed

alternately. In order to manage such a dock cycle, it
is necessary to prepare all the blocks until the
erections begin. Because of the short period of
erection, the shipyard has Stockyard #1/ Stockyard
#2 to store blocks of two ships, that is, 120 blocks.
Also it has 8 block assembly lines as conveyers. The
related parameters are shown in Table1. Each line is
able to fabricate blocks with a constant pitch.
2.3 Scheduling by Solving Model Predictive

Equation[5]
Assume that a linear system representation of
production system to be scheduled is obtained as
(3). From the state equation, the following relation
is derived.
1
1
1
1
:
Using the output equation, (5) becomes
1 ,
:
where in the case of 1-step-ahead prediction,
in the case of 2-step-ahead prediction,

1
.
1
Figure 7. Block Assembly Lines and Stockyards
1
8
Table1. Parameters on Block Assembly Lines
Now let the delivery times be

1 .
It is required to determine the arrival times of parts

satisfying
1 ,
.
m'
line#
m'
7/8
21
6/4
10
7/8
5/9
22
7/8
5/7
15
10
7/10
34
4/4
10
m: number of stock places assigned to conveyer
11
12
which does not have the solution

in general.
Therefore, it is required to obtain the maximal
of (12) satisfying
solution
c: pitch of conveyer
(10) becomes
m: number of workplaces on conveyer
10
In the case that the following condition holds,
line#
Fig.8 shows the actual numbers of large-scale

blocks to be fabricated in 2009 to 2010. There are 7
cycles in which each cycle has around 220 blocks
for four ships, that is, around 60 blocks for one ship.
13
In the case that (11) does not hold, it is impossible to

obtain the appropriate
. Then it is necessary to
change (6) itself by reducing the processing times.
3. SCHEDULING OF LARGE-SCALE
BLOCK ASSEMBLY LINES
The half of the shipyard layout considered in the
paper is depicted in Fig.7. There is a dock which can
include four ships. One pair of Ship A/ Ship B and
Figure 8. Numbers of Blocks Fabricated in 2009 to

2010
191
For example, the conventional scheduling plan for

Ship C in the Cycle 2 is shown in Fig.9. As for the
other Ship A,B,D in the other Cycle 1,3,4,5,6,7, we
have the same figures. The red bars indicate the
election dates for Ship C. Also the yellow bars, the
blue bars and the green bars indicate pre-election
periods, stock periods and assembly periods
respectively. We can reveal the idling days by
taking minimum stock periods which is necessary
for pre-fitting and painting as in Fig.10. Our
scheduling problem is to reduce the idling days as
many as possible such that the facility constraints on
conveyers and stockyards are satisfied as in Fig.11.
The scheduling plan for Line #1 is made as

follows. So are the other lines. At first, consider all
the blocks for Ships A,B,C,D fabricated in Line #1
in Fig.12-LHS (Left-Hand Side). Then sort them
according to the dates given by subtracting the
minimum stock periods from the pre-erection dates
as shown in Fig.13-LHS. Then make the sorted
scheduling plan based on the max-plus method in
Sec.2.3 as in Fig.13-RHS in which the equation (12)
with the parameters (7) of 1-step-ahead prediction is
solved. Finally obtain the max-plus scheduling plan
in Fig.12-RHS by sorting reversely. Note that the
idling days are reduced from 58 to 14.8.
Figure 9. Conventional Scheduling for Ship C
Figure 12. Scheduling for Line#1 in Cycle 2
Figure 10. Revealing Idling Days in Fig.9
Figure 13. Sorted Scheduling for Line#1 in Cycle 2
Figure 14. Counting Blocks Stored in Stockyards
Figure 11. Max-plus Scheduling for Ship C
192
In Table 2, note that Line #13 and #15 are of

feedback type. This means that some sub-blocks are
utilized to fabricate the other sub-blocks.
Specifically, sub-blocks in Line #13 and #15 are fed
back to Line #15. The scheduling plans for these
lines are made as follows. At first, they are planned
for sub-blocks without feedback as shown in Fig.17
and Fig.18, in which the target dates are taken as the
start dates of the corresponding large-scale blocks.
From these results, the target dates for sub-blocks
fed back in Line #13 and #15 are determined. They
must be pushed in scheduling on Line #15 as shown
in Fig.19.
Figure 15. Counting Workplaces Occupied in Line#1
In order to check that the facility constraints on
conveyers and stockyards are satisfied, Fig.14 and
Fig.15 are presented. Fig.14 shows the counting
results of blocks stored in stockyards, and Fig.15
shows the counting results of workplaces occupied
in Line#1. Although there are some violations in the
conventional scheduling, the max-plus scheduling
have no violation.
4. SCHEDULING OF MEDIUM-SCALE
BLOCK ASSEMBLY LINES
The another half of the shipyard layout considered
in the paper is depicted in Fig.16. The large-scale
blocks are made from several medium-scale subblocks which are fabricated in
5 sub-block
assembly lines as conveyers. The related parameters
are shown in Table2.
Figure 17. Sorted Pre-Scheduling for Line#13
Figure 18. Sorted Pre-Scheduling for Line#15
Figure 16. Sub-block Assembly Lines & Stockyards

Table2. Parameters on Sub-block Assembly Lines
line#
m'
11
4/8
10
12
24
7/24
31
13*
10
4/10
13
14
4/9
12
15*
10
4/10
14
Figure 19. Sorted Scheduling for Line#13 and #15
* lines with feedback flows
193
The sorting is done for all sub-blocks fabricated

in Line #13 and #15 according to the dates given by
subtracting the minimum stock periods from their
target dates. Fig 20 shows the conventional
scheduling plans for Line#13 and #15, where the
numbers of workplaces occupied in Line #13 and
#15 are larger than the limitations. Fig.21 shows the
max-plus scheduling plans for Line#13 and #15,
where the idling days are increased compared with
the conventional scheduling ones, but the counting
results of sub-blocks stored in stockyards are not so
deteriorated as shown in Fig.22.
5. CONCLUSION
The contributions of the research at present are
summarized as follows:
1. pull-type scheduling methodology of block
assembly lines and stockyards based on conveyer/
stock models to satisfy facility constraints and the
slack dispatching rule to determine the order of blocks
fabricated.
2. pull-type scheduling methodology of sub-block
assembly lines with the feedback flow, that is, some
sub-blocks are utilized to fabricate the other subblocks
The further research will be concerned with a
multi-cycle scheduling problem which is necessary
for discussing a overlapping problem between two
cycles, and workload balancing problems[6]
REFERENCES
[1] Kajiwara, H., Nakao, Y.: An approach to
scheduling problems of shipbuilding lines based on
discrete-event system theory, Proc. of ICCAS2007
Figure 20. Conventional Scheduling for Line#13, #15
[2] Kajiwara, H., Hitoi,Y., Nakao, Y.: Max-plusalgebra based scheduling of a ship building line, Proc.
of ICCAS2009
[3] Heidergott, B., Olsder, G.J., Woude, J.: Max Plus
at Work: Modeling and Analysis of Synchronized
Systems: A Course on Max-Plus Algebra and Its
Applications, Princeton University Press, 2005
[4] Stanczyk, J.: Max-Plus Algebra Toolbox for
MATLAB and GNU Octarve , 2005
[5] Goto, H., Masuda, S.: Online Model Predictive
Control for Max-plus Linear Systems with Selective
Parameters, IEICE Trans. on Fundamentals of
Electronics Communications and Computer Sciences
E87A (11) pp.2944-2949, 2004
Figure 21. Max-plus Scheduling for Line#13 and #15
[6] Hassan, K., Kajiwara, H.: On max/min-plus

algebra approach for shipyard scheduling problems,
Proc. of JASNOE2010
Fig.22. Counting Sub-Blocks Stored in Stockyard
194

OPTIMIZATION OF THE PRELIMINARY DESIGN PARAMETERS AND COST

EVALUATION OF CONTAINER VESSEL FOR THE INLAND WATERWAYS OF
BANGLADESH.
Goutam Kumar Saha1, Golam Mortuja2 and Md. Emdadol Haque3
1
Associate Professor
Bangladesh university of Engineering and Technology
E-mail: goutamkumar@name.buet.ac.bd
2,3
Undergraduate Student
Bangladesh university of Engineering and Technology
E-mail: mo.navalarch@gmail.com
ABSTRACT
Demand of Containership for the inland waterways system in Bangladesh augmented on a large scale which
forced to construct new containerships. The selection of the best design parameters and perform the necessary
economic evaluations to ensure that the ship will make profit, is a great task for a Naval Architect. This paper
represented an integrated methodology for the preliminary optimum design parameters considering the
existing constraints provided by the Bangladesh Inland Water Transport Authority. The work is tackled from
the ship owners point of view. The required freight rate is chosen as the objective function because the most
important thing that concerns the ship owner is whether the ship will make a profit or not, and if so, how much
profit it can make. It is very important for the users to formulate carefully the optimization problems so that it
gives a stable and reasonable solution.
Keywords: containership, inland waterways, optimization, required freight rate, Bangladesh.
Wm
Lut
NOMENCLATURE
RTotal
RF
RAPP
RW
RB
RTR
RA
L
B
T
Cp
TEUb
TEUd
Sb
Sd
Lb
Ld
Db
TNd
Wh
Woh
Total resistance of a ship

Frictional Resistance
Resistance of appendages
Wave-making and wave breaking resistance
Additional pressure resistance of bulbous bow
nears the water surface
Additional pressure resistance of immersed
transom stern
Model-ship correlation resistance
Length on the waterline
Molded Breadth on the waterline
Draft of hull
Prismatic coefficient
Container number below deck
Container number above deck
Storage factors for containers below deck
Storage factors for containers above deck
Container number below deck along the
length direction
Container number above deck along the
length direction
Container number below deck along the
depth direction
Tier number for containers above deck
Hull steel weight
Hull and outfit weight
Pwt
St
Machinery weight
Time for loading and unloading containers
per round Trip
Port waiting time per round trip
Time spent at Sea
1. INTRODUCTION
Bangladesh is a maritime nation with 166,000 sq.km
area of sea, abundant with living and no living
resources. Bangladesh also has a considerable
domestic merchant shipping fleet for the
transportation of passengers, food grains, machineries
and oil products etc. Shipbuilding industries are
essential to the requirement to support this fleet.
The basic ship design aims to determine major ship
characteristics affecting cost and performance. Basic
design includes the selection of ship dimensions, hull
form, and power, preliminary arrangement of hull,
machinery arrangement, and major structure. During
the preliminary Ship Design phase, important
decisions with significant impact on the vessels
performance have to be made by the designer, usually
based on very limited information [4]. In such cases,
the design has to rely on his experience and
engineering judgment, occasionally supported by the
exploration of relevant data from past designs.
195
Optimization methodology, facilitating the fast

exploration of a series of design alternatives would be
of great assistance to designer, in search of the
optimum solution subject to specific owners
requirements [8]. Proper selections assure the
attainment of the mission requirements such as good
seakeeping performance, maneuverability, desired
speed, endurance, cargo capacity, and deadweight.
Furthermore, it includes checks and modifications of
the achievement of required cargo handling
capability, quarters, subdivision and stability
standards, and freeboard and tonnage measurements,
usually defined as constraints. The derivation of a
feasible technical design will take the form of an
iterative process of analysis and synthesis; i.e. is a
repetitive process whereby the design is resolved into
simple elements and relevant calculations made, after
which the elements are combined into the total ship
design [2]. The overall flow path shown in Figure 1 can
thus be established. The preliminary design process will
normally take the form of a techno-economic appraisal,
using a fundamental engineering economy approach.
Operating Environments and Owner

Requirements
Feasible Technical Design
Kj(X) 0
XiL
Xi
j = 1, NCON
XiU
i = 1,NDV
Inequality Constraints
Side Constraints
Where X is the vector of design variables, NCON is

the number of constraints, NDV is the number of
design variables, XiL and XiU are the lower and
upper bounds on the design variables. To solve and
optimization problem, it is usually started with an
initial design point, X0. Then, the next design point
is determined by:
Xi+1 = Xi + .Si
(2)
Where I is the iteration number, is the step length

and S is the searching direction.
3. PERFORMANCE EVALUATION
MODULE
The objective for the performance evaluation
module is to calculate the technical and economical
performance of the container ship for inland
waterways of Bangladesh. Holtrop and Mennens
method is used for the resistance estimation.
Container number calculated for a ship was also
calculated. Some empirical formulae are used for
weight and cost calculation.
3.1 Resistance Calculation

Holtrop and Mennens method [5] is used to
predict the resistance of the ship. The total resistance
of a ship is divided into components as follows:
Estimates of Building and Operating

Costs and Revenues
Economic Evaluation of Alternatives
Choice
Detailed Design
Construction
Figure1: Overall flow path.
2. GENERAL FORMULATION OF
OPTIMIZATION PROBLEM
RTotal = RF (1+K1) + RAPP + RW+ RB+ RTR+ RA (3)

As the containership is for the inland route of
Bangladesh from Dhaka to Chittagong, bulbous bow
was not provided to the ship. Here the frictional
resistance, wave-making and wave breaking
resistance and model-ship correlation resistance was
calculated. Other components of total resistance
such as appendage resistance (RAPP), resistance due
to bulbous bow (RB), transom resistance (RTR) is
ignored in the preliminary design calculation since
they are very small percentage of total value. The
ship is within the range of the condition for using
Holtrop and Mennens formulae. Condition is as
follows [7]:
The target of optimization is the objective

function or criterion of the optimization. It is subject
to boundary conditions or constraints. Constraints
may be formulated as equations or inequalities. All
technical and economical relationships to be
considered in the optimization model must be
known and expressed as functions. The optimization
problem is formally stated as follows [9]:
0.55 < Cp < 0.85

3.9 < L/B < 15
2.1 < B/T < 4.0
Holtrops formulation is based on a statistical
analysis of resistance data. All the component of the
resistance was calculated for fixed speed of the
vessel 11 knots. Effective Power (EHP) and Shaft
Power (SHP) are also determined using empirical
formulas.
Minimize or maximize:
3.2 Container Number Calculation
F(X)
Subject to:
Objective Function
(1)
Container ships may be classified as capacity or

space determined designs and their size are generally
defined by their container capacity [2] e.g. 1500,
3000, 6000 or 10000 TEU. (Container sizes
196
20 8 8.5 height or 40 8 8.5 ). TEU =

Twenty foot Equivalent Unit (e.g. 40 container = 2
TEUs). The number of Container carried by a vessel
was calculated by the following expression [9]:
TEU = TEUb + TEUd
(5)
and Sea time was calculated to estimate the number of

trips that can be made by a ship throughout the year
considering 15 off-hire days. It can be formulated as
follows:
TEUb = Sb Lb Bb Db
(6)
Therefore the number of trips per year is:
TEUd = Sd Ld Bd TNd
(7)
Rtt = Lut + Pwt + St
NT =
Where TEU is the total number of containers

(Twenty-foot Equivalent Unit). The actual number of
Container that could be carried by the ship is the
integer number.
(10)
(11)
Where, Ot = on-hire time per year in days. In this

work it is defined as 350 days.
3.3 Weight Estimation

Weight estimation can be done based on empirical
regressions for structure, machinery, outfit and
deadweight items other than cargo. The Naval
Architect will always attempt to make the lightweight
as low as possible without endangering the safety and
strength of the new vessel [3]. The total weight of the
ship is divided into components as follows:
Wtotal = Wcon + Wlight + Wfuel + Wmisc
(8)
In this work, each TEU is considered as 14 TONNE.

The lightship weight is further divided into subcomponents as follows:
Wlight = Wh + Woh + Wm
(9)
Hull structure includes the main hull structure,

superstructure, deck houses and all internal divisional
bulkheads over one eight inch thick. Outfit constitutes
hull insulation, joiner bulkheads, hawse pipes, deck
fittings, cargo booms, hatch covers, anchors, rudder
and stock and gallery equipment. The propulsion
machinery weight is determined by the shaft
horsepower of the main engine. Fuel weight is
determined by the shaft horse power SHP, the service
range, the speed of the ship, and specific fuel
consumption of the engine with an additional tenpercent allowance.
3.4 Round Trip Time

The distance between the target ports (Dhaka
Chittagong) is about 307 Km (166 n. miles). Figure 2
shows the route between Dhaka and Chittagong with
blue line on the map. On the route between Dhaka and
Chittagong, there is the coastal channel across
sandbars from Sandwip to South Hatia, where the
shallowest part in the channel is deeper than 4.7m.
About 2.0m of wave height is estimated at the coastal
area but during cyclones, 1.5m to 4.4m of wave height
were observed. The Navigational waterways between
Dhaka and Chittagong or Mongla and the depth of
waterways are to be controlled and guaranteed for
vessel navigation by proper dredging throughout the
year so that the vessel of 3.4m draft can operate all the
seasons. Loading unloading time, Port waiting time
Figure 2: Map of Dhaka- Chittagong Route

(Blue line)
3.5 Cost Estimation

3.5.1 Building Cost
The building cost comprises the cost for
shipbuilding, including machinery, steel hull, hull
engineering and outfit. The general idea for building
cost is to calculate man hours required based on the
weights of the related components and use the manhour estimate to determine the labor cost [9]. The
estimation of labor man-hours necessary for ship
production, as a part of shipbuilding cost, has usually
evolved at two stages of detail. The early stage
provides only a preliminary estimate before any
details of ship design and production processes are
considered. Such preliminary estimates is made
usually using empirical equations based on the ship
weight, size and other general design parameters. A
more detailed man-hours estimate starts after signing
the contract, as the information of the project
increases parallel with detail ship design, so as to
make suitable planning and scheduling for
shipbuilding process [1]. The material cost is
calculated as a function of the weight. All the cost is
taken from the market present value and in terms of
local currency BD Taka.
197
3.5.2 Operating Cost

Operating Cost is the most important part after
building the ship. Wages for the crews, stores and
supplies, Insurance, Maintenance and Repair cost,
Port expenses, Annual Fuel cost results the Annual
operating cost [9]. In this case, Interest rate is defined
as 15% and the ship life is defined as 30 year. Fuel
cost is related to the SFC and the number of Trips that
a ship can make.
4.3 Design variable
3.5.3 Required Freight Rate

The supply of sea transport is influenced by freight
rates. This is the ultimate regulator which the market
uses to motivate decision-makers to adjust capacity in
the short term, and to find ways of reducing their
costs in the long-term. Required Freight rate is the
amount the owner charges the customer to breakeven. A more precise definition can be provided
saying The required freight rate for a given rate of
utilization produces net profits which exactly cover
the operating costs inclusive of calculated interest on
the invested capital; i.e. profit and rate of return are
nil. [9] In this work the required freight rate,
expressed in TAKA per TEU per Trip. Therefore the
required freight rate, RFR, is given by
RFR =
number of container module. The Length of the ship

was given 71m< L < 75m. The inadequate ships
proportion of length/breadth will raise an increase of
resistance for propulsion which means an increase of
horse power of main engine to keep the specified
speed The Breadth of the ship should be within 12.5m
< B <13.5m and the Cb is to be 0.70< Cb <.78. There
was also a draft limit of 3.4 m and the speed of the
vessel is fixed to 11 knots, although the program can
be used for different draft and speed condition.
(12)
4. OPTIMIZATION PROCEDURE
Completing the general formulation for the
optimization, each specific optimization problem was
linked between each other in the Microsoft Excel, and
the SOLVER add-in of Excel has been used to find
out the optimum result. The formulation of the
optimization problem includes three parts: selection of
the design variables, formulation of the objective
function and formulation of the constraints.
4.1 Objective function

The goal of this work is to develop a tool to help
the designer find the best suitable ship design
parameter which will make profit for its owner. The
objective is to minimize the required freight rate with
a fixed range between two ports (Dhaka- Chittagong).
The objective function of this work is formulated as
follows:
There are several principles for choosing the design

variables in an optimization problem. First, a design
variable should have direct influence on the objective
function and constraints. Second, the number of
design variables should be kept as small as possible.
Third, the design variables should be kept as
independent as possible [9]. In this work Length,
Breadth and Block coefficient used as design variable.
5. RESULTS & DISCUSSIONS

The work is done in two ways separately. A
computer program in C++ was developed to give the
maximum number of combination within the range
and export the result in Microsoft Excel. Then the
sorting was done for maximum container number,
minimum resistance etc. A list of best 21 vessels
dimension was obtained for the graphical
presentation. The speed and draft of the vessel was
fixed for the inland waterways system. Speed was 11
knots and draft was 3.4 meter. On the other hand,
whole formulation was also putted into four Microsoft
Excel sheets. Each sheet was inter-connected. In the
last sheet named Optimization, objective function,
Variable and Constraints were defined. With the help
of the SOLVER (Excel Add-in) the optimization
process performed. Quadratic approach is used to
obtain initial estimates of the basic variables in each
one-dimensional search. Quadratic uses quadratic
extrapolation, which can improve the results on highly
nonlinear problems. Table 1 shows the lower and
upper boundary of design constraint used by program
as well as the optimum result found by the SOLVER.
Table1: Optimization result with respond to Limit.
Objective function:
Minimize F(X) =
(13)
where:
RFR0: Required freight rate at the initial design point
RFR
Cb
TEU
(Taka)
(m)
(m)
(m)
Lower
Bound
8502.05
71
12.5
0.70
96
Upper
Bound
8733.48
75
13.5
6.25
0.78
116
Result
8496.84
71
12.5
0.74
98
4.2 Constraints
For the inland waterways for Bangladesh, the
length, Breadth, and Depth are very much limited.
The length to be required first is for container hold
which is somewhat automatically decided by the
Four graph was produced for easy visualization of

the changes occurred. In those entire four graphs it
can be seen that between the length 74 and 75 there
are some trough and crest shape. Each graph also have
198
corresponding Contour plot, which can be used for

determining the values for the intermediate Length,
Breadth, TEU, RFR, SHP and AAC.
8730
8720
8710
8700
8690
8680
8670
8660
8650
8640
8630
8620
8610
8600
8590
8580
8570
8560
8550
8540
8530
8520
8510
Figure 7: Required Freight Rate (BD Taka) Vs Length

(m) & Breadth (m).
Figure 3: Total Container number (TEU) Vs

Length, L (m) & Breadth, B (m).
B readth (m )
13.5
13
12.5
71
71.5
72
72.5
73
73.5
74
74.5
75
Length (m)
Figure 8: Contour Plot of Required Freight Rate (BD

Taka) Vs Length (m) & Breadth (m).[ Contour Plot]
Figure 4: Total Container number (TEU) Vs

Length (m) & Breadth (m). [Contour plot]
Figure 5: Shaft Power (SHP) Vs Length, L (m) &

Breadth, B (m).
Figure 9: Annual Average Cost, AAC (BD Taka) Vs

Length, L (m) & Breadth, B (m).
B readth (m )
13.5
13
12.5
71
71.5
72
72.5
73
73.5
74
74.5
75
Length (m)
Figure 6: Shaft Power (HP) Vs Length (m) & Breadth

(m). [Contour plot]
Figure 10: Annual Average Cost (BD Taka in million)

Vs Length (m) & Breadth (m).[Contour plot]
199
Figure 3 and 4: establishes relationship among Total

Container number, Length (m), and Breadth (m).It can
be seen from the Figures that the number TEU
increased with increasing length and breadth.
Between the length 74 and 75 there was sudden
trough and crest indicating the changes in the TEU.
From the corresponding Contour plot, Total Container
number can be found for a ship having Length
between 71 to 75 meter and breadth between 12.5 and
13.5 meter. In Figure 5 and 6 shows a relationship
among the Shaft Horsepower, Length and Breadth are
established. A dramatic behavior in the Figure can be
seen for the length between 74 and 75 meter.
Maximum shaft horsepower is needed when the
length and breadth is maximum. From the
corresponding contour plot, values of probable SHP
for the intermediate length and breadth can be
estimated. In Figure 7 and 8 displays the most
important relationship, RFR, Length and Breadth is
established. RFR is calculated in terms of BD TAKA.
From the Figure it can be observed the up and down
of RFR for different values of length and breadth.
With different length and breadth, optimum or desired
RFR can be found from the corresponding contour
plot. Final two Figure 9 and 10 shows the relationship
among the Annual Average Cost (BD Taka), Length
and Breadth of the ship. AAC increased for the
maximum length and breadth but optimum point can
be found from the contour plot of the corresponding
relationships. Those Figures give the designer the
preliminary idea of the important design parameters
before going to the drawing table.
6. CONCLUSION
An integrated methodology for the basic
preliminary design evaluation and optimization of
Containership for inland waterways of Bangladesh
has been presented. The reverine condition of
Bangladesh has limited the option of thinking about
large vessel. The main difficulty in most optimization
problems does not lie in the mathematics or methods
involved, i.e. whether a certain algorithm is more
efficient or robust than others. The main difficulty lies
in formulating the objective and all the constraints [6].
The optimization problem has been carefully
formulated to give a stable solution. The design
variables are chosen after careful inspection of every
module. The constraints are based on the performance
requirements. The objective function is normalized
with respect to its initial value. For the cost evaluation

process, empirical formulas along with the practical
data were used. For the detailed cost evaluation,
requires a separate work. From the result of the
optimization we have conclusions as follows:
We have only one global minimum point
within our design space.
The optimum container ship tends to be as
big as possible.
ACKNOWLEDGMENTS
organizers for providing the opportunity of
presenting the research work.
REFERENCES
[1] Ahmad M. Rashwan, Estimation of Ship
production Man-hour, Alexandria Engineering
Journal, Vol. 44, No. 4, July 2005.
[2] Anthony
F.Molland,
THE
MARITIME
ENGINEERING REFERENCE BOOK A
Guide to Ship Design, Construction and
Operation, Butterworths, 2008.
[3] Dr C.B. Barrass Ship design and performance
for master and mates, , Butterworths, 2004
[4] Hernani L. Brinati, Oscar B. Augusto, Mardel B.
de Conti, Learning Aspects of Procedures for
Ship Conceptual Design Based on First
Principles, International Conference on
Engineering Education ICEE 2007
[5] Holtrop, J. (1984), A Statistical Re-analysis of
Resistance and Propulsion Data, International
Shipbuilding Progress, Vol. 31. (P 272- 276)
[6] H.Schneekluth, Ship Design for efficiency and
economy, Butterworths, 1987
[7] Hullspeed User Manual, version 13, Maxsurf
package.
[8] S.Skoupas, G.Zaraphonitis, An optimization
Procedure for the preliminary design of highspeed RORO-Passenger Ships, National
Technical University of Athens, Greece.
[9] Ying Chen, Formulation of a Multi-Disciplinary
Design Optimization of Containerships, Virginia
Polytechnic Institute and State University, 1999.
200

PASSING SHIP EFFECTS ON A MOORED SHIP: A NUMERICAL STUDY

Md. Nazrul Islam1, M. Rafiqul Islam2 and Md. Sadiqul Baree3
1
Bangladesh Shilpa Bank
Dhaka, Bangladesh.
E-mail: nazrul.islam.md@gmail.com
2,3
Professor
Bangladesh University of Engineering and Technology
Dhaka - 1000, Bangladesh.
E-mail: rafiqis@gmail.com
ABSTRACT
Sailing a ship generates wave. Not only this generated wave reaches the moored ship nearby but also the
presence of sailing ship changes the flow around moored ship and cause pressure gradient along moored ship.
The reached wave from sailing ship plus the pressure distribution along the hull of moored ship significantly
causes forces on moored ship. Mentioned forces results motion of moored ship and influences loading /
discharging operation and also causes to damage the mooring system. These hydrodynamic forces are low
frequency and sometimes known as suction forces. These forces and resulting moments depend on hull size,
speed of the passing ship, lateral distances between the ships, water depth, and incident wave direction, period
and height. In the present paper, by using linear wave theory the mathematical model of dynamic interaction
between two ships is derived. The mathematical model is solved by using 3D source distribution panel method.
A general overview and findings of some numerically computed forces and motions of moored ship is presented..
Keywords: Linear Wave Theory, 3D source distribution technique, Green Function.
1. INTRODUCTION
Moored or Floating structures are subjected to
hydrodynamic forces and motions due to other ships
sailing nearby. These hydrodynamic actions may
results to high forces and unacceptable motions, and
these hinder the loading and discharging operations
and also cause damage to the hull and mooring
systems. The problems between two moving ships in
waves studied by Gung-Rong Chen and Ming-Chung
Fang (2000) [1] using Three-dimensional potentialflow theory based on the source distribution
technique. The numerical solution was compared with
experimental results and strip theory. They did not
investigate the effect of moving ship on moored ship.
J. A. Pinkster and P. Naaijen (2003) [2] studied the
effects of sailing ships taking into account free surface
effects for both slow and fast moving ships. In case of
large and slow moving ships, free surface effects are
due to the surrounding harbor geometry. In case of
fast moving ships, free surface effects are due to the
wash waves generated by the sailing ship that
propagate into the harbor creating more or less
complicated incoming wave system for the moored
ship. The hydrodynamic interactions that can
influence Seakeeping was studied by Kevin
McTaggart, David Cumming, C. C. Hsiung and Lin Li
(2003) [3]. The numerical predictions include the
influence of interaction effects on hydrodynamic
forces for two ships in waves. The numerical
predictions and experiments showed that the presence
of a larger ship can significantly influence the motions
of a smaller ship in close proximity. The

hydrodynamic interaction forces/moments acting on a
moored ship due to the passage of another ship in its
proximity is researched by K. S. Varyani and P.
Krishnankutty (2006) [4] by considering the influence
of ship form against the idealized approach of the use
of parabolic sectional area distribution. Moving ships
also create impact on moored ships and coastal
structures. The forces and moments are more
significant on the moored ship than on the moving
ship.
2.1 Coordinate system
Ship is considered as a rigid body having six
degrees of freedom. It is subjected to hydrodynamic
forces due to incident waves and radiated and
diffracted waves due to other ship(s). The right hand
coordinate systems are defined in Figure 1.
201
Figure 1 Mooring arrangement
2.2 Mooring particulars
2.4 Boundary condition
The mooring particulars of the study are furnished in

Table-1.
On the mean wetted surface area of each body S,

the above linear velocity potentials must satisfy the
Laplace equation and also the following boundary
conditions:
The steady motion potential:
Table 1 Mooring particulars

Coordinates on
moored ship
X1
Y1
Z1
-125
-48
10
-125
-48
10
125
-30
10
125
-30
10
Connect
or
Number
1
2
3
4
Stiffnes
s
(KN/m)
1325
1325
1325
1325
Coordinates on Jetty
X2
-155
-110
110
155
Y2
-52
-52
-52
-52
Z2
8
8
8
8
W.n = 0 on S
where
The fluid is assumed to be incompressible, inviscid

and irrotational and the vessel is assumed to be
moored in open water. Then there exists a velocity
potential satisfying Laplace equation together with
boundary conditions on the free surface, on the body,
and at the bottom, and the radiation condition in the
far field. The time dependence of the fluid motion to
be considered here is restricted to simple harmonic
motion and accordingly the flow field can be
characterized by the following velocity potential:
= 0ei t + 7mei t + ie X mj j
e
2.4.1 Linearised free surface condition:

2
[( i e U
) +g
] = 0 at z = 0
x
z
2.4.2 Boundary condition on the sea floor:
= 0 on z = h
z
2.4.3 Body boundary conditions:

Due to linearization, this boundary condition may be
applied on the wetted surface of the floating bodies
in their equilibrium position
0 7m
i
+
= 0 on S and i = 1...N
n
n
jm
m
= i e n mj + U M mj on S
n
(1)
(2)
n =1 j =1
e =
Where,
0 =
ig
2U
g
jm
= 0 on S i (i m).
cos
cosh [k ( z + h ) ] ik ( x cos
e
cosh kh
U = s
2.3 Assumptions
( x , y , z , t ) = U x + s +
W .n = U x
(4)
In which
n mj is the direction cosine on the surface of
the body 'm ' in the j-th mode of motion and has the
following form:
+ y sin )
n1m = cos(n m , x m ) : n2m = cos(n m , y m )
S = Steady velocity potential due to mean
n3m = cos(n m , z m )
Forward speed of the ship
0 = Incident wave potential,
n4m = ( y m yGm )n3m ( z m zGm )n2m
7m = Diffraction wave potential on body
n 5m = ( z m z Gm )n1m ( x m x Gm )n 3m
jm = Potential due to motion of the body in j-th
n 6m = ( x m x Gm )n 2m ( y m y Gm )n1m
M 1m = M
mode
e = Encountering frequency and is defined as:

= Circular frequency of incident wave
M 5m
= Wave heading angle from X -axis

m= number of ships
N= number of panels
The differential equation governing the fluid
motion follows from the application of the continuity
equation, which yields the Laplace equation.
n3m , M 6m
= M
m
4
=0
x Gm , y Gm , z Gm = co-ordinate of the centre of gravity of

the body 'm '
m
x ,
y , z = investigating point on the wetted

surface of the body 'm '
The radiation condition of the potentials j , in

m
which in polar co-ordinate:
The individual potentials are the solutions of the

following Laplace equation:
=0
m
3
n2m
= M
where,
a = Incident wave amplitude,
m
2
(3)
202
lim (r 2 (
2
i ( ) )) = 0
g
r
Where, Fkm denotes the k-th component of wave
2.5 Velocity Potentials

There is no analytical solution for
m
7
and
m
j
exciting forces or moments, Fkjm denotes the k-th
, so the problem should be solved numerically.

According to the 3-D source sink method, the
potentials
7m
and
jm
component of force arising from the j-th component

of motion of the body m. Moreover, it is customary
to decompose the hydrodynamic forces resulting from
motion of the bodies into components in phase with
the acceleration and velocity of the rigid body
motions. These yield the added mass and damping
coefficients respectively. These coefficients can be
expressed from equation as:
can be expressed in terms of
well known Green functions that can be expressed by

the following equation :
mj ( x, y, z ) =
1 N
m
j ( , , )G (x, y, z; , , ) ds
4 n =1 s
where, ( , , ) denotes a point on surface S and

( , , ) denotes the unknown source distribution.
The integral is to be carried out over complete
immersed surface of the object. The Green function
G (source potential) must in order of the
representation in the above equation to be valid,
satisfy all the boundary conditions of the problem
with the exception of the body boundary conditions
and have a source like behavior. As a result,
boundary conditions are reduced only to on wetted
surfaces of the bodies. So, the wetted surfaces should
be subdivided into panels to transform integral
equations to a system of algebraic equations to
determine unknown source density over each panel.
The appropriate Green function used in this paper to
the boundary value problem posed is given by
Wehausen and Laitone [6]. After getting the source
density, the velocity potentials on each panel can be
obtained using the above equation.
a kjmn = . Re s (
b kjmn =
m
j
m
j
). n km . ds
). n km . ds
m
m
m
m m
m
m
( M kj + akj ) X&& j + bkj X& j + CX j = Fk
j =1
(k = 1.....6)
Where,
M kjm = inertia matrix in k - mode due to the motion in
(5)
j - mode,
Now after putting the value of in the equation (5),

the following expression is obtained,
akjm = added mass coefficient matrix of kj ,
P
1
1
2
= ie + {(Ux +s)} +(Ux +s). + ()2 + gz
2
2
bkjm = damping coefficient matrix of kj ,
By neglecting the higher order terms, we can write
U S
gz) + (ie +U )
P = (
(6)
x
2 x
As first part of equation (6) is associated with the
hydrostatic and steady forces, so neglecting this part,
the first order wave exciting forces or moments and
oscillatory forces and moments caused by the
dynamic fluid pressure acting on the body can be
obtained from the following integrals:
mj m
Fkjm .e i e t = e i e t s e 2 X j mj i e X jU
.nk .ds
x
m
j
After solving the above exciting forces, added

mass and damping coefficients, the motions of two
ships can be solved by the following coupled
equations of motions. To describe the motion of the
multiple floating bodies, two co-ordinate systems, one
fixed to the body and the other fixed to the space have
been introduced. The two co-ordinate systems are
shown in figure 1. The equation of motion is
expressed by the time varying relation between these
two co-ordinate systems. The equation of motion will
be coupled dynamically because of hydrodynamic
interaction and mechanical connections between
them. So the equation can be considered by using the
following matrix relationship:
Once the velocity potential is obtained, the

hydrodynamic pressure at any point on the body can
be obtained from the linearized Bernoullis equation
and can be written as:
U m
Fkm .eie t = ieeiet s 0 + 7m i ( 0 + 7 ).nkmds
x
e x
. Im s (
2.7 Equation of motion in frequency domain
2.6 Forces and Moments
1
P
+ ( ) 2 +
+ gz = 0
t
2
m
j
C = hydrostatic restoring force coefficient matrix of
X mj =
kj ,
vector containing the three translational and
three rotational oscillations about the co-ordinate

axes in j - mode.
The suffices, k, j = 1, 2, 3, 4, 5, 6 represent surge,
sway, heave, roll, pitch and yaw modes, respectively
and m =1, 2 number of ships.
3 RESULTS AND DISCUSSIONS
In this study General Cargo Ship is assumed
moored ( U = 0 ). The mooring arrangement is shown
in Figure 1 and the mooring particulars are given in
Table 1. Barge shaped Ship has a forward speed.
In the Figures, results for single body means
203
1.0
Non-dimensional Heave Force
Cargo Ship moored on its own and double body

means it moored in the presence of Barge Ship which
is sailing nearby. The lateral separation distances
(side to side), period of incident wave, Froude number
are varied. To examine the effects of sailing ship for
the motion of moored ship, a computer program has
been developed to carry out the computation for the
pair of ships mentioned above (Table 1) at head sea
(180 degree) and in a depth of 260 m. The
computation model is expected to be validated by the
model tests. But since the tests have not been carried
out yet, the results obtained from computation have
been compared with published results; in this case the
works of Cheng and Fang [1]. This comparison has
been done in Islam M.R. etal.[5] for heave force for
Fn=0.14 and separated distance (centerline to
centerline) 32.476m.
Present Work
Chen and Fang (2001)
Fn = 0.14, = 45 deg
Ship : Mariner
0.8
0.6
0.4
0.2
0.0
0
/L
Figure 2 Comparison of heave force

[Islam 2008]
25
30
Double Body Fn=0.1
25
Single Body
20
Double Body Fn=0.2
20
Double Body Fn=0.1
Surge
Roll
Single Body
15
Double Body Fn=0.2
15
10
10
5
0
0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
/L
/L
3.5
3
2.5
4.5
Double Body Fn=0.1
4
Double Body Fn=0.1
Single Body
3.5
Double Body Fn=0.2
Single Body
Double Body Fn=0.2
Pitch
Sway
1.5
1
2.5
2
1.5
1
0.5
0.5
0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
/L
/L
4.5
3.5
3.5
3
Single Body
2.5
Double Body Fn=0.2
2.5
Double Body Fn=0.1

Double Body Fn=0.2
Single Body
yaw
Heave
Double Body Fn=0.1
2
1.5
1.5
1
0.5
0.5
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
/L
/L
Figure 3 Non-dimensional motion of Moored Ship With/without sailing ship based on different /L.
( = 180o, and Water depth=260m)
204
Figure 2 indicates that the two results are similar

in trend. However, the actual values differ and this
may be due to the use of different Green function. In
their work, Cheng and Fang [1] used zero speed
Green function, whereas in the present work, Green
function is computed with speed dependent Green
function.
In Figure 3, the motions of moored ship in absence
and presence of sailing ship have carried out based on
/L.,
The graphs show the maximum pick of sway heave,
pitch and yaw on Fn= 0.1 take place at /L=1.11
(T=14.6s); whereas the pick occur at /L=1.69
(T=18s) at surge and roll motions.
In comparison, there is no considerable increase

on the other graph at low /L, related to Fn = 0.2,
except sway motion. The mentioned motion pick
occur at /L=0.775 (T=12.2s).
The motions of Cargo Ship are plotted against
separated distance in Figure 4, at an angle of incident
wave 180 degree, in a water depth of 260m, period
10s and Froude number 0.1. It is seen in Figure 4, the
motions on Cargo Ship reaches the first pick at
separate distance 12m.
Figure 5 shows the motions of moored ship by the
effect of sailing ship at different Froude numbers.
1.60
0.25
1.40
0.20
1.20
1.00
Roll
Surge
0.15
0.10
0.80
0.60
0.05
0.40
0.20
0.00
5
10
15
20
25
30
35
40
45
50
55
60
65
70
0.00
75
Distance
10
15
20
25
30
35
40
45
50
55
60
65
70
45
50
55
60
65
70
75
Distance
0.60
0.80
0.70
0.50
0.60
0.40
Sway
Pitch
0.50
0.40
0.30
0.30
0.20
0.20
0.10
0.10
0.00
0.00
10
15
20
25
30
35
40
45
50
55
60
65
70
10
15
20
25
30
35
40
75
Distance
75
1.80
0.45
1.60
0.40
1.40
0.35
1.20
0.30
1.00
0.25
yaw
Heave
Distance
0.80
0.20
0.60
0.15
0.40
0.10
0.20
0.05
0.00
0.00
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
10
15
20
25
30
35
40
45
50
55
Distance
Distance
Figure 4 Non-dimensional motions of Moored Ship based on different distance.

(T=10s, Fn=0.1 , =180o, and Water depth=260m)
205
60
65
70
75

0.4
0.05
0.35
0.04
0.3
0.25
Surge
Roll
0.03
0.02
0.2
0.15
0.1
0.01
0.05
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0
0.04
0.4
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
Froud No.
0.16
0.25
0.2
0.12
Pitch
Sway
0.15
0.1
0.08
0.04
0.05
0
0.04
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
0.08
0.12
0.16
0.4
0.08
0.3
0.06
0.2
0.04
0.1
0
0.04
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
yaw
Heave
Froud No.
0.02
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0
0.04
0.4
Froud No.
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
Figure 5 Non-dimensional motions of Moored Ship based on different Froude Number. (T=8s, Distance =
8.64m, =180o, and Water depth=260m)
There are a dramatically increscent until Fn=0.07 and

all motions reach a pick at Fn=0.07 except yaw
motion that reach its pick at Fn=0.06. The reason of
this increasing is generation of Long-wave by sailing
ship at low speeds.
The results in Figure 6 carried out at pick points of
Figure 3 (Distance =12m) and Figure 4 (/L= 1.11,
T=14.6s). Graphs show the motions increased
considerably when the period is 14.6s (near natural
period) and distance is 12m. Comparing the results in
Figures 3 and 6 , show that the treatment of curves are
similar.
CONCLUSION
A method for and results of computational

hydrodynamics studies of interaction between two
ships have been presented. Preliminary results

indicate that the effects of interaction on a moored
ship when another ship sailing nearby is quite
significant. For such interaction, motions have
significant influence, which may hinder loading and
unloading operation. In this paper, the model is
validated only with published results with speed but
the model need to be validated by model experiment.
Also computations need to be carried out for various
depths and different incident angles. The
computations also need to be carried in calm water
condition so that the effect of generated waves can be
easily understood. However, the program developed
for computation of motions for a moored ship while
other ships sailing nearby numerically expected to be
able to predict satisfactorily.
206
16
2
1.8
14
1.6
12
1.4
10
T=8s Distance=8.68m
Roll
Surge
1.2
T=14.6s Distance=12m
1
0.8
T=8s Distance=8.68m
0.6
4
0.4
2
0.2
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0
0.04
0.4
0.08
0.12
0.16
Froud No.
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
10
4.5
4
3.5
3
Sway
T=8s Distance=8.68m
T=8s Distance=8.68m
2.5
1.5
2
1
0.5
0
0.04
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.08
0.12
0.16
0.4
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
Froud No.
2.25
1.2
1.75
T=8s Distance=8.68m
T=8s Distance=8.68m
1.5
0.8
yaw
Heave
1.25
1
0.75
0.6
0.4
0.5
0.2
0.25
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0
0.04
0.4
Froud No.
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
Froud No.
Figure 6 Non-dimensional motions of Moored Ship at Pick Points of Figures 3 and 4 based on different
Froude Number. (T=14.6s, Fn=0.07, =180o and Water depth=260m)
REFERENCES
[1] Gung-Rong Chen and Ming-Chung Fang,
Hydrodynamic interactions between two ships
advancing in waves, Ocean Engineering, Vol.
28, pp. 1053-1078 (2001).
[2]
J. A. Pinkster and P. Naaijen, Predicting the effect of

passing ships, International Workshop on water
Waves and Floating Bodies (2003).
[3] Kevin McTaggart, David Cumming, C. C.

Hsiung and Lin Li, Seakeeping of two ships in
lose proximity, Ocean Engineering, Vol. 30, pp.
1051-1063 (2003).
[4] K. S. Varyani and P. Krishnankutty,

Modification of ship hydrodynamic interaction
forces and moments by underwater ship
geometry, Ocean Engineering, Vol. 33, pp.
1090-1104 (2006).
[5] Islam MR, Islam Nazrul and Maimun Adi,
Effect of Vessel Interaction On Hydrodynamic
Forces While Advancing Closly In Waves,
Marine Technology Conference (MARTEC
2008), Indonesia (2008).
[6] Wehausan and Laitone, Surface Waves,
Encyclopedia of Physics, Vol. 9, SpringerVerlag, Berlin (1960).
207

EFFECT OF DIAMETER OF HEATED BLOCK ON MHD FREE CONVECTION

FLOW IN A CAVITY
Salma Parvin and Rehena Nasrin
Department of Mathematics, Bangladesh University of Engineering
and Technology, Dhaka-1000, Bangladesh.
E-mail: salpar@math.buet.ac.bd
ABSTRACT
The present work executes a numerical study of magnetohydrodynamic (MHD) free convective flow and heat
transfer characteristics inside a square cavity with a uniformly heated solid circular block placed at the centre.
The left wall is at a constant temperature and the rest walls of the cavity are considered to be adiabatic. Finite
element method based on Galerkin weighted residual approach is used to solve two-dimensional governing
mass, momentum and energy equations for steady state, natural convection problem in presence of magnetic
field. Results are presented in terms of streamlines, isotherms and average Nusselt number (Nu) at the heated
surface for different values of Hartmann number (Ha) and diameter (D) of the block. The results exhibit that the
flow structure and the heat transfer rate depend significantly on the mentioned parameters.
Key words: MHD, natural convection, block, cavity, finite element method.
1. INTRODUCTION
Several numerical and experimental methods have
been developed to investigate cavities with and
without obstacle because these geometries have
practical engineering and industrial applications, such
as in the design of solar collectors, thermal design of
building, air conditioning, cooling of electronic
devices, furnaces, lubrication technologies, chemical
processing equipment, drying technologies etc. Many
authors have recently studied heat transfer in
enclosures with partitions, fins and obstacles which
influence the convection flow phenomenon. Free
convection flow of an electrically conducting fluid in
a cavity in the presence of magnetic field is of special
technical significance because of its frequent
occurrence in many industrial applications such as
geothermal reservoirs, cooling of nuclear reactors,
thermal insulations and petroleum reservoirs. These
types of problems also arise in electronic packages,
micro electronic devices during their operations.
House et al. [6] studied the effect of a
centered, square, heat conducting body on natural
convection in a vertical enclosure. They showed that
heat transfer across the cavity enhanced or reduced by
a body with a thermal conductivity ratio less or
greater than unity. Garandet et al. [4] analyzed the
buoyancy driven convection in a rectangular
enclosure with a transverse magnetic field. The
geometry considered in the numerical study of Oh et

al. [9], was a cavity with a heat generating conducting
body. Under these situations, it was shown that the
flow was driven by a temperature difference across
the cavity and a temperature difference caused by the
heat-generating source. Roychowdhury et al. [10]
analyzed the natural convective flow and heat transfer
features for a heated cylinder placed in a square
enclosure with different thermal boundary conditions.
Natural convection in a horizontal layer of fluid with
a periodic array of square cylinder in the interior were
conducted by Ha et al. [5], in which they concluded
that the transition of the flow from quasi-steady up to
unsteady convection depends on the presence of
bodies and aspect ratio effect of the cell. Lee et al. [8]
considered the problem of natural convection in a
horizontal enclosure with a square body. Braga and de
Lemos [1] investigated steady laminar natural
convection within a square cavity filled with a fixed
volume of conducting solid material consisting of
either circular or square obstacles. They used finite
volume method with a collocated grid to solve
governing equations. They found that the average
Nusselt number for cylindrical rods was slightly lower
than those for square rods. Lee and Ha [7] considered
a numerical study of natural convection in a
horizontal enclosure with a conducting body. Natural
convective heat transfer in square enclosures heated
from below was investigated by Calcagni et al. [2].
Sarris et al. [11] studied MHD natural convection in a
209
laterally and volumetrically heated square cavity.

They concluded that the usual damping effect of
increasing Hartmann number was not found to be
straightforward connected with the resulting flow
patterns. Roy and Basak [12] analyzed finite element
method of natural convection flows in a square cavity
with non-uniformly heated wall(s).
In the light of the above literature, it has been
pointed out that there is no significant information
about MHD free convection processes in an enclosure
with a heated obstacle. The present study addresses
the effect of size of block and magnetic field which
may increase or decrease the heat transfer on natural
convection in a square cavity. Numerical solutions are
obtained for different diameter of the obstacle and
Hartmann number while other parameters such as
Rayleigh number and Prandtl number are kept fixed.
The numerical results are presented graphically in
terms of streamlines and isotherms. Finally the
average Nusselt number at the heated surface is
calculated.
incompressible and there is no viscous dissipation.

The gravitational force (g) acts in the vertically
downward direction and radiation effect is neglected.
The governing equations under Boussinesq
approximation are as follows
u v
+
=0
x y
(1)
2u 2u
u
p
u
+ v =
+ 2 + 2
x
y
x
y
x
2v 2v
v
p
v
+ v =
+ 2 + 2
x
y
y
y
x
+ g (T Tc )
u
B 02
(2)
(3)
2T 2T
T
T
+v
= 2 + 2
x
x
y
y
(4)
where = k C p is the thermal diffusivity of the

fluid. The boundary conditions are
u (x, 0 ) = u (x, L ) = u (0, y ) = u (L, y ) = 0
v(x, 0 ) = v(x, L ) = v(0, y ) = v(L, y ) = 0
T (0, y ) = Tc
2. PHYSICAL CONFIGURATION
A schematic diagram of the system considered in
the present study is shown in Figure 1. The system
consists of a square cavity with sides of length L and a
heated circular solid block of diameter d is located at
the centre of the enclosure. A Cartesian co-ordinate
system is used with origin at the lower left corner of
the computational domain. The left wall of the cavity
is kept at a constant temperature Tc while the other
walls are considered to be adiabatic. The uniform
temperature of the obstacle is assumed to be Th. Here
Tc is less than Th. A magnetic field of strength B0 is
applied horizontally normal to the side walls.
T (x, 0 ) T (x, L ) T (L, y )

=
=
=0
y
y
x
At
the
circular
body
surface
u ( x, y ) = v ( x, y ) = 0 , T ( x, y ) = T h
The local Nusselt number at the heated circular
body in the square enclosure is evaluated by the
following expression in dimensional form as
Nu local = hd / k , where h is the convective heat

transfer coefficient.
The above equations are non-dimensionalized by

using the following dimensionless dependent and
independent variables
x
Tc
Th
B0
uL
vL
d
y
x
, Y = , D= , U =
, V=
,
L
L
L
T Tc
pL2
, =
P=
Th Tc
2
X=
After
substitution of the above variables, equations (1) to
(4) transformed into the following non-dimensional
equations
L
adiabatic
U V
+
=0
X Y
Figure 1. Schematic diagram of the problem
3. MATHEMATICAL FORMULATION
U
In the present problem, it can be considered that the

flow
is
steady,
two-dimensional,
laminar
210
2U 2U
U
U
P
+V
=
+ Pr
+
2
X
Y
X
Y 2
X
2V 2V
V
V
P
+V
=
+ Pr
+
X 2 Y 2
X
Y
Y
modified nonlinear equations are transferred into

linear algebraic equations by using Newtons method.
Finally, these linear equations are solved by using
Triangular Factorization method.
+ Ra Pr Ha 2 Pr V
U
2 2
+
+V
=
X
Y X 2 Y 2
Pr =
where
Ra =
g (Th Tc )L3
4.1 Grid Refinement Check
is
Prandtl
number,
is Rayleigh number and Ha is
Hartmann
number
2 3
B
L
0
.
Ha 2 =
which
is
defined
as
The corresponding boundary conditions then take

the following form
At the inside and on the wall of enclosure fluid
pressure P = 0
In order to determine the proper grid size for this

study, a grid independence test is conducted with five
types of mesh for Ha = 50, Pr = 0.7, Ra = 105 and D =
0.25. The extreme value of Nu is used as a sensitivity
measure of the accuracy of the solution and is selected
as the monitoring variable. Considering both the
accuracy of numerical values and computational time,
the present calculations are performed with 39284
nodes and 5936 elements grid system.
Table 1. Grid Sensitivity Check at Ha = 50, Pr = 0.7,
Ra = 105 and D = 0.25.
U ( X , 0) = U ( X , 1) = U (0, Y ) = U (1, Y ) = 0
V ( X , 0) = V ( X , 1) = V (0, Y ) = V (1, Y ) = 0
(0 , Y ) = 0
( X , 0) ( X , 1)
=
Nodes
(elements)
Nu
(1, Y )
=
=0
X
Time (s)
7432
(1096)
11988
(1784)
26536
(3992)
39284
(5936)
79500
(12080)
0.578491
0.596590
0.598085
0.598592
0.598592
226.265
292.594
388.157
421.328
627.375
At the circular body surface

U (X , Y ) = V (X , Y ) = 0 , (X , Y ) = 1
The average Nusselt number at the body of the
enclosure may be expressed as
Ls L
Nu =
ds
n
2

=
+
and s is the
n
X
Y
coordinate along the circular surface.
where
4. NUMERICAL TECHNIQUE
The numerical procedure used in this work is based
on the Galerkin weighted residual method of finite
element formulation. The application of this technique
is well described by Taylor and Hood [13] and
Dechaumphai [3]. In this method, the solution domain
is discretized into finite element meshes, which are
composed of non-uniform triangular elements. Then
the nonlinear governing partial differential equations
(i.e. mass, momentum and energy equations) are
transferred into a system of integral equations by
applying Galerkin weighted residual method. The
integration involved in each term of these equations is
performed by using Gausss quadrature method. The
nonlinear algebraic equations so obtained are
modified by imposition of boundary conditions. These

Effect of the physical parameter (D) and Hartmann
number (Ha) on heat transfer and fluid flow inside the
cavity has been analyzed in the present study. Heat
transfer rate in terms of the average Nusselt number at
the heated block in the cavity has been also discussed.
The ranges of D and Ha for this investigation vary
from 0.15 to 0.5 and 0 to 50 respectively whereas
other parameters are fixed at Ra = 105 and Pr = 0.7.
The influence of diameter D for a fixed Ha (= 50)
on the flow field is depicted in Fig 2 (a). The flow
with D = 0.15 has been affected by the buoyancy
force, thus creating two rotating cells. The right top
one is small which vanishes with increasing D. The
left one covers almost the whole cavity including the
block. It contains a small vortex. The size of this
vortex increases due to the increasing values of D. For
D = 0.50, the existing recirculation region becomes
larger and two inner vortices are developed. Fig. 2 (b)
illustrates the temperature field in the flow region.
The stratified isothermal lines concentrated near the
left wall and the heated body. The concentrated
temperature region near the block becomes denser for
larger D and consequently the isothermal lines almost
disappear from the right top portion of the cavity.
The influence of Hartmann number Ha (from 0 to
70) on streamlines as well as isotherms for the present
211
Ha = 20
D = 0. 35
Ha = 70
D = 0. 25
Ha = 50
D = 0.15
configuration at Ra = 100000, Pr = 0.7, D = 0.25 has

been demonstrated in Fig. 3. The flow with Ha = 0
creates two vortices near both side of the heated body
due to the buoyancy force. The vortex on the right top
corner losses its strength for higher Ha and finally
disappeared. The corresponding isotherm patterns are
shown in Fig. 3 (b). The high temperature region is
concentrated near the circular obstacle for Ha = 0.
The lines become less bend from the left top corner
because of increasing values of Ha. The concentrated
temperature region near the heated surface becomes
thin for larger Ha.
Ha = 0
(a)
(b)
D = 0. 5
Fig. 3: (a) Streamlines and (b) Isotherms for

varying of Ha with Ra = 105, D = 0.25 and Pr = 0.7
(a)
(b)
Fig. 2: (a) Streamlines and (b) Isotherms for varying
of D with Ra = 100000, Ha = 50 and Pr = 0.7
In order to evaluate how the presence of magnetic
field and diameter of the obstacle affect the heat
transfer rate along the heated surface, the average
Nusselt number is plotted as a function of Hartmann
Fig. 4: Effect of Ha and D on Nu while Ra = 105

and Pr = 0.7
212
number as shown in Fig. 4 for four different diameter

( D = 0.15, 0.25, 0.35 and 0.5) while Ra = 105 and Pr
= 0.7. It is observed that Nu decreases with the
increasing Ha and increases with the growing D. The
maximum heat transfer rate is obtained for the lowest
Ha. This is because the magnetic field retards the
flow. Also the average Nusselt number is found
highest for the largest value of D because larger body
having larger surface is capable to transfer more heat.
6. CONCLUSION
A finite element method is used to make the present
investigation for steady-state, incompressible, MHD
free convection flow in a cavity with a heated body.
The major conclusions have been drawn as follows:
The diameter of the body has a significant effect on
the flow and temperature fields. Buoyancy-induced
vortex in the streamlines increased and thermal layer
between the left wall and the heated surface become
denser for increasing values of D.
The influence of Magnetic parameter Ha on
streamlines and isotherms are remarkable. The vortex
in the streamlines decreased and thermal layer near
the heated surface becomes thin with increasing
values of Ha.
The average Nusselt number Nu at the heated
surface increases for larger values of Ra and D where
as it decreases with the increasing values of Ha.
[6]
House, J. M., Beckermann, C., and Smith, T.

F., Effect of a centered conducting body on
natural convection heat transfer in an
enclosure, Numerical Heat Transfer A-Appl.,
Vol. 18, pp. 213225 (1990).
[7]
Lee, Jae Ryong, and Ha, Man Yeong, A

numerical study of natural convection in a
horizontal enclosure with a conducting body,
Int. J. of Heat and Mass Transfer, Vol. 48, pp.
33083318 (2005).
[8]
Lee, J. R., Ha, M.Y., Balachandar, S., Yoon, H.

S., Lee, S. S., Natural convection in a
horizontal layer of fluid with a periodic array
of square cylinders in the interior, Phys.
Fluids, Vol. 16, pp. 10971117 (2004).
[9]
Oh, J. Y., Ha, M. Y., Kim, K. C., Numerical

study of heat transfer and flow of natural
convection in an enclosure with a heat
generating conducting body, Numerical Heat
Transfer A-Appl., Vol. 31, pp. 289304 (1997).
[10]
Roychowdhury, D. G., Das, S. K.,

Sundararajan, T. S., Numerical simulation of
natural convection heat transfer and fluid flow
around a heated cylinder inside an enclosure,
Heat Mass Transfer, Vol. 38, pp. 565576
(2002).
[11]
Sarris, I. E., Kakarantzas, S. C., Grecos, A. P.,

Vlachos, N. S., MHD natural convection in a
laterally and volumetrically heated square
cavity, Int. J. of Heat and Mass Transfer, Vol.
48, pp. 34433453 (2005).
[12]
Roy S. and Basak Tanmay, Finite element

analysis of natural convection flows in a square
cavity with non-uniformly heated wall(s), Int.
J. of Engineering Science, Vol. 43, pp. 668680
(2005).
REFERENCES
[1]
Braga, E. J., and De Lemos, M. J. S., Laminar

natural convection in cavities filed with
circular and square rods, Int. Commun. Heat
Mass, Vol. 32, pp. 12891297 (2005).
[2]
Calcagni, B., Marsili, F., and Paroncini, M.,

Natural convective heat transfer in square
enclosures heated from below, Applied
Thermal Engineering, Vol. 25, pp. 25222531
(2005).
[3]
Dechaumphai, P., Finite Element Method in

Engineering,
2nd
ed.,
Chulalongkorn
University Press, Bangkok, (1999).
[4]
Garandet, J. P., Alboussiere, T., and Moreau,

R., Buoyancy driven convection in a
rectangular enclosure with a transverse
magnetic field, Int. J. of Heat Mass Transfer,
Vol. 35, pp. 741-748 (1992).
[5]
Ha, M. Y., Yoon, H. S., K. S. Yoon, S.

Balachandar, I. Kim, J. R. Lee, and H.H.,
Chun, Two dimensional and unsteady natural
convection in a horizontal enclosure with a
square body, Numerical Heat Transfer AAppl., Vol. 41, pp. 183210, (2002).
[13] Taylor, C., Hood, P., A numerical solution of
213
the Navier-Stokes equations using finite

element technique, I, Computer and Fluids,
Vol. 1, pp. 7389 (1973).

The Internatiional Conference on
o Marine Technoology
11-12 Decem
mber 2010, BUET, Dhaka, Bangladeesh
AN EN
NVIRONME
ENT FRIEN
NDLY MET
THOD OF DRIVING
D
S
SHIPS WITH
H AIR JET
T
PROP
PULSION
Mohammad Jiaur Rahm
man
Naval Architect and Marine Engin
neer
Email: jia011_buuet@yahoo.com
ABSTRA
ACT
This paperr presents a new
w method of driving
d
marine vessels with air
a jet propulsioon and thereby
y reducing greeen
house effecct. Air jet can be produced using
u
compresssor, which can be passed throough pipes or nozzles
n
installeed
at stern off the ship. Theeoretical relatioonship among the ship speeed and other parameters suchh as compresssor
pressure, displacement
d
of
o ship, air inllet velocity etcc. has been esstablished. Thee elastic collission between air
a
particles an
nd water particcles doubles thhe ship speed which is not possible
p
in case of water jet propulsion. Thhe
method is applied
a
for calculating speed
d of a patrol boaat and satisfacttory result has been obtained. Moreover, CO
O2
present in the
t air jet is disssolved into seea water and thhereby reduced green house effect.
Key word
ds: Green housse gas, Air jet, ship propulsioon, Elastic colliision, Joule-Thhomson effect.
NOMENCLATURE
Densityy of air
A is the crooss section area of each pipe or nozzle
` Densitty of water
d Diameteer of pipe or noozzle
Viscositty of gas
n Numberr of pipes use for

fo passing air jet
j
Viscosity of water
h` Depth of
o pipes from free
f surface of water
1. INTR
RODUCTION
N
h Distancce of compressoor centre from datum line
At pressent United Naation, Europeann Union and reest

of the wo
orld try to keeep environmennt free from air
a
pollution. Because the temperature of the world is
increasingg day by day, ddue to increasee of green houuse
gas e.g. carbon di oxxide (CO2)[1]. Figure 1 & 2
t
increase of
o CO2 in atm
mosphere. Manny
express the
scientists have reporteed on globall warming annd
suggestedd to reduce thee use of CFC, carbon di oxidde,
carbon moono oxide etc [2].
[
ho Distancce of the centree of pipes from

m datum line
g Accelerration due to grravity
P Pressuree of the compreessor
P` Pressuree of gas in a coompartment
Po Water pressure
p
(h`g)) + atmospheriic pressure
m Mass of
o air passing th
hrough one pippe per unit timee
M is the mass
m of the shipp = Displacemeent of the shipp
()
k Proporttional constant
k` Anotheer proportional constant
Rn Reyno
olds number
V Velocity
y of air draw
wn from the out
o side of thhe
compressoor or environmeent
Vo Air vellocity at the end
d of pipe
V` Velocitty of ship whicch created due to air thrust onn
ship body
o ship
V`` Resulltant velocity of
Velociity gradient.
T Temperrature of gas
Figure 1:: The Keelingg Curve of attmospheric CO

O2
concentraations measured at Mauna Loa
L Observatorry.
[3, 4]
215
Proceeding
gs of MARTEC 20010
incident on water byy compressorr pressure annd

sometimees it is conveert into turbulent flow whicch
depends on
o Reynolds number
n
and Reynolds
R
numbber
define as defined by
Rn=
Turbuleent flow is suittable for mixinng any gas (heere

gas is CO
O2) with sea waater. This inciddent of air helpps
to mix CO
C 2 with ocean water easiily than naturral
process. In this proceess CO2 abso
orption will be
b
j
accelerateed than naturaal process. Therefore air jet
propulsion
n help to reduuce CO2 i.e. green
g
house gas
g
from surroounding.
Figure 2: Yearly incrrease of atmosspheric CO2: Inn

the 1960s, the average annnual increase was
w 37% of thhe
2000-20077 average. [1]
70% porrtion of the wo
orld is coveredd by water andd
therefore, water
w
ways veh
hicle play veryy important rolle
on transpoortation. All these
t
vessels are moved byy
hydrocarbo
on (oil, gas ettc) which prodduce carbon ddi
oxide, carbbon mono ox
xide, etc as well
w
as help too
increase temperature off the world. The chemicaal
i
reaction off hydrocarbon is
Cx Hy + (x+ ) O2 = x CO2 +
Here,
2. MATH
HEMATICAL
L ANALYSIS
S OF AIR JET
In Figure 3 let a compressor or a series of
compressor is used to suupply air throuugh the n numbber
of pipes installed
i
at thee stern of the ship. The crooss
section arrea at the end oof the pipe is A and pressure Po
[which is
i atmospheric pressure +hg
+
(watter
pressure)]]
The work
ks of the
H2O
Cx Hy is Hydroocarbon or fuell.
C 2 is carbon di
CO
d oxide.
H2O is Water.
From thhe above equ

uation it is seen that largge
quantity off CO2 is produuced from hyddrocarbon. Fueel
consumptio
on of high speeed vessel is reelatively higheer
than low or
o medium speeed vessel. On the other handd
high speedd vessel has low efficienccy due to slipp,
cavity etc as well as shhaft losses occcur at time oof
propeller revolution.
r
Alll these lossess increase fueel
consumptio
on i.e. greenn house gas. But air jeet
propulsionn is free from these losses. If high speedd
vessel can be run by air jet propulsionn, then fuel cann
be saved. Therefore it helps
h
to decreaase green house
gas. Now a days many ship designerr try to reduce
green housse gas by desiign optimizatioon such as hulll
form optim
mization to reduce
r
frictioonal resistancee,
changing the shape of bridge deck, funnel, wheeel
r
wind reesistance by making
m
aerofoiil
house to reduce
section annd so on [5]. But this optiimization is noot
sufficient too reduce green hoouse gas. From this research iit
is shown thhat CO2 absorp
rption is acceleerated by air jeet
propulsionn in sea water.
Sea wateer absorbs CO
O2 and produces carbonic acid
with bi carbonate
c
andd carbonate ions
i
[6]. Thiis
absorption occurs natuurally but thhis absorptionn
increase byy air jet propullsion. In this method,
m
air jet iis
d
of a ship
s
and pipinng
Figure 3:: Schematic diagram
system with
w compressor or series of compressor orr a
device whhich can perfoorm according to the propossal
of this theeory.
Compreessor is drawn air from the environment
e
annd
drive out through the ppipes or nozzlles with a greeat
velocity Vo. Let the insiide pressure off the compresssor
is P and drawn
d
velocityy of air to the compressor
c
froom
environm
ment is V. The ddistance of the compressor annd
the end of
o the pipe froom a datum linne are h and ho
respectiveely. So the masss of the air paasses through thhe
each pipe per unit time is
i m=AVo
hole mass of the air passess through the n
The wh
number of pipes is nm = nAVo
(1).
216
Proceeding
gs of MARTEC 20010
Applying
g the equation
n of continuityy at the end oof
each pipe, then it becomees
A o=k . (22)
AV
Accordinng to Newton``s third law off motion, whenn
air jet is coming
c
out fro
om the stern of ship, it willl
move forw
ward.
Let the speed
s
and masss (displacemeent in tonne) oof
the ship arre V` and M respectively.
r
Therefore
T
it cann
be written from the conseervation of momentum
M V`=nmVo
Or, MV`=AnVo2
from equattion (1)]
[S
Substituting thhe value of nm
m
Or, V`= A
An Vo2/M
(3)
In Equattion (3) n, , A & M are co
onstants. So thhe
ship speed
d is directly prroportional to the squared oof
air velocityy Vo2;
i.e. , V`=kk' Vo2 [where, k' is another constant]
Or V`
Vo2
For supp
plying air jet,, there is a neeed continuouus
flow of aiir through com
mpressor and the end of thhe
pipe. So ap
pplying the Berrnoulli equatioon [15]
Vo =
..
(4)
That means the density of water app

proximately 8000
times greeater than air. The air consiists of so manny
gases parrticles such as nitrogen, oxyygen, inert gasses
etc and acccording to gas kinetic theorry these particlles
are alwayys moving. Sinnce the air consists of differeent
gas particcles, adhesive fforce is act amo
ong them. Agaain
the rangee of adhesive force of gas particles
p
is veery
small and
d approximatelyy 1EXP-10m. But
B according to
gas kinettic theory, a small distancce always exiist
among thhe gas particlees. For these reasons
r
adhesivve
force of gas
g particles iss very small. But
B the cohesivve
force of water particle is very muuch larger thaan
adhesive force of air parrticles. From gas
g law we knoow
that temp
perature and pressure of gas are relateed
proportionnally (P` T) when volumee and number of
moles of gas
g are constannt.
In the compressor gaas temperaturee is high due to
high presssure but in Figure-5 the temperature of
portion 1 is normal, beecause in this region pressuure
energy off gases must bbe dropped andd converted innto
velocity energy
e
accordinng to Bernoullli equation. Heere
gas density is very mucch lower than compressor gas
g
T
the ggas particles posses
p
a distance
density. Therefore
among thhem in this reggion 1 and this kinetic energgy
gain by absorbing gaas temperaturee and pressuure
ng to Joule-Thhomson effect.). Again theese
(Accordin
pipes aree always kept in the water or cold regioon
comparedd with compresssor gas temperrature. For thoose
reasons thhe temperature of this region remains normaal.
At normal temperature, viscosity of gaases is low.
Combining
C
Equuations (3) & (4
4)
V`=
Figure 4: Schematic diaagram of air jeet when produce

thrust in the
t water.
.. (5)
But Po = atmospherric pressure + h``g.

h
This is the
t air jet veloocity which is protruded from
m
end of pip
pe or end of nozzles. Butt the collisionns
between air
a & water particles are almost elastiic
collision. Since
S
the mass of water parrticles are veryy
large than air particles. The density of air and wateer
1
kg/m at
a
are respecctively 1.293 kg/m and 1000
temperaturre
normal
and
pressuree.
Figure 5: The atmospheeric pressure acct on sea surface

which inccrease inertia of water.
217
Due to low viscosity of gas particles, air jet or air

particles can impinges on water with high velocity.
The viscous force of water particle is very much
larger than air particle since
Viscosity of water at 20C is 1EXP-3 Ns/m
Estimation for a patrol boat in case of ideal

fluid:
Let for a patrol boat moving through ideal fluid
n=100 ,
=1.293 kg/m , M=140000 kg
Or,140tonnes , P=250atm , h-ho=3m
V=5m/s
, A=r
Viscosity of air at 20C is1.8EXP-5 Ns/m

Again atmospheric pressure and water pressure
(h``g) act opposite to the direction of gas flow on the
end of nozzles and these features increase the
elasticity of water.
Finally following features of water and air can be
considered:
1.
Mass of water particles are very large than air

particles.
2.
Cohesive force of water particles are greater than

adhesive force of air particles.
3.
Viscosity of water also greater than air.
4.
Atmospheric pressure and water pressure (hg)

help water to increase inertia.
5.
Mass of air particles is very much smaller than

that of water particles.
6.
Adhesive force of gas particles is small.
7.
Viscosity of air is smaller than that of water.
From the above decisions it can be said that the

collision between air particle and water particle must
be elastic collision. Therefore reaction thrust of water
gives equal velocity (V`) to air jet .
If r=1cm, A=0.000314 m
Put this value in the equation (5) then velocity
become V`=11.169m/s
From the equation (6a) the resultant velocity of the
ship is V``=2*11.169=22.34m/s = 80.424 km/h
V``= 43.47 knots without resistance.

Figure 6 shows the variation of ship velocity (V``)
with nozzle or pipe number (n). This figure is
constructed from equation (8) and others parameters
of this equation are considered as a constant. This
figure produce a straight line y = mx whereas x is an
independent variable, y is a depended variable and m
is slope. Figure 6 shows that the line passes through
origin. That means when nozzle number zero (n=0)
then ship velocity is zero (V``=0). But ship velocity
linearly increases with increase of nozzle numbers.
Therefore designer should provide more nozzles to
increase ship speed. But complexity will come to
install large number nozzle since it needs more air
supply, more compressor pressure and so on.
Therefore designers have to be chosen optimum
number of nozzles and ship speed.
100
Therefore the resultant ship speed becomes
80
Ship Velocity (m/s)
V``=V`- (-V`).. (6)

Because water behaves like a movable wall which
always exist against air jet flow.
Or, V`` =2V`. (6a)
.. (8)
100
150
200
250
Pipe number
Figure 7 shows the variation of ship velocity (V``)

with nozzle diameter (d). From equation (8) it shows
that ship velocity is squarely proportional to nozzle
diameter. Therefore this figure is slightly exponential.
Others parameters of equation (8) are also considered
as a constant. From the equation of continuity it is
clear that velocity of air jet and area of nozzle are vice
218
50
Figure 6: Variation of ship speed (V``) with different

number of pipes (n).
(7)
But M is equat to ship displacement and substitute

by , then equation (7) becomes
V``=
40
20
Substituting the value of V` from equation (5) in

equation (6a) then it becomes
V``=
60
versa. Since equation of continuity represent a reverse

relation between cross section area and velocity of
flow (AVo=k). If cross section area increases then
same jet velocity can be maintained by increasing
supply of air. However designer can easily set any
parameter from working formula (Equation-8) to
maintain compressor pressure with large diameter.
100
compressor pressure change spontaneously and others

parameters are fixed. But for zero compressor
pressure there is a reverse velocity. This reverse
velocity produces for water pressure and atmospheric
pressure which is already been mentioned. Due to the
velocity, water will go in pipe and compressor and
other devices if they are situated below water line. To
protect this problem designer should be given one
way bulb in the end of nozzles or in suitable position
of nozzles.
50
60
40
40
Ship Velocity (m/s)
Ship Velocity (m/s)
80
20
0
0.5
1.0
1.5
2.0
2.5
3.0
30
20
10
0
Pipe diametre (m)
-10
0

nozzle or pipe diameter (d).
Figure 8 shows the variation of ship velocity with
displacement. Rests of the parameters of equation (8)
are fixed. It shows that ship velocity decrease with
increasing displacement. Initially ship velocity is
decreased at a higher rate with increasing
displacement but latter decrease at a lower rate.
Therefore the process is comparatively suitable for
large displacement. But displacement related with
resistance, inertia of ship and these parameters are
related with propulsion power. Therefore large
displacement need large propulsion power and
propulsion is depended on compressor pressure. The
compressor pressure and others parameters have to be
capable to gain these situations.
60
55
Ship Velocity (m/s)
50
45
40
35
30
25
20
15
10
15
20
25
Displacement (Kg) X 10
30
35

displacement ().
Figure 9 shows the variation of ship velocity with
respect to compressor pressure. Ship velocity is
increasing with increase of compressor pressure. Here
100
150
200
250
300
Compressor Pressure (atm)

compressor pressure (P).
4. CONCLUSION
This paper seek new type of propulsion system to
enrich the ship propulsion as well as to reduce the
green house gas. This propulsion system will be
environment friendly with respect to other propulsion
devices used presently. If the compressor can be
designed according to expectation, then this
propulsion can be implimented practically. Now a
days some air cusion vheicles (ACV) have been
developed and these air cusion is produced by
compressor[10]. This propulsion absorbs CO2 by the
sea from environment but this CO2 produce carbonic
acid with bi carbonate and carbonate ions which have
already been mentioned. Some of this CO2 are
minimized by photosynthesis of organism present in
sea water. But most of the carbonic acid adversely
attack to sea alkalinity and this effect seriously harm
organism as well as decrease shell forming [2]. These
problems of acidity can be removed by adding extra
dilute alkali substance since acid-base reaction
produce water and salt which neutralize the acid or
base density (Ph) of a solution [11]. But a big
challenge will come in this way to mix the alkali
substance and it may be overcome by moving of ship.
If alkali solution is passed at stern of ship, it will react
with sea acidity. After that it will be easily mixed and
neutralized by producing water and salt since most of
the ship is run by propeller and propeller produce
uniform flow with circulation which is very helpful to
mix solution. But further research is needed to do this
because there is may be come various problems and
impact on sea environment. In spite of these
219
50
difficulties, this method can be used to reduce green

house gas to save our environment.

Technology has been acknowledged for his valuable
suggestion for the study.
ACKNOWLEDGEMENT
Prof. Dr Mashud Karim, Head of department of

Naval
Architecture
and
Marine
Engineering,
REFERENCE
1.
2.
8.
Garrison, Tom (2004). Oceanography: An

Invitation to Marine Science. Thomson Brooks.
pp. 125. ISBN 0534408877
"Climate and CO2 in the Atmosphere".

http://earthguide.ucsd.edu/virtualmuseum/climate
change2/07_1.shtml. Retrieved 2007-10-10.
9.
Martini, M. (1997). "CO2 emissions in volcanic

areas: case histories and hazaards". in Raschi, A.;
Miglietta, F.; Tognetti, R.; van Gardingen, P.R.
(Eds.). Plant responses to elevated CO2:
Evidence from natural springs. Cambridge:
Cambridge
University
Press.
pp. 6986.
ISBN 0521582032.
10.
ACS Demonstrator, which features a groundbreaking energy-saving air cavity system integral
to her hull. International journal of the Royal
institution
of
naval
architects.
www.rina.org.uk/tna. / Green ship technology/
CAD/CAM update/ Vietnames shipbuilding/
Norway`s marine industries January 2009. P-2631.
11.
Miessler, L. M., Tar, D. A., (1991) "Inorganic

Chemistry" 2nd ed. Pearson Prentice-Hall p165.
12 .
Doney, Scott C.; Naomi M. Levine (2006-11-29).

"How Long Can the Ocean Slow Global
Warming?". Oceanus.
http://www.whoi.edu/oceanus/viewArticle.do?id=
17726. Retrieved 2007-11-21
"Volcanic Gases and Their Effects".

http://volcanoes.usgs.gov/Hazards/What/VolGas/
volgas.html. Retrieved 2007-09-07
13.
Climate Change 2007: Synthesis Report, IPCC
14.
A Ground Test Rocket Thrust Measurement

System Mary Fran Desrochers, Gary W. Olsen,
and M. K. Hudson
PMEL Ocean Acidification Home Page.

http://www.pmel.noaa.gov/co2/OA/
15.
Halliday, Resnick, Krane, Thrust on a Rocket,

John Wiley & Sons, Physics, Volume1, Fifth
edition. P-358. P 360-361.
3.
Mauna Loa CO2 annual mean data from NOAA.

"Trend" data was used. See also: Trends in
Carbon Dioxide from NOAA.
4.
"Annual Mean Growth Rate for Mauna Loa,

Hawaii". Trends in Atmospheric Carbon Dioxide.
NOAA Earth System Research Laboratory.
http://www.esrl.noaa.gov/gmd/ccgg/trends/#mlo_
growth. Retrieved 28 April 2010.
5.
6.
7.
Department of Applied Science and The Graduate

Institute of Technology University of Arkansas at
Little Rock, Little Rock, AR 72204 USA
Pieter (2008) "Annual CO2 mole fraction increase

(ppm)" for 1959-2007 National Oceanic and
Atmospheric Administration Earth System
Research Laboratory, Global Monitoring
Division (additional details).
Swedish research body (2008). Optimized hulls

lower gas emission. International journal of the
Royal institution of naval architects.
www.rina.org.uk/tna.Vietnames
shipbuilding/Norway`s marine industries
CAD/CAM update/ Green ship technology/
January 2008. P-73.
220

MOTION RESPONSES AND INCIDENT WAVE FORCES ON A MOORED SEMI

SUBMERSIBLE IN REGULAR WAVES
Hassan Abyn1, Adi Maimun2, M. Rafiqul Islam3, and Dr. Allan Magee4
1
Department of Marine Technology
Universiti Teknologi Malaysia,
Skudai Johor Bahru, 81310 Malaysia
E-mail: abynhassan@gmail.com
3
Department of Naval Architecture and Marine Engineering.
BUET,
Dhaka 1000, Bangladesh
E-mail: rafiqis@name.buet.ac.bd
2
Department of Marine Technology
Universiti Teknologi Malaysia,
Skudai Johor Bahru, 81310 Malaysia
Email : adi@fkm.utm.my
4
Manager R&D
Offshore Facilities Product Line & Technology
Technip Malaysia
Email: amagee@technip.com
ABSTRACT
Nowadays, floating structures play an important role for exploring the oil and gas from the sea. With evergrowing needs for oil and gas resources, the ocean engineering has been paid much attention to the world
engineering community not only in coastal regions but also in deepwater. Unlike seagoing ships, moored
floating offshore structures such as semi submersibles and tension leg platforms (TLP) are usually positioned at
a given location at sea and their motion is externally constrained by the moorings. Motion response of a floating
structure should be kept adequately low to guarantee the safety of risers and umbilical pipes as most important
components in the equipment of oil production. The objectives of this study is to investigate the hydrodynamic
forces and motions of both free floating and moored semi submersible under incident, scattered and radiated
waves numerically. In numerical formulation, hydrodynamic problems are solved by using three-dimensional
source distribution method, within the scope of linear wave theory and using frequency domain. Since
experimental test have not been done yet, so results obtained from computations were validated with the results
obtained using commercial software MOSES and WAMIT and other published papers.
Key words: semi submersible, floating structures, 3D source density distribution technique
1. INTRODUCTION
Floating structures such as ship, semi-submersible,
FPSO, TLP, breakwater and other free floating or
moored structures, are subjected to wave, wind and
current at sea. They have six-coupled degrees of
freedom of motions. Namely, linear motions are
surge, sway and heave, and angular motions are roll,
pitch and yaw. Oscillation of floating structure affects
the loading and offloading operation systems.
There are different theories for studying motion of
floating structure such as strip theory and potential
theory. In this paper 3D source density distribution
technique is used to get the potential over the floating
structure. Having flow velocity potentials on and off
the panels, hydrodynamic coefficients of floating
structure can be determined. Using Bernoullis
equation leads to calculation of pressure distribution
and forces over the floating structure. A mathematical
model is mathematical structure that can be used to
describe and study a real situation. A second-order
linear differential equation for coupled six degree of
freedom can describe the hydrodynamics of floating
structures; consist of added mass, damping

coefficient, stiffness coefficient, forces and motions in
six directions.
J. L. Hess and A. Smith [1] studied on the
calculation of non-lifting potential flow about
arbitrary 3D bodies. They utilized a source density
distribution on the surface of the body and solved for
distribution necessary to take the normal component
of fluid velocity zero on the boundary. Plane
quadrilateral source elements were used to
approximate the body surface, and the integral
equation for the source density is replaced by a set of
linear algebraic equations for the values of the source
density on the quadrilateral elements. By solving this
set of equations, the flow velocities both on and off
the surface were calculated.
Oortmerssen [2] dealt with the hydrodynamic
forces between two structures floating in waves by
using a three-dimensional linear diffraction theory and
the results agree well with experiments..
Wu et al. [4] studied the motion of a moored semi
in regular waves and wave induced internal forces
221
numerically and experimentally. In the mathematical

formulation, they modeled the moored semi as an
externally constrained floating body in waves, and
derived the linearized equation of motion.
Yilmaz and Incecik [3] analyzed the extreme
motion response of moored semi-submersible. They
developed and employed two different time domain
techniques since there are strong nonlinearities in the
system due to mooring line stiffness and damping and
viscous drag forces. First one is for simulation of
wave frequency motions in which the first-order wave
forces are the only excitation forces. First-order wave
forces acting on semi-submersibles are evaluated
according Morison equation, current effect is taken
into account by altering the drag term in Morison
equation. Second one is to simulate the slowly
varying and steady motions under the excitation of
slowly varying wave, current and dynamic wind
forces. Slowly varying wave forces are calculated
using the mean drift forces in regular waves and
applying an exponential distribution of the wave force
record in irregular waves.
2. MATHEMTICAL MODEL
2.1 Coordinate System
The individual semi submersible is treated as a
rigid body having six degrees of freedoms. It is
subjected to hydrodynamic forces due to incident
waves and radiated and diffracted waves due to other
vehicle(s). Two right hand coordinate systems are
defined in Figure 1. One is fixed to the space on water
surface and the other one is fixed to the centre of
gravity.
The fluid is assumed to be incompressible, inviscid
and irrotational and the vessel is assumed to be freely
floating in open water. Then there exists a velocity
potential satisfying Laplace equation together with
boundary conditions on the free surface, on the body,
and at the bottom, and the radiation condition in the
far field. The time dependence of the fluid motion to
be considered here is restricted to simple harmonic
motion and accordingly the flow filed can be
characterized by the following velocity potential:
Soylemez [5] developed a prediction technique to

simulate the motion response of damaged platform
under wave, wind and current forms. The equation of
motion was obtained using Newtons second law and
the numerical solution technique of non-linear
equations of motion is explained for intact and
damaged cases. The analysis technique employs large
displacement non-linear equations of motion.
Solutions were obtained in the time domain to predict
the motion characteristics.
Clauss et al [8] analyzed numerically and
experimentally the sea-keeping behavior of a semi
submersible in rough waves in the North Sea. They
used
panel
method
TiMIT
(Time-domain
investigations, developed at the Massachusetts
Institute of Technology) for wave/structure
interactions in time domain. The theory behind TiMIT
is strictly linear and thus applicable for moderate sea
condition only.
Sujatha and Soni [9] used diffraction theory to
formulate the boundary value problem and analyzed
forces and motion responses of moored offshore
floating structures like semi submersible. They solved
the interaction problem involves the motion of
structure by using constant triangular elements on
indirect BEM.
In this study the impact of regular waves is
investigated for the semi submersible of the type of
GVA 4000 which is characterized by favorable seakeeping behavior. This offshore structure is designed
for world wide operation, especially for the harsh
conditions in the North Sea.
Figure 1: Definition of co-ordinate system.
= Re[ ( x, y, z )e it ]
(1)
j =1
= i (0 + 7 ). a + ( X j j )
0 =
ig a cosh[k ( z + h)] ik ( x cos + y sin )

e
cosh kh
(2)
(3)
Where,
0=incident wave potential
7= diffraction wave potential on body
j= potential due to motion of the body in j-th mode
= circular frequency of incident wave
a=incident wave amplitude
= wave heading angle from X -axis
The differential equation governing the fluid
motion follows from the application of the continuity
equation which yields the Laplace equation. The
222
individual potentials are the solutions of the following

Laplace equation:
2 = 0
(4)
In order to ensure that the velocity potential has the

correct amplitude behavior in the far field, the
radiation and scattering potentials must satisfy the
radiation (Summerfield) condition:
k i 2
k = 0 , as r k = 1,2,...,7
r
L
2.2 Boundary Condition

On the mean wetted surface area of body S, the
above linear velocity potentials must satisfy the
Laplace equation and also the following boundary
conditions:
-linearized free surface condition:

+
=0
g
z
2
at
z=0,
(5)
-boundary condition on the sea floor:
=0
z
on z = h ,
(6)
k
= ink , k = 1,2,...,7
n
2.3 Velocity Potential

However, there is no analytical solution for 7 and
j, so the problem should be solved numerically.
According to the 3-D source sink method, the
potentials 7 and j can be expressed in terms of well
known Green functions that can be expressed by the
following equation:
j ( x, y, z) =
Another boundary condition is the kinematics

boundary condition on wetted surface of the floating
bodies. Due to linearization, this boundary condition
may be applied on the wetted surface (S) of the
floating body in its equilibrium position:
(8)
1 N
j ( ,, )G(x, y, z; ,, )ds
4 n=1 s
(9)
Where, (,,) denotes a point on surface S and

(,,) denotes the unknown source distribution. The
integral is to be carried out over complete immersed
surface of the object. The Green function G (source
potential) must in order of the representation in
equation (9) to be valid, satisfy all the boundary
conditions of the problem with the exception of the
body boundary conditions and have a source like
behavior. As a result, boundary conditions are reduced
only to on wetted surfaces of the bodies. So, the
wetted surfaces should be subdivided into panels to
transform integral equations to a system of algebraic
equations to determine unknown source density over
each panel. The appropriate Green function used in
this paper to the boundary value problem posed is
given by Wehausen and Laitone [7]. After getting the
source density, the velocity potentials on each panel
can be obtained using the equation (9).
(7)
1
= i cos( n , x )
n
2
= i cos( n , y )
n
3
= i cos( n , z )
n
4
= i[( y yG ) cos( n, z ) ( z zG ) cos( n, y )]
n
2.4 Forces and Moments
5
= i[( z zG ) cos( n, x ) ( x xG ) cos( n, z )]
n
Once the velocity potential is obtained, the

hydrodynamic pressure at any point on the body can
be obtained from the linearized Bernoullis equation
and can be written as:
6
= i[( x xG ) cos(n, y ) ( y yG ) cos(n, x)]
n
P
1
+ ( ) 2 + + gz = 0
t 2
= 0
n
n
xG , yG , zG = Co-ordinate of the centre of gravity
(10)
Now after putting the value of in the equation

(10), the following expression is obtained,
of the body
x, y, z = Investigating point on the wetted surface
= i +
1
( ) 2 + gz
2
(11)
By neglecting the higher order terms, we can write:
of the body
P = gz + i
223
(12)
As first part of equation (12) is associated with the

hydrostatic and steady forces, so neglecting this part,
the first order wave exciting forces or moments and
oscillatory forces and moments caused by the
dynamic fluid pressure acting on the body can be
obtained from the following integrals:
Fk .e it = ie it 0 + 7 .nk ds
s
Fkj .eit = e it 2 X j j .nk .ds

s
(13)
(14)
Where, Fk denotes the k-th component of wave

exciting forces or moments, Fkj denotes the k-th
component of force arising from the j-th component
of motion of the body. Moreover, it is customary to
decompose the hydrodynamic forces resulting from
motion of the bodies into components in phase with
the acceleration and velocity of the rigid body
motions. These yield the added mass and damping
coefficients respectively. These coefficients can be
expressed from equation as:
akj = . Re j .nk ds
bkj = . Im j .nk ds
S
(15)
3. VALIDATION
To obtain the wave exciting force and motion
responses of moored semi submersible, a computer
program has been developed. The computation model
expected to be validated by the model tests. But since
the tests have not been carried out yet, the results
obtained from computation of a box (Table 1) for
wave exciting forces have been compared with results
obtained from WAMIT-MOSES [6]. Motion results
have been compared with Sujatha and Sonis [9]
results. From these comparisons, it is seen that the
surge wave exciting forces are very good agreement
with results obtained from WAMIT- MOSES (Fig. 2).
On the other hand for roll wave exciting forces, it is
seen that higher wave exciting moments are obtained
in resonance frequency region (Fig. 3). In the
resonance frequency range difficult to compute the
forces. Fig.4 shows comparison of pitch motion of
another box (Table 2) in reference [9]. In figure 4
characteristic dimension of the structure (a) is
considered 90m. Overall, a very good agreement has
been obtained in all the cases.
Table 1. Principal Particulars of the BOX [6]
(16)
L (m)
B (m)
T (m)
Displacement Ton
2.5 Equation of Motion
(M
j =1
kj
Surge Wave Exciting Force

Wave Heading 45 Deg
2,000
Wamit
MOSES
Present Study
1,500
Fx/a
After solving the above exciting forces, added mass

and damping coefficients, the motions of two ships
can be solved by the following coupled equations of
motions. To describe the motion of the multiple
floating bodies, two co-ordinate systems, one fixed to
the body and the other fixed to the space have been
introduced. The two co-ordinate systems are shown in
figure 1. The equation of motion is expressed by the
time varying relation between these two co-ordinate
systems. The equation of motion will be coupled
dynamically because of hydrodynamic interaction and
mechanical connections between them. So the
equation can be considered by using the following
matrix relationship:
200
40
28
229640
1,000
500
0
0
12
16
20
24
28
32
36
40
44
T(s)
Figure 2: Comparison of Surge wave Exciting force.
+ akj ) X&& j + bkj X& j + CX j = Fk ,
Table 2. Principal Particulars of the BOX [9].
k=1, 2 6, j=1, 2 6
(17)
Where,
Mkj=inertia matrix in k mode due to the motion in j
mode
akj=added mass coefficient matrix of kj
bkj=damping coefficient matrix of kj
C=hydrostatic restoring force coefficient matrix of kj
Xj=vector containing the three translational and
three rotational oscillations about the co-ordinate axes
in j - mode.
The suffixes k, j=1, 2, 3, 4, 5, 6 represent surge,
sway, heave, roll, pitch and yaw modes, respectively.
224
L (m)
90
B (m)
90
T (m)
20
Displacement (Ton)
166050
CG(x,y,z) (m)
0, 0, 8.82
Gyration Radii(xx,yy,zz) (m)
37.32, 33.30, 40.08
GMT=GML(m)
14.93
Table 3. Principal particulars of the Semi

submersible.
Roll Wave Exciting Force

Wave Heading 45 Deg
70,000
50,000
Mx/a
Pontoon length (m)

Pontoon depth (m)
Pontoon beam (m)
Pontoon centerline separation (m)
Column longitudinal space (centre)
Column diameter (m)
Draft (m)
Water depth (m)
Number of Columns
Wamit
MOSES
Present Study
60,000
40,000
30,000
20,000
10,000
0
0
12
16
20
24
28
32
36
40
44
T(s)
66.78
6.3
13.3
45.15
45.58
10.59
16.73
175
4
Table 4. Mooring Specification.
Figure 3: Comparison of Roll wave Exciting

Moment.
No
.
1
2
3
4
Pitch Motion of Box (90x90x20m) =180

14
x1
y1
z1
x2
y2
z2
120
120
-120
-120
120
-120
-120
120
6.3
6.3
6.3
6.3
25
25
-25
-25
29
-29
-29
29
-14
-14
-14
-14
Stiffness
(KN/m)
206
206
206
206
Pitch RAO
12
Present Study
10
Wave Exciting Force of SEMI =180
Refrence [9]
0.30
0.25
Surge
0.20
Fx/ a
2
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
^2a/g=ka
0.15
0.10
0.05
Figure 4: Comparison of Pitch Motion.
0.00
5
10
15
20
25
30
T(s)
Figure 5: Surge Wave Exciting force
225

0.30
0.25
Fz/ a
0.20
Heave
0.15
0.10
0.05
0.00
5
10
15
20
25
30
T(s)
Figure 6: Heave Wave Exciting force

0.30
0.25
M y/ a
Computations have been carried for wave exciting

forces and motion responses of a semi submersible
(Table 3) moored by four connector with stiffness of
206 KN/m (Table 4). In this study the wave exciting
forces and motions of a moored semi submersible are
plotted against wave period in Figures 5 to Figure 10,
at a head sea and water depth of 175 meter. Figure 5,
shows non-dimensional surge wave exciting force that
drops to 0.02 at 7.4s period, increases dramatically at
11s, then decreases smoothly at 30s. In Figure 6,
heave wave exciting force, reaches to 0.27 at 9.2s,
falls down rapidly to zero at 15.5s, and then rises
slowly to get 0.17 at 30s. Depends on wave period,
sometimes wave passes directly through under the
structure and excites a little. Wave exciting moment
on pitch, Figure 7, decreases slowly from 0.06 at 8.4s,
reaches zero at 20s and does not change until 30s.
Figure 8, grows up dramatically and gets peaks
13.8 and 16.4 respectively at 12.2 and 16s and reaches
0.3 at 13.8s. At the other periods surge motion falls
down. There is a coupling between pitch RAO (Figure
10) and Surge RAO. This motion peaks at the same
periods to get 24.5 and 41. Heave motion has constant
amounts of 0.5 until 15s, falls to 0.1 at 15.5s, rises up
to 2.7 at 16.5s sharply and falls down to 0.9 to stay
constants at the bigger periods.
0.20
0.15
Pitch
0.10
0.05
0.00
5
10
15
20
25
30
T(s)
Figure 7: Pitch Wave Exciting Moment
20.0
REFERENCES
Surge Motion of Moored SEMI =180
[1]
16.0
RAO
12.0
[2]
Surge
8.0
4.0
[3]
0.0
5
10
15
20
25
30
T(s)
[4]
Figure 8: Surge Motion of Moored SEMI

[5]
Heave Motion of Moored SEMI =180

20.0
RAO
15.0
[6]
Heave
10.0
[7]
5.0
0.0
5
10
15
20
25
[8]
30
T(s)
Figure 9: Heave Motion of Moored SEMI

[9]
Pitch Motion of Moored SEMI =180

50.0
RAO
40.0
Pitch
30.0
20.0
10.0
0.0
5
10
15
20
T(s)
25
30
Figure 10: Pitch Motion of Moored SEMI
5. CONCLUSION
A method for and results of computational
hydrodynamic studies of wave exciting forces and
motion responses of a moored semi submersible have
been presented. Wave exciting forces lead to motion
of floating/moored structure, which has significant
influence on loading and unloading operation. In this
paper, the model is validated only with published
results but it needs to be validated by model
experiment.
Also computations need to be carried out for
various depths and different incident angles. However,
the program developed for computation of wave
exciting forces and motion responses for a freely
floating and moored semi submersible numerically
expected to be able to predict satisfactorily.
226
John L. Hess and A.M.O. Smith (1964).

Calculation of Nonlifting Potential Flow About
Arbitrary 3D Bodies. Journal of Ship Research
Oortmerssen
G.
(1979).
Hydrodynamic
Interaction between Two Structures Floating in
waves. Proceeding Boss79, Second International
Conference of Behavior of Offshore Structures,
London.
S.Wu, J.J.Murray and G.S.Virk. (1997). The
Motion and Internal Forces of a Moored Semi
Submersible in Regular Waves. Journal of Ocean
Engineering Vol. 24, No. 7
O. Yilmaz and A. Incecik (1996). Extreme
Motion Response Analysis of Moored Semi
Submersibles. Journal of Ocean Engineering Vol.
23,No. 6
Muhittin Soylemez (1998). Motion Response
Simulation of Twin-Hulled Semi Submersible.
Journal of Ocean Engineering Vol. 25, No. 4-5
WAMIT-MOSES
Hydrodynamic
Analysis
Comparison
Study.
www.ultramarine.com/hdesk/document/wamit_m
oses.pdf
Wehausen, J. V. and Laitone (1969). Surface
Waves. Handbuch der physik, 9 (Springer-Verlag,
Berlin).
Gunther F. Clauss et al (2002). Time-Domain
Investigation of a Semi Submersible in Roguest
Waves.
Proceeding of OMAE 2002, 21
International Conference on Offshore Mechanics
and Arctic Engineering, June 23-28, Oslo,
Norway.
S. Sujatha and R.Y. Soni (1997). Wave Induced
Response of Floating Structures Using BEM.
Proceeding of the Conference on Ship and
Offshore Technology (SHOT97), March 5-7,
1997. Indian Institue of Technologh, Karagpar.

PREDICTION OF MANOEUVRING BEHAVIOUR OF AN OFFSHORE SUPPLY

VESSEL BY USING SIMULATION PROGRAM
Che Wan Mohd Noor1, Abdul Majeed Muzathik1, Wan Mohd Norsani Wan Nik1
Mohammad Fadhli Ahmad1, Khalid Samo1 and Rajoo Balaji2
1
Department of Maritime Technology

Faculty of Maritime Studies and Marine Science
University Malaysia Terengganu
21030, Kuala Terengganu
Malaysia.
Akademi Laut Malaysia

PO Box 29, Wakaf Tengah
21030, Kuala Terengganu
Malaysia.
ABSTRACT
Manoeuvring ability of Offshore Supply Vessel (OSV) is a very critical aspect. An early prediction of vessel
behaviour will definitely help to improve upon the design. The regular methods available for manoeuvring
prediction such as free running model test, captive model test etc., are found to be expensive and time
consuming. As an alternative, the current approach tries a numerical simulation method with parameters
determined from a database. This study presents the manoeuvring prediction of an OSV which includes the
development of time domain simulation program by using Matlab Simulink software. Three degrees of
freedom were considered and applying the Newtonian laws, the equations of motion were framed. Further,
forces on hull, forces and moments induced by propeller and rudder were also taken into reckoning. Results
were obtained with inputs of vessel speeds, engine revolutions etc. Validation of the prediction results was
also carried out by comparing the results with full-scale sea trial data. The prediction results show a good
agreement with the sea trial data. Applying approximate numerical formula for manoeuvring prediction is seen
to be a reliable and economic prediction tool at early design stages of such vessels.
Keywords: Manoeuvring prediction, Manoeuvring characteristics, Offshore Supply Vessel
1. INTRODUCTION
Manoeuvring characteristic of an Offshore Supply
Vessel (OSV) is a very important aspect as her
mission of operation demands high manoeuvrability
in order to make sure that it can be operated in
various situations and locations. In harbour and
offshore operations and tow and tug activities the
risk of collision and grounding are high for such
vessels. A successful manoeuvring is interpreted as
the ability of the ship to go anywhere, from straight
ahead without any rudder action to tight turning
with significant rudder action.
In general, low speed vessels with high block
coefficient such as OSVs are known to have bad
manoeuvring characteristic because of the full hull
form with small length to beam ration [1]. The ship
particulars of OSV are shown in Table 1.
In December 2002, International Maritime
Organization (IMO) has adopted the Resolution
MSC.137 (76), Standards for Ship Manoeuvrability.
These standards were developed to ensure safe
operation of ships at sea. In order to comply with the
IMO manoeuvring standards, the ability to predict
the manoeuvrability of OSV at the design stage is
important. The fact that it is too late to effect design
changes after the vessel has been built makes such
predictions imperative. The complexity in analysing

manoeuvrability characteristics is due to many
variables.
Table 1: Ship particulars
Particulars
Vessel
LOA
60.8 m
LBP
54.0 m
14.8 m
5.7 m
4.6 m
Displacement
2373 tonnes
Speed
12.5 knots
Block coefficient
0.705
Prismatic coefficient
0.72
Firstly, the area and shape of the rudder may be

mentioned. The flow around the rudder is affected
by a combination of variables such as the eddy
currents created by ships motion and the propeller
races.
Being located partially or as a whole, the rudder
experiences basic hydrodynamic phenomena of
stall, cavitation and aeration. During the stall,
separation occurs resulting in reduction of normal
flow pressures. These factors are difficult to predict
227
as they aree governed by different laws of similitude

and furtheer affected by the ships turrning actions.
Secondly, the steering action
a
of the ship itself is
b many facto
ors. The hull section just
affected by
forward of
o the propelller affects thhe directional
stability. Larger
L
the areaa of this sectio
on, greater is
the resistannce to turning. In the case of OSVs, where
greater maanoeuvrability is desired, th
hese areas are
reduced. Factors beyond the control off design could
be the deptth of water, nuumber of propeellers, relative
position off rudders and propellers.
p
Lasttly, the forces
on the shhip while the turning occuurs affect the
manoeuvraability. The iniitial turning phhase causes an
inward heeeling which may
m change to outward heel
due to heeeling momennts caused byy the forces.
During thee final steady tuurning phase, thhe position of
the rudderr is important in correct the heel and the
ship-keepin
ng as well.
2. IMO MANOEUVR
M
RING STANDA
ARDS
The problem
p
of ship manoeuuvrability hass
consideraably grown inn the last deccade, both forr
merchant and naval shipps. With regardds to merchantt
ships,
The
Interim
m
Standardss
of
Shipp
Manoeuvrrability A.751 (18) was adoppted finally ass
IMO Resolution MSC.1137 (76), Stanndards for Shipp
Manoeuvrrability in 76thh MSC of IMO
O in Decemberr
2002. IM
MO establishedd minimum manoeuvrability
m
y
standards to ensure safetty of all seagoiing ships.
The curreent IMO standaards for ship manoeuvrability
m
y
contain fo
our distinct crriteria: turning ability, initiall
turning ab
bility, stoppingg ability, yaw
w-checking andd
course-keeeping abilitiess. These criteriia are listed inn
Table 2 while
w
the manoeuvring indicees are shown inn
Figure 1.
Tab
ble 2: IMO Evaaluation criterria in final
sta
andards for sh
hip manoeuvra
ability [2]
While measuring
m
the manoeuvring performance
of vessels,, the regular tests would be
b on turning
circles, pulll-back manoeeuvres, zig-zag
g manoeuvres,
changing of
o vessels head
ding, and turniing / stopping
/ direct-revverse tests. IMO
O standards prrovide criteria
on the shiip turning ability, yaw checcking ability,
course keeeping ability and
a stopping ability.
a
Of the
several meethods availablee to assess the manoeuvring
criteria of a vessel, sea trrial and free ruunning model
tests are the straightforrward methodds which are
regularly employed. Hoowever these methods are
expensive.
Ability
Turning
ability
Initial
turning
Ability
Stopping
ability
Simulatiion methods coould also be ussed to analyze

manoeuvriing criteria eiither by incorrporating the
hydrodynaamic coefficien
nts obtained from captive
model tests or empiriccal formulae. The captive
urs heavy exxpenditure in
model tesst often incu
equipment, testing time and cost. In practice,
p
both
time and costs
c
are limiteed in the early design stages
and so thee execution of
o extensive model
m
test on
every shipp is practicallly not achievvable. As an
alternative, nowadays, the numericaal simulation
method wiith parameters determined froom a database
is the prinncipal approacch employed. Further, the
reliability and accuracy of
o numerical simulation
s
are
required to
o be validated with
w full-scale sea trial data.
The estab
blishment off cheaper and
a
accurate
manoeuvriing prediction tools based on
o numerical
approach is essentially im
mportant in ord
der to comply
m
s
standards.
with IMO manoeuvring
Test
Turning
test with
max.
Rudder
angle
(35o.)
10/10
Z-test
Criteria
Advance <4.5 L
Tactical Diameeter <5.0 L
Stopping
test with
full
astern
10/10
Z-test
Track reach < 15L
Coursekeeping
and
Yawchecking
ability
20/20
Z-test
Distance ship run

r before 2nd
rudder execution < 2.5L
1st Overshoot
<10
(L/U<10s)
<(5+0.5 L/U) (10s<L/U<30s)
<20
(30s<L/U)
2nd Overshoot
(L/U<10s )
<25
L/U))
<(17.5+0.75(L
(10s<L/U<30s)
<40
(30s<L/U)
1st Overshoot. <25
This stuudy describes a manoeuvrin

ng prediction
study baseed on numericaal simulation method
m
which
did not reequire a seriess of model tannk tests. The
objective of
o this study is to predict manoeuvring
performancce of an Offsh
hore Supply Vessel
V
(OSV)
by using numerical appproach simulattion program
mulation results with the fulland then vaalidate the sim
scale sea trrial data.
Figure 1:
1 Manoeuvrinng Characteristiic Indices for
Turningg Ability [3]
228
The mathematical model for manoeuvring motion
can be structured from the equations of motion with
reference to the co-ordinate system, whose origin is
the ships centre of gravity as shown in Figure 2.
x0
u, X
r, N
The superscript {'} in the equations refers to the

non-dimensional quantities.
CG
v, Y
y0
Figure 2: Co-ordinate System [4]

As shown in Figure 2, (U) is the actual ship
velocity that can be resolved into advance velocity
(u) and transversal velocity (v). The ship has also a
rotation velocity with respect to the z-axis. This axis
is normal to the XY plane and passes through the
ship centre of gravity (C.G). () is the angle between
U and the x-axis and it is called drift angle. () is
the ship heading angle and () is the rudder angle.
In this study, manoeuvrability will be approached
as a bi-dimensional phenomenon. Two reference
systems will be used, one of them fixed (Xo, Yo) and
the other moving with the ship, with its origin at the
centre of gravity. Yaw motion is supposed to occur
around this point. In the moving reference, X-axis is
positive forward and Y is positive starboard. For
both systems, moving and fixed, angles are positive
in the clockwise sense.
3.1
Equations of Motion
Once the reference systems have been defined, the

ship is considered as a solid with three degrees of
freedom: surge, sway and yaw. In accordance with
these three degrees of freedom, the Newtons
Second Law equation is applied to the moving
reference system of Figure 2 for each motion [4].
Interrelating both axes and differentiating the
Newtonian equations of motion, the dynamical
model which governs the motion of the ship, can be
simply written as follows:
L U&
(m + mx ) cos & sin + m + my r sin = X
U U
L U&
m + my sin & cos + (m + mx )r cos = Y
U U
2
U
L U&
+ J zz
) r + r& = N
(I zz
L
U L
where: X, Y: Dimensionless Surge and Sway force

acting on a ship, N: Dimensionless Yaw moment
acting on a ship, m, mx my: Dimensionless mass of
ship, and added mass in x- and y-directions, Izz:
Dimensionless moment of inertia of ship in z-axis,
Jzz: Dimensionless added moment of inertia of ship
in z-axis, : Drift angle at the center of gravity C.G.
[ = - sin-1(v/U)], r : Dimensionless Turning rate [
r = d/dt ], L, T, U : Ship Length, Ship Draught and
Ship Speed respectively.
(1)
The left hand-side of equations (1) represents the

inertial terms. The right hand-side represents the
external forces and moments that act on the ship.
These forces and moments can be described
separately into the following components from the
viewpoint of the physical meaning. The subscripts
H, P, and R symbolize ship hull, propeller, and
rudder respectively according to the concept of
MMG [5] [6].
X = XH + XR + XP
Y =YH + YR
(2)
N = NH + NR
3.2
Forces and Moments Acting on Hull
XH, YH and NH are approximated by the

following polynomials in terms of and r' at the
amidships. The coefficients of the polynomials are
called hydrodynamic derivatives. The longitudinal
component of hydrodynamic force (XH), the lateral
force (YH), and yaw moment (NH) acting on the
ship hull are expressed as follows [4]:
XH = { Xr r sin + Xuu cos2 }
YH = { Y + Yr r + Y + Yrr rr+
(Yr + Yrrr) r }
NH = { N + Nr r + N + Nrr rr+
(Nr + Nrrr) r }
(3)
XH is consist of hydrodynamic derivates which
expresses the change of ship resistance due to drift
angle = tan-1(-v/u), yaw rate r, and Xuu which is
the ship resistance in forward straight motion. Y,
Yr , ...., Nr are also the hydrodynamic derivatives.
The Kajima formulae are based on the functions
of ship length (L), breadth of hull (B), draught (d),
block coefficient (CB), form factor (K) and aspect
ratio of ship hull k (k = 2d/L).These formulae were
derived based on a database involving 15 kinds of
ships (container, bulk carrier, tanker, general cargo,
VLCC, car-ferry etc.) and their 48 loading
conditions [7]. The formulae of individual
hydrodynamic derivatives for sway and yaw motion
can be deduced to following equations:
229

X r = (1.66CB
C + 1.5)my
4. SIMU
ULATION PRO
OGRAM
X uu = RT /(0.5LTU 2 )
Y = 0.5k + 1.9257(CB / L ) a ,
The vessels swept ppath was obtainned by doublee

integratinng the acceleraation mathemaatical equationn
in surge, sway
s
and yaw as shown in eq
quation (7).
Yr (m + mx ) = 0.25k + 0.052ea 0.457,

Y = 1.199CB
C a + 1.05,
Yrr = 0.225(dC
CB / B)ea 0.12,
Yrr = 7.1256{d (1 CB) / B},
x=
Y r = 10.443[{{d (1 CB) / B}ea ]2 9.374{d (1 CB) / B}ea + 1.227,

N = k[(150.6668{d (1 CB) / B}ea K ) 2 23.819{d (1 CB
C ) / B}ea K + 1.802],
]
= 43.857{d (1 CB) / B}ea K ) 2 3.671{d (1 CB) / B}ea K + 0.086],

N
N rr = 0.4086
6CB + 0.27,
r = 0.826{d (1 CB) / B}ea 0.026
N
(4)
Forces & Mom

ments Induced by Propeller
The hyydrodynamic forces generaated by the

propeller are
a expressed as below [8][9
9], neglecting
lateral forrce and yaw moment thatt affects the
stopping manoeuvre:
m
XP =[ Ctp ( 1 tP ) n2 DP4 KT ( JP )]/0.5
5LdU2
YP = 0
NP = 0
K T ( J P ) = C1 + C 2 J P + C3 J P
J P = U cos (1 wP ) / (nDP )
t1
t2
t1
X m + m r sin
U
y
&
+ sin
n
dt
cos
(m + mx ) L
(7)
N r = 0.54k + k 2 0.0477ea K + 0.0368,

N rr = 0.15K 0.068,
3.3
t2
Simulatioon program was written in graphicall

programm
ming languagge and com
mpiled usingg
MATLAB
B Simulink coompiler. The programs aree
developedd in three maain blocks wh
hich are inputt
block, pro
ocess block andd output block..
In inpu
ut block, the pprogram will read
r
the inputt
data suchh as rudder angle and hydrodynamicc
coefficiennts. These inpuut data will th
hen be used inn
the moduule process block
b
to calcuulate the hull,,
rudder an
nd propeller forces. The layouts of thee
process block
b
module aare shown in Figure
F
3. Hulll
modules are
a divided into three sub-bloocks which aree
called surrge, sway and yyaw sub-block. The propellerr
and ruddder forces aree also calcullated in theirr
respectivee sub-blocks.
(5)
where: tp: thrust reducction coefficieent in straight

forward moving,
m
Ctp: constant, N:
N propeller
revolution,, Dp: propelleer diameter, wp: effective
wake fraction coefficiennt of propelleer in straight
running, KT: thrust coeffficient of a prropeller force,
Jp : advannce coefficientt, C1, C2, C3: constants for
propeller open
o
characteriistics.
3.4
Forces and Moments Induceed by Rudder
The hyd
drodynamic forrces generated by rudder are
described below [8] [9],, in term of ru
udder normal
force FN, rudder ang
gle and rudder-to-hull
r
interactionn coefficient tR, aH and xH ( = xH/L).
XR = (1 tR )FN sin
YR = (1 + aH )FN cos
NR = ( xR + aH xH )FN cos
Figure
F
3: Layoout of Process Block
B
The equuation of motiion will be douuble integratedd
to obtainn the translatiion of motionn in x and y
directionss. Lastly, the output
o
results of ships pathh
are displaayed in graphiccal form as sh
hown in Figuree
4.
(6)
where: xR: The distan

nce between the
t centre of
gravity of ship and centree of lateral forcce, ( xR = x'R .
L ) & xR represents
r
the location of ruudder ( = L /
2), xH : Th
he distance beetween the cenntre of gravity
of ship andd centre of lateeral force, ( xH = xH .L ), :
Rudder an
ngle, FN : Dim
mensionless ruudder normal
force, tR : Coefficient
C
forr additional draag, aH : Ratio
of addition
nal lateral forcee.
Figure 4: Outpuut result of Shipps Path
230
5. RESUL
LTS AND DIS
SCUSSION
5.1
Siimulation
Simulatiion for Turnning Circle and

a
Zig-Zag
manoeuvree was run at constant
c
ship speed
s
of 12.5
knots. Thee propeller waas assumed too be working
ahead withh constant RPM
M, neglecting the transient
behaviourss of the mainn engine causeed by torque
limit, goveernor and so on.
o The prediccted result of
Turning Circle to Port and
a Starboard are
a plotted in
a 6, respectiv
vely.
Figures 5 and
predicted results of Z
Zig-Zag 10/10 and Zig-Zagg
20/20 aree recorded in T
Table 4. It sho
ows that the 1stt
nd
and 2 overshoot anglee of Zig-Zag 10/10
1
and Zig-MO criteria. Thus,
T
the OSV
V
Zag 20/20 meet the IM
had a goood yaw checkinng ability. Thee details of thee
predicted time historiees simulationn for Zig-Zagg
d Zig-Zag 20//20 are plotted
d in Figures 7
10/10 and
and 8, resspectively.
Table 3:
3 Prediction results of Turrning Circle
(Port)
Turning
((Starboard)
MO
IM
Crriteri
a
Result
Advance
Distance
3.1 L
(188m)
3.2 L
(194m)
4 L
< 4.5
Comply
Tactical
Diameter
2.7 L
(164m)
2.8 L
(169m)
5 L
< 5.0
Comply
Parameter
Turning
Table 4: Prediction reesults of Zig-Z

Zag 10/10 and
220/20
Zig-Zag 10/10
Zig-Zag 20//20
Parameter
Figuree 5: Prediction result of Turniing Circle

Trajectoories to Port
Figuree 6: Prediction result of Turniing Circle

Trajectoriees to Starboard
Resultt
Predi.
IMO
Predi
I
IMO
1st
overshoot
(deg)
6.5
10
15.7
25
Comply
2nd
overshoot
(deg)
10.
6
25
23.6
Comply
Figurre 7: Predictionn result of Zig--Zag 10/10
Table 3 shows the sum

mmary of predicted results
for turninng circle to port and staarboard. The
Advance Distance
D
(Ad) and Tactical Diameter
D
(Td)
for Port an
nd Starboard turning
t
are fouund to satisfy
IMO criterria. Turning manoeuvre
m
is carried
c
out to
determine the turning abiility and the reesponse of the
ship to thee deflection off rudder at 35o angle. Thus,
the results show that OSV
V has a good turning
t
ability
based on IM
MO criteria.
On the otther hand, thee Zig-Zag maanoeuvre was
carried ou
ut for purposee of yaw cheecking ability
which reprresents the inhherent effectivveness of the
two rudderrs in making ch
hanges of ship heading. The
Figurre 8: Predictionn result of Zig--Zag 20/20
231
5.2
Validation with Full-Scale Sea Trial
A full-scale manoeuvring sea trial of an OSV was

carried out and the ship path data was recorded by
using FURUNO GPS/WAAS Navigator model GP32. Ship heading angle are displayed on magnetic
compass which are mounted onboard. Sea trial data
for Turning Circle to Port and Starboard are
recorded in Table 5 and 6 respectively. It shows that
the sea trial results are smaller than predicted value
for Advance Distance and Tactical Diameter.
The discrepancies between predicted and
measured results can be shown in terms of deviation
percentage which was calculated as follow:
Deviation (%) =
(Pr ediction Measuremen t )

X 100
Measuremen t
(8)
Table 5: Comparison between Prediction and Sea

trial results for Turning Circle to Port
Turning
(Port)
Prediction
Sea
trial
Deviation
(%)
188
174
8.0
164
160
2.5
Advance
Distance
(Mtr)
Tactical
Diameter
(Mtr)
Table 6: Comparison between Prediction and Sea

trial results for Turning Circle to Starboard
Turning
(Stbd)
Prediction
Sea
trial
Deviation
(%)
194
176
10.2
169
157
7.6
Advance
Distance
(Mtr)
Tactical
Diameter
(Mtr)
The deviation percentage of Advance Distance

and Tactical Diameter for Port Turning Circle was
8.0 % and 2.5 % respectively. Meanwhile the
deviation percentage for Advance Distance and
Tactical Diameter for Starboard Turning Circle was
10.2 % and 7.6 % respectively. The predicted results
for Port Turning Circle are more accurate than the
Starboard Turning Circle with the smaller deviation
percentage.
6. CONCLUSION
Manoeuvring assessment of an Offshore Supply
Vessel had been successfully performed by using
numerical simulation method which was developed
on Matlab Simulink software. The prediction result
shows that manoeuvring characteristic of an OSV
meet the requirements stipulated by IMO resolution
MSC.137 (76), International Maritime Organization
standards for ship manoeuvrability. Validation with
full-scale sea trial data gives small deviation to the
predicted result. This discrepancy may due to
inaccuracy of estimated hydrodynamic derivatives

and others external forces which are not taken into
account. However, the numerical method developed
in this study can be used as a primary tool in order
to access OSV manoeuvring characteristic at early
design stages. This method may be considered as an
economical and reliable manoeuvring prediction
method which can estimate manoeuvring
characteristic without relying on model tests
(captive or free running test). Although the results of
prediction tools show a good agreement to the sea
trial data, the authors suggest that for future work,
the study should be continued with further
validation process with other full-scale sea trials
data and captive model test data.
REFERENCES
[1] Lee H.Y., Shin S.S (1998). The Prediction of
Ships Manoeuvring Performance in Initial
Design Stage. Hyundai Maritime Research
Institute, R&D Division, HHI, Ulsan, Korea.
[2] International Maritime Organization (2002).
Standards for Ship Manoeuvrability, Report of
the Maritime Safety Committee on its SeventySixth
Session-Annex
6
(Resolution
MSC.137(76)), London, UK: International
Maritime Organization.
[3] Lewis, E. V. ed. (1989). Principles of Naval
Architecture, Volume 3, Jersey City, USA:
SNAME.
[4] Kijima, K. (2003). Some Study on the
Prediction
for
Ship
Manoeuvrability.
Proceeding of the International Conference of
Ship Simulation and Ship Manoeuvrability,
Kanazawa, Japan.
[5] Ogawa, A., Kasai, H. (1978). On the
Mathematical Model of Manoeuvring Motion of
Ship. ISP, Vol. 25, No. 292.
[6] Kose K. (1982). On a Mathematical Model of
Manoeuvring Motions of a Ship and its
Applications.
International Shipbuilding
Progress, Rotterdam, Netherlands, Vol. 29, No.
336. August 1982.
[7] Kijima et. Al. (2000). On the Prediction
Method for Ship Manoeuvrability. Proc. Intern.
Workshop on Ship Manoeuvrability. Hamburg
Ship Model Basin, Germany.
[8] Kijima, K., Yasuaki, N. and Masaki, T. (1990).
Prediction Method of Ship Manoeuvrability in
Deep and Shallow Water. Proceedings of the
Marsim & ISCM 90 Conference. June 4-7.
Tokyo, Japan.
[9] Kijima, K. and Tanaka, S. (1993). On the prediction
of ship manoeuvrability characteristics. Proceeding
of the International Conference of Ship Simulation
and Ship Manoeuvrability. September 15-17.
London.
232

VARIOUS ASPECTS OF SOUND SHIP RECYCLING IN SOUTH ASIA:

A COMPROMISE NOT A CONFRONTATION!
Arun Kr Dev1
1
School of Marine Science and Technology, Newcastle University, United Kingdom

Newcastle University Marine International (NUMI) - Singapore
E-mail: a.k.dev@ncl.ac.uk
ABSTRACT
The significance of ship recycling may not even be lesser than shipbuilding. The recent studies reveal that the
global disposal of ships touches the number of 7 billion LDT every year. Total percentage of ship recycling in
South Asia amounting is almost more than 80% of the total figure in the world. The ship dismantling yards are
not always well established. So, naturally it calls for the safety, health and environment (SHE) aspects and they
are of increased attention in these industries. Thus it is a cause of concern for the world in terms of workers
plight in this so-called dangerous industry. On the contrary, socio-economic state and level of living standard in
the other factors are contributing to the current condition of ship dismantling industries. Low economic strength
and unemployment problem leads to the lower rate of scrapping prices in these countries. The current market
organization for disposing of obsolete or damaged ships does not favour improvements. Dismantling the ships in
a responsible way would, under current market conditions, increase the price that the scrapper needs from the
steel industry to a level, which will make ship dismantling uncompetitive, or reduces the price of a ship an
owner gets for his end-of-life ship. A techno economic analysis may prove helpful to reduce the impact of the
problem. Identifying problems, analyzing them from various points of view, planning a long term strategy and
finally a compromise among all parties is the way forward for these very delicate and sensitive issues related to
ship recycling industries in South Asian countries mainly Bangladesh, India and Pakistan. If the problems
rather challenges can properly be identified and the solutions be found, an environmentally sound and safe
recycling system can be established, although the current recycling condition is unfavourable for both people
involved and the environment. And hence a safe & environmentally sound recycling of ships will be possible
which will be economically acceptable.
Keywords: sound ship recycling, ship recycling facilities, safety, health, environment, national and international
legislation, socio-economic impacts, LDT, South Asia, ship dismantling, ship disposal.
1. INTRODUCTION
Ship scrapping is a function of breaking up ships
for further use in some forms. It is getting increased
significance as the number of vessels to be disposed
off is increasing. Every year, the shipping industry
sends about 600 end-of-life ships to disposal yards.
The present paper exposes the theoretical
background of the ship recycling problems and their
solution matrix, analyzing different existing disposal
methods with their techno-economical advantages and
disadvantages. The current condition of ship recycling
is not favourable for both health and safety of the
people involved and for the environment. It is evident
that frequent death occurs either from occupational
disease from ship borne hazardous substances like
asbestos or PCBs, or from explosions, fires and other

accidents.
Bangladesh is currently the largest ship dismantling
country, having processed more than 60% of all large
ocean-going vessels in 2007.
The environment is being affected very highly by
the ship dismantling activities. Old ships contain
many toxic substances, asbestos, pollutants and
hazardous materials. In the ship dismantling yards in
Asian countries there are no waste treatment system.
The toxic substances are being released into the
environment.
For the purpose of overcoming these problems, it is
now necessary to find the reasons and the main bottle
233
necks of the problem. One of the main reasons might

be the technical unavailability to deal with the
situation in a responsible manner. Economic factors
associated with these industries may be another
reason for irresponsible ways of recycling world fleet.
2. BACKGROUND INFORMATION AND

PROBLEM DEFINITION
2.1 Background
Dismantling or recycling of ocean going merchant
ships has come under increased scrutiny of
international organizations such as IMO [1], ILO [2],
UNEP [3] (Basel Convention [4]). The main reason
for this is that dismantling normally takes place under
difficult conditions, both for the people involved in
the actual scrapping process as well as for the
environment which suffers from toxic residues which
emerge and escape during this process.
The current market organization for disposing of
obsolete or damaged vessels does not favour
improvements in both the situation of men and
environment. The latter will provide an incentive to
avoid clean and safe ship breaking. Thus the
additional costs of clean and safe operations form the
main bottleneck is in its introduction.
2.2 Problem Definition

Ships are not being dismantled as much as
expected, so there are increasing numbers of old
ships. One view is that the tonnage of vessels
available for recycling will match actual world
recycling capacity estimated at 12 million LDT per
year.
Health and Safety
The ship recycling industry is one of the most
deadly in the world. Workers dismantle the ships with
their bare hands. Ships are scrapped by poor workers.
Working condition is unsafe. They are unprotected
from toxic substances, explosions, falling objects and
working condition hazards and accidents. Machinery
parts, oil rags, open fires and leaking barrels are all
over the beach including asbestos. Workers with torch
cutters dismantle the huge sea-carriers.
Environmental Pollution
Old ships contain a lot of toxic substances and other
hazardous materials. The vessels are broken on Asian
beaches and along rivers. The toxic substances are
released into the environment. The soil, sea, rivers and
air are heavily polluted.
2.3 State of Industry

Existing Recycling Procedures
There are no recycling facilities at all in the South
Asian countries. Pristine beaches are being used to
break the huge vessels. No docking facilities.

Ships are dismantled by poor workers earning
between US $1.5 and 2.5 a day. They break the ships
with torch cutters and their bare hands - unprotected
from toxic substances, explosions and falling steel.
Well over 100,000 men and women work at ship
dismantling yards worldwide. They are desperate for a
job. Unknowingly they dry asbestos plates in the sun
and sell the pieces. Even in their sleeping quarters the
toxic substances are all over the place.
Most of the ships dismantled today were built in the
1970s. That is to say prior to the banning of many
hazardous materials. On the Asian beaches, these
toxic substances are released into the environment and
the workers bodies.
2.4 Typical Hazards

Fire and Explosion
In recycling yards in Asia rarely a ship is cleaned
properly or gas freed before it is dismantled. The
workers are not trained and aware of fire and
explosion.
Toxic Substances
In Europe hazardous substances are subject to
special monitoring. The disposal is highly regulated.
But at the ship breaking yards of Asia waste is burned
in the open area. They cause fumes with very
dangerous dioxins and PAHs (Polycyclic Aromatic
Hydrocarbons). These are so called Persistent Organic
Pollutants (POPs). The typical toxic materials found
on board ship are POPs (Persistent Organic
Pollutants), PVC, Dioxins, PAHs, Organotins, PCBs,
Asbestos.
Heavy Metals
Shipwrecks are big heaps of toxic scrap', built with
lots of materials that contain toxic substances. Heavy
metals can be found in paints and coatings. They are
also in anodes, insulation, batteries and electrical
compounds.
2.5 Regulatory Bodies

For the better understanding of safe and
environmentally sound disposal/recycling, it is
necessary to review existing regulation. Following are
the regulatory bodies in the subject area.
International Maritime Organisation (IMO)
The IMO is the United Nations body that regulates
the shipping industry.
Green Peace
Green Peace lobby group has a rational view of
recycling but may not always offer practicable or
234
economic solutions. The issue of ship breaking has

been brought to the attention of IMO, and Green
Peace [5] is the leading lobby group in this area.
Often it is dominated by Flag of Convenience (FOC)
countries, whose first interest is to get high numbers
of ships in their ship registry.
Basel Convention
The Basel Convention (1989) was initiated by the
UN Environmental Programme. Under it, exporting
nations and polluting industries have to take care of
their own toxic waste. This promotes a shift to clean
production. The Convention places the onus on the
exporting countries rather than on the importers.
scraped is increasing. Price of steel scrap is also to see

rise due to the increasing demand of steel and steel
price. So it is obvious that the ship recycling activity
in no doubt is existing strong. But the concern is it
needs to be done in a safe way and without harming
the environment which incurs cost. Hence a detailed
economic analysis is necessary to suit the safe and
environmentally sound recycling to the existing
scrapping function.
3. SHIP DISPOSAL
3.1 Where and Why
EWSR
The EU regulates the export of waste in the
European Waste Shipment Regulation (EWSR [6]) or
Council Regulation 259/93/EEC. This incorporates
the Basel Convention.
The ship recycling industry has been historically

footloose industry. The ship recycling industry is
concentrated in a very limited number of countries
and those countries level and state of economic
development makes it an activity that suits their
needs perfectly.
International Labour Organization (ILO)

International Labour Organization is the most
important body to concern about the welfare of
workers, their safety and health. ILO also has its role
in ship breaking. It has published "Safety and health
in ship breaking: Guidelines for Asian countries and
Turkey.
It has in the past moved from Europe to the Far

East, in particular to Taiwan, from where it moved on
to the current four foremost ship breaking nations:
India, Bangladesh, Pakistan and China. These four
countries handle an estimated 85% of the world scrap
tonnage in terms of light ship weight.
2.6 Market Review and Cost Consideration

There is an urgent need to generate ideas for the
development of an economic system that would result
in a safe and environmentally sound system of ship
recycling. An estimation of the costs of
environmentally sound recycling is essential in the
introduction of such a financing system. Scrap price
and demand are the main factors, which influence the
scrap market. Scrap price trend shows that it was
always increasing in the past although there is a down
ward trend in the recent years. Scrap demand in the
dismantling nations is still high in line with the
population growth and other socio-economic change
in condition.
Cost of dismantling is another factor leading to the
decision making process of the scrapers. If the cost of
the total process is higher than the profit gain no one
is interested to step into it, doing it in a proper manner
is far more unrealistic. The total cost of dismantling a
ship can be the sum of operational cost and labour
cost. Labour costs in the Asian countries are relatively
much lower than western countries. And probably due
to the stringent labour, environment and safety
regulation in western countries has stimulated a
geographical shift of recycling activities to the less
developed countries. But labour cost is dominant as
the dismantling process is highly labour intensive.
Future trend of the scrapping activity can be
assumed to be inclining as the number of ships to be
The reasons why ship disposal takes place in those

regions of the world are as follows: 1. Sea access, 2.
Scrap demands, 3. Unemployment problems, 4. Low
cost of dismantling, 5. Lack of safety standards, and
6. Lack of environmental control.
3.2 Significance of the Problem
Ships are not being scrapped as much as expected,
so there are increasing numbers of old ships. The
tonnage of vessels available for recycling will match
actual world recycling capacity estimated at 12
million LDT (Light Displacement Tons) per year.
At the end of the decade (2010) some 4.000 ships
with an aggregate gross tonnage of 24 million should
be recycled every year. The replacement of 4,000
ships per annum with an aggregate gross tonnage of
24 million by the year 2010 may result in a littering
of the oceans, beaches, rivers and ports if these ships
are not recycled in a responsible way [5].
3.3 The Problems Associated with Disposal
The problems arising within the major ship
recycling nations are concentrated around three
themes: Safety, health and Environment.
Safety and Health of People Involved
The most important concern about ship dismantling
is safety of human being involved in the dismantling
process. Hundreds of people at the ship recycling
yards endure hard physical labour. They work in
235
permanent danger. Accidents are very common in the

scrapping yards. On board gasses cause fires and
explosions. Main causes of death are fire/explosion,
hit by materials falling, falls from height, suffocation
and inhaling CO2. The following table shows an
overview of high rate of accidents in the dismantling
process.
Table 1: No. of major and minor accidents in
Bangladesh ship breaking yards
Statement of accidents by ILO in twenty ship
breaking yards in Bangladesh
Year
Minor
Accidents
Major
Accidents
Fatal
Accidents
1996
147
21
1997
201
10
1998
203
11
Total
551
42
The low level of economic development does not

allow the acquisition and maintaining of costly,
energy-consuming mechanised means. Same reasons
hinder the building of new infrastructures (slipways,
graving docks, waste collection, treatment networks
and transport links).
Fire and explosions

Toxic substances
Hazardous materials
POPs, PVCs, Dioxins and PAHs
Organotins, PCBs and Asbestos
Falling from heights and falling objects
Working condition hazards and accidents
Environmental pollution and contamination
Costs
The problem is one of cost Disposal costs are high.
Due to risk associated with the ship recycling and
uncertainty of the business investors are not interested
to invest.
Lack of Rules and Regulations for Implementation
Hazardous waste materials from the ships are
widely distributed. In most ship breaking nation's
proper waste management is absent. There are no
rules and regulations. And if there are, very often
there is a great lack of control.
Role of IMO in Ship Recycling
International treaties can help protect the
environment and the workers' health. In March 2002 it
started to debate its role in ship breaking. IMO
Regulations on Ship Recycling are as follows:
Green Passport
Phase out of single hull oil tanker
Prior cleaning before dismantling
Each of these types varies in relation to secondary

characteristics, infrastructures and outlet markets for
recoverable products. However there are economic
and human aspects which appear to be universal.
This is a labour-intensive process. This process uses

a strong labour force, which is found in the Indian
subcontinent. This type of process is found in a social
and economic environment offering abundant, cheap
labour and within a civil society that is little sensitive
to the social and environmental aspects of this activity
and of other industrial activities in general.
The following hazardous conditions are commonly

observed in a ship scrapping yard
1.
2.
3.
Different dismantling models exist depending upon

the local industrial and economic environment, the
mechanical level, available equipment, infrastructures
and the outlet markets for recoverable products,
secondary raw materials that are collected from
recycling. Basically, three types of dismantling can be
defined globally: 1. The Non-mechanised process [8],
2. The Highly mechanised process [8] and 3. The
Intermediate Process [8].
4.1 The Non-Mechanised Process
Source: ILO Report on ship scrapping [2]
4. CURRENT RECYCLING
The ships are progressively cut into vertical

sections using blowtorches, starting with the front
parts and the ship sections are then gradually hoisted
onto the shore away from tide waters. These sections
are subsequently cut into strips of metal which can
each be carried by a team of twenty or so workers
having no handling equipment, these strips being
sized mostly to meet the requirements of the re-rolling
workshops located in direct, close vicinity of the
dismantling site.
This industrial process starts by grounding the ship
on the shoreline, if possible benefiting from an
exceptionally high tide so as to place the working area
away from the effects of rising and ebbing tides.
Simultaneously with the disembarking of all
equipment and materials which are easy to handle, the
front of the ship is cut with a blowtorch on the beach
itself by a strong labour force. The dismantling of the
ship then progresses in vertical sections towards the
stern of the ship. The sections that have been cut are
brought higher up onto the beach so that cutting can
be completed into smaller parts.
The work in these places is conducted under high
temperature conditions, but the temperatures are not
high enough to the destruction of pollutants which
236
still remain on the scrap metal. The conditions are

similar to forge working conditions. Workers work
with no protection means for their safety and health.
The steelworks are not all equipped with fume
separation or filtering installations capable of trapping
these pollutants. This scrap is treated in the same way
as common land scrap. As a result, this entire nonmechanised process generates high environmental and
social affects which are neglected by the economic
players concerned since they do not directly have to
take these in charge.
4.2 The Highly Mechanised Process

This process uses a small labour force, which is
found solely in Western countries.
This type of highly mechanised process may be in
an economic setting in which labour is expensive, and
in which health, safety and environmental
requirements are important by all economic players
concerned. Dismantling does not require either heavy
investments (with the exception of acquiring and
adapting a site) or highly skilled labour. This labour
does need to have specific experience which is
achieved by several successive ship dismantling
operations. This knowledge is a combination of
industrial demolition techniques and shipbuilding
techniques.
The actual dismantling operation can be initiated
through the rapid use of mechanical means for
horizontal cutting of the ship, so that large metal
blocks can be removed by crane onto a quayside of
the dismantling site. The employment of cutters
armed with blowtorches is just reserved for situations
where large mechanised means are not applicable. The
dismantling process is interrupted whenever necessary
to take general protective measures.
The end non-recyclable wastes are isolated from the
recycled metals in the same way as land scrap treated
in the electric furnaces. The environmental, social and
human reliability of this entire process is based on the
standardised work methods, employing a reduced
number of operators and using equipment which is
able to control environmental and human risks.
4.3 The Intermediate Process

This process uses some equipment but still
maintaining a significant labour force, which can be
found in Turkey, China or even at some dismantling
facilities in the United States.
This type of process, which employs abundant
relatively cheap labour, also uses mechanised
equipment. This model is based on the existence of
low-cost national or immigrant labour. The
dismantling methods are manual, but the partial use of
equipment such as cranes overcomes the need for
heavy handling work and extensive manual cutting of

the ship, thereby reducing risks and the harshness of
work conditions.
Additionally, the existence of specifically built
slipways or wet berths allows the use of standardised,
stable methods allowing better safety of worker and
the environment. This economic model remains viable
in a social environment whose standards of worker
protection and environmental protection are close to
European standards (Turkey, USA). In this type of
dismantling process, the general location and
infrastructure is almost same as in highly mechanised
Western dismantling sites. This type of process tends
to treat pollutants and hazardous substances in order
to reduce the impact on the environment and on
workers safety. The fact that practically no direct
conversion facilities exist for recovered metals. This
reduces the dismantling times and initial cleaning
phase.
The steps of this intermediate process start with an
inventory to list the hazardous substances on board.
The ship is then taken to an industrial site having the
necessary infrastructures (quayside, slipway, wet
berth). Dismantling is essentially manual (using
blowtorches), proceeding in horizontal layers similar
to the method used in the highly mechanised process.
Non-accessible hazardous materials which could not
be removed during the initial phase are removed as
and when demolition progresses. The remaining
bottom shell of the ship is cut manually in vertical
sections, either in an insulated wet berth or on a
slipway after pumping off effluent.
This intermediate process involves the direct reuse
of some equipment recovered on board, but not in a
manner as in the Indian subcontinent. On the other
hand, practically no direct reuse is made of secondary
raw materials, which means that all the scrap is
directed towards large electric steelworks to undergo a
complete cycle which provides a better guarantee for
the removal of waste still remaining on this scrap.
5. ECONOMIC ANALYSIS
The purpose of this study is to give an overview of
economic issues involved in a safe and
environmentally sound system of ship recycling. The
economic phase of a recycling system can be divided
into the following categories: 1. Cost consideration, 2.
Market review and 3. Future trend
5.1 Cost Consideration

An estimation of the costs is essential in the
introduction of such a safe & environmentally sound
scrapping. This study helps to estimate the additional
costs of environmentally sound scrapping. The
following costs are associated with the safe recycling:
237
1.
2.
3.
4.
introducing recycling regulation might lead to

substantially lower costs for environmentally sound
scrapping. These costs reductions are on the long term
basis only (some thirty years from now).
The costs for removing structural

components
The costs for removing operational waste
The costs for increasing the capability of the
ship recycling yard
The costs for adjustments at the ship
recycling yard to deal with the on-board
generated waste.
6. POSSIBLE SOLUTIONS
But, any additional cost involved can be handled by

means of the following:
1.
2.
3.
4.
Increasing recycling instead of scrapping as

much as possible.
Increase the price of ship to be scraped.
Raising ship recycling fund.
Increasing productivity- A well developed
recycling yard is more profitable
Ship recycling fund is to be contributed by the ship

owner at the operational life of ship as the
responsibility of safe recycling goes to the ship owner
itself.
5.2 Market Review and Future Trends

Scrap Demand and Price
It can be seen from the trends presented bellow that
although the recent scrap price in some countries are
in a down turn, the demand of scrap is still significant.
A little reduction in scrap price now might be due to
the recent decline in the world economy. But it may
not be the same in the near future as was not in the
past.
Available Number of Ships to be Recycled
Ships are not being scrapped as required. Huge
numbers of vessels are waiting to be scraped.
According to E.A. Gibson Shipbrokers [7], ship
owners extract an average US$ 1.9 million per End of
Life Vessel. Every year around 600-700 larger sea
vessels are taken out of service and brought to
scrapping yard. In the 1990s they had an aggregate
tonnage of around 15 million dwt a year. However,
the scrap market increases and will increase
substantially the following years. In 2001 the total
608 numbers of vessels were sold for scrap already
totalled a figure of 28 million dwt. This marks a year
on year growth of nearly 25%.
Future Trends
Current costs estimations are linked to ships on the
scrapping market which have been built 20 to 30
years ago. Ships becoming available for scrapping in
the future might show a different cost pattern as
technical developments have progressed. Market
information indicates that cost reductions are not
expected to be large at the short term. However,
current initiatives aimed at building cleaner ships and
6.1 Implementation of Ship Recycling /

Disposal Regulations
Various Local and International regulations are in
effect worldwide. Those regulations are based on the
analysis of past and current recycling system and
focused to protect the life of workers and prevent the
environment from pollution. If the rules and
regulations are implemented, the safe and
environment friendly recycling can be achieved.
IMO is the role playing in the international context.
It has come out with the concept of Green Passport to
minimize the hazardous substances on board ship and
prepared the guide lines on ship recycling. ILO came
out with Safety and health in ship breaking:
Guidelines for Asian countries and Turkey. The other
significant organisations relevant to ship recycling are
EWSR and Green Peace. United Nations UNEP is in
action from the Basel Convention for control of
hazardous wastes and its disposal which is applicable
in the ship recycling industries.
Port State Control will play a role in checking on
what ships contain, so that cleaning and disposal can
be planned.
6.2 Developing Recycling Facilities

Design of Ship Recycling Yards
As in the case of design for recycling, it is now
possible to design safe and environmentally sound
recycling facility. A number of concept designs can be
introduced depending upon the type of ships or
structures to be recycled, since there are not many
changes in future design.
Design of Recycling yards to conduct safe,
environment friendly and economically profitable
recycling operations can be developed.
1.
2.
3.
For the future, it is difficult to see changes in ship

design. The ship has to be watertight and meet
strict requirements such as port restriction and
canal regulation.
Bans on hazardous materials are in place.
In operations, there are requirements for
maintenance to ensure the ship is not hazardous.
A well developed recycling yard is normally

equipped with the necessary equipment for the
operation and with necessary facilities to deal with
238
the hazardous wastes in a responsible manner. In

good scrapping yards workers are well protected
against hazardous materials. These may ensure the
safety of the people & also help protect the
environment. There should have a waste reception
and treatment facility.
6.4 Ship Design for Recycling

Ship design for Production and Operation is existed
in ship design. Ship designs for Safe Recycling can
also be developed since it is difficult to see changes in
ship design in the future. Ship has to be designed
following strict requirements. Cars are designed for
disposal. Similarly ship can be designed considering
safe and sound disposal or recycling.
A simple recycling yard with waste reception

facilities and only two cranes for material lifting,
using beaching method may take a view of the figure
below.
6.3 Safe Practice to Protect Workers and

Environment
By introducing safe working practice, it is possible
to minimize the rate of accidents. In a recycling yard
in South Asia currently there is no standard practice of
recycling. Introduction of standard work practice
which is suited for the current facilities will not
increase the cost very much. This standard work
practice may consist of a set of procedure of recycling
which include the following:
The shipping industry must build clean ships

without toxic materials. Ship can be designed for easy
dismantling without health and environmental risks.
Encourage naval architects and shipbuilders to take
due account of the ships ultimate disposal when
designing and constructing a ship by doing the
following:
1.
2.
3.
4.
1.
2.
3.
4.
Recycling plans
Methods of work
Step by step work details
Heavy lifting procedures
Using materials which can be safely recycled

Minimizing the use of materials known to be
potentially hazardous to health and environment
Limiting the use of composite materials such as
sandwich panels which are difficult to recycle
Taking measures to facilitate the removal of such
materials
Recycling is more economic than disposal. Cars are

designed for disposal. Ships also can be designed for
recycling. By applying this concept disposal risk to
human and environment pollution can be minimized.
Safety and environment management systems

suitable for the particular scrapping yard which will
establish safe working procedures and assure the safe
working condition. This can be done by introducing a
permit to work system where any hazardous work
commencement requires a permit issued by relevant
and appropriate body such as certificate for hot work,
permit to entry into confined space etc.
In good scrapping yards workers are well protected
against hazardous materials. Protective clothing, hand
gloves to prevent hand injuries, use of face musk to
protect from fumes and gases, hard hat/ helmet to
protect head, safety boot can protect legs from
injuries, safety goggles to protect eyes and ear plug to
prevent from hearing losses.
Training of the workers is very important phase of
the safe practices of the scrapping yard. They should
be trained to follow a plan of work, and use protective
appliances and to know how to work safely.
There must be proper waste management to prevent
contamination of the environment. The enforcement
of the use of protective clothing and working
appliances should be present in the national level. The
appropriate and relevant government body has to
perform checks to the recycling yard whether the
implementation of safe work procedure is present and
use of protective appliance is in practice.
6.5 Owners Approach

The shipping industry is responsible for its own
vessels. They should take immediate measures to
prevent pollution of the environment and protect the
workers that dismantle their ships. They can do that
by decontaminating and degassing their ship-forrecycle before sending to dismantling yards.
Ship owners can help recycling their ship in safe
and sound way. They need to maintain ships as clean
as possible. Prior cleaning before sending the ship to
recycling yards is important and should be done by
the ship owners in operation. Ship owners should also
gradually make their vessels still in operation cleaner.
During maintenance and survey ships, hazardous
materials should be replaced with clean alternatives. A
list of hazardous materials found on board ships to be
included.
6.6 Inventory
An inventory of all materials on board ship will
help the recyclers to take necessary precaution for
recycling. The knowledge of the hazardous material
present will help them planning operation in safe and
sound way. Before sending the ship for dismantling,
ship owners should clean their ships of hazardous
materials and present a complete inventory of the
remaining hazardous materials to recycling yards.
239
6.7 Disbursement of Funds for

Environmentally Sound Ship Recycling
Currently the ship breaking yards are not
established for environmentally safe recycling.
Development of recycling yard is costly. To help
development of yard or adjustment to current yards
ship recycling fund can be raised. Ship owners and
relevant national and international bodies can
contribute to the fund.
6.8 Financing R&D on Clean and Safe Ship

Recycling
Research & Development of recycling will help
improve the current condition in this sector. R & D
should be financed too. The ship recycling fund can
be used to support R & D activities.
7. CONCLUSIONS
It is true that the contribution of ship recycling
industries to the environment pollution and human
health is significant. The current problem is complex
and cannot be solved overnight. It will take time. The
risks and associated effects on the environment can be
minimized.
1.
Implementation of ship recycling regulation is

one of the solutions. For example green passport
concept will minimise the use of hazardous
material for shipbuilding and an inventory of all
materials on board will help safe and
environment friendly recycling. Other regulations
such as wearing protective clothing and waste
treatment before releasing into environment are to
minimise accidents and prevent environment
pollution. But the current regulation demands a
drastically change in the situation which is not
practical. It has to be enforced gradually.
2.
Ship design for recycling is another solution to

this problem in the long run. That is it can be
helpful some 30 years or later. If a ship is built
today giving consideration for recycling it will
have to be recycled some 25 to 30 years later. So
ship design for recycling concept is for the future.
It does not solve the current problem.
3.
The other solutions e.g. facilities development,

safe practices may prove helpful if some form of
subsidy or funding is provided, although a huge
investment is not practical. Practice of safe
procedures is the most important to address the
current problem. It requires minimum effort both
with respect to technical consideration and
economic practicability. Suitable practices can be
regulated by the national and international bodies.
Certification of recycling yards and control of
working practices, creation of an inventory for
4.
the new ships and the ships in operation can be

regulated.
However an initial financing is required.
Disbursements of funds and research and
development have to be financed. Financing can
be made in a variety of ways locally and
internationally by shipping industries as well as
with the help of the environmental agencies.
Initialisation can be in a small scale and then
grown slowly.
Although the current situation is unfavourable, ship

recycling can be green industry if everybody
concerned comes forward to it. Changes can only
come about if it is a mutual desire and there is much
hope that there will come a day when the horrible
image of the ship recycling can be forgotten and the
new age of clean and green ship recycling will be the
future of a dream come true.
ACKNOWLEDGEMENT
The Author is extremely grateful to the following
ex-students Mohmmad Shariful Alam [8], Jerri
Ealham [9] and Sasheena d/o Thanasingam [10] of
School of Marine Science and Technology (MAST),
University of Newcastle upon Tyne, Newcastle
University Marine International (NUMI) - Singapore
for their works of final year projects.
REFERENCES
[1] IMO (International Maritime Organization);
http://www/imo.org
[2] ILO (International Labour Organization);
http://www.ilo.org
[3] UNEP (United Nations Environment
programme); http://www.unep.org
[4] Basel Convention; http://www.basel.int/
[5] Greenpeace; http://www.greenpeace.org
[6] European Community law on waste shipments;
www.ec.europa.eu/environment/waste/shipments
/legis.htm
[7] E.A. Gibson; http://www.gibsons.co.uk/
[8] Alam, Mohammad Shariful, Analysis of Safe
and Environmentally Sound Ship Recycling,
BEng Hons Thesis 2009, School of Marine
Science and Technology (MAST), Faculty of
Science, Agriculture and Engineering (SAgE),
Newcastle University, UK.
[9] Ealham, Jerri, Ship Recycling in Bangladesh
and its Socio Economic Effects, BEng Hons
Thesis 2010, School of Marine Science and
Technology (MAST), Faculty of Science,
Agriculture and Engineering (SAgE), Newcastle
University, UK.
[10] Sasheena d/o Thansingam, Ship Recycling in
Third World Countries, BEng Hons Thesis
2010, School of Marine Science and Technology
(MAST), Faculty of Science, Agriculture and
Engineering (SAgE), Newcastle University, UK.
240

ROLE OF ACIDS IN THE DEGRADATION OF ALKANOLAMINE DURING CO2

AND H2S REMOVAL PROCESS
Md. Sakinul Islam1*, Rozita Yusoff1, Brahim Si Ali1
1
Department of Chemical Engineering

Faculty of Engineering, University of Malaya
Kuala Lumpur, Malaysia
Email: sakin.chem.buet@gmail.com
ryusoff@um.edu.my
brahim@um.edu.my
ABSTRACT
The formation of organic acid and inorganic acids in CO2 and H2S removal process are common phenomenon.
Amines are basic in nature, which react easily with the acidic components to form heat stable salts and other
degradation products. Heat stable salt and degradation products are problematic in the operation of CO2, H2S
removal unit, as they reduce the solvent capacity to absorb CO2 and H2S, cause corrosion, foaming and solvent
losses. This paper gives an insight on the degradation of amine solvent due to these acidic components and
highlights the latest findings and research work carried out in this area of study.
Keywords: acid; degradation; amine; alkanolamine; absorption; heat stable salt
1. INTRODUCTION
Amines and alkanolamines are widely used in the
natural gas refining for the sweetening of sour gas by
absorbing the CO2 and H2S acidic gases. During the
absorption process amine are highly degraded by this
acidic oxide as a result amine loses its capacity to
absorb these acidic gases. All acidic gases of natural
gas composition play an important role to form amine
to aldehyde and acids. Oxygen directly oxidized the
amine and converted into carbonyl compounds like
aldehyde or ketone or acid compound. Aldehydes
react with water; in aqueous solution of formaldehyde
is hydrated [1]. Aldehydes also react readily with
primary amines. These reactions would not be
significant with tertiary amines. In aqueous solution
of monoethanolamine, formaldehyde will be present
as an equilibrium mixture primarily in the form of
imine with some hydrated aldehyde, carbinolamine,
and aminal. Similar reactions can occur with other
aldehydes such as hydroxyacetaldehyde. The imine
formed from primary or secondary amines such as
MEA may be subject to similar attack by radical
oxidants. The presence of formaldehyde may also
facilitate the oxidation of the amine molecule. Other
less compounds easily hydrated to aldehydes, such as
acetaldehyde and hydroxyacetaldehyde, react at room
temperature with oxygen to produce the
corresponding organic acid [2, 3]. The oxidation
ususally takes place by a free radical chain
mechanism. An initiator extracts a H from the
carbon to produce a free radical. The radical reacts
with O2 to produce a peroxyacid radical. The
peroxyacid radical reacts with aldehyde to propagate
the reaction with the production of the aldehyde
radical and a peroxyacid. So the prevention of amines

loss is the major focus of interest for the chemical
engineer. The prevention of amine losing by acid is
the one of the condition for successful plant operation.
This prevention of amine loss is also the key of
economical benefit. Hence, to identify the
alkanolamines degradations are important. Amine
degradation
compounds
are
quantified
by
chromatographic
methods.
Identification
of
degradation products is performed with several
analytical methods like GC/MS, IC and HPLC etc. In
some cases, synthesis of molecules is necessary to
confirm structure of degradation products. The
highest molecular weight compounds are identified
with a high resolution mass spectrometry technique
(FT-ICR/MS) with electrospray ionization (ESI). In
some cases, NMR analyses are useful to determine
compound structure. Formic, glycolic, acetic and
oxalic acids, nitrite and nitrate are quantified by ion
chromatography. The role of these acids and
degradation path way was critically reviewed and a
brief discussion was given by the following
representation.
2. AMINE BASED ABSORPTION AND
STRIPPING SYSTEM
For over seventy years the alkanolamine process
has been considered the best approach in removing
H 2 S and CO2 for the purpose of purification and
separation. It is based on the reaction between weak
acid (H2S and/or CO2) and weak base (alkanolamines)
to give a water soluble amine acid gas salt. An amine
based sour gas sweetening plant was configured with
absorber and regenerator. The temperature range is
241
Proceeedings of MARTE
EC 2010
40-60C annd 100-120C at the absorbeer and the stripper

respectivelly. Mainly amine
a
solventts are used as
absorbentss for sweetenin
ng natural gass and for proloong
use of thiss solvent that is
i transferred to
t the stripper by
means off regenerating
g absorbed gases. Figurre-1
represents a sweetening process
p
of sourr gas.
degradaation productss are ethyleneediamine derivvatives

(THEED
D) [6], which can form wheen CO2, COS or O2
are in an amine system. The basic
b
chemisttry of
degradaation of alkanoolamine is expplained in thee open
literaturre [7] and thesee are given bellow. In the folllowing
chemicaal reaction it iss clear that amiines occur due to the
protonaation and carbaamate formatiion reaction brrought
about byy the hydrolysis of water andd acids.
H+ + OH
H2O
H+ + HS
H2S
HCO
O3 + H+
CO2 + H2O
RNH
H3+
RNH2 + H+
RNH2 + CO2
RN
NHCOO + H+
Amines/Alkanolaminnes + Sour Gas
The same reactionn is also explaiined in the liteerature

[8-10] for DEA deggradation. Theese reactions are
a as
follows. Hydrogen suulfide emittedd hydrogen ionn (Eq.
B.6) in water and caarbon dioxide as well as carrbonyl
sulfide and producess COS and H2S respectivelyy (Eq.
B.7). This
T
reaction causes additioonal reactionss with
DEA molecule.
m
Sww
wet Gas
Natural Gas Com

mponents:
Absorber Chemicals:
CH4, C2H6, C3H8, C4H10, CO
Sweet Gas + DGP
Aminnes/Alkanolamines, Wateer
C5H12, CO2, N2, H2, H2S, O2,

H2O
D
DGP=
Degradation Produccts; = Absorber/ Strippeer Temperature
H+ + HS
H H2S
COS + H2O
CS2 + H2O
CO
OS + H2O
CO2 + H2S
H+ + HCO-3
H2O + CO2
Figure. 1: Typical CO
O2 absorption and
a desorptioon
prrocess [4]
Sour gas components
c
reeact with aminne solvent durring
absorption-desorption cy
ycles graduallly at higher and
a
lower tempperature and some
s
of these reactions (A.11 to
A.11) are given in the following
f
to shhow how DGP
P is
formed [5].
HO-C2H4NH
H2 + HCl
DEA
HO-C2H4NH
H 2 + H 2S
DEA
R-NH2 + O2
MEA + CO22
MEA + COSS
DEA + CO2
DEA + COS
O2
MDEA + CO
DGA + CO2
DGA + COS
DIPA + CO2
HO-C
H 2H4NH3+Cl
HSS
HO-C
H 2H4NH3+HS
HSS
CnH2n+1COOH / Imidazolidonne
OZD, HEI,, HEED
DEA Urea
HEOD, BH
HEEP, THEED
MEA, HEI, BHEED
EG, HMP, DEA, TEA, BHEP
BHEEU
BHEETU
HMPO
B.1
B.2
B.3
B.4
B.5
A.11
A.2
A.33
A.44
A.55
A.66
A.77
A.88
A.99
A.110
A.111
3. BASIC DEGRADATION CHEMIS

STRY
Degraddation products are contaminnants in soluttion
that are deerived from thee breakdown of
o the base am
mine
molecule and
a often causeed reversibly and
a irreversiblyy to
form totallly different chemical speciies. Examples of
B.6
B
B
B.7
B
B.8
B
B.9
DEA molecule directly

d
reactts with CS2 and
successively producees DEA thioccarbamate saltt (Eq.
B.10 & B.11). On thee other hand it produces carbbamate
O2 (Eq. B.13). DEA carbamaate reacts withh MEA
with CO
and DE
EA and forms BHEED
B
and THEED
T
respecctively.
These tw
wo DGP furtheer form BHEI and BHEP witth CO2
respectiively. HEOD is one of the faamiliar DGP off DEA
and thiss is also from DEA carbamaate. Protonatedd DEA
moleculle reacts withh CO2 and converts into MEA,
ethanol and water. Thhe formation of
o MEA is bettter for
DEA syystem because the author assuume that it has better
absorptiion capacity. It
I is also assuume that any amine
system where higher molecular
m
weigght compoundd break
mine like MD
DEA or
down innto lower molecular mass am
DEA to MEA that process willl rich at thee high
absorptiion rate. Alkaanolamines aree oxidized andd form
aldehydde or ketone thhen formed accid. This acid finally
f
forms HSAS.
H
So to protect
p
the form
mation of HSA
AS the
intermeediate reactionn should be controlled at first.
Howeveer there is insufficient innformation inn this
connecttion in the oppen literature. Ethanol can easily
degradee to aldehyde and
a then an accid. This acid forms
various acetate and ester
e
compoundds in amine syystem.
EA are
An oveer view of thee reaction patthways of DE
242
EC 2010
to form a salt that is not

n capable of being
b
regeneraated by
the adddition of heat and are thus referred to ass Heat
Stable Salts. Not onlly do they tiee up the aminne and
therebyy reduce the accid gas carryinng capacity, buut they
are alsoo considered corrosive.
c
How
wever, there is still a
consideerable amount of confusion in the industryy over
the terrm Heat Staable Salt. Some
S
propertiies of
common HSS is menttioned in Table -01 below.
mentioned below.
D dithiocarbamate
DEA+CS2 DEA
DEA dithiocarbaamate + DEA DEA dithiocarbamate
d
salt
B.10
DEA + H+
B.12
B.13
DEA + CO2
DEAH+ + CO2
DEAH+
DEACOO-H+
MEA + C2H4OH + CO2
B.11
B.14
DEACOO H + MEA BHEED + COO2 + H2O
B.15
DEACOO-H+ +DDEA THEED + CO2 + H2O
B.16
- +
THEED + CO2
BHEP + CO2 + H2O
DEACOO-H+
HEOD + H2O
Table-001: Common HSS

H in amine system [14]
Name of
mical
Origiin of
Lim
mit
Chem
Formu
HSS
I
Ions
ppm
m
la
anions
Nitratee
NO3
Inorgganic
-
B.17
B.18
Nitritee
NO2
Inorgganic
Formaate
CHO2
Orgaanic
5000
Oxalatte
C2O4
Orgaanic
2500
Acetatte
C2H2O2
Orgaanic
10000
Sulfatee
SO4
Inorgganic
5000
4. ACID FORMATION
F
N REACTION IN AMINE
SYSTEM
Sulfitee
SO3
Inorgganic
5000
Phosphate
PO4
Inorgganic
The reaaction in this sccheme was prooposed to account

for the presence of all th
he anions obserrved in this stuudy.
This acid formation
f
mechanism could be
b applied to any
a
alkanolamiine. Also, it was
w noted that although
a
the acid
a
form of accetic, formic, glycolic, glyooxalic, and oxaalic
acids are shown
s
in the fo
ollowing reactiion, each of thhese
anions are almost complletely ionized to
t the amine heat
h
stable salt in basic solutiions. The literaature [11] cited a
oxidation reaction
r
of ME
EA which are ass follows.
Thiosuulfate
S2O3
Inorgganic
10,000
Thiocyyanate
CNS
Inorgganic
10,000
Glycolate
C2H3O3
Orgaanic
5000
Malonnate
C3H3O3
Orgaanic
5000
Succinnate
C4H5O4
Orgaanic
10000
Chloriide
Cl
Inorgganic
5000
- +
MEACOO H + MEA
BHEED + CO2
HEI + H2O
BHEI + H2O
B.19
B.20
6. SPEC
CIFIC DEGRA
ADATION PRO
ODUCTS BY ACID
A
Whiile Heat Stable Salts (HSS) are common to all
amine solvents, deggradation prooducts are soolventA descriiption of com
mmon
specificc [6, 15-17]..
degradaation productss found in various
v
gas trreating
solventss follows.
5. TYPICA
AL HSS IN AM
MINE SYSTE
EM
Heat sttable salt (HSS
S) formation iss another form
m of
active amiine loss. The amine
a
and an acid form a salt
s
that cannoot be regeneraated in the stripper [12, 13].
1
Strong acidd anions such as formate, accetate, thiosulfaate,
thiocyanatee, and chloridee can tie up ann amine moleccule
mides
Formam
N-foormyl aminess (formamides) are degraadation
productts generally foound in gas treeating solutionns that
are baseed on primary and secondaryy amines [18]. Under
certain conditions all primary andd secondary amines
a
react with
w the formic acid in solutioon (via dehydrration)
to form
m n-formyl amines.
a
The data from process
p
solutionns containing formate
f
as a HSS
H anion show
w that
the following equilibbrium relationnship exists in
i the
solutionn between thee formate HSS
S and the n-fformyl
MEA (M
MEA-F) [19].
Form
mic Acid + ME
EA MEA-F + Water
Since thhe above equaation represennts equilibrium
m, it is
also possible to hydroolyze MEA-F back into ME
EA and
243
formic acid. The heat and water present in the stripper

of the amine unit will generate a new equilibrium if
removing one of the above components disturbs the
balance of the equation.
HEED
HEED is a degradation product of MEA from
reactions with CO2. There is a wealth of literature on the
reaction mechanisms and the corrosive nature of HEED
[20]. While much of the literature has focused on MEA
in CO2 service only, HEED has been found in many
combined systems treating H2S and CO2.
HEEU
HEEU is a degradation product of MEA that is not
that well known since most of the literature has focused
on degradation in CO2 service. HEEU is formed via the
same reaction pathway as HEED when COS is present
in the gas [21].
Polymers
We generally find polymers when HEED, is present
in MEA systems treating H2S and CO2. These polymers
are formed from the reaction of HEED with MEA
molecules to make longer chained ethylenediamines
[22]. We also find polymers when THEED is present in
DEA systems treating H2S and CO2 [22].
BHEEU
BHEEU is an inert degradation product of DGA
which formed in the presence of COS and CO2 [21].
.Formation of BHEEU can be reversed by thermal
reclaiming.
Morpholine
This is an inert degradation product of DGA that,
while rare, can occur when the solution is subjected to
high temperatures.
THEED
THEED is a well-known degradation product of
DEA from reactions with CO2. There is a wealth of
literature on the reaction mechanisms and the corrosive
nature of THEED [20, 21]. While much of the literature
has focused on DEA in CO2 service only, THEED has
been found in combined systems treating H2S and CO2.
Bis-HEP
Bis-hydroxyethyl piperazine (Bis-HEP) is a wellknown degradation product of DEA from reactions with
CO2. There is a wealth of literature on the reaction
mechanisms of bis-HEP [20, 21]. While much of the
literature has focused on DEA in CO2 service only, bisHEP has been found in combined systems treating H2S
and CO2.
MEA
In the presence of certain chemical compounds
(oxygen) or intermediates, it is possible to degrade or
break down the DEA molecule to simpler amines.
Monoethanolamine (MEA) is one of the simpler amines
that may be formed from DEA degradation. It is
important to monitor the level of MEA in the circulating
DEA system due to Amine Stress Corrosion Cracking
(ASSC) concerns associated with MEA. MEA will also
generally further degrade in the system.
Bicine
Bis-(hydroxyethyl) glycine (Bicine) is a degradation
product formed in the presence of DEA and unstable
chemical intermediates, and is considered corrosive [23,
24].
DIPA-OX
Hydroxypropylmethyloxazolidone (HPMO or DIPAOX) is a well-known degradation product of DIPA from
reactions with CO2. There is a wealth of literature on the
reaction mechanism of DIPA-OX [20, 21, 25]. The
literature also states that DIPA-OX is the endpoint of
DIPA degradation in CO2 service (due to steric
hindrance), so polymers of DIPA are not supposed to be
formed like they are with MEA and DEA.
MDEA Fragments
MDEA in Tail Gas Treating units (TGTU), Acid Gas
Enrichment units (AGE), and in some main amine
systems should be monitored for MDEA fragments.
These include MMEA, DEA, Bicine and C2+ HSS
anions [26]. It is also important to note that as
ethanolamines (from MDEA degradation) accumulate in
the solvent, they will generally undergo further
degradation to compounds listed in the above sections.
The presence of these primary and secondary amines
may affect the gas treating selectivity of the amine
solution.
Dimers
The literature states that steric hindrance prevents
further degradation of DIPA-OX to diamines
(dimers). However, we generally do find dimers in
Sulfinol-D systems treating H2S and CO2. These
dimers are similar to the diamines found in MEA and
DEA systems, and are likely formed via the same
pathway.
TIPA
TIPA may be found in systems treating H2S and CO2
utilizing Sulfinol-D, when evidence of oxygen
degradation is present. Literature shows that when
amines degrade they generally form simpler amines, but
sometimes these simpler amines react with the base
amine molecule (or others) to form a more complex
amine [27].
244
EC 2010
7. COMMO
ON DEGRADAT
TION PRODUCTS AND EFFEC
CT
All alkkanolamine deg
gradation produucts are not saame
in behavior, some of these are acidic and
a some of thhese
are alkalinne based. Dep
pending on thhe nature of the
degradationn products thesse are classifieed as follows.
7.1 Acidic degradation product
p
Acidic degradation prroducts are higghly corrosive for
plant unit, thus metallic unit
u of the gas sweetening pllant
is highly interactive by
y this product. So the aciidic
degradationn products are
a
very harm
mful among all
degradationn products. Heeat stable saltss are the comm
mon
degradationn products off amine strippping system and
a
these are potentially co
orrosive. All of these aminnes
degraded and
a formed caarboxylic acidss. These acids are
capable of
o causing fu
urther chemicaal reaction with
w
miscellaneeous degradatio
on products. Byy this reaction the
concentratiion of contam
minants increaases continuouusly
during abssorption-desorp
ption cycles and finally shoows
operational problems.
d
product
p
7.2 Basic degradation
The alkkaline degradaation products are
a less corrossive
than acidicc or some of these
t
are noncoorrosive but both
still have a bad impact on plant operaation or even the
neutral DG
GP has the sam
me effect too.. It increases the
viscosity and
a as a resultt it highly inhhibits the soluttion
flow rate. This lower flow rate shhould affect the
efficiency of the plant [28]. To accelerate the pllant
lower flow
w rate to satisfactory flow rate,
r
high enerrgy
consumptioon is required. For this reasoon no degradattion
products arre expected.
d
prroduct
7.3 Ionic degradation
The deegradation product which coontains chargee at
the chain end
e and it may
y be positive or negative thhese
are definedd as ionic degrradation produccts. The literatture
is citied itss effect on thee plant perform
mance [8, 29, 30].
3
Ionic degrradation produccts are more porn
p
to affect the
plant perfoormance becau
use it is easilly react with the
componentts and form miiscellaneous reaction productts.
7.4 Polymeric degradatiion product
Many articles refleect a lot of information on
polymeric degradation products but there is no speciific
polymer formation deegradation reeaction [31-333].
Polymeric degradation products
p
identtified by GC-M
MS
detection may be neutrral or living. Living polym
mer
means it coontains a functtional group at the chain endss of
the moleccule which arre capable off causing furtther
chemical reaction with accidic oxide or any reactive unnit.
The literaature cited very inadequuate informattion
regarding the polymerization reacttion mechaniism
though it has a satisfaactory detectioon report. Wide
W
ranges off research on
n polymerizatiion reaction are
suggestted to recover this literaturee gap. Literatuure [7]

focusedd that THEED
D, Bicine, HSS,
H
HE-Sarccocine,
HEED, HEOD, BHE
EED and TH
HEED are corrrosive
compouunds [31, 34]. Actually these
t
are corrrosive
becausee all of these are acidic orr basic or conntain a
living functional
fu
grouup or oxazolidoone ring at thee chain
end or in
i its structure.
utral Degradation product
7.5 Neu
Neuutral degradatiion products having no active
functionnal group are very harmfull for fouling effect.
These degradation
d
products are inaactive and just move
towardss the entire sollution direction. It also makkes the
solutionn more viscouus. Neutral degradation prooducts
having no movementt so there is a great probabiility to
adherinng with plant pipelines, as a result that place
easily corrode
c
or desstroy. Carl andd Haws studieed the
effect of neutral degraadation productts [35]
7.6 Volaatile degradattion product
Volaatile degradatioon products arre highly harm
mful to
the envvironment. Geenerally these gases cause green
house effect,
e
world climate changge and so onn. The
effects of the aminne were well conducted in
i the
literaturre [36] last yeaar.
mmon effect of
o DGP
7.7 Com
The foaming behaavior of degradation productts was
conductted by Thitaakamol and Veawab [30,, 37].
Corrosion behavior of degradatiion products were
conductted by Dawoodu and Meiisen [31] andd also
conductted by the reseearchers [34, 366]. The recent article
[34] poiinted out severral problems which
w
occurred due to
irreverssible degradatioon reaction andd mentioned beelow.
Reductionsolutio
oncapacityandp
plant
performance
Foamingeffectofdegradationproducts
Capital/MaterialofConstructionIsssues
Corrosioneffectofdegradation
products
Environmentaleffectof
degradationproducts.
Foulingeffecto
ofdegradation
products
NCLUSION
8. CON
Durring the sour gas sweeteninng process a lot of
solvent are losses by amine degradation. The
245
degradation products arises various difficulties and as a

result the remediation of this problem has become a
major focus of interest for the last few decades. HSAS
neutralization
techniques
(vacuum
distillation,
electrodialysis, ion exchanges) are still being developed
to improve plant efficiencies and reduce operation
problems.
THEED
The literature mostly cited the heat stable salt and its
minimization. HSAS are basically the reaction products
of acid and amine. Amine is the main solvent; on the
other hand an acid is the by product, if we control the
acid formation in amine system then automatically
reduces the HSAS formation. Acids can also be
converted to acid amides by adding some additives.
Because amides contain CONH2 functional group this is
capable to capture more acid gases. So we need further
research on it to recover the research gap. The author
suggested the acid remediation by converting into
amides compound. Author also would like to put
emphasis for the minimization of the acidic content
inside the solution. The above mentioned contents of
this review paper highlighted very important and recent
information regarding alkanolamines degradation by
acid. Knowledge of the reaction mechanism aids the
development of procedures to minimize degradation or
purify partially degraded solutions. Therefore, the
overall representation would help to understand the role
of both inorganic and organic acid on amine
degradation.
HPLC
BHEI
BHEEU
NMR
GC-MS
IC
FTIR
HPLC-RID
REFERENCES
[1]
[2]
[3]
ACKNOWLEDGEMENT
The authors would like to acknowledge Ministry of
Science, Technology and Innovation (MOSTI),
Malaysia, for the project: RG003/09AET and the
University of Malaya for the financial support and for
providing us with an excellent research environment.
[4]
[5]
[6]
ABBREVIATIONS
NG
DIPA
DEA
MDEA
MEA
DGP
DGA
HSS
HSAS
AMP
HEP
BHEED
BHEP
HEOD
Natural gas
Di-isopropanolamine
Diethanolamine
Methyldiethanolamine
Monoethanolamine
Degradation product
Diglycolamine
Heat stable salts
Heat stable amine salts
Aminomethyl propanol
Hydroxyethyl piperazine
N,N-Bis(hydroxyethyl) ethylnediamine
N,N-Bis(hydroxyethyl)-piperazine
Hydroxyethyl-oxazolidone
N,N,N-Tris(hydroxyethyl)ethylenediamine
N,N Bis (hydroxyethyl)imidazolidone
N,N-Bis(hydroxyethoxy-ethyl)urea
Nuclear magnetic resonance
Gas chromatography mass
spectrometer
High performance liquid
chromatography
Ion chromatography
Fourier transforms infrared resonance
spectrometer
High performance liquid
chromatography-refractive index
detector
[7]
[8]
[9]
246
Singh, V. N., Gangwar, M.C., Saxena, B.B.L.,

Singh, M.P.,. "Kinetics and Mechanism for the
Oxidation of Formaldehyde by Hexacyanoferrate
(III) Ion,". Can. J. Chem. Eng. 1969; 47 10511056.
Denisov, E. T., Mitskevich, N.I., Agabekov, V.E.,.
"Liquid-Phase Oxidation of Oxygen-Containing
Compounds," Consultant Bureau, New York.
1977.
Sajus, L., Seree De Roch, I.,. The Liquid Phase
Oxidation of Aldehydes, in Comprehensive
Chemical Kinetics, Liquid-Phase Oxidation.
Elsevier Scientific Publishing Co., Amsterdam.,
1980; 16 89-124.
Abdi, M. A.,Khakdaman, H. R.,Zoghi, A. T.
Predicting amine blend Performance,. 2005.
Haws, R. Contaminants in Amine Gas Treating
Retrieved on 13 July 2010. Avalable at website: .
2010.
Liu, H. J., Dean, J. W., and Bosen, S. F.,.
Neutralization to Reduce Corrosion from Heat
Stable Amine Salts, Corrosion95, Paper No. 572,
1995.
Wenten, I. G. Chemical solvent base, Innovative
Solution in CO2 Nowadays, oral presentation,
retrieved on February,2010, available on the
website
http://www.iatmicirebon.org/ver.2/control/Materi/wenten_chemica
l.pdf. 2010.
Dawodu, O., F.,,Meisen, A. Degradation of
aqueous diethanolamine solutions by carbon
disulfide. Gas. Sep. Purif 1996; Vol. 10, No. 1 111.
Dawodu, O. F.,Axel Meisen, A. Mechanism and
kinetics of COS-induced diethanolamine
degradation,. Ind. Eng. Chem. Res 1994; 33 480487.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
Choy, E., T.,,Meisen, A. Gas chromatographic

detection of diethanolamine and its degradation
products. Journal of Chromatography 1980; 187
145-152.
Rooney, P. C.,Daniels, D. D. Oxygen solubility in
various alkanolamine/water mixtures,. Gas
technology 1998.
The Dow Chemical Company. Gas Conditioning
Fact Book, Midland, Michigan, The Dow
Chemical Company, 1962.
Kennard, M., L.,,Allelsen, A. Mechanisms and
Kinetics of Diethanolamine Degradation. Ind.
Eng. Chem. Res. Fundam. 1985; lQ85,24 129140.
Stewart, E. J.,Lanning, R. A. Reduce amine plant
solvent losses,. 1994.
Meisen, A., et. al.,. Degraded Amine Solutions:
Nature, Problems and Distillative Reclamation,
Proceedings of the 1996 Laurance Reid Gas
Conditioning Conference. 1996.
Cummings, A. L., Veatch, F.C., and Keller, A. E.,.
Corrosion and Corrosion Control Methods in
Amine Systems Containing H2S, Corrosion97,
Paper No. 341, 1997.
Rooney, P. C., Bacon, T. R., and Dupart, M. S., .
Effect of Heat Stable Salts on MDEA Solution
Corrosivity, Hydrocarbon Processing, March
1996.
Koike, e. a. N-Formyldiethanolamine: A New
Artefact (sic) in Diethanolamine Solutions,
Chemistry & Industry, 1987. 1987.
HCTT. Huntsman Corporation Technical
Training(HCTT), Evidence of Empirical
Equilibrium Between Formate and n-Formyl
Amines in Gas Treating Application, 1999.
1999,.
McCullough, J. G., and Nielsen, R. B.,.
Contamination and Purification of Alkaline Gas
Treating Solutions, Corrosion96, Paper No. 396.
1996.
Kohl, A., and Nielsen, R., . Gas Purification
5th Edition, Gulf Publishing, Houston,. 1997;
pp. 236-239, 261-262.
Holub, P. E., et. al.,. Amine Degradation
Chemistry in CO2 Service, Proceedings of the
1998 Laurance Reid Gas Conditioning
Conference. 1998.
Cummings, A. L., et. al., . Corrosion and
Corrosion Control Methods in Amine Systems
Containing H2S, Corrosion97, Paper No. 341.
1997.
Howard, M., and Sargent, A.,. Operating
Experiences at Duke Energy Field Services
Wilcox Plant with Oxygen contamination and
Amine Degradation, Proceedings of the 2001
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
247
Laurance Reid Gas Conditioning Conference.

2001.
van Grinsven, V., et al., . DIPA as the Preferred
Solvent for Amine Treatment in Refinery
Applications, Proceedings of the 1999 Laurance
Reid Gas Conditioning Conference. 1999.
Critchfield, J. E. a. J., J. L.,. Evidence of MDEA
Degradation in Tail Gas Treating Plants,
Petroleum Technology Quarterly, Spring 1999.
1999.
Chakma, A., and Meisen, A.,. Methyldiethanolamine Degradation Mechanism and
Kinetics,. The Canadian Journal of Chemical
Engineering, October 1997 1997.
Haws, R.,Jenkins, J. Contaminant Reporting in
Amine Gas Treating Service,. 2000.
Supap,
T.,Idem,
R.,Tontiwachwuthikul,
P.,Saiwan, C. Analysis of Monoethanolamine and
Its Oxidative Degradation Products during CO2
Absorption from Flue Gases: A Comparative
Study of GC-MS, HPLC-RID, and CE-DAD
Analytical Techniques and Possible Optimum
Combinations. Ind. Eng. Chem. Res 2006; 45
2437-2451.
Thitakamol, B.,Veawab, A.,Aroonwilas, A.
Foaming in amine-based CO2 capture process:
experiment, modeling and simulation. Energy
Procedia 1 2009; GHGT-9. 1381-1386.
Dawodu, O., F.,,A. Meisen, A. Effect of
composition
on
the
performance
of
alkanolamines blends for gas sweetening. .
Chemical engineering communications 1996; 144
103-112.
Lepaumier, H.,D. Picq, D.,Carrette, P., L., .
Degradation Study of new solvents for CO2
capture in post-combustion. . . Energy Procedia
2009; 1 893900.
Lepaumier, H.,Martin, S.,Picq, D.,Delfort,
B.,Carrette, P.-L. New Amines for CO2 Capture.
III. Effect of Alkyl Chain Length between Amine
Functions on Polyamines Degradation,. Ind. Eng.
Chem. Res., 2010,.
Bedella, S. A. Oxidative degradation mechanisms
for amines in flue gas capture. Energy Procedia
2009; 1 771-778.
Carlson, S.,Canter, S.,Jenkins, J. Canadian Gas
Treating Solvent Quality Control Unique
Challenges,. 2001.
Eide-Haugmo, I.,Brakstad, O., G.,,Hoff, K.,
A.,,Srheim, K., R.,,Silva, E., F.,D.,,Svendsen,
H., F.,. Environmental impact of amines. nergy
procedia 2009; 1 1297-1304.
Thitakamol, B.,Veawab, A. Foaming Behavior in
CO2 Absorption Process Using Aqueous
Solutions of Single and Blended Alkanolamines.
Ind. Eng. Chem. Res 2008; 47 216-225.

ASSESSMENT OF SHIP MANOEUVRABILITY IN SHALLOW WATERWAYS

Osman Md. Amin1 and Kazuhiko Hasegawa2
1
Department of Naval Architecture and Ocean
Engineering,
Graduate School of Engineering, Osaka University,
2-1 Yamadaoka, Suita, Osaka 565-0871, Japan
E-mail: osmanamin@naoe.eng.osaka-u.ac.jp
2
Department of Naval Architecture and Ocean
Engineering,
Graduate School of Engineering, Osaka University,
2-1 Yamadaoka, Suita, Osaka 565-0871, Japan
E-mail: hase@naoe.eng.osaka-u.ac.jp
ABSTRACT
Navigation of ships in inland waterways faces many difficulties in terms of ship manoeuvrability. While the
ships are operating in the restricted waterways such as channels, canals or rivers, manoeuvring characteristics
of ships get significantly changed from those in deep and wide waterways. The reduction in water depth
contributes significantly to the loss of turning ability of ships. In this research, the manoeuvring characteristics
of a particular type of ship have been investigated considering the shallow water influence. Mostly through
numerical simulations, the particular behavioral change of the navigating ship at different depths of water has
been thoroughly analyzed. The simulations were carried out on the basis of a mathematical model which
describes the ship motion properly at different conditions of ship navigation. From theoretical point of view,
some proposals also have been made on the basis of previous research works which may provide reliable tools
for the prediction of manoeuvrability of ships in shallow water.
Key words: Shallow water, ship manoeuvrability, hydrodynamic interaction coefficients, MMG model, advanced vortex
method, slender body theory.
1. INTRODUCTION
Manoeuvrability of ships in shallow water is one of
the most sought after problems in the field of ship
motion. Ships, operating in harbor and port areas,
encounter
a
significant
change
in
their
manoeuvrability characteristics because of the change
in the water depth in restricted areas. The main
intention of this paper is to introduce the readers with
several techniques, which are available now-a-days, to
predict the manoeuvrability of ships in shallow
waterways. The focus would be directed upon the
vertical constraints that the ship faces in those areas.
On the basis of experiment data, Fujino [7] had shown
that the water depth to ship draft ratio when exceeds
2.5, the shallow water effect becomes less remarkable.
And he concluded that at intermediate water depths
ship becomes very unstable, whereas in very shallow
water it again becomes stable. As the main concern of
predicting manoeuvrability is to measure the
hydrodynamics forces and moments acting on the
ship, a simplified simulation technique serving this
purpose would provide a great tool to the navigators
for safe operation of ships in narrow waterways.
Hull induced forces and moment are considered to
comprise the biggest portion of the total forces acting
on the ship while in motion, whereas the interaction

between hull, propeller and rudder has a significant
bearing on the manoeuvrability of ships. On this
premise, here a special attention has been put forth in
determining those interaction effects through
numerical simulations.
2. VALIDATION OF SHIP
MANOUEVRABILITY
Manoeuvrability of a ship signifies the
predictability and controllability of the motion of
ships at different situations of sea conditions.
Qualitatively the manoeuvrability of a ship is
measured through some manoeuvring tests, which
when performed on the ship would provide us with a
comprehensive way to understand the behavior of the
operating ship. In this context, the quantitative
measurement is done through the determination of the
incurred upon hydrodynamic forces and moments by
varying flow patterns around the ship. Either
theoretical approach or numerical simulation - none
could so far provide a reliable method to simulate
appropriately the manoeuvring characteristics of ships
in shallow water without conducting the captive
model tests. Several of these treatments are explained
below.
249
2.1 Analytical Methods

Slender body theory is one of pioneering theory for
ship hydrodynamics. It exploits the slenderness of the
ships (longitudinal scale is far greater than the beam
and depth) in transforming the boundary conditions
(B.C.) to simpler forms. In this approach, perturbation
technique [6] is applied to obtain the first order and
second order equations of the B.C. [16] on the basis of
the perturbation parameter (in this case, the
slenderness ratio - the ratio of ship beam to length).
This would require a matching of the solutions in
outer region and inner region of the body, which
through the application of distinct B.C. for different
regions would render a unique solution of the velocity
potential. Then on the basis of this outcome the
pressure distribution on the body can be achieved.
E.O.Tuck [31] has shown that, the trim and sinkage
predicted in shallow water using the above mentioned
method provides a significant lack of dissimilarity
between theory and experiment, which may have
occurred due to the linearization of the B.C. On this
premise, Yeung [36] has proposed a general theory of
motion to provide a six D.O.F. motion model of the
slender body ship. A chronological survey done by
Maruo [17] shows that, the non-effectiveness of this
method to predict ship motion is fixed in the nonlinear
B.C. For this reason, the authors have taken measures
to solve those boundary conditions, keeping the
nonlinearities intact, using a non-perturbative
mathematical method, named Homotopy Analysis
Method. Homotopy is a concept related to the field of
topology (set theoretic form of geometry). Liao [15]
has proposed this method to solve the differential
equations by taking into consideration of the
nonlinearities. Systematic solution of the B.C. using
this method, as converging series, would provide a
suitable base to rely on for further analytical
derivations involving ship motion in shallow water.
Some of the primitive techniques like Prandtl's
lifting-line theory have been implemented by
Newman [18] to take account of the lateral restrictions
along with the vertical one in narrow channel. Hess
[8] using similar technique of slender body/porous
wing model has shown that the rudder force increases
with decreasing water depth and rudder moment gets
reduced due to a shift in center of pressure toward
amidships. This result conforms to the experimental
data concerning the hull-rudder interaction of ship in
shallow water.
Most of the analyses, as mentioned before, consider
the double hull model to take account of the free
surface B.C. That means, the free surface is
considered to be rigid on assumption of low speed
manoeuvring in shallow water. Chen et al [4] figured
out that the ship moving in a narrow channel may
produce waves which would counteract the bank
suction force, transforming it into a repulsive one at

supercritical speed. Although Ohkusu [20] has
pointed out that, so far the number of experiments
carried out to verify the above mentioned theoretical
approaches are very few to justify the applicability of
those methods.
Another powerful concept is the strip theory, which
assumes that the fluid flow in a plane perpendicular to
the longitudinal axis of the ship is identical to the twodimensional fluid motion around a cylinder of infinite
length with the same cross section as that of the strip
of the ship. For low frequency range of motion this
method cannot predict well the flow field because of
the lack of consideration of interaction between the
strips (three dimensional effect). The earliest strip
theory concept (dubbed 'Ordinary strip theory') [34]
didn't consider the forward speed in a rational way,
which was later modified by Tasai [28] using
conformal mapping technique to take account of the
change in flow field due to the presence of forward
speed. Keil [11] extended this technique for different
water depths. Strip theory considers the dimensions of
the strips to be much smaller than axial length, which
in turn shows the equivalency of this method with the
slender ship theory.
2.2 Simulation using Mathematical Models

The assessment of manoeuvrability at the initial
design stage of a ship requires a good mathematical
description of the manoeuvring motion of the ship.
Apart from the theoretical formulation, the
hydrodynamic forces acting on the ship can be
represented mathematically in many forms, from the
fairly simple Abkowitz [1] derivatives for prediction
of first quadrant manoeuvres, to a full four-quadrant
deep and shallow water simulator model.
The modular mathematical model, developed by
MMG group of JTTC [19], was designed to express
the stream behavior around the ship by considering
the individual open water characteristics of the hull,
propeller and rudder and the interaction effects among
them. This model was intended to provide a rational
way of deducing several hydrodynamic coefficients
from Captive Model Test and also to establish an
analytical model/full-scale correlation. This is why,
the proper determination of the interference effect
between hull, propeller and rudder comprises a
significant proportion of the manoeuvrability
prediction. Significant amount of research works have
been done so far to confirm the validity of this model
in case of shallow water manoeuvring. Among them
Hirano et al [9] proposed a simplified method to
predict the linear hull hydrodynamic coefficients on
the basis of low aspect ratio wing theory. A more
practical hull force prediction method was proposed
by Kijima et al [13] through the formulation of some
250
The Authors have extended the MMG model to

take into account of water depth change, as faced by
the ship in narrow waterways [2]. Some semiempirical formulae to predict the hydrodynamic
interaction coefficients, which comprise the MMG
model, have been proposed for single-propeller
single-rudder ship in that paper on the basis of
available experimental data for shallow water
manoeuvring. Several simulations have been carried
out for a VLCC ship (Esso Osaka [5]) to judge the
manoeuvrability at different depths of water. Figure1
shows the body plan of the tanker and Figure 2
depicts the turning circles for 35-degree turning test at
different depths of water along with the experimental
data. In these cases the initial speed is considered to
be 0.3167 m/s. From Figure 2 it is apparent that the
simulated turning trajectories closely resemble that of
the experimental one.
H/d-6.0 (experiment)
H/d-6.0 (simulation)
H/d-2 (simulation)
5
4
X/L
semi-empirical formulae to derive hull hydrodynamic

coefficients. So far this method is considered to be the
most congruent one in terms of its practical
applicability. Other researchers [14, 33, 37]
concentrated more on specific types of ships for the
formulation of manoeuvrability characteristics in
shallow water on the basis of MMG model. The
dominant hydrodynamic forces acting on ship, when
operating at low speed in shallow water are the
inertial forces. Kan et al [10] used low-aspect-ratio
wing theory and method of images to calculate twodimensional added mass coefficients for body section
in a channel of finite depth. Van Oortmerssen [32] has
shown that the dependency of added mass, in case of
shallow water manoeuvring, on frequency of motion
should be taken into account for frequencies higher
that 0.1s-1. Since the ships do not face that much high
frequency of disturbance, while operating in calm
shallow channels, the added mass and moment of
inertia can be considered to remain constant
throughout its operation in those areas. Sadakane et al
[24] have proposed some semi-empirical formulae to
predict the added masses and moment of inertia in
terms of ship principle particulars. And they proved
those equations on the basis of experiment and
potential flow theory.
3
2
1
0
-4
-2
Y/L
-1
-2
-3
-4
-5
Figure 2. 350 turning circles at different depths of

water
Kijimas regression model for hull hydrodynamic
forces along with the proposed model for interaction
coefficients have been adopted in this analysis to
establish a generalized mathematical model for the
prediction of ship manoeuvrability in shallow water.
2.3 Numerical Techniques

The analytical methods explained in section 2.1
consider the flow around ship to be inviscid and
incompressible. To take account of the viscosity and
to simulate the flow field in detail around the ship
several numerical techniques are so far being tried out
by many researchers. CFD technique such as RANS
(Raynolds Averaged Navier-Stokes) code is becoming
more and more popular to predict the wave resistance
and manoeuvrability fairly well in deep water.
Although the treatment of hull, propeller and rudder
interaction still seems be far fetched from the point of
view of CFD simulation [22, 25, 27, 30]. Progresses
are still under way to predict the flow around the ship
properly and to measure the propeller-hull interaction
at full scale Reynolds number.
There are plenty of other numerical techniques
which rely on large computational power. One of
them is the LES (large eddy simulation) of the wake
field behind the moving ship. The LES method
considers the large eddies created in the wake to be
composed of tiny discrete elements, such as vortex
blobs, vortex particles etc., and simulates the flow
field on the basis of shed vortices from the body of
the ship. The advantage with this method compared to
the RANS code is that the flow field does not have to
be discretized to simulate its characteristics [12]. The
Authors have implemented a certain Vortex shedding
method [21, 26] to replicate the flow pattern around
Figure 1. Body Plan of VLCC tanker (Esso Osaka)
251
the Esso Osaka hull at certain drift angle, as can be

seen in Figure 3.
Figure 3 . Flow pattern around Esso Osaka at drift

angle =120,Ut/L=0.5, boundary layer thickness=0.005
To take account of the interaction effect in shallow
water, this method is adopted on the hull along with a
sink model for propeller as proposed by Yamazaki et
al [35] and also with a lifting surface for rudder.
The principle concern with the methodology
involved with computational fluid dynamics is the
lack of availability of mathematical model to
formulate the turbulent flow around ship. Turbulence
is considered to be a feature of the flow, not of the
fluid. On this basis the instantaneous characteristic
quantity of the turbulent flow is decomposed into the
mean average value and a fluctuating part, which
allows us to develop a statistical treatment of
equations of motion. This extra random part gives rise
to the additional turbulent stress, which arises from
convective derivative of the Navier-Stokes equation.
Several methods are available in deriving this
turbulent stress, such as RANS method and spectral
methods [3], which provides us a way to figure out
the energy exchange between the different sized
eddies. Another approach is to find a 'hidden'
parameter besides other parameters, like Reynolds
stress, the viscous dissipation rate etc, which would
influence the property of turbulent flow [23]. As
mentioned before, the LES is another concept which
has huge popularity in simulating the turbulent
behavior of flow around ship. Despite of all of these
efforts, still there remained a lack in the understanding
of the fundamentals behind the generation of
Turbulence. As Lumley [29] pointed out - "In
turbulence, the equations do not give the entire story.
One must be willing to use simple physical concepts
based on experience to bridge the gap between the
equations and actual flows".
3. CONCLUSION
The extensive amount of techniques which are
available for the prediction of ship manoeuvrability in
shallow water are thoroughly explained in this paper.
The requirement of fast time simulation in ship
handling simulators has put on a challenge for the
researchers to find out an appropriate method to
predict the manoeuvrability of ships in narrow
waterways. In this context, only the mathematical
model approximation of the ship manoeuvring
dynamics provides practical way of simulating the

ship behavior, where captive model tests are necessary
to verify the model. Although CFD and other
numerical techniques are gaining in popularity
because of the enhancements in the computing
powers, the model tests are still the most reliable
means to predict the ship behavior in shallow
waterways.
REFERENCES
[1] Abkowitz,
M.-A.,
"Lectures
on
Ship
Hydrodynamics - Steering and Maneuverability,"
Hydro-Og Aero-dynamisk Laboratorium, Report
Hy-5, Lyngby, Denmark, pp.113 (1964).
[2] Amin, O.-M., Hasegawa, K., "Generalized
Mathematical Model for Ship Manoeuvrability
considering Shallow Water Effect," Conference
Proc. of Japan Society of Naval Architects and
Ocean Engineers, Vol.10, pp.531-534 (2010).
[3] Canuto, C., Hussaini, M.-Y., Quarteroni, A.,
Zang, T.A., "Spectral Methods - Fundamentals
in Single Domains," Springer - Verlag Berlin,
Heidelberg (2006).
[4] Chen, X.-N., Sharma, S.-D., " Nonlinear theory
of asymmetric motion of a slender ship in a
shallow channel," 20th Symp. on Naval
Hydrodynamics, pp.386-407 (1994).
[5] Crane, C.-L.-Jr., "Maneuvering Trials of a
278000 DWT Tanker in Shallow and Deep
Waters," Transactions of SNAME, Vol.87,
pp.251-283 (1979).
[6] Dyke, V., "Perturbation Methods in Fluid
Mechanics," The Parabolic Press, Stanford,
California, (1975).
[7] Fujino, M., "Studies on Manoeuvrability of
Ships in Restricted Waters," Journal of Society
of Naval Architects of Japan, Vol.124, pp.157184 (1968).
[8] Hess, F., "Rudder Effectiveness and Coursekeeping Stability in Shallow Water: A
Theoretical Model," International Shipbuilding
Progress, Vol.24, No.176, pp.206-221 (1977).
[9] Hirano, M., Takashina, J., Moriya, S.,
Nakamura, Y., "An Experimental Study on
Maneuvering Hydrodynamic Forces in Shallow
Water," Transactions of West-Japan Society of
Naval Architects, pp.101-110 (1984).
[10] Kan, M., Hanaoka, T., "Analaysis for the Effect
of Shallow Water upon Turning," Journal of
Society of Naval Architects of Japan, Vol.115,
pp.49-55 (1964).
[11] Keil, H., "Die Hydrodynamische Krfte bei der
periodischen Bewegung zweidimensionaler
Krper an der Oberflche flacher Gewasser,"
Bericht Nr. 305, Institut fr Schiffbau der
Universitt Hamburg, Deutschland, (1974).
252
[12] Kijima, K., Furukawa, Y., Yukawa, K., "On a

prediction method of hydrodynamic forces
acting on ship hull including the effect of hull
form," Proc. of Marine Simulation and Ship
Manoeuvrability, Rotterdam, pp.411-418 (1996).
[13] Kijima, K., Nakiri, Y., "On the Practical
Prediction Method for Ship Manoeuvrability in
Restricted Water," Transactions of West-Japan
Society of Naval Architects, pp.37-54 (2003).
[14] Kobayashi, E., "The Development of Practical
Simulation
System
to
Evaluate
Ship
Manoeuvrability in Shallow Water," Proc. of
International Symposium on Practical Design of
Ships and Other Floating Structures, pp.1.7121.723 (1995).
[15] Liao, S.J., "Beyond perturbation: introduction to
the homotopy analysis method," Boca Raton:
Chapman & Hall/CRC Press (2003).
[16] Maruo, H., "Application of the Slender Body
Theory to the Longitudinal Motion of Ships
among Waves," Bulletin of the Faculty of
Engineering, Yokohama National University,
Vol.16, pp.28-61 (1967).
[17] Maruo, H., "Evolution of the Theory of Slender
Ships," Schiffstechnik Bd., 36-3, pp.107-133
(1989).
[18] Newman, J.N., "Lateral motion of a slender
body between two parallel walls," Journal of
Fluid Mechanics, Vol.39, part 1, pp.97-115
(1969).
[19] Ogawa, A., Kasai, H., "On the Mathematical
Model of Manoeuvring Motion of Ships,"
International Shipbuilding Progress, Vol.25,
pp.306-319 (1978).
[20] Ohkusu, M., "Validation of Theoretical Methods
for Ship Motions by Means of Experiment,"
22nd Symp. on Naval Hydrodynamics, pp.341358 (1999).
[21] Ojima, A., Kamemoto, K., "Numerical
Simulation of Unsteady Flow around Three
Dimensional Bluff Bodies by an Advanced
Vortex Method," Proc. of JSME International
Journal, The Japan Society of Mechanical
Engineers, pp.127-135 (2000).
[22] Raven, H.-C., Ploeg, V.-D., Strake, A.-R., Eca,
L., "Towards a CFD-based Prediction of Ship
Performance - Progress in Predicting Full-Scale
Resistance and Scale Effects," Proc. of
International RINA MARINE CFD conference
(2008).
[23] Reynolds, W.-C., Kassinos, S.C., "A one point
model for the evolution of the Reynolds stress
and structure tensors in rapidly deformed
homogenous turbulence," Proc. of Osborne
Reynolds Centenary Symposium, University of
Manchester Institute of Technology, (1994).
Inertia of Ship in ShallowWater," Journal of

Japan Institute of Navigation, pp.11-20 (2001).
[25] Simonsen, C.-D., Frederick, S., "RANS
Maneuvering Simulation of Esso Osaka with
Rudder and a Body-Force Propeller," Journal of
Ship Research, Vol.49, No.2, pp.98-120 (2005).
[26] Suzuki, K.,
Hirakawa, R., Ojima, A.,
Kamemoto, K., "Calculation of Viscous Flow
around Ship Models by Means of Advanced
Vortex Method," Proceedings of CDF workshop
Tokyo 2005, NMRI, Tokyo, pp.453-479 (2005).
[27] Tahara, Y., Wilson, R.-V., Carrica, P.-M., Stern,
F., " RANS simulation of a container ship using
a single-phase level-set method with overset
grids and the prognosis for extension to a selfpropulsion simulator," Journal of Marine
Science and Technology, Vol.11, pp.209-228
(2006).
[28] Tasai, R., "Improvements in the Theory of Ship
Motions in Longitudinal Waves," Proc. of 12th
ITTC, pp.677-687 (1969).
[29] Tennekes, H., Lumley, J.L., "A First Course In
Turbulence," The MIT Press, (1972).
[30] Toxopeus, S.-L., "Deriving mathematical
manoeuvring models for bare ship hull using
viscous flow calcualtions," Journal of Marine
Science and Technology, Vol.14, pp.30-38
(2009).
[31] Tuck, E.O., "Shallow water flows past slender
bodies," Journal of Fluid Mechanics, Vol.26,
part 1, pp.81-95 (1966).
[32] Van Oortmerssen, G.., "Influences of the Water
Depth on the Manoeuvring Characterisitcs of
Ships," Proc. of Symposium on Ship Handling,
Wageningen, The Netherlands, (1973).
[33] Vantoree, M., "Manoeuvring Coefficients for a
Container Carrier in Shallow Water: An
Evaluation of Semi-Empirical Formulae," Proc.
of Mini Symposium on Prediction of Ship
Manoeuvring Performance, pp.71-81 (2001).
[34] Weinblum, G., Denis, S.-M., "On the motion of
ships at sea," Transactions of SNAME, Vol.58,
pp.184-248 (1950).
[35] Yamazaki, R., Nakatake, K., "Free-Surface
Effect on The Hull-Propeller Interaction," 15th
Symp. on Naval Hydrodynamics, pp.463-479
(1985).
[36] Yeung, R.W., "Applications of slender body
theory to ships moving in restricted shallow
water," Proc. of Symposium on aspects of
navigability, pp.1-16 (1978).
[37] Yoshimura, Y., "Mathematical Model for the
Manoeuvring Ship Motion in Shallow Water,"
Journal of Japan Society of Naval Architects and
Ocean Engineers, pp.41-51 (1986).
[24] Sadakane, H., Toda, Y., Lee, Y.-S., "The

Simplified Formulas to Predict the Coefficients
of Added Mass and Yaw Added Moment of
253

PROBABILISTIC EVALUATION OF PLATE EFFECTIVENESS IN TERMS OF EFFCTIVE

WIDTH OF ATTACHED PLATING
S. C. Vhanmane1 and B. Bhattacharya2
1
Research and Rule Development Division, Indian Register of
Shipping, Powai, Mumbai 400 072, Maharashtra, India.
E-mail: suhas.vhanmane@irclass.org
2
Department of Civil Engineering, Indian Institute of
Technology, Kharagpur 721 302, West Bengal, India.
E-mail: baidurya@iitkgp.ac.in
ABSTRACT
A ship hull is a complex structure consisting of unstiffened/stiffened plates, longitudinals, frames,
transverses, etc. and is subject to longitudinal bending, transverse bending and torsion. A primary load effect in
ship hull components is axial compression induced due to longitudinal bending. The stiffened plate which is the
backbone of the primary support members in a ship hull is widely modeled to consist of the stiffener with an
attached plating of a certain effective width. This is called the beam-column model and its behaviour largely
depends on its cross-section properties and hence on the effective width of the attached plating. Many
researchers in the past assumed a convenient constant value of effective width. However, fabrication related
initial imperfections in the form of initial deflection and welding induced residual stresses always develop in the
plate elements between two longitudinals and they do affect plate behaviour. Field measurements indicate that
plate design parameters such as material yield strength, initial deflection, welding residual stresses as well as
corrosion loss are random in nature. Growing concern towards ship safety and reliability through Goal Based
Standards (GBS) highlights the need to employ probabilistic methods in studying the mechanical behaviour of
stiffened plates. In this paper, the effective width of the attached plating is assessed using probabilistic methods
considering the effect of random material yield strength, initial deflection, welding residual stresses and
corrosion loss in the plating between two longitudinals of a bulk carrier and a oil tanker. The results obtained
are presented by platting histograms of effective width ratio for various cases. The results obtained in this study
can be very useful for reliability analysis of ship plated structures.
Key words: Effective width, random, average stress-strain relationship, reliability analysis.
1. INTRODUCTION
A ship hull is a complex structure mainly consisting
of
continuous
unstiffened/stiffened
plates,
longitudinals, frames, transverses, etc. and is subject
to longitudinal bending, transverse bending and
torsion. A primary load effect is axial compression
induced due to longitudinal bending. Behaviour of
structural element under axial compressive load is
very important. The plate effectiveness can be reduced
by lateral deflection that may arise due to (i) initial
deflection and welding residual stresses due to
welding process, (ii) lateral pressure due to cargo or
sea load, and (iii) buckling due to axial compression.
Hence, the plate effectiveness is the function of initial
imperfections (initial deflection and welding residual
stresses) and the applied load actions.
Depending upon the applied load actions, Paik [1]
measured the effectiveness of the plate by using three
different concepts: (i) effective breadth, (ii) effective
width and (iii) effective shear modulus. Effectiveness
of the deflected plate is evaluated by considering an
equivalent flat plat but with reduced effectiveness and

this reduced effectiveness is measured by the three
different concepts as stated above depending upon the
applied load action. The term effective breadth is
typically used when the lateral deflection is caused by
out-of-plane or lateral pressure action in association
with shear lag effect. The term effective width is
typically used when the lateral deflection is caused by
in-plane compressive loads and the term effective
shear modulus is used when a plate is predominantly
bucked by edge shear action. In this paper, we have
concentrated on the evaluation of effectiveness using
the term effective width.
A equation originally presented by von
Karman et al. (1932) for effective width (beff) of plates
as
beff / b = cr / Y ; where b is the full width of
the plate,
cr is the critical buckling stress and Y is
the material yield stress. This expression was

reasonably accurate for thin plates but is found to be
relatively conservative in case of thick plates.
255
The classification societies use the following

formula for effective width proposed by Faulkner [2]
for merchant ship plating that implicitly considers a
moderate level of initial deflection but does not
consider welding residual stresses:
1.0
= C1 C2
b 2

beff
for < 1,
for 1.
(0.1)
where C1 and C2 are constants and depends on the

applied boundary conditions of the plate (C1 = 2.0 , C2
= 1.0 for all edges simply supported; C1 = 2.25 , C2 =
1.25 for all edges clamped), = (b / t P ) Y / E is
loads, the membrane stress distribution inside the

plate becomes non-uniform provided the unloaded
edges (intersection at which stiffener and plating are
welded) of the plate remain straight. The membrane
stresses, compressive in nature, become maximum at
the plate edges and minimum in the middle region
associated with membrane tensile stresses due to the
plate deflection. The plate width that supports the
maximum membrane stresses is called the effective
width of the plate. Immediately after buckling, the
maximum membrane stress becomes larger than the
average value of non-uniform membrane stresses in
longitudinal direction as shown in Fig. 1.
the plate slenderness ratio (b = plate breadth or

stiffener spacing, tP = plate thickness and E = Youngs
modulus). It can be seen that Faulkners expression
does not consider the effect of the applied load: under
axial compressive load, the effective width should
naturally decrease as the load increases. Also,
Faulkner neglected the presence of the welding
residual stresses induced during the manufacturing
process.
The presence of initial imperfection reduces the
effective width of attached plating (analytically
demonstrated by Paik et al., [3]). Higher the
magnitude of initial imperfection (both, initial
deflection and welding residual stresses), lower the
value of effective width/ultimate strength of plating.
Hence, it is obvious that the cross-sectional area of
combined attached plating and stiffener structure
(stiffened plate element) will be reduced, which
subsequently reduces the ultimate load carrying
capacity of individual stiffened plate element. In turn,
this reduces the ultimate strength of the hull girder.
The aim of the present paper is to study the effect
of influential parameters in the evaluation of plate
effectiveness using the term effective width. Further,
randomness in the influential parameters (material
yield stress, initial imperfections due to welding
process and aging effect due to corrosion) is
considered to propose the effective width as a random
variable. Four different candidate plates (two of a bulk
carrier and two of a oil tanker) are analysed in this
study. The results are presented through histoghrams
and are compared with Faulkners formula [2] and
IACS CSR formula [4].
2. CONCEPT OF EFFECTIVE WIDTH

Generally, the stiffened plate is represented by
beam-column approach [5-9] which is an assembly of
stiffener and attached plating. If the stiffened plate is
subjected to axial compressive load actions, the extent
of attached plating is defined by the term effective
width of attached plating.
As the plate buckles under uniform compressive
Figure 1. Membrane stress distribution, (a) before

buckling, and (b) after buckling
After buckling, the maximum membrane stress
becomes larger than the average stress (Fig. 1-b). In
this case, the ratio of effective width to the full width
is same as the ratio of the average stress (avg) to the
maximum stress (max) [1] and can be presented as:
b /2
be
=
b
x dy
=
b /2
max
avg
max
(0.2)
It is traditional to compare the reduced strength of

the plate with full width (b) by equating it to the
strength of another plate that has an effective width
(beff) and collapses at the material yield stress.
Therefore, the effective width becomes equal to the
ultimate strength [9]. Hence, ratio of effective width
(beff) to the full width (b) is same as the ratio of
ultimate strength (u) to the yield stress (Y) as:
256
beff
b
u
Y
(0.3)
Analytical methods, experimental tests, empirical

approaches and non-linear finite element method have
been used widely to determine the buckling and
ultimate strengths of ship platings and significant
progress has been achieved. However, some important
aspects of this subject remain areas of active research:
In the following sections, the probabilistic

modeling of initial imperfections (initial deflection
amplitude and welding residual stresses), material
yield strength and corrosion is described. Following
that the statistics of effective width of candidate plates
are investigated.
- development of analytical formulas,

- development of simplified methods,
- assessment of effects of initial imperfections,
- assessment of effects of fatigue cracks.
An analytical formulation to determine the
effective width or ultimate strength of plating between
two stiffeners, considering presence of initial
imperfections (initial deflection and welding residual
stresses), is presented in our earlier publication [10]
by deriving average stress-strain relationship under
axial compressive load action as prescribed. The
average stress-strain relationship is established by
taking advantage of two different methods in
combination: the membrane stress method by
involving large elastic deformation theory and the
rigid plastic collapse mechanism theory. Using the
methodology, it shown that the effective width and/or
ultimate strength of plate under compressive load is a
function of applied load, material yield strength (Y)
and magnitude of initial imperfections (A0 and r) in
Fig 2-4. The circles on the curves shown in Fig. 4 are
effective width ratio when the plate is on the verge of
yielding and is important in the determination of
ultimate strength of plating.
Figure 3. Variation in the effective width ratio with

different magnitudes of welding residual stresses
Figure 4. Effective width compared with Faulkners

formula
3. INITIAL IMPERFECTIONS

different magnitudes of initial deflection
If the magnitudes of material yield strength and
initial imperfection are random, the effective width of
attached plating, and hence the properties of the
stiffened plate are random in nature. Thus, the hull
girder ultimate strength should be described as a
random variable as well.
The initial imperfections (initial deflection, A0 and

welding residual stresses, rc) are introduced into the
plating because of welding process used in ship
building during fabrication process. When the welded
zone cools down, it shrinks but can not regain its
original shape and thus tensile stresses develop.
Consequently, compressive stresses develop in the
adjacent colder parts. The plate effective width at the
ultimate strength is different depending on the level of
initial deflection. Smith et al. [11] and Paik and
Thayamballi [12] have defined three different levels
of initial imperfections as slight, average and severe
as:
257
Table 1. Initial Imperfection Levels [11-12]

Welding residual
Initial deflection
Level
(A0)
stress (rc)
Slight
0.025 tP
0.05y
Average
0.1 tP
0.15y
Severe
0.3 tP
0.3y
2
2
The above model is based on data measured on plates

for b / t P 120 , the majority being in the range of
20 b / t P < 120 . It can be seen that the mean
However, Faulkner formula [2] explicitly does not

take into account the effect of initial deflection as a
parameter of influence. Most of the shipyards apply
manual welding processes and so there is possibility
of variation in the welding process parameters. Due to
such variations which are inherent and uncertain, the
initial imperfections are random in nature. This is
supported by actual statistical data collected by [13]
and [14].
We now assume that the initial deflection
amplitude (A0) is a random variable while the pattern
of the initial deflection (w0) over the plate, as shown
in Fig. 5, remains the same:
w0 = A0 sin
y
m x
sin
a
b
increases while the c.o.v. decreases with increasing

b/tP.
In the idealized welding induced residual stress
distribution used in the present method (previously
shown in Fig. 6), the tensile residual stresses, rt, are
developed at the edges of the plating i.e. along the
welding line; the residual compressive stresses, rc,
are developed in the middle of part of the plating. The
breadth of the tensile residual stress may reach the
yield stress but a somewhat reduced 80% of the yield
stress tensile residual stress may be used.
(0.4)
where, A0 = defined is the initial deflection amplitude.
Figure 6. Idealized welding residual stress distribution

of plating between stiffeners
The model of uncertainty in welding residual stress
proposed by [14] is:
r
2
=
Y b
2
tP
With reference to [14], a Normal probability model

for A0 normalized by b ( A0 / b ) is used in the present
study to demonstrate the random initial deflection
amplitude and thus the random initial deflection of
plating between two stiffeners. Hence,
(0.9)
and a standard deviation given by
= 0.04b / t P
(0.10)
4. RANDOM YIELD STRENGTH

(0.6)
and the coefficient of variation (c.o.v.) given by:
c.o.v. = 0.675 0.004b / t P
E[ ] = 1.20 + 0.06b / t P
(0.5)
where, mean () is given by
= 0.12b Y / t P E
where, r = welding residual stress, Y = material

yield strength, = bt / tP = non-dimensional width of
tensile stress block (as shown in Fig. 6) having a
Normal distribution with mean
Figure 5. Initial deflected shape of plating between

the stiffeners
A0 / b ~ Normal( , )
(0.8)
(0.7)
The hull girder ultimate strength is rather sensitive

to the variation in the yield strength. The material
yield strength of each structural member (stiffener,
plate panels or hard corner) in the mid-ship section of
hull girder is considered to be random. According to
the existing literature on the statistics of yield strength
of structural steel used in the marine industry, the bias
258
(i.e., mean/ nominal) is between 1.00 and 1.10 (Table

2) while the coefficient of variation (c.o.v., i.e.,
standard deviation / mean) varies between 6% and
10%. In this study, it is assumed that the material
yield strength is lognormal with c.o.v. 6% and mean
value of 1.1 times the nominal value.
6. NUMERICAL EXAMPLES AND

DISCUSSION
Table 2. Randomness in yield strength

Source
DnV [15]
Guedes Soares and
Kmiecik [16]
Paik and Frieze [17]
Harada and Shigemi [18]
Distribution
Type
Lognormal
1.1
c.o.v.
(%)
6.0
Lognormal
1.0
10.0
Normal
Lognormal
1.0
1.1
10.0
6.0
Bias
demonstrate the effect of random corrosion wastage

on effective width of attached plating, we assume
value of B to be equal to 1. The corrosion rate, A is
taken from [19].
5. RANDOM CORROSION WASTAGE

The corrosion growth of the marine structures
depends on the loss of coating effectiveness and
corrosion rate. Generally, it is assumed that corrosion
starts immediately after coating loss. Paik et al. [19]
proposed a time dependent corrosion wastage model
wherein the transition between the coating breakdown
and corrosion initiation ( t ) was considered and is
In this section, the effective width of the attached

plating to the stiffener is determined under the effect
of random initial imperfections (as described in
Sections 3), random yield strength (as described in
Section 4) and random corrosion wastages in Section
5. The candidate deck and bottom plates used in these
examples are taken from two ship structures (a bulk
carrier and double hull tanker) from a benchmark
study of ISSC Technical Committee VI.2 [20] and the
details are given in Table 3 and Table 4 shows the
effective width for each plating calculated
deterministically using Faulkar [2] and IACS CSR [4]
formula
Table 3. Plating details
shown as (Fig. 7):
Plating
a (mm)
b (mm)
tP (mm)
(0.11)
Tanker bottom plate
4950
830
22.5
1.44
313.6
where: A = corrosion rate (mm/year); B = constant;

t = present age of the ship (years); c = coating life
Tanker deck plate
4950
830
20.0
1.62
313.6
Bulk carrier bottom plate
2610
880
18.5
1.85
313.6
Bulk carrier deck plate
5220
880
24.5
1.57
392
d (t ) = A(t c t ) B
(years);
t =
transition
between
the
coating
breakdown and corrosion initiation (years).
Table 4. Effective width of plating
Figure 7. Paik corrosion wastage model

In the present study, it is assumed that the corrosion
starts immediately after loss of coating and hence
transition between the coating breakdown and
corrosion initiation is ignored (t = 0). It is assumed
that the coating life, c follows normal distribution
with mean of 7.5 and c.o.v. of 0.4 [19]. For corrosion
of marine structure, B may be typically vary in the
range of 0.3 1.5. For the sake of simplicity and to
Plating
Fualkner (1975)
IACS (2008b)
Tanker bottom plate
0.90
0.98
Tanker deck plate
0.85
0.98
0.79
0.80
0.87
0.98
As stated, effective width is the function of

applied load and decreases with increase in the load
action. The same is demonstrated in Fig.8. Three
different cases are shown: (i) random yield strength
combined with slight level of initial imperfection, (ii)
random yield strength combined with average level of
initial imperfection and (iii) random yield strength
combined with severe level of initial imperfection.
259
Monte-Carlo simulation is performed to obtain

statistics of effective width ratio. Firstly, we have
considered random material yield strength associated
with deterministic levels of initial imperfections as
mentioned in Table 1. Figure 9 shows the histogram
of effective width ratio of tanker bottom plate when
random yield strength is combined with slight level of
imperfection. Mean and c.o.v. of effective width ratio
is given. Figure 10 and Fig. 11 shows the histogram of
same plate when random yield is combined with
severe level of imperfections and random yield
strength and random initial imperfections in
combination respectively.
Figure 9. Tanker bottom plate random yield strength

+ slight initial imperfection
Figure 10. Tanker bottom plate random yield

strength + severe initial imperfection
applied uniaxial compressive load
It is observed that the mean is of effective width
decreases with increase in the level of initial
imperfection from slight to severe and c.o.v. increases.
However, in case of combined random yield and
random initial imperfections the mean is higher than
combined random yield strength severe initial
imperfection but is lower than combined random yield
strength and slight initial imperfection. The c.o.v. is
higher than both the cases. Remaining plots for tanker
deck plate (Fig. 12-14), for bulk carrier bottom plate
(Fig. 15-17) and for bulk carrier deck plate (Fig. 1820) also reinforce the same findings. The c.o.v. which
is the measure of uncertainty is upto 7 to 8 % when
both yield strength and initial imperfections are
random. Comparing the mean effective width ratio
with Faulkners formula [2] and IACS CSR formula
[4] in Table 4, it is observed that the effective width
obtained by Faulkners formula resembles with the
one obtained when combined random yield strength
and random initial imperfections are accounted in the
analysis.
Later, random corrosion wastage (as described in
Section 5) is considered in the evaluation of effective
width of the candidate plates. The corrosion rate, A
assumed to have Weibull distribution as given in [19]
and Table 5 shows the mean and c.o.v. of A for each
of the candidate plates.
260
Figure 11. Tanker bottom plate random yield

strength + random initial imperfection
Figure 12. Tanker deck plate random yield strength

+ slight initial imperfection

+ severe initial imperfection
Figure 17. Bulk carrier bottom plate random yield


+ random initial imperfection
Figure 18. Bulk carrier deck plate random yield

strength + slight initial imperfection

strength + slight initial imperfection



261
20%) and increase in the c.o.v. upto 3.87 % is

observed. Similar trends in the mean and c.o.v. is
obtained in case of deck plate of tanker as given in
Table 7. Surprisingly, c.o.v. in case of severe initial
imperfection is lower than average level initial
imperfection.
Table 5. Probabilistic details of the corrosion rate

Plating
Mean (mm/yr)
c.o.v. (%)
Tanker bottom plate
0.0597
0.9901
Tanker deck plate
0.0581
0.8262
0.0497
0.9557
0.1188
0.9217
Table 7. Effective width ratio- Tanker deck plate

t = 20 yrs
t = 25 yrs
t = 30 yrs
c.o.v.
c.o.v.
c.o.v.
c.o.v.
0.97
0.62
0.97
1.17
0.96
2.03
0.95
2.91
2*
0.87
1.13
0.86
2.02
0.86
2.64
0.85
3.30
3*
0.75
1.05
0.75
1.64
0.74
2.10
0.74
2.58
Figure 21 shows the variation in the effective

width ratio with applied load at various values of
present age (t). This is evident that the effective width
decreases with increasing time from new built
structure.
t = 15 yrs
*Initial imperfection level: 1= Slight; 2=Average; 3=Severe
In Table 8, the results of bulk carrier bottom plate

are given. The mean is reduced by 22% from slight
level to severe level imperfection in all cases and
c.o.v. is increased by 38 to 46 % at t =15 yrs to t = 30
yrs respectively. There is reduction of around 21~22
% in mean of bulk carrier deck plate as given in Table
9 with increasing level of initial imperfection and
c.o.v. follows the same trend as obtained in case of
tanker deck plate i.e. c.o.v. at any particular age
increases from slight to average level but it is lower in
case of severe level than average level.
Table 8. Effective width ratio- Bulk carrier bottom
plate
Figure 21. Variation in the effective width ratio at

different age (t) values
In different cases, random corrosion wastage is
combined with the three levels of initial imperfections
as mentioned in Table 1. For each case, 1000 samples
of effective width ratio are obtained by Monte-Carlo
simulation. Table 6 presented the statistical
parameters of effective width ratio of the tanker
bottom plate for different assumed ages of 15, 20, 25
and 30 yrs.
Table 6. Effective width ratio- Tanker bottom plate
t = 15 yrs
t = 20 yrs
t = 25 yrs
t = 30 yrs
c.o.v.
c.o.v.
c.o.v.
c.o.v.
1*
0.99
0.28
0.99
0.59
0.98
1.40
0.98
2.11
2*
0.92
1.11
0.92
1.72
0.91
2.46
0.91
3.66
0.80
1.33
0.79
1.94
0.79
2.65
0.78
3.87
t = 15 yrs
t = 20 yrs
t = 25 yrs
t = 30 yrs
c.o.v.
c.o.v.
c.o.v.
c.o.v.
0.91
1.55
0.90
2.35
0.89
3.43
0.88
4.65
0.80
1.40
0.79
2.17
0.79
3.05
0.78
4.12
3*
0.70
0.95
0.70
1.45
0.69
1.93
0.69
2.47
Table 9. Effective width ratio- Bulk carrier deck plate

t = 15 yrs
t = 20 yrs
t = 25 yrs
t = 30 yrs
c.o.v.
c.o.v.
c.o.v.
c.o.v.
1*
0.97
1.38
0.94
3.27
0.95
5.06
0.94
7.10
0.88
2.18
0.86
3.58
0.85
5.46
0.84
6.88
0.76
1.93
0.75
3.09
0.74
4.00
0.73
5.09
There not a significant change in the mean of

effective width ratio. But, the c.o.v. increases with the
time for a particular level of initial imperfection. At
any particular age, reduction in the mean (around
Now, both corrosion wastage and initial

imperfections are combined and are random. The
mean and c.o.v. for each of the candidate plate are
given in Table 10 with four different values of
assumed ages. The mean effective width reduces by
2~3 % from t =15 yrs to t = 30 of age. The c.o.v. is
dependent on plate details and can not be commented
in generalized manner.
262
Table 10. Effective width ratio- combined random

corrosion and random initial imperfection
t = 15 yrs
t = 20 yrs
t = 25 yrs
t = 30 yrs
c.o.v.
c.o.v.
c.o.v.
c.o.v.
0.92
5.60
0.91
6.09
0.90
6.60
0.90
6.89
0.87
7.25
0.86
7.44
0.85
7.78
0.84
8.17
2
3* 0.80 8.09
0.80 8.32
0.79 8.53
0.78
4* 0.87 7.22
0.86 8.45
0.85 8.79
0.84
*Candidate plate: 1= Tanker bottom plate; 2=Tanker deck
3=Bulk carrier bottom plate; 4=Bulk carrier deck plate
8.86
9.88
plate;
7. CONCLUSION
In this paper, effectiveness of the deflected plate is
assessed using the term effective width. It
demonstrated that the effective width is a function of
initial imperfections, yield strength, corrosion loss and
applied compressive load. Next, the randomness in
initial imperfections, in material yield strength and in
corrosion wastage is discussed in the determination of
effective width of the four candidate plates from a
bulk carrier and a oil tanker. Monte-Carlo simulation
is performed to obtain statistics of effective width and
the results are presented by histograms different cases.
It is observed that the mean is of effective width
decreases with increase in the level of initial
imperfection from slight to severe and c.o.v. increases
at random yield strength. However, in case of
combined random yield and random initial
imperfections the mean is higher than combined
random yield strength severe initial imperfection but
is lower than combined random yield strength and
slight initial imperfection. The c.o.v. is higher than
both the cases.
Aging effect is considered by reduction in the plate
thickness due to random corrosion. The obtained
results show that there not a significant change in the
mean of effective width ratio. But, the c.o.v. increases
with the time for a particular level of initial
imperfection. At any particular age, reduction in the
mean (around 20 22 %) and increase in the c.o.v.
upto 6-7 % is observed. When both random corrosion
wastage and random initial imperfections are
combined, the mean effective width reduces by 2~3 %
from t =15 yrs to t = 30 of age. This information
seems to be important for reliability analysis of unstiffened plates. However, to make a general
conclusive remark a series of cases need to be
analysed and that remains our future work.
REFERENCES
[1] Paik, J. K., Some recent advances in the
concepts of plate-effectiveness evaluation, ThinWalled Structures, Vol. 46 , Issues 7-9, pp. 10351046 (2008).
[2] Faulkner, D., A Review of effective plating for
use in the analysis of stiffened plating in bending
and compression, Ship Research, Vol. 19, No. 1,

pp. 1-17 (1975).
[3] Paik, J. K., Thayamballi, A. K., Kim, D. H.,An
analytical method for the ultimate compressive
strength and effective plating of stiffened panels,
Constructional Steel Research, Vol. 49, pp. 43-68
(1999).
[4] IACS CSR. Common structural rules for bulk
carriers,
International
Association
of
Classification Societies, London, UK (2010).
[5] Hughes, O.F., Ship Structural Design: A
Rationally-Based, Computer-Aided Optimization
Approach, SNAME, New Jersey (1988).
[6] Cho, R.S., Choi, B.W. and Song, I.C.,Post
ultimate behaviour of stiffened panels subjected
to axial compression, Proceedings of the
International 2nd Conference on Thin-Walled
Structures (ICTWS'98), Singapore, pp. 433-440.
(1998).
[7] Yao, T. and Nikolov, P.I.,Progressive collapse
analysis of a ships hull under longitudinal
bending, Journal of the Society of Naval
Architects of Japan, Vol. 170, pp. 449-461
(1991).
[8] Yao, T. and Nikolov, P.I.,Progressive collapse
analysis of a ships hull under longitudinal
bending (2nd report), Journal of the Society of
Naval Architects of Japan, Vol. 172, pp. 437-446
(1992).
[9] Gordo,
J.M.
and
Guedes-Soares,
C.,
Approximate load shortening curves for
stiffened plates under uniaxial compression,
Integrity of offshore structures 5, Engineering
Materials Advisory Services (EMAS) Ltd.,
Warley, UK. pp. 189-211 (1993).
[10] Vhanmane, Suhas and Bhattacharya, Baidurya,
On improved analytical method on stress-strain
relationship for plate elements under axial
compressive load, Ship and Offshore Structures,
Vol. 2, No. 4, pp. 347-353 (2007).
[11] Smith, C.S., Davidson, P.C., Chapman, J.C. and
Dowling, P.J., Strength and stiffness of ships'
plating under in-plane compression and tension,
Transactions of RINA, Vol. 130, No. pp. 277-296
(1988).
[12] Paik, J. K., and Thayamballi, A., Ultimate limit
state design of steel-plated structures, John
Wiley & Sons, Chichester, UK. (2003).
[13] Kmiecik, M., Jastrzebski, T. and Kuzniar, J.,
Statistics of ship plating distortions, Marine
Structures, Vol. 8, No. 2, pp. 119-132 (1995).
[14] Chryssanthopoulos, M.K., Probabilistic buckling
analysis of plates and shells, Thin-Walled
Structures, Vol. 30, No. 1-4, pp. 135-157 (1998).
[15] DnV., Classification Notes No. 30.6, Structural
Reliability Analysis of Marine Structures, Det
Norske Veritas, Hovik, Norway, (1992).
263
[16] Guedes Soares, C. and Kmiecik, M., Simulation

of the ultimate compressive strength of
unstiffened
rectangular
plates,
Marine
Structures, Vol. 6, No. 5-6, pp. 553-569 (1993).
[17] Paik, J.K. and Frieze, P.A., Ship structural safety
and reliability, Progress in Structural
Engineering and Materials, Vol. 3, pp. 198-210
(2001).
[18] Harada, M. and Shigemi, T., A method for
estimating the uncertainties in ultimate
longitudinal strength of cross section of ships
hull based on nonlinear FEM, Proceedings of
the 10th International Symposium on Practical
Design of Ships and Other Floating Structures
(PRADS' 07), Houston, USA (2007).
[19] Paik, J.K., Thayamballi, A.K., Sung, K.K. and
Soo, H.Y., Ship hull ultimate strength reliability
considering corrosion, Ship Research, Vol. 42,
No. 2, pp. 154-165 (1998).
[20] Yao, T., Astrup, O.C., Caridis, P., Chen, Y.N.,
Cho, S.R., Dow, R.S., Niho, O. and Riggo, P.,
2000. Committe VI.2, Ultimate Hull Girder
Strength, Proceedings of the 14th International
Ship and Offshore Structures Congress (ISSC
2000), Nagasaki, Japan, 2-6 October (2000).
264

TECHNO ECONOMIC STUDY OF POTENTIAL USING SOLAR ENERGY AS A

SUPPORTING POWER SUPPLY FOR DIESEL ENGINE FOR LANDING CRAFT
O. Sulaiman1, H.Aron2, , A. H. Saharuddin3 ,W.B. Wan Nik4, A.S.A.kader5 and M.F. Ahmad6
5
Department of Mechanical Engineerng, University Technology
Malaysia, Skudai, Johor. Malaysia
1,2,3,4,6
ABSTRACT
Increasing price of oil and the concern of maintaining clean environment have become the main reasons for
a shipping company to focus on the vessel fuel consumption. Fuel economy has become a major concern due to
the increasing of oil price. The availability of fossil fuels is limited and depleting day by day as the consumption
is increasing very rapidly. Environmental impact by diesel exhaust is also taken into consideration as diesel
exhaust is a complex mixture of thousands of gases and fine particles (commonly known as soot) that contains
more than 40 toxic air contaminants. These include many known or suspected cancer-causing substances, such
as benzene, arsenic and formaldehyde. As for these matters, renewable energy is said to have potential in
replacing fossil fuels in providing energy and is more environmental friendly. Solar energy represents one of
the best practice available alternative energy to mitigate these challenges. Solar energy can assist in
supporting auxiliary power for the instrument on board vessel such as landing craft. This study will determine
the potential of using solar PV system onboard a landing craft by analyzing the reduction usage of fuel and
diesel exhaust as well as carrying out economical analysis. The power requirement for the vessel was
calculated. Fuel and money saved by using solar panels were obtained and comparison between data of using
diesel alone and using solar panels as supporting power supply were analyzed. Economical analysis were
obtained by determining data of cost estimation, cash flow diagram, Annual Average Cost (ACC), and return
investment of both vessel with the support of solar PV system and without solar PV system.
Key words: Solar PV system, landing craft, auxiliary power supply, fuel reduction, Annual
(ACC)
1. INTRODUCTION
A lot of studies had been carried out in order to
find an alternative energy for fossil fuel with other
renewable energy resources. One of the most
promising renewable energy is the use of solar
energy. Solar energy can save sailing vessel and
powerboat owners money by reducing their
dependence on shore power and or gas/diesel
generated electricity. Solar power has become an
important part of ocean-cruising by providing
recreational and commercial boaters a secure, safe,
sustainable and renewable energy source. Solar
energy is holding out promise for reduced usage of
fossil fuels, with consequent economic and
environmental benefits [1,2].
This paper present result of techno-economic
analysis of the study that deal with reduction in usage
of fossil fuel and economic aspects analysis. These
Average Cost
points will be the evidence of whether there are any

potential of using solar energy as a supporting power
supply for diesel engine of a landing craft owned by
Dickson Marine Co. Sdn. Bhd. There is currently the
need of an alternative energy to support the power
supply of a marine vessel and also find ways to
minimize the impacts of diesel fuel to the
environment.
This paper describe the potential of using solar
energy with diesel engine on a landing craft by using
other means of supplying energy and reducing diesel
exhaust pollution. The paper discussed assessment of
effectiveness of using solar energy as a supporting
power supply in reducing fuel consumption and diesel
emission. The paper also discussed the economic
analysis that the effectiveness of installation of solar
panels for auxiliary purposes [3,4].
265
2. METHODOLOGY
Data from the landing craft, Dickson 33, which
have report, logbook and plan, is taken to obtain data
on the fuel and power consumption, voyaging route
and its principle particular. The auxiliary power of the
vessel consumed is taken into consideration to
determine the solar panels which are suitable for this
vessel and can supply enough power as supplied by
diesel. Data on solar radiation at certain route is also
taken into consideration. The route of the ship is
limited in Malaysia only. Other data and facts
required include the cost of purchasing diesel fuel and
exhaust of diesel from the vessel. The cost of
purchasing diesel is obtained from the company. The
cost of diesel is determined to compare the cost of
using diesel and the cost of using solar panels. Other
data collected for the solar power which will be used
on the vessel would be the market price of the panels
[5, 7, 10].
Then the total power requirement for 24V DC is
calculated to find the power needed for the solar
panels to supply that will be used in the system. For
this project, solar panels will be used to supply power
of 24 volts DC from 2 banks of 24 volts 200 ampere
hour batteries which runs electronic equipments such
as alarms, emergency lights, radio navigational aids,
navigational lights and other emergency loads on
board the vessel.
Assume the system requires 2 kilowatts for 24
hours per day. As natural losses are also taken into
consideration, the result must be multiplied by 1.2,
assuming 80% of efficiency. The corrected load will
be determined by multiplying power required by
system with 24 hours. Then this load will be
multiplied by 1.2. It is assumed that the solar panels
received 8 hours of solar radiation. So the power is
divided by 8 hours [6, 11].
There are several valid economic criteria that relate to

the profibility of competing ship design [8, 9].

It is known that the north-east monsoon season
is between the months of November to March. So for
5 months, the generator must produce 4/8 to 7/8 of the
system needs. For the dry season, it is assumed that
the solar panels will receive more solar radiation
where the solar PV system can be used to the max.
During the dry season, the generator is estimated to
only produce 1/8 to 4/8 of the system needs. Tables 13 show that, more power is supported by the
generator during wet season, if the solar panels are
used to the max during the dry season, it would
reduce the total annual generator output thus reducing
the fuel consumption of the generator [12, 1, and 17].
Table 1: 5 Months of wet season (November
March)
Hours
Power
produced
generator
by
Power requirement for 5

months
4/8
4/8 x 58.1 = 29.1 kWh
29.1 kWh x 151 days =

4394.1 kWh
5/8
5/8 x 58.1 = 36.3 kWh
36.3 kWh x 151 days =

5481.3 kWh
6/8
6/8 x 58.1 = 43.6 kWh
43.6 kWh x 151 days =

6583.6 kWh
7/8
7/8 x 58.1 = 50.8 kWh
50.8 kWh x 151 days =

7670.8 kWh
Table 2: 7 Months of dry season (April October

Hours
Power
produced
generator
1/8
1/8 x 58.1 = 7.3 kWh
7.3 kWh x
1562.2 kWh
214days =
Cost Estimation
When alternative power source such as solar
panels is to be installed to a vessel, some costs need to
be taken into consideration. The costs are
investment/initial cost, maintenance cost, operation
cost and salvage value. Comparison between the cost
for vessel with solar PV system and vessel without
solar PV system are discussed.
2/8
2/8 x 58.1 = 14.5 kWh
14.5 kWh x
3101.0 kWh
214days =
3/8
3/8 x 58.1 = 21.8 kWh
21.8 kWh x
4665.2 kWh
214days =
4/8
4/8 x 58.1 = 29.05 kWh
29.5 kWh x
6216.7 kWh
214days =
2.2 Economical Analysis
Wet season
Dry season
Total
Annual
Generator Output
4/8 = 4394.1 kWh
4/8 = 6216.7 kWh
10610.8 kWh
5/8 = 5481.3 kWh
3/8 = 4665.2 kWh
10146.5 kWh
6/8 = 6583.6 kWh
2/8 = 3101.0 kWh
9684.6 kWh
7/8 = 7670.8 kWh
1/8 = 1562.2 kWh
9233.0 kWh
2.1
Economical aspect is one of the main points that

would be the evidence of whether solar power has the
potential to be used on board the vessel as an assisting
power supply for the electrical instruments to reduce
fuel consumption thus saving cost of purchasing
diesel fuel. For this project, Annual Average Cost
(AAC) between actual vessel which is operated by
main engine and solar assisted vessel are studied.
by
Power requirement for 7

months
Table 3: 7 months of wet season (April October)
266
Without the PV system, the generator would have

to provide 58.1 kWh per day for the whole 365 days
in a year. So if 58.1 multiplied by 365 days would
results in 21206.5 kWh of annual generator output.
For the maximum use of PV system during dry
season, the PV system saves 21206.5 kWh 9233.0
kWh = 11973.5 of generation by generator.
It is estimated that 1 kWh uses 0.3 liters of fuel.
For 9233.0 kWh generator output, the generator is
using 2769.9 liters of fuel per year. If the vessel is
using the generator for the whole power supply, the
total generator output multiplied by total days in a
year, results in 21206.5 kWh.
As 1 kWh uses 0.3 liters fuel, 21206.5 kWh x
0.3 = 6361.95 liters of fuel. To determine the liters of
fuel saved by using solar panels for 1//8 and 7/8
fraction, the total fuel used by generator without solar
panels is deducted by the total fuel used by generator
with support of solar panels. By using solar panels to
the max, the generator can save 3592.1 liters of fuel
per year (See Table 4 and 5).
Table 4: Assume that 1 liter of diesel today cost RM2
( 4 RM = 1USD)
Generation
generator
by
Total
Annual
Generator
Output
Energy
saved
by using
solar
panels
Fuel
saved
(liters)
Money
saved
(RM)
Wet
season
Dry
season
(5
months)
(7
months)
4/8=
4394.1
kWh
4/8
=
6216.7
kWh
10610.8
kWh
10595.7
kWh
3178.7
3178.7
x 2.0 =
6356.2
5/8
=
5481.3
kWh
3/8
=
4665.2
kWh
10146.5
kWh
11060.0
kWh
3318.0
3318.0
x 2.0 =
6636.0
6/8
=
6583.6
kWh
2/8
=
3101.0
kWh
9684.6
kWh
11521.9
kWh
3456.6
3456.6
x 2.0 =
6913.2
7/8
=
7670.8
kWh
1/8
=
1562.2
kWh
9233.0
kWh
11973.5
kWh
3592.1
3592.1
x 2.0 =
7184.2
Table 5: Percentages of fuel savings

Generation
generator
Wet
season
(5
months)
4/8=
4394.1
kWh
5/8
=
5481.3
kWh
6/8
=
6583.6
kWh
7/8
=
7670.8
kWh
by
Dry
season
(7
months)
4/8
=
6216.7
kWh
3/8
=
4665.2
kWh
2/8
=
3101.0
kWh
1/8
=
1562.2
kWh
Total
Annua
l
Gener
ator
Output
10610.
8 kWh
Energy
saved by
using
solar
panels
Fuel
saved
(liters)
Savings
percentage
10595.7
kWh
3178.7
50%
10146.
5 kWh
11060.0
kWh
3318.0
52%
9684.6
kWh
11521.9
kWh
3456.6
54.3%
9233.0
kWh
11973.5
kWh
3592.1
56.5%
The solar panels are used to supply power to be

stored in the batteries to support power for alarms,
emergency
lights,
radio
navigational
aids,
navigational lights and emergency loads. These
equipments are very important even though they do
not use a high amount of power. With the support of
solar panels, the power will be stored in the batteries
where the vessel will continuously receive power
supply even though both of the generator sets broke
down. These equipments are crucial during
emergencies at sea [14, 15, 16].
3.1 Economic Results
Economical analysis is done to make comparison
between the original vessel and the vessel with a solar
PV system attached to it to support the generator in
supplying power for auxiliary purposes. The analysis
is carried out to determine Annual Average Cost
(AAC) and the capital investment recovery. The
analysis also discussed on the sensitivity due to oil
price.
3.2 Cash Flow Diagram

Cash flow diagram is determined to show the
flow of the costs for a project. The positive direction
shows that the profit (debit) and negative direction
shows that expenditure (credit).
For overall fuel consumption by both main

engine and generator, the percentage of the fuel
reduced is determined by obtaining the difference
between the amount of fuel used by the vessel without
PV system and the amount of fuel used with the
support of PV system. The difference is divided with
the normal amount of fuel used without the support of
PV system and the value is then multiplied with 100
to get the percentage. The percentage of fuel reduced
is 0.66%. Even though the percentage of saved fuel is
small, for a long term consideration, it would really
help in saving the environment due to the saved fuel
as lesser fuel used would reduce the exhaust.
267
Figure 1: The costs for the vessel without solar

(including burn cost)
From Figure 2 the cash flow diagram shows

the cash flow for the vessel without PV solar system
including the burn cost which is the cost to build the
vessel. The maintenance cost will go through 1.5
percent reduction each year. The operation cost and
annual income are estimated to be the same each year.
3.3
Annual Cost
Annual Average Cost (AAC) between the

original vessel and the solar assisted vessel will be
analyzed in this section.
The calculation of AAC can be obtained by using the
following formula.
AAC (NPV)
(equation 1)
i = 10%, n =20 year
NPV = total net present value
The result from the Figure, both cases under
consideration for the vessels are found to be
profitable due to the positive direction according to
the cash flow charts. The ACC for vessel with solar
PV system is found to be lower than the ACC for the
original vessel. The result shows that using solar PV
system on the vessel is more economical rather than
to use diesel generator alone. The analysis may get a
better result if the cost of the installation of the PV
system is much cheaper and more efficient in
supporting the power supply. With a lower cost and
the ability to save more fuel, the result may come out
much better where using solar panels would give
more benefit and profit rather than disadvantages.
Figure 2: Cash Flow Diagram for vessel without solar

Figure 3 shows the cash flow diagram for
vessel without solar. In this figure, all the costs are
sum up to get the present cost except for the salvage
value as it is profitable only after 20 years of usage.
3.4
Return Investment
Return investment is carried out to determine

the numbers of years that the investment will be
recovered. The capital investment recovery can be
determined as shown;
Investment cost for boat with solar PV system
= RM97280.30
Cost saved = RM7184.20
Investment recovered = RM97280.30/RM7194.20
=13.5year= 14 year
4.
Figure 3: Cash flow for vessel with solar PV system

In Figure 3, the graph shows the cash flow
diagram for vessel with solar PV system. The graph is
similar to the 2 graphs above but this graph includes
the investment cost on solar PV system.
CONCLUSION
The use of solar panels to supply some of the

power required by instruments and lightings onboard
the vessel reduced fuel consumption and thus
reducing the diesel exhaust. With the instability of
fuel price nowadays, reduction in fuel consumption
would help in saving the companys money and
would somehow increase the income. The reduction
in diesel exhaust might be small but for a long term
period, the reduction of those exhausts would really
help in saving our environment from pollution. The
solar panels would also helps in keeping the power
supply continuously if both generators face
breakdown or difficulties. As the solar energy provide
power for 24V DC, it will ensure there will be power
to operate the alarms, emergency lights, radio
268
navigational aids, navigational lights and other

emergency loads. Even though these instruments use
a small amount of power, their functions are very
important in case of emergencies and also for
navigational purposes.
In conclusion the use of solar panels will give
benefits to the vessel and also to the environment.
Reduction of fuel consumption and diesel exhaust are
proven to show that there are potential in using solar
energy to support the power requirement for the
landing craft as the installation of the solar hybrid
system give benefits to the company and also the
environment.
4.1
- Design of a prototype of the system is encouraged as

it would give more accurate results.
-Different types of solar panels can be taken into
consideration to improve the performance of the
system. There are more improved and hi-tech solar
panels nowadays that can be used for the system and
may give better results.
-It is also advisable to reduce the installation cost by
looking for cheaper yet efficient equipments for solar
PV system especially the solar panels. By reducing
the installation price it would give a much better
results in the economic analysis for the vessel with
solar PV system.
-Comparisons with other hybrid system are also
recommended to find the best system that can be
applied on a vessel.
REFERENCES
[2]
[3]
Regulations for the Prevention of Air Pollution

from Ships. (1997). Available at website
(http//:www.imo.org)
[6]
M. Asif and T. Munee, Energy supply, its

demand and security issues for developed and
emerging
economies,
Renewable
and
Sustainable Energy Reviews, (2007), pp. 13881413.
[7]
IEA. International energy agency statistics, 2002
[8]
BP statistical review of world energy. British

Petroleum, 2003.
[9]
M. Salameh, Oil crises. Historic perspective.

Encyclopedia of energy vol. 4, Elsevier,
Amsterdam (2004).
Recommendations
There are a lot of aspects that could be

investigated and improve in the future for a much
better results. Some of the suggestion and
recommendation are as follows [17, 18].
[1]
[5]
Yoshihiro HAMAKAWA, Recent Advance In

Solar Photovoltaic Technology and Its Role
for Environment Issue, Renewable Energy,
(1994), pp.34-43.
Health Effect of Diesel Exhaust, A Fact Sheet
by Cal/EPAs Office of Environmental Health
Hazard Assessment and American Lung
Association of California. Available at
website (http//:www.oehha.ca.gov)
[10] U.S. Department of Energy, Energy Efficiency

and Renewable Energy, transatlantic21.org,
solarsailor.com, news.bbc.co.uk
[11] Roger Messenger and Jerry Venture, PV System
Examples,
Photovoltaic
Systems
Engineering,(2000), pp 119-121.
[12] David Roche, Antony Schinckel, John Storey,
Clive Humphris and Michelle Guelden, Chapter
6 Solar Arrays, Solar Racing Cars: 1996 World
Solar Challenge, Australian Government
Publishing Service, Canberra, (1996). PP 95-97.
[13] Dubois A, et al. Design and operation of a
solarphotovoltaicwind energy cogeneration
system for a group of houses. Advances in
Solar
Energy
Technology.Vol.
PergamonPress, 1988. pp. 441446.
[14] E.D. Gomez et al., State of the Marine
Environment in the East Asian Region. UNE
Regional Seas Reports and Studies No 126
(1990).
[15] Azlan Musa. Feasibility study of incorporating
wind and solar assisted technology in ships.
Master Thesis, University Teknologi Malaysia.
[16] Solar
Panel
Selection
Guide.
http://www.ecodirect.com/Sharp-NU-U230F3
230-Watt-20-Volt-p/sharp-nu-u230f3.htm Date:
(8 February 2010)
[17] O.Sulaiman et. al., Multihybrid Energy for
Marine System Powering.Port Congress. (2010),
Perth
[18] Kamaruzzaman Sopian and Mohd. Yusof Hj.
Othman, (1991) Estimates of monthly average
daily global solar radiation in Malaysia.
Bos, E., et al., World Population Projections,

1994-95, The Johns Hopkins University Press
Baltimore, MD, 1994.
[4] Benny Joseph, K G Gopichandran, and V K

Vaidyan, The Sun - Major Energy Resource for
the Future.(1997)
269
NUMERICAL INVESTIGATION ON NATURAL CONVECTION FROM AN OPEN

RECTANGULAR CAVITY CONTAINING A HEATED CIRCULAR CYLINDER
Sheikh Anwar Hossain1 and Dr. M. A. Alim2
1.Department of Mathematics, Jagannath University, Dhaka
E-mail: ahossain4044@yahoo.com
2. Department of Mathematics, BUET, Dhaka

E-mail: a0alim@gmail.com
ABSTRACT
The problem of natural convection heat transfer in a square open cavity containing a heated and conducting
circular cylinder at the centre is analyzed in this paper. As boundary conditions of the cavity, the left vertical
wall is kept at a constant heat flux, bottom and top wall are kept at different high and cold temperature
respectively. The remaining side is open. Two dimensional laminar steady state natural convection is considered.
This configuration is related in the design of electronic devices, solar energy receivers, uncovered flat plate
solar collectors, geothermal reservoirs etc. The fluid is concerned with different Prandtl numbers; Grashof
numbers and the properties of the fluid are assumed to be constant. The development of Mathematical model is
governed by the coupled equations of continuity, momentum and energy and is solved by employing Galerkin
weighted finite element method. Flow field and heat transfer were predicted for fluid with Pr = 0.72 , 1.0 ,7.0;
Gr = 103,104, 105 ,106; inclination angles of the cavity are = 0o, 15o,30o, 45o and cylinder diameter to cavity
length ratio dr = 0.2.The average Nusselt number increases as the increases of inclination angle of the cavity
for lower Pr and lower temperature at bottom wall. The average Nu increases mainly for higher inclinations
and for higher Gr. Various vortexes and recirculations are formed into the flow field for higher Pr and higher
temperature at the bottom wall.
Key words: Heat transfer, finite element method, natural convection, square open cavity, heated circular
cylinder.
NOMENCLATURE
1. INTRODUCTION
Cfx Local skin friction coefficient

Cp Specific heat at constant pressure
f
Dimensionless stream function
g
Gravitational acceleration (ms2)
Gr Grashof number
K Thermal conductivity of the fluid (Wm1K1)
H Height
L Length of the cavity (m)
D Diameter of the cylinder
Nu Nusselt number
P
pressure (Nm2)
P non-dimensional pressure
Pr Prandtl number, /
Q Heat flux (Wm2)
dr cylinder diameter to cavity length ratio
T Temperature of the fluid in the cavity (K)
Ts Temperature at the surface (K)
T Temperature of the ambient fluid (K)
Th High temperature at the bottom wall (K)
Tc Cold temperature at the upper wall (K)
Nuav Average Nusselt number
u,v Velocity component (ms-1)
U Free stream velocity(ms-1)
U0 Reference velocity
U,V non-dimensional velocity components
x,y Cartesian co-ordinates (m)
X,Y non-dimensional Cartesian co-ordinates
Tcl temperature of the cylinder
The study of natural convection in open cavities

has been the subject of many experimental and
numerical investigations during the past two decades.
The previous researchers have investigated the effect
on flow and heat transfer for different Rayleigh
numbers, aspect ratios and tilt angles. As a result a
number of studies have been done for natural
convection heat transfer over a cylinder placed inside
a closed enclosure.
Showole
and
Tarasuk
(1993)
studied
experimentally and numerically, the steady state
natural convection in a two dimensional isothermal
open cavity. They found the experimental results for
air, varying the Rayleigh number from 104 to 5.5105,
cavity aspect ratios of 0.25, 0.5 and 1.0 and
inclination angles of 0o, 30, 45o and 60 (for 0, the
wall opposite the aperture was horizontal and the
angles were taken clockwise). They calculated the
numerical results for Raleigh numbers between 104
and 5.5105, inclination angles of 0 and 45 and an
aspect ratio equal to one. In the result it is seen that
for all Rayleigh numbers, the first inclination of the
cavity caused a significant increase in the average
heat transfer rate, but a further increase in the
inclination angle caused very little increase in the heat
transfer rate. In another result it is observed that, for
0, two symmetric counter rotating eddies were
271
formed, while at inclination angles greater than 0, the

symmetric flow and temperature patterns disappear.
Momentum
Chan and Tien (1985a) investigated a square open

cavity numerically, which had an isothermal vertical
heated side facing the opening and two adjoining
adiabatic horizontal sides. To obtain the satisfactory
solutions in the open cavity the boundary conditions
for field were approximated.
U
U
P
1 2U 2U
+ sin
+V
=
+
+
X
Y
X
Gr X 2 Y 2
V
V
P
1 2V 2V
+V
=
+
2 + 2 + cos
X
Y
Y
Y
Gr X
Chan and Tien (1985b) investigated shallow open

cavities and made a comparison study using a square
cavity in an enlarged computational domain. In the
result they observed that for a square open cavity
having an isothermal vertical side facing the opening
and two adjoining adiabatic horizontal sides.
Satisfactory heat transfer results could be obtained,
especially at high Rayleigh numbers.
Energy equation
equations:
1 2 2
+V
=
+
X
Y Pr Gr X 2 Y 2
Tc
Mohammad (1995) investigated inclined open

square cavities, by considering a restricted
computational domain. The gradients of both velocity
components were set to zero at the opening plane in
that case which were different from those of Chan and
Tien (1985a). In the result he found that heat transfer
was not sensitive to inclination angle and the flow
was unstable at high Rayleigh numbers and small
inclination angles.
Th
Rahman et al. (2009) investigated on Mixed

Convection in a rectangular cavity with inlet, outlet
and a heat conducting Horizontal Circular Cylinder.
They found that both the heat transfer rate from the
heated wall and the dimensionless temperature in the
cavity strongly depends on the governing parameters
and configurations of the system such as size,
location, and thermal conductivity of the cylinder and
location of the inflow, outflow opening.
L
(a) Horizontal cavity
L
Tc
2. MODEL SPECIFICATION
2.1 Physical Model

The physical model considered here along with
necessary geometric parameters. In this study heat
transfer and the fluid flow in a two-dimensional open
rectangular cavity of length L was considered, which
is shown in the following schematic diagram.
U V
+
=0
X Y
D
Th
2.2 Mathematical Formulation

The flow inside the cavity is assumed to be twodimensional, steady, laminar, and incompressible and
the fluid properties are to be constant. The radiation
effects are taken negligible and the Boussinessq
approximation is used. The dimensionless governing
equations describing the flow are as follows:
(b) Inclined cavity

Figure 1. Schematic diagram of the physical problem
2.2.1 Boundary Conditions
At the bottom wall: U = V = 0; = 1

At the top wall: U = V = 0 ; = 0
At the left wall: U =V = 0 ; heat flux q = 500
At the right side & open side:
272
Convective Boundary Condition (CBC), P = 0
nodes, 11413 elements; 96030 nodes, 12356 elements;

892450 nodes, 13686 elements. In this case 982450
nodes, 13686 elements can be chosen through the
processing to optimize the relation between the
accuracy required and the computational time.
Non-dimensional scales:
X =
x
y
u
v
, Y= , U=
, V=
,
L
L
U0
U0
P P
T T
v
g TL
, =
, Pr = , Gr =
2
U0
t
2
D
qL
t = (Ts T ) , t =
dr = ,
L
K
P=
2.3 Used Heat Transfer Parameter

Since the flow field in natural convection is
governed by the dimensionless Grashof Number Gr.
Where the Grashof number represents the ratio of the
buoyancy force to the viscous force acting on the fluid
and the reference velocity Uo is related to the
buoyancy force term and is defined as
2.5 Code Validation

Due to lack of availability of experimental and
numerical results published earlier on this regard,
validation of this prediction could not be done. The
present numerical result is verified against
documented numerical study, namely; Finite element
simulation of natural convection from an open
rectangular cavity containing adiabatic circular
cylinder the M. Phil. Thesis in Mathematics by
Koabra (2008) BUET, Dhaka.
Table1. Results for the different Grashof numbers
with Pr = 0.72, dr = 0.2
Gr
U 0 = g L (Ts T )
The Nusselt number Nu is also an important nondimensional parameter to be computed for heat
transfer analysis in natural convection flow. The
Nusselt number for natural convection is a function of
the Grashof number only. The local Nusselt number
Nu can be obtained from the temperature field by
applying
the
function
Nu =
1
(0, Y )
The overall or average Nusselt number was

calculated by integrating the temperature gradient
over the heated wall as follows:
1
1
dy
(0, Y )
0
Nuav =
Since the dimensionless Prandtl Number Pr is the

ratio of kinematic viscosity to thermal diffusivity. So
Pr is a heat transfer parameter in the flow field of
natural convection.
2.4 Grid Independence Test

To obtain grid independent solution, a grid
refinement study is performed for a rectangular open
cavity with Gr=106 and dr=0.2.The test shows the
convergence of the average Nusselt number Nu at the
heated surface with grid refinement. It is observed
that grid independence is achieved with 13686
elements where there is insignificant change in Nu
with further increase of mesh elements. The grid
refinement tests are done by taking six different non
uniform grids with the following number of nodes and
elements: 27342 nodes, 4818 elements; 49335 nodes,
7663 elements; 72782 nodes, 10365 elements; 73542
103
104
105
106
Nuav
Present work
Koabra(2008)
3.3282
3.4014
4.4181
5.8227
3.2473
3.3020
4.1825
5.5521
3. RESULTS AND DISCUSSIONS

The results Nuav are obtained for Grashof number
from 103 to 106 at Pr = 0.72, 1.0, 7.0 while inclination
angles 0o to 45o and different high temperatures 320k,
350k , 375k at bottom wall with constant physical
properties. Here the parametric analysis for a wide
range of governing parameters shows consistence
performance of the present numerical approach to
obtain as streamlines and temperature profiles. The
obtained results show that the heat transfer coefficient
Nu is strongly affected by Grashof number. An
empirical correlation can be developed using Nusselt
number and Grashof number. In the effect of Grashof
number, the isotherm figures show that as the Grashof
number and the inclination angle increases the
buoyancy force increases and the thermal boundary
layers become thinner in Figures 2 and3. The Figures
2 and 3 of streamlines and isotherms for Pr = 0.72,
1.0, 7.0, high temperature Th =320k at the bottom
wall, cold temperature Tc = 276 k at the upper wall
and Tcl =300k temperature of the cylinder show that
the fluid moves from the bottom wall of the cavity,
circulates in a clockwise direction around the cylinder
and moves toward the upper part of the cavity and
fresh cool fluid enter the cavity through the bottom
wall. The streamlines patterns are very similar for
different inclinations while Gr = 103. The fluid moves
faster for G=104 to 106 and the upper thermal
boundary layer becomes thinner. One vortex is formed
at the centre of the opening side for each of Gr =103 to
273
106, because of the effect of the heated circular

cylinder. One vortex is formed near the cylinder for
each of Gr = 104, 105 and two vortexes are formed for
Gr =106 near the bottom wall. Streamlines show that
as the inclination angle of the heated wall increases,
the velocity gradient remained almost same as before
for Pr = 0.72, 1 and 7.0. The variations of the average
Nusselt number and the average temperature are also
presented. The results are obtained for a Grashof
numbers ranging from 103 to 106 and for inclination
angles range from 0o to 45o of the cavity. In figures of
streamlines are almost laminar for Gr = 103 while Pr =
7.0. Due to the effect of the heated cylinder, a vortex
is formed at the open side of the cavity for each of Gr
= 103 to 106. Due to increase in Th at the bottom wall,
the fluid rises faster from the bottom wall and one
vortex is created for each of Gr = 104, 105 near the
cylinder, one small vortex is formed near the bottom
wall while Gr = 105. Two vortexes are formed near the
bottom wall for Gr =106 .The streamlines and
isotherms for dr = 0.2, Pr = 1.0, Th = 375 K, Tc =
=278 K and for inclination angles = 0o, 30o, 45o are
shown in Figure 4 and Figure 5 respectively for the
variation of Grashof number. One vortex is created for
each of Gr = 103, 104, 105, 106 on the open side. This
unsteady streamlines and isotherms are occurred due
Pr =0.72,
to higher Prandtl number Pr = 7.0. Since the relation

between Nu and Pr is non-linear and one vortex is
formed near the bottom wall for Gr = 106. The
isotherms patterns of these Figures are similar.
For Grashof number 103, the streamlines are almost
laminar but one vortex is created for each of Gr = 104,
105; but two recirculation cells are formed near the
cylinder and bottom wall for Gr = 106. One small
vortex is formed at the bottom wall. One vortex is
created near the cylinder for each of Gr = 103, 104.
Two recirculation are formed for Gr = 105,106 one of
these is near the cylinder and another is near the
bottom wall. One vortex is formed on the open side
for each of Gr =103, 104, 105 and 106. This is due to
increase in temperature at the bottom wall and
presence of the highly heated circular cylinder and
cylinder redirect the fluid. One vortex is formed on
the open side for each of Gr =103, 104, 105 and 106.
This is due to increase in temperature at the bottom
wall and presence of the highly heated circular
cylinder and cylinder redirect the fluid. Some
streamlines and isotherms patterns of this problem are
shown in the following Figures from Figure 2 to
Figure 5.
= 0o
Pr = 1.0,
Gr = 103
Gr =104
Gr =105
274
=30o
Pr = 7.0, = 45o
Gr = 106
Figure 2: Streamlines for dr = 0.2, Gr = 103, 104 , 105, 106 and Th =320k
Pr =0.72,
= 0o
Pr = 1.0,
=30o
Pr = 7.0, = 45o
Gr = 103
Gr =104
Gr =105
Gr = 106
Figure 3: Figures of Isotherms for dr = 0.2, Gr = 103, 104, 105, 106 and Th = 320
275
Table 2. Average Nusselt number Nuav for different

inclination angles and Grashof numbers with Pr =
0.72 and dr = 0.2.
5. CONCLUSION
The flow field and heat transfer are studied in a square
open cavity containing a heated circular cylinder of
diameter ratio dr = 0.2. The effect of Pr , Gr and high
temperature at the bottom wall are studied for various
inclination angle for the cavity. In this study we
observed that the heat transfer rate depends on Prandtl
number and Grashof number. Heat transfer rate
increases as Prandtl number increases. Thermal
boundary layer thickness becomes thinner for higher
Grashof number. The heat transfer rate decreases for
Gr =103 and increases gradually for increasing of
Grashof number. The heat transfers rate Nu increases
as inclination angle of the cavity and Grashof number
increase. Various vortexes and recirculation are
formed into the flow field and a boundary vortex at
the centre of the cavity is seen in the streamlines.
Nuav
Gr=10
Gr=10
Gr=105
Gr=106
0o
3.2725
3.29
4.07
5.4947
15o
3.2756
3.37
4.2837
5.8547
30o
3.3525
3.5443
4.5646
5.9657
45o
3.4125
3.4956
4.7543
5.9757
Table 3. Average Nusselt numbers (Nuav) for different

Prandtl numbers and Gr taking Pr = 0.72, 1.0 and 7.0,
angle =0o and dr =0.2
Pr
Nuav
Gr =103
Gr =104
Gr =105
Gr =106
0.72
3.2825
3.2975
4.2575
5.6747
1.00
3.2734
3.3595
4.3465
5.8857
7.00
3.3255
4.4231
4.8753
8.8447
REFFERENCES
[1] Chan, Y. L. and Tien, C. L., A Numerical study
of two-dimensional natural convection in square open
cavities, Numerical Heat Transfer, Vol. 8, pp. 65-80,
1985b.
[2] Chan,Y. L. and Tien, C. L., A Numerical study
of two-dimensional laminar natural convection in a
shallow open cavity, Int. J. Heat Mass Transfer, Vol.
28, pp. 603-612,1985a.
= 0o
Gr = 103
Gr =104
276
=30o
= 45o
Gr =105
Gr = 106
Figure 4. Streamlines patterns for dr = 0,2 , Pr = 1.0.and Th = 375 k
= 0o
Gr = 103
Gr =104
Gr =105
277
=30o
= 45o
Gr = 106
Figure 5: Isotherms patterns for dr = 0,2 , , Pr = 1.0. and Th = 375 k

[3] Fariza Tul Koabra Numerical investigation on
Two-dimensional laminar steady state natural
convection in a rectangular open cavity containing a
adiabatic circular cylinder. Thesis paper for the
degree of M. Phil.in Mathematics, BUET, Dhaka
(2008).
[4] Mohamad, A., Natural convection in open
cavities and slots, Numerical Heat Transfer, Vol. 27,
pp. 705-716, 1995.
[5] Rahman Md. M., Alim M. A., Mamun M.A H.

Finite Element Analysis of Mixed Convection in a
Rectangular Cavity with a HeatConducting
Horizontal Circular Cylinder, J. of Non Linear
Analysis: Modeling and Control, 2009, vol-14, No 2.
[6] Showole, R.A. and Tarasuk, J.D., Experimental
and numerical studies of natural convection with flow
separation in upward-facing inclined open cavities,
Journal of Heat Transfer, Vol. 115, pp. 592-605,
1993.
278

A NUMERICAL STUDY ON HYDRODYNAMIC INTERACTION FOR A SMALL

3-D BODY FLOATING FREELY CLOSE TO A LARGE 3-D BODY IN WAVES
Mir Tareque Ali1, Md. Mashiur Rahaman2 and Tasnuva Anan3
Associate Professor
Marine Engineering, BUET
mtarequeali@name.buet.ac.bd
Assistant Professor
2nd E-mail
M. Sc. student
3rd E-mail
ABSTRACT
This paper investigates the first order wave exciting forces and motion responses due to hydrodynamic
interaction between two unequal-sized freely floating three dimensional rectangular boxes in regular waves.
The three dimensional source distribution method has been adopted to carry out the numerical investigation.
The validation of the computer code developed for this purpose has been justified by comparing the present
results with that of the published ones for simple geometrical shaped floating bodies. The numerical
computations have been carried out for different wave heading angles and separation distances (gaps) between
the floating rectangular boxes. To illustrate the hydrodynamic interaction phenomena, the computed results for
an isolated body case are also presented along with the multi-body results. Finally some conclusions have been
drawn on the basis of the present analysis.
Keywords: Hydrodynamic interaction, 3-D rectangular boxes, Source distribution technique.
1. INTRODUCTION
Recently, there is a significant increase in the
number of activities that involves applications of
offshore structures, where two or, more bodies are
floating in sufficiently close proximity. As a result of
the wave motion, these bodies will experience the
hydrodynamic interaction. The hydrodynamic
analysis of multiple floating bodies is different from
that of a single floating body, because one body is
situated in the diffracted wave field of others. The
bodies will experience incident as well as scattered
waves impinges upon them. Now if these waves
arrive in phase then there will be a considerable
escalation in the magnitude of the wave exciting
forces on the floating bodies compared to a body in
isolation. On the other hand, if these waves arrive out
of phase then there will be a significant reduction of
wave force. Moreover, each body will also experience
radiated waves due to the motion of other bodies. The
actual importance of interaction effect depends on the
configuration of the multi-body system, which means
the size and shape of the floating bodies and the
separation distances (gaps) between them.
Understanding hydrodynamic interaction is
particularly important while a small body is floating
close to a large body in waves. As for example, a

supply vessel is operating close to a large ocean going
vessel at sea. The wave exciting forces and the motion
responses of the small vessel will be affected
significantly due to the presence of neighboring large
vessel. The interaction between multiple floating
bodies will play an important role for the safety and
performance of the floating bodies. For this reason, it
is necessary to have an insight into the mechanism of
the influence of gap between the floating bodies as
well as the wave heading influence on the wave
exciting forces and motion responses of the bodies
floating at close vicinity.
There are many investigations related to the
hydrodynamic interaction between multiple floating
bodies in waves. Ohkusu [1] extended the classical
solution for a single heaving circular cylinder to the
case of two cylinders in a catamaran configuration.
Faltinsen & Michelsen [2] used panel method for
direct numerical solution of wave effects on 3-D
floating bodies. The panel method was further
extended for two independent bodies by van
Oortmerssen [3]. Matsui and Tamaki [4] calculated
hydrodynamic coefficients and wave exciting forces
for simple vertical floating cylindrical bodies in close
proximity. Choi & Hong [5] used the higher-order
279
boundary element method to solve the interaction

problem between multiple floating bodies. Recently,
Wang et al. [6] investigated the problems experienced
during tandem offloading operations of Floating
Production, Storage and Offloading (FPSO) using a
Time-domain simulation technique considering
hydrodynamic interaction between the vessels.
In this paper, the 3-D source distribution method
has been adopted to calculate the hydrodynamic
coefficients and wave-exciting forces for two
unequal-sized rectangular boxes by taking into
account the effect of hydrodynamic interaction among
the two bodies and the coupled equations of motions
are solved directly. Based on the formulation, a
computer code has been developed to investigate the
hydrodynamic interaction phenomena. In order to
justify the validity of the code, some published results
have been verified for two freely floating vertical
cylindrical bodies. Then the numerical computations
are performed for the case of two floating boxes as
well as for an isolated box case, varying the wave
headings and the separation distances (gaps). It is
observed from the present study that the wave
exciting forces and motion responses due to
hydrodynamic interaction vary significantly with
wave headings and gaps between the floating boxes.
= i ( 0 + 7 ) a + ( X mj mj )
where
is the incident wave potential,
diffracted wave potential,
jm
radiation wave potentials, X
is the
represent potentials
due to motion of body 'm' in

m
j
(3)
m =1 j =1
j -th mode i.e.,
is the motion of body
'm' in j -th mode and a is the incident wave

amplitude. The incident wave potential can be
expressed as
0 =
g cosh[k ( z + h)] ik ( x cos + y sin )

e
cosh kh
2
where
(4)
is the wave heading angle measured
+ X -axis, h is the depth of water, g is the

acceleration due to gravity and k is the wave
from
number. The individual potentials are all solutions of

the Laplace equation, which satisfy the linearized free
surface condition and the boundary conditions on the
sea floor, on the bodys surface and at infinity.
2.2 Source Density and Velocity Potentials
2. FORMULATION OF THE PROBLEM
The potential function at some point ( x, y, z ) in

the fluid region in terms of surface distribution of
sources can be written as:
2.1 Assumptions and Boundary Conditions

Consider a group of N 3-dimensional bodies of
arbitrary shape, oscillating in water of uniform depth.
The amplitudes of the motions of the bodies and
waves are assumed to be small, whereas the fluid is
supposed to be ideal and irrotational. Two righthanded Cartesian coordinate systems of axes with
origin at the free surface are considered. In regular
waves a linear potential , which is a function of
space and of time, can be written as a product of
space-dependent term and a harmonic time-dependent
term as follows:
( x, y, z; t ) = ( x, y, z ) . e it
The wave circular frequency
2
T
(1)
can be written as
mj ( x, y, z )
=
1
4
( ,, ) G(x, y, z; ,, ) dS
n =1 S n
m
j
(5
)
( , , ) is a point on surface S and

( , , ) is the unknown source density. The
where
solution to the boundary value problem is given by

Equation (5), which satisfies all the boundary
conditions. And since Greens function ( G ) satisfies
these conditions, applying the kinematics boundary
condition on the immersed surface yields the
following integral equation:
(2)
where T is the wave period. The potential function

can be separated into contributions from all modes
of motion of the bodies and from the incident and
diffracted wave fields as follows:
280
mj ( x, y, z )
1
= mj ( x, y, z )
n
2
1 N
+
mj ( , , )
G ( x, y, z; , , ) dS
4 n =1 S
n
n
(6)
for m n , the coefficients are due to the motion of

other bodies.
2.3 Numerical Evaluation of Velocity

Potentials
A numerical approach is required to solve the
integral Equation (6), as the kernel
G
n
is complex
and it does not permit any solution in closed form.

The wetted surface of body is divided into l number
of
quadrilateral
panels
of
area
2.5 Equations of Motions in Frequency

Domain
The equations of motion can be expressed by using
the following matrix relationship:
slm (l = 1........E n ) and the node points are

considered at the centroid of each panel. The
continuous formulation of the solution indicates that
Equation (6) is to be satisfied at all points
( x , y , z ) on the immersed surface but in order to
obtain a dicretized numerical solution it is necessary
to relax this requirement and to apply the condition at
only N control points and the location of the control
points are chosen at the centroids of the panels.
Consequently, discretization process allows Equation
(6) to be replaced as
1
1
( mj ) l +
4
2
( )
En
m
j )k
n =1 k =1 S n
G
(l , k )dS
n
( M + a ) X&& + bX& + cX = F
(11)
where M is the inertia matrices, a is the added

mass matrices, b is the fluid damping matrices, C is
the hydrostatic stiffness matrices, F is the wave
exciting force vector and X is the motion response
vector. The above equations of motion are established
at the centers of gravity of each body of the multibody floating system. Since each body is assumed
rigid and has six degrees of freedom, each matrix on
the left-hand side of Equation (11) has a dimension of
(6 N 6 N ) and X and F are (6 N 1) column
vectors for N floating body system.
0
, j=7
n l
nm
, j = 1,2....6
j l
= (if panel ' l ' belongs to body ' m' )

0
, j = 1,2....6
(if panel ' l ' does not belong to body ' m' )
( )

(7)
3.1 Two Freely Floating Vertical Cylinders

The diameter and draft of each cylinder is 40.0 m
and 10.0 m respectively and the gap between them is
20.0 m. The water depth is considered as 200.0 m.
The wetted surface of each cylinder is divided into
o
234 panels. For 0 wave heading, Body 1 and Body 2
represent the lee side and weather side cylinder
respectively as shown in Figure 1.
2.4 Hydrodynamic Coefficients and Wave

Exciting Forces and Moments
Once the velocity potentials have been determined,
mn
then the added-mass coefficients ( a kj ), the fluid

mn
damping coefficients ( bkj ) and the first order wavem
exciting forces ( Fk ) can be calculated as follows:
akjmn = e nj nkm dS
S m
(8)
bkjmn = m nj nkm dS
S m
(9)
Fkm = a 2 e it (0 + 7 )nkm dS
Sm
(10)
For m = n , the added mass and damping

coefficients are due to bodys own motion, whereas
281
Figure 1. Two freely floating vertical cylinders.
The non-dimensional surge wave exciting forces

on Body1 and Body2 are shown in Figure 2. Figure 3
presents the surge motion responses of Body1 and
Body2. The results are plotted against ka , where k
and a denote the wave number and radius of each
cylinder respectively. Figure 2 and Figure 3 also
make a comparison between the present results with
the published results of Goo and Yoshida [7] and the
agreement was found quite satisfactory.
o
( Wave heading = 0 and gap = 20.0 m)

Body1 (lee side)
Ja-sam Goo et al.
Present calculation
Body2 (weather side)
Ja-sam Goo et al.
Present calculation
2.5
F1/ga a
2.0
1.5
1.0
0.5
0.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
ka
Figure 2.
Figure 4. In this paper, three different gap sizes

between the two boxes are chosen i.e., 25m, 50m
and 75m. On the other hand, the wave heading is
taken as 0o, 180o and 270o.
Figure 5 and Figure 6 show the non-dimensional
surge wave exciting forces against wave frequency
for the smaller box at 0o and 180o wave heading
respectively. The smaller body is located in the
weather side for 0o wave heading, whereas it is
positioned as lee side body for 180o wave heading. In
Figure 5, the peak magnitude of the surge wave
exciting force is amplified almost two times for a gap
of 75 m when compared to an isolated box result. The
peak magnitude shows a gradual increasing tendency
as the gap is reduced to 50m and then to 25m. This
amplification of peak magnitude may occur due to the
radiated waves from the near-by larger body. It can be
seen from the figure that the overall magnitude of
wave exciting forces is higher than that of an isolated
box result. For 180o wave heading, no amplification
of peak magnitude is observed and the overall
magnitude of the surge wave exciting forces on the
smaller box is reduced considerably due to the
shielding of the incident wave by the larger box.
Surge wave exciting forces on two

vertical floating cylinders.
o
(Wave heading = 0 and gap = 20.0 m)

1.5
Body1 (lee side)
Ja-Sam Goo et el.
Present calculation
Body2 (weather side)
Ja-Sam Goo et al.
Present calculation
|XA| /a
1.0
0.5
0.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
ka
Figure 3. Surge motions of vertical floating cylinders
3.2 Two Freely Floating Rectangular Boxes

Two unequal-sized rectangular floating boxes are
considered for the numerical analysis. The ratio of
the volume of displacement between the smaller and
the larger box is taken as 1:8. The length, breadth
and draught of the smaller box are 109.70 m, 101.40
m and 30.0 m respectively and its C. G. is 0.20 m
below M. W. L. Consequently, the length, breadth
and draught of the larger box will be two times
those of the smaller box. The dimension of the
smaller box is similar to that of the box model
considered by van Oortmerssen [3]. The water depth
is considered as 100.0 m. The wetted surface of
each box is divided into 300 panels as shown in
Figure 4. Mesh arrangement of wetted surface of two

unequal-sized floating boxes.
Figure 7 shows the heave wave exciting forces on
the smaller box for 0o wave heading. Since the
smaller box is located in the weather side, the
magnitude of heave exciting force is amplified
initially followed by a sharp fall and afterwards it
follows the isolated body result. Figure 8 and Figure
9 show the pitch wave exciting moments on the
smaller box for 0o and 180o wave heading
respectively. The overall tendency of pitch exciting
moment is found similar to the surge exciting force
results. For surge and pitch mode, when the gap is
wider the period of trapped wave is getting longer
meaning the gap effect will be located at a lower
frequency than a narrower gap. Figure 10 and
Figure 11 show the surge responses for the smaller
282
box for 0o and 180o wave heading respectively.

Initially, the motion response for both the wave
heading is similar to an isolated box response.
0.20
1 Body
2 Body (gap=25m)
2 Body (gap=50m)
2 Body (gap=75m)
F1/(gVa/L)
1 Body
2 Body (gap=25m)
2 Body (gap=50m)
2 Body (gap=75m)
0.25
F5/(gVa)
Wave heading = 0
Wave heading = 0
0.30
0.15
0.10
0.05
0.00
0.0
0.5
1.0
1.5
(rad/s)
1
Figure 8.Pitch exciting moment on smaller box
0
0.0
0.5
1.0
1.5
(rad/s)
Wave heading = 180
0.30
Figure 5. Surge wave exciting forces on smaller box

Wave heading = 180
0.20
F1/(gVa/L)
F5/(gVa)
1 Body
2 Body (gap=25m)
2 Body (gap=50m)
2 Body (gap=75m)
1 Body
2 Body (gap=25m)
2 Body (gap=50m)
2 Body (gap=75m)
0.25
0.15
0.10
0.05
0.00
0.0
0.5
1.0
1.5
(rad/s)
1
Figure 9. Pitch exciting moment on smaller box
0
0.0
0.5
1.0
1.5
(rad/s)
o
Wave heading = 0
Figure 6. Surge wave exciting forces on smaller box

3
Wave heading = 0
1Body
2Body (gap=25m)
2Body (gap=50m)
2Body (gap=75m)
1 Body
2 Body (gap=25m)
2 Body (gap=50m)
2 Body (gap=75m)
XA/a
F3/(gVa/L)
0
0
0.0
0.0
0.5
1.0
1.5
0.5
1.0
1.5
(rad/s)
(rad/s)
Figure 7. Heave wave exciting force on smaller box
Figure 10. Surge motion response of smaller box
It can be seen from both the figures that near 0.24

rad/s a sharp fluctuation occurs due to the formation
of standing waves and afterwards the motion
responses of the smaller box increases for lee side
position and decreases for weather side position.
Figure 12 and Figure 13 show the heave motion
responses for smaller box for 0o and 180o wave
heading respectively. The peak magnitude is
amplified significantly with sharp and rapid
fluctuations for weather side location of the smaller
box. However the motion response reduces
significantly for lee side position of the smaller box.
Due to coupling of surge, heave and pitch mode,

fluctuation of motion responses is observed near 0.24
rad/s for all of these modes. Finally, Figure 15 shows
the surge motion response of the smaller box for 270o
wave heading. Although, the motion response in the
surge mode is absent for an isolated body case, yet,
due to asymmetric hydrodynamic forces to wave
direction surge motion emerges for two box case.
283
Wave heading = 180

3
Wave heading = 180

3
1Body
2Body (gap=25m)
2Body (gap=50m)
2Body (gap=75m)
1Body
2Body (gap=25m)
2Body (gap=50m)
2Body (gap=75m)
XA/a
ZA/a
0.0
0.5
1.0
0.0
1.5
0.5
1.5
Figure 13. Heave motion response of smaller box
Figure 11. Surge motion response of smaller box

Wave heading = 0
1.0
(rad/s)
(rad/s)
Wave heading = 270
0.6
1Body
2Body (gap=25m)
2Body (gap=50m)
2Body (gap=75m)
1Body
2Body (gap=25m)
2Body (gap=50m)
2Body (gap=75m)
0.4
ZA/a
XA/a
0.2
0.0
0.0
0.0
0.5
1.0
0.5
1.0
1.5
(rad/s)
1.5
(rad/s)
Figure 14.Surge motion response of smaller box

Figure 12. Heave motion response of smaller box
4. CONCLUSION
Using 3-D source distribution method, the
hydrodynamic interaction for two unequal-sized
freely floating rectangular boxes in regular waves is
studied. Hydrodynamic interaction causes rapid
changes in hydrodynamic loads and responses along
the wave frequencies. The amplitude of motion
responses and wave exciting forces for the smaller
box can be increased or, reduced depending upon the
wave heading. As the gap between the two bodies is
shortened, the peak frequencies due to interaction
move to higher frequencies. Therefore, hydrodynamic
interaction should be taken into consideration to
evaluate and design the safety and performance of a
multi-body floating system.
REFERENCES
[1] Ohkusu, M., On the heaving motion of two
circular cylinders on the surface of a fluid,
Reports of Research Institute for Applied
Mechanics, XVII, No. 58, pp. 167-185(1969).
[2] Faltinsen, O. and Michelsen, F., Motions of
large structures in waves at zero Froude
number, Proc. Intl. Symp. on the Dynamics of
Marine Vehicles and Structures in Waves,
London, pp. 91-106 (1974).

[3] van Oortmerssen, G., Hydrodynamic interaction
between two structures, floating in waves,
Proc. 2nd. Intl. Conf. on Behaviour of Offshore
Structures, London, pp. 339-356 (1979).
[4] Matsui, T. and Tamaki, T., Hydrodynamic
interaction between groups of vertical
axisymmetric bodies floating in waves, Int.
Symp. on Hydrodynamics in Ocean Engineering,
Trondhem, pp. 817-836 (1981).
[5] Choi, Y. R. and Hong, S. Y., An analysis of
hydrodynamic interaction of floating multi-body
using higher-order boundary element method,
Proc. 12th Int. Offshore and Polar Engineering
Conference, Kitakyushu, pp. 303-308 (2002).
[6] Wang, Q., Sun, L. P. and Ma, S., Time-domain
analysis of FPSO-tanker responses in tandem
offloading operation, Journal of Marine
Science and Application (Pub. Harbin
Engineering University), Vol. 9, No. 2, pp. 200207 (2010).
[7] Goo, J. and Yoshida, K., Hydrodynamic
interaction between multiple three dimensional
bodies of arbitrary shape in waves, Journal of
the Society of Naval Architects Japan, Vol. 165,
pp. 193-202 (1989).
284

AN INVESTIGATION OF SHIP PERFORMANCE IN SEAS

M.S. Baree
Department of Naval Architecture and Marine Engineering,
Bangladesh University of Engineering and Technology,
E-mail: baree@name.buet.ac.bd
ABSTRACT
Numerical calculations of total added resistance both in regular and irregular long crested waves for
Series 60 ships, parent form, have been carried out. The total added resistance in regular waves has been
determined by adding separately Maruos method for ship motion to that of Fujii-Takahashis method for wave
reflection. Computed results have been compared with experimental ones. Average added resistance in a
seaway is obtained from the analytically obtained mean response curve for added resistance in regular waves
and the ISSC spectrum applying linear superposition technique. In order to determine the performance, two
situations have been considered. They are: power increase at constant speed and speed loss at constant power.
The performance of series 60 ships, parent form, has then been presented in graphical forms for prediction in a
seaway. Effect of ship speed, wave direction and block coefficient on performance has been examined and the
percentage added power or speed loss with respect to calm water characteristics at different sea states has been
presented.
KEY WORDS: Added resistance, Added power, Ship performance
1. INTRODUCTION
The knowledge of hydrodynamic behavior of a
ship is important because the success of a ship design
eventually depends on its performance in a seaway.
The horsepower of a ship predicted by a tank test is
usually the value in a calm sea and a considerable
power should be added to this when a ship is steaming
in a seaway. The prediction of ship motions,
resistance and power in a realistic seaway is such a
complex problem that the added power required in a
seaway is traditionally accounted for by increasing a
certain percentage of its calm water characteristics.
While this method has been adequate, it would
certainly be advantageous to the designer if he could
predict accurately the increase in power of a ship in a
seaway. Though the increase of resistance of a ship
due to the waves has been recognized for a long time,
there are different opinions as to the cause of excess
resistance. Presently there is large number of methods
available for the calculation of resistance increase in
waves. The object of the paper is not to make a
comparative study of the different methods for
prediction of added resistance rather the object is to
investigate the performance of ship in seas due to
waves only by applying certain theories and methods
without considering the effects of wind and current. In
order to do this, numerical calculations of total added
resistance both in regular and irregular long crested
waves for Series 60 ships, parent form, have been

carried out. Of the several alternative approaches,
total added resistance in regular waves has been
determined by adding separately Maruos method [1]
for resistance increase due to ship motion to that of
Fujii-Takahashis method [2] for wave reflection. A
number of researchers have proposed similar methods
for predicting total added resistance, reference [8].
Average added resistance in a seaway is obtained
from the analytically obtained mean response curve
for added resistance in regular waves and the ISSC
spectrum applying linear superposition technique. In
order to determine the performance, two situations
have been considered. They are: power increase at
constant speed and speed loss at constant power. The
performance of series 60 ships, parent form, has then
been presented in graphical forms for prediction in a
seaway. Influence of ship speed, wave direction and
block coefficient on performance has been examined
and the percentage added power or speed loss with
respect to calm water characteristics at different sea
states has been determined.
2. PREDICTION METHOD
As pointed out earlier, total added resistance in
regular waves has been determined by adding
separately Maruos method [1] for resistance increase
285
due to ship motion to that of Fujii-Takahashis method

[2] for wave reflection. In regular waves, the added
resistance RAW is computed for different waveencounter frequencies to obtain an accurate
representation of the response amplitude operator for
2
a is
the wave
amplitude. Average added resistance in irregular long

crested waves is then predicted from the mean curve
of response amplitude operator for added resistance in
regular waves and energy spectrum of the sea
recommended by ISSC by following linear
superposition technique.
g a2 B 2 / L
9
8
7
6
5
4
3
2
1
0
0.5
1.0
1.5
2.0
2.5
/L
Fig.1: Comparison of added resistance between

numerical and experimental results for CB = 0.70 at a
wave heading of 170o
Cal.:Maruo+Fujii-Takahashi
Exp.:Vossers et al. [1960]
Fn=0.10, cal.
Fn=0.15, cal.
Fn=0.20, cal.
Fn=0.25, cal.
Fn=0.10, exp.
Fn=0.15, exp.
Fn=0.20, exp.
Fn=0.25, exp.
= 130
Added Resistance, RAW/(ga B /L)
10
8
7
6
5
4
3
2
1
0
0.0
0.5
1.0
1.5
2.0
2.5
/L

wave heading of 130o
From Figs. 1-4, in head and bow waves, at lower
speeds, the agreement is found to be not satisfactory,
while relatively better agreement is found at higher
speeds. While in beam and following seas,
comparatively better agreement is obtained. From the
overall comparisons, prediction seems to give better
agreement especially in oblique directions. However
it may be noted that there are discrepancies between
various laboratory measurements and various theories
as pointed out by different investigators, for example;
reference [7].
It should be noted again that the wavelength to
ship length ratios covered in the experimental results
range from 0.6 to 1.8 and as such at shorter
wavelength ratios below 0.6, it was not possible to
show the comparison, which is important especially
for fuller ships.
6
Fn=0.10, cal.
Fn=0.15, cal.
Fn=0.20, cal.
Fn=0.25, cal.
Fn=0.10, exp.
Fn=0.15, exp.
Fn=0.20, exp.
Fn=0.25, exp.
= 90
5
Figs. 1-4 show the comparison of non-dimensional

added resistance with the thrust increase coefficient of
experimental results for Series 60 hull, parent form,
with block coefficient CB = 0.70 at four different
Froude numbers with each figure representing for
individual wave heading. It should be noted that the
experimental results of thrust increase coefficient is
where
Fn=0.10, cal.
Fn=0.15, cal.
F =0.20, cal.
F =0.25, cal.
F =0.10, exp.
F =0.15, exp.
F =0.20, exp.
F =0.25, exp.
-1
Numerical calculations have been carried out for

three Series 60 ships, parent form, reference [4] in
order to predict the added resistance both in regular
and irregular waves. The computed results in regular
waves are compared with the thrust increase
coefficient of experimental results for Series 60 ships.
The comparisons have been shown in Figs. 1 to 4. It
should be noted that the experimental results of added
resistance in oblique waves are very rare. Fortunately
extensive model tests of Series 60 forms in oblique
long crested waves were conducted in a unique
seakeeping facility at MARIN as reported by Vossers
et al. in reference [6]. Fourteen models were tested in
five different wavelengths and five different oblique
directions at 170o, 130o, 90o, 50o and 10o. Four speeds
were used in the range of Fn = 0.10 to 0.25. Among
other things the extensive measurements made
included thrust increase coefficient. The results were
presented graphically. Among these, the results of the
model with block coefficient CB = 0.70 and L/H=
17.50 conform to that of the parent form.
= 170
10
0.0
3. COMPUTATIONS AND COMPARISONS
11
-1
Hosoda [3] pointed out that the contribution of the

lateral motions to the added resistance in oblique
waves is almost negligible and according to this
conclusion, the added resistance in oblique waves is
calculated on the basis of prediction for head sea
waves, for example, Fujii and Takahashi [2]. It should
be noted that the motion responses have been obtained
by applying Ordinary Strip Method (OSM).
expressed by
12
Added Resistance coeff., RAW/(ga B /L)
added resistance, RAW/ a , where
13
T is the
4
3
2
1
0
-1
0.0
0.5
1.0
1.5
2.0
2.5
/L
thrust increase. It may be noted that Hosoda [3] has

given a similar comparison between theoretically
obtained resistance increase coefficients with that of
experimentally obtained thrust increase coefficients.

wave heading of 90o
286
= 10
-1
0.0
0.5
1.0
1.5
2.0
2.5
/L

wave heading of 10o
local speed exponent of the resistance curve in the

region of interest. In the determination of speed loss at
constant power, it is seen that the essential controlling
feature is the slope of the power-speed curve in the
range of speed of interest. Some typical values for
n may be found in reference [9] as follows:
n =2.886
Fn=0.25
Ship (CB =0.60)
Fn=0.15
n =2.153
Ship (CB =0.80)
4. PERFORMANCE RESPONSE TO
ADDED RESISTANCE
The response of a ship due to added resistance
from waves will depend upon engine control system
and any intervention by the ships operator. However
two situations have been considered for prediction of
performance of the vessels under study; one is the
power increase at constant speed and the other is the
speed loss at constant power.
4.1 POWER INCREASE AT CONSTANT

SPEED
In Fig. 5, variation of mean added resistance in

irregular waves are shown against wave headings at
different significant wave heights for Series 60 ship
with CB=0.70 at Fn=0.20. It is seen that the added
resistance increases with increase of significant wave
heights at all wave headings. Head sea waves are seen
to give maximum added resistance, while from
quartering to following seas, it increases again with
the increase of significant wave heights. At a wave
heading of approximately 60 to 70 degrees, the added
resistances are found to be minimum and independent
of significant wave height.
2/3
Added Resistance Coeff.x10 , RAWI/(VOL) V
From reference [10], power increase at a given

speed when the ship experiences an added resistance
can be expressed by
R
R0
P
=
1
0
P0
1+
( 0 ) 0
1+
P is the increase in power P0 to maintain

speed when the resistance is increased by R .
0 is the change in propeller efficiency
where
0 consequent upon the change in propeller loading.

Reference [9] provides a diagram of the relationship
between
0
R
and
of series 60 ship propeller.
0
R0
This diagram is reproduced as Figure 12. Whilst it is

easily possible to calculate the relationship for a
particular propeller and ship, if the data is available, it
is considered adequate to estimate appropriate values
from figure 12.
4.2 SPEED LOSS AT CONSTANT POWER

Reference [11] gives:
V
1 P
=
V0
n + 1 P0
power is
(n + 1) 1 times the power increase at constant speed.

Here V0 is the ship speed at calm water and n is the
Fn=0.10, cal.
Fn=0.15, cal.
Fn=0.20, cal.
Fn=0.25, cal.
Fn=0.10, exp.
Fn=0.15, exp.
Fn=0.20, exp.
Fn=0.25, exp.
or the speed loss fraction at constant

5
18
H1/3= 7.0 m
16
6.0 m
14
12
5.0 m
10
Fn = 0.20
CB=0.70
4.0 m
3.0 m
4
2
2.0 m
0
-2
-20
20
40
60
80
100
120
140
160
180
200
Wave Heading, (degree)
Fig.5.Variation of Added Resistance against wave

headings for Series 60 ship with C B =0.70 and
LBP=121.95 m at different significant wave heights
In Figures from 6 to 11, the percentage added
resistance with respect to calm water resistance of
three Series 60 ships with CB=0.60, 0.70 and 0.80
have been shown at different Froude numbers. The
three block coefficients are chosen so that it
represents fine, medium and full ship forms. Fig. 6
and 7 represent the percentage added resistance for
CB=0.60 at Fn= 0.20 and 0.25 respectively. Fig. 8 and
9 represent the same for CB=0.70 at Fn= 0.20 and 0.25
respectively, while Fig. 10 and 11 are for CB=0.80 at
Fn=0.15 and 0.20 respectively. The Froude numbers
are chosen so that it represents the region around
service speeds for this type of ships. As can be seen
from the figures, the results are presented against nondimensional sea state defined by H1/3/LBP. The use of
non-dimensional sea states for the calculation has the
287
200
=0
=30
=60
R/R0(%)
150
140
R/R0(%)
120
100
80
Fn=0.20
CB=0.70
=30
=60
=90
=120
=150
=180
60
40
20
0
-20
0.00
0.01
0.02
0.03
0.04
0.05
0.06
H1/3/LBP
Fig. 8. Variation of percentage added resistance for

Series 60 ship with CB=0.70 at Fn=0.20
120
=0
100
80
Fn0.25
CB=0.70
=30
=60
=90
o
o
=120
=150
=180
60
40
20
0
0.00
0.01
0.02
0.03
0.04
0.05
0.06
H1/3/LBP
=90
100
=120
240
=150
220
=0
=180
200
=30
180
=60
=90
160
=120
140
=150
=180
R/R0(%)
50
0.00
=0
160
Fig.9. Variation of percentage added resistance for

Fn=0.20
CB=0.60
180
R/R0(%)
advantage that it ties the importance of sea state to the

length of the ship. In this way, a description of the
ships behavior in all seaways of interest is possible. It
may be noted that calm water resistance is calculated
from reference [5] for naked hull only. Considering a
non-dimensional sea state of 0.025, a Series 60 ship
with CB=0.60 is seen to experience an added
resistance of approximately 1 to more than 50% at
Fn=0.20 depending upon the wave heading, while at
Fn=0.25, the range is from 0 to more than 30%. At the
same non-dimensional sea state, a Series 60 ship with
CB=0.70, is seen to experience an added resistance of
1 to more than 40% at Fn=0.20 and from 1 to more
than 20% at Fn=0.25. Series 60 ship with CB=0.80, is
seen to experience an added resistance from 0 to
approximately 55% at Fn=0.15 and from 0 to 25% at
Fn=0.20 at the same non-dimensional sea-state
depending upon different wave headings. It may be
noted that at lower speeds, the percentage increase of
added resistance is more than that at higher speeds.
The lower values of percentage increase of added
resistance is due the fact that calm water resistance is
a function of square of ship speed, whereas the added
resistance is not that speed sensitive [8].
0.01
0.02
0.03
0.04
0.05
0.06
120
100
Fn=0.15
CB=0.80
80
60
40
H1/3/LBP
20
0

Figures 13 to 16 have been prepared for ships with
block coefficient 0.60 and 0.80, for prediction of
power increase at constant speed and speed loss at
constant power. The figures therefore represent the
performance of fine and full ships in terms of
penalties for power or speed. The Froude numbers are
chosen so that it represents the service speed for these
types of ships.
140
=0
120
=60
=90
80
=120
=150
60
=180
0.01
0.02
0.03
0.04
0.05
0.06
H1/3/LBP

Fig. 13 represents the percentage increase in power
due to waves for the ship with CB=0.60 at Fn=0.25
while Fig.14 represents percentage speed loss for the
same ship. Fig. 15 & 16 repeat Fig. 13 & 14 for ship
with CB=0.80 at Fn=0.15. As can be seen from the
figures, the results are presented against nondimensional sea state defined by H1/3/LBP for different
headings.
120
R/R0(%)
R/R0(%)
100
Fn=0.25
CB=0.60
=30
-20
0.00
40
20
Fn=0.20
CB=0.80
110
=0
100
=30
90
=60
80
=90
70
=120
60
50
o
o
=150
=180
40
30
0
-20
0.00
20
10
0.01
0.02
0.03
0.04
0.05
0.06
-10
0.00
H1/3/LBP
0.01
0.02
0.03
0.04
0.05
0.06
H1/3/LBP

Fig.11. Variation percentage added resistance for

288
1.35
340
Ship with Cb=0.60, Fn=0.25

Ship with Cb=0.80, Fn=0.15
1.30
320
300
1.25
=150
0
0
=120
280
=90
260
=60
240
=30
=180
0
o
=150
=0
220
1.20
200
Cb=0.80
P/P0 (%)
1/(1+0/0)
=180
1.15
Cb=0.60
1.10
180
160
140
=120
120
100
1.05
80
o
=0
60
40
1.00
=30
o
=90
20
0
0.95
0.010
=60
0.015
0.020
0.025
0.90
0.030
0.035
0.040
0.045
0.050
0.055
H1/3/LBP
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
1+R/R0
Fig.12. Change in open water efficiency due to

increase in resistance of Ships with block coefficient
0.60 and 0.80
Fig.15. Percentage Increase in Power for Series 60

ship with CB=0.80 at Fn=0.15
110
100
180
=180
90
=150
=120
80
=90
=60
=30
=180
=150
160
=120
=180
=90
=60
140
=30
=150
=0
70
V/V0 (%)
P/P0 (%)
120
100
80
=0
= 180
= 150
60
50
o
= 120
40
o
=120
60
30
40
20
=0
20
=30
10
o
o
=90
o
=60
0
0.020
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0.050
0.055
Fig.13. Percentage Increase in Power for Series 60

ship with CB=0.60 at Fn=0.25
=180
=150
=120
= 180
=90
=60
= 150
=30
=0
V/V0 (%)
30
20
= 120
10
0.025
0.030
0.035
0.040
0.045
0.035
0.040
0.045
0.050
Fig.16. Performance of Speed for Series 60 ship with

CB=0.80 at Fn=0.15
Now assuming the ships are encountering a nondimensional significant wave of 0.025 on their route,
the penalties in power and that of speed of the ships
appear as follows:
0
0.020
0.030
The nature of the curves of Fig.13 to 16, for

obvious reason, is similar to those of Fig. 6 to 11. As
expected, it is seen that the percentage increase in
power increases with increase of non-dimensional
significant wave heights at almost all wave headings.
Head and bow seas are seen to give maximum added
power as usual, while quartering and following seas
are seen to give minimum added power. At a wave
heading of 60 degree, the added power is found to be
minimum and almost independent of non-dimensional
significant wave heights, the reason for which is
evident from Fig.5.
50
40
0.025
H1/3/LBP
H1/3/LBP
0.050
H1/3/LBP
Fig.14. Performance of Speed for Series 60 ship with

CB=0.60 at Fn=0.25
At head and bow seas for the ship with CB=0.8,

added power requirement is approximately 60~80%,
at beam seas approximately 14%, while in quartering
289
and following seas approximately 0~18% at a

constant speed represented by Fn=0.15. On the other
hand, at the same wave strength, the same ship at
constant power is expected to experience a speed loss
of approximately 20~25% at head and bow seas, 4%
at beam seas and 0~5% at quartering and following
seas.
In a similar situation, for the ship with CB=0.6,
added power requirement is approximately 30~45% at
head and bow seas, approximately 5% at beam seas,
while in quartering and following seas approximately
0~10% at a constant speed represented by Fn=0.25.
On the other hand, at the same wave strength, the
same ship at constant power is expected to experience
a speed loss of 8~11% at head and bow seas, 1.5% at
beam seas and 0~3% at quartering and following seas.
[6]
Vossers, G., Swaan, W.A. and Rijken, H.,

Experiments with Series 60 Models in
Waves,Trans.SNAME,Vol.68,(1960).
[7]
Nakamura, S., Added Resistance and

Propulsive Performance of Ship in Waves,
ProceedingsontheInternationalSeminaron
Wave Resistance, Tokyo, Japan, pp.199216,
(1976).
[8]
Baree, M.S., Inoue, Y., Islam, M. R., An

Investigation of Added Resistance of Ships in
Oblique Seas, Transactions of the Honkong
Institution of Engineers, Vol.13, Number 2,
June (2006).
Baree, M. S., Performance Prediction of Two
Series 60 Ships due to weather and Hull
Roughness, The Guardian, A national
monthly, August (2001).
Berlekom, W.B. Van, Wind Forces on Modern
Ship Forms - Effects on Performance, Trans.
NECIES 97, (1981).
Baree, M. S., An Approach for Speed
Reduction or Power Increase due to Added
Resistance of a Ship, The Journal of NOAMI,
Vol.3, No.2, December, (1986).
[9]
[10]
6. CONCLUSIONS
From the computational results and analysis for
Series 60 ships, parent form, it may be concluded that
for this type of ships, the sea has significant effect on
the performance at all speed range. However the
effect is more dominant at lower speeds than that at
higher ones. Head and bow seas have more dominant
effect on performance than other headings. Moreover
fuller ships are expected to experience less penalties
for power or speed than finer ships. .For a moderate
sea represented by non-dimensional significant wave
height of 0.025, the finer ships are expected to
experience a power increase of 1~45% and a speed
loss of 0~11%, while the fuller ships are expected to
experience a power increase of 0~80% and a speed
loss of 0~25% at their service speeds depending upon
different wave headings with the minimum penalty
being at a wave heading of 60 to 70 degrees.
[11]
REFERENCES
[1]
[2]
[3]
[4]
[5]
Maruo,H., Resistance Increase in waves, 60th

Anniversary Series, Vol.8, The Society of
Naval Architects of Japan. (1963).
Fujii, H, Takahashi, T., Experimental Study on
resistance Increase of a Ship in regular Oblique
waves, 14th ITTC Proceedings, pp 351-360,
Seakeeping Committee (1975).
Hosoda, R., The Added resistance of Ships in
Regular Oblique Waves, Vol.133, SNAJ,
Japan, (1973).
Todd, F.H., Some Further Experiments on
Single Screw Merchant Ship Forms-Series 60,
Trans. SNAME, (1953).
Todd, F.H., et al.,Series 60 The Effect upon
Resistance and Power of Variation of Ship
Proportions,Trans.SNAME,(1957).
290

MHD-CONJUGATE FREE CONVECTION FLOW FROM AN

ISOTHERMAL HORIZONTAL CYLINDER WITH STRESS WORK
NHM. A. Azim1, ATM. M. Rahman2 and M. K. Chowdhury3
1
School of Business Studies, Southeast
University, Dhaka, Bangladesh,
Email: nhmarif1@gmail.com
2
Department of Computer Science and
Engineering,, Dhaka International
University, Dhaka, Bangladesh
Email: atmmrahman2005@gmail.com
3
Department of Mathematics,
Bangladesh University of Engineering
and Technology, Dhaka, Bangladesh
Email: mkamal@math.buet.ac.bd
ABSTRACT
The effects of conduction and stress work on Magnetohydrodynamic(MHD) free convection flow from an
isothermal horizontal cylinder along the outer surface from the lower stagnation point to the upper stagnation
point are investigated. The developed governing equations and the associated boundary conditions are made
dimensionless using a suitable transformation then the non-dimensional governing equations are solved using
the implicit finite difference method with Keller box-scheme. Numerical outcomes are found for different
values of the magnetic parameter, stress work parameter and conjugate conduction parameter. Results for the
details of the velocity profiles and the temperature distributions as well as the skin friction coefficients and the
rate of heat transfer are shown graphically and discussed.
Keywords: Free convection, Conduction, MHD, Stress work, Horizontal cylinder, Finite difference method.
1. INTRODUCTION
Many researchers investigated natural convection
flow from a horizontal cylinder [1-4] under diverse
surface boundary conditions (isothermal, uniform heat
flux and mixed boundary conditions) using different
mathematical technique. The conjugate heat transfer
process formed by the interaction between the
conduction inside the solid and the convection flow
along the solid surface has a significant importance in
many practical applications. Gdalevich and Fertman
[5] studied the conjugate problems of natural
convection. Miyamoto et al. [6] analysed the effects
of axial heat conduction in a vertical flat plate on free
convection heat transfer. Pozzi et al. [7] investigated
the entire thermo-fluid-dynamic (TFD) field resulting
from the coupling of natural convection along and
conduction inside a heated flat plate by means of two
expansions, regular series and asymptotic expansions.
Moreover, Kimura and Pop [8] analysed conjugate
natural convection from a horizontal circular cylinder.
MHD flow and heat transfer process are now an
important research area due to its potential application
in engineering and industrial fields. A considerable
amount of research has been done in this field. Wilks
et al. [9] studied MHD free convection about a semiinfinite vertical plate in a strong cross field. Aldoss et
al. [10] analysed MHD mixed convection from a
horizontal circular cylinder. El-Amin [11] found out
the combined effect of viscous dissipation and Joule
heating on MHD forced convection over a non-
isothermal horizontal circular cylinder embedded in a

fluid saturated porous medium. The influence and
importance of viscous dissipation and stress work
effects in laminar flows have been examined by
Gebhart [12]. Later the numerical solution of the
effect of viscous dissipation and pressure stress work
in natural convection along a vertical isothermal plate
studied by Pantokratoras [13] without any
approximation.
The objective of the present paper is to obtain the
numerical result of MHD-conjugate free convection
flow from an isothermal horizontal circular cylinder
considering stress work effect with a complete
discussion.
2. MATHEMATICAL ANALYSIS
Let us consider a steady natural convection
flow of a viscous incompressible and electrically
conducting fluid from an isothermal horizontal
circular cylinder of radius a placed in a fluid of
uniform temperature T. The cylinder has a heated
inner region of temperature Tb and the thickness of
the circular cylinder is b with Tb >T. A uniform
magnetic field having strength B0 is acting normal
to the cylinder surface. The x -axis is taken along
the circumference of the cylinder measured from the
lower stagnation point and the y -axis is taken
normal to the surface. It is assumed the fluid
properties to be constant and the induced magnetic
field is ignored. The effects of stress work in the
291
flow region and conduction from inner surface to

the outer surface considered in the present study.
1 2
u + u
+v
=
+
T
T
x
y Pr y 2
Where
2
M = ( a 2 B0 ) /( Gr1/ 2 ) is
parameter,
= ga / c p
is
the
(7)
the magnetic
stress
work
parameter and Pr = c p / is the Prandtl number.

The boundary conditions (4) can be written as in the
following dimensionless forms:
u = v = 0, 1 = p

on y = 0, x > 0
y
(8)
u 0, 0 as y , x > 0
Where
Fig. 1: Physical Model and coordinate system
Under the balance laws of mass, momentum and

energy and with the help of Boussinesq
approximation for the body force term in the
momentum equation, the equations governing this
boundary-layer natural convection flow can be written
as:
u
x
u
v
y
(1)
=0
x B 2 u
2 u
u
u
= 2 + g (T f T )sin 0
+v
y
x
y
a
2
T f T u p
=
+
+v
u
2
cp x
y cp y
x
T f
T f
(2)
f + ff f 2 Mf +
u = v = 0, T f = T ( x, 0)
(4)
at y = 0, x > 0
y
f 0, 0 as y , x > 0
Principle physical quantities, the shearing stress and

the rate of heat transfer in terms of skin friction
coefficient C f and Nusselt number Nu respectively
C f Gr1 / 4 = x f ( x,0)
(14)
= ( x,0)
(15)
Nu Gr
Where is the dimensionless temperature. The non

dimensional forms of the equations (1)-(3) are as follows:
u
u
2u
+ v + Mu = 2 + sin x
x
y
y
(13)
can be written as:
x
y
ua 2
, y = Gr 4 , u =
Gr
a
a
1
T f T
va
v = Gr 4 , =
Tb T
conjugate
In the above equations primes denote differentiation

with respect to y. The corresponding boundary
conditions (8) take the following form
x=
u v
+
=0
x y
the
f
sin x
f
= x f
f (11)
x
x
x
f = f = 0, 1 = p
The governing equations and the boundary conditions

(1)-(4) can be made non-dimensional, using the
Grashof number Gr = (g a 3 (Tb T ) )/ 2 which is
assumed large and the following non-dimensional
variables:
is
f (12)
1
T

+ f x f xf = x f
x
Pr
T
T
x
u 0, T f T as y , x > 0
conduction parameter. The present problem is

governed by the magnitude of p. The values of p
depends on b/a, kf /ks and Gr. The ratios b/a and kf /ks
are less than one where as Gr is large for free
convection. Therefore the value of p is zero (b=0) or
greater than zero.
To solve equation (5)-(7), subject to the boundary
condition (8), we assume following transformation:
(9)
= x f ( x, y )
Where is the stream function usually defined as
(10)
u = / y , v = / x
Substituting (10) into the equations (5)-(7), the new
forms of the dimensionless equations (6) and (7) are
(3)
The physical situation of the system suggests the

following boundary conditions
on y = 0, x > 0
T f / y = s (T f Tb ) / b f
p = b f Gr 1/ 4 / (a s )
1 / 4
The results of the velocity profiles and temperature

distributions can be calculated by the following
relations respectively.
(5)
u = f ( x, y ) , = ( x, y )
(6)
292
(16)
decreases for increasing stress work parameter. The

decreased temperature for increasing stress work
parameter within the boundary layer reduced the
temperature difference between the boundary layer
region and the core region eventually increases heat
transfer rate as illustrated in Fig.9.
3. METHOD OF SOLUTION
Equation (11) and (12) are solved numerically
based on the boundary conditions as described in
equation (13) using one of the most efficient and
accurate methods known as implicit finite difference
method with Keller box scheme [15, 16].

The conjugate heat transfer analysis from an
isothermal horizontal circular cylinder considering
stress work effect is the main purpose of the present
work. The Prandtl number is considered 1.0 for the
simulation which corresponds to steam.
A comparison of the local Nusselt number and the
local skin friction factor obtained in the present work
with M = 0.0, = 0.0, p = 0.0 and Pr = 1.0 and
obtained by Merkin [1] and Nazar et al. [14] have
been shown in Tables 1 and 2 respectively. There is
an excellent agreement among these three results.
Figs.2 and 3 illustrate the velocity and temperature
distribution against y for different values of the
magnetic parameter and the skin friction coefficient
and the heat transfer rate against x for varying
magnetic parameter with p =1.0 and =0.01 are
depicted in figs.4 and 5 respectively. The magnetic
field opposes the fluid flow as a result the peak
velocity decreases with the increasing M as shown in
fig. 2. Consequently, the separation of the boundary
layer occurs earlier and the momentum boundary
layer becomes thicker. From Fig. 3 it can be observed
that the increasing value magnetic parameter
increases temperature within the boundary layer, this
is expected as there is an interaction with magnetic
field with fluid flow. Thus, the magnetic parameter
increases the thickness of the thermal boundary layer.
Temperature at the interface also varies with different
M since the conduction is considered within cylinder.
The Magnetic force opposes the flow, as
mentioned earlier, and reduces the shear stress at the
wall as illustrated in Fig. 4. Moreover, the heat
transfer rate also decreases as temperature difference
between solid surface and boundary layer region is
reduced as revealed in Fig. 5.
The velocity profiles, temperature distributions,
local skin friction coefficients and the heat transfer
rate for different values of stress work parameter are
presented in Fig. 6, Fig. 7, Fig. 8 and Fig. 9,
respectively with p =1.0 and M =0.1. Increasing value
of the stress work parameter containing gravitational
force g work against the buoyancy force as a result
the motion of the fluid motion is decreased as plotted
in Fig.6. The reduced velocity decelerates fluid flow
which ultimately decreases the shear stress at the wall
which is observed from fig.8.
The velocity profiles and temperature distributions

for different values of conjugate conduction
parameter p are presented in fig.10 and fig.11,
respectively with M = 0.1 and = 0.01. It is observed
that both the velocity profile and temperature
distribution decrease for increasing p. It is expected
because increase value of conjugate conduction resists
heat conduction from the solid to the boundary layer.
Fig.12 and fig.13 depict the skin friction
coefficient and the heat transfer rate for different
values of conjugate conduction parameter p,
respectively. It can be noted from these two figures
that, both the skin friction coefficient and heat
transfer rate decrease as the values of conduction
parameter increase.
5. TABLES AND FIGURES

Table 1: Numerical values of (x,0) for different
values of x while Pr =1.0, M = 0.0, = 0.0 and p =
0.0.
Nu Gr 1/ 4 = ( x,0)
x
0.0
/6
/3
/2
2/3
5/6
Merkin
[1]
0.4214
0.4161
0.4007
0.3745
0.3364
0.2825
0.1945
Nazar et
al. [14]
0.4214
0.4161
0.4005
0.3741
0.3355
0.2811
0.1916
Present
0.4216
0.4163
0.4006
0.3741
0.3355
0.2811
0.1912
Table 2: Numerical values of x f (x,0) for different

values of x while Pr = 1.0, M = 0.0, = 0.0 and p =
0.0.
On the other hand from fig.7 it could be concluded

that the temperature within the boundary layer
293
C f Gr 1/ 4 = x f ( x,0)
x
Merkin [1]
0.0
0.0000
0.4151
0.7558
0.9579
0.9756
0.7822
0.3391
/6
/3
/2
2/3
5/6
Nazar et al.
[14]
0.0000
0.4148
0.7542
0.9545
0.9698
0.7740
0.3265
Present
0.0000
0.4139
0.7528
0.9526
0.9678
0.7718
0.3239
0.30
0.4
0.20
Heat transfer, Nu x
Velocity,
f'
M = 0.10
M = 0.30
M = 0.50
M = 0.70
0.10
M=0.10
M=0.30
M=0.50
M=0.70
0.3
0.2
0.1
0.00
0.0
1.5
3.0
4.5
6.0
0.0
0.0
7.5
Fig.2. Variation of velocity profiles against y for varying of

M with Pr = 1.0, = 0.01 and p =1.0.
0.5
1.0
1.5
2.0
2.5
3.0
Fig.5. Variation of rate of heat transfer against x for varying

of M with Pr = 1.0, = 0.01 and p =1.0.
0.30
0.80
= 0.01
= 0.05
= 0.10
= 0.20
'
Temperature,
0.60
M=0.10
M=0.30
M=0.50
M=0.70
Velocity, f
0.40
0.20
0.10
0.20
0.00
0.0
1.5
3.0
4.5
6.0
0.00
0.0
7.5
Fig.3. Variation of temperature distributions against y for

varying of M with Pr = 1.0, = 0.01 and p =1.0.
4.5
0.60
Temperature,
0.60
0.45
0.30
0.15
0.5
1.0
1.5
2.0
2.5
7.5
=0.01
=0.05
=0.10
=0.20
0.40
M=0.10
M=0.30
M=0.50
M=0.70
6.0
Fig.6. Variation of velocity profiles against y for varying of
0.80
Skin friction , Cfx
3.0
with Pr = 1.0, M = 0.1 and p = 1.0.
0.75
0.00
0.0
1.5
0.20
0.00
0.0
3.0
Fig.4. Variation of skin friction coefficients against x for

varying of M with Pr = 1.0, = 0.01 and p=1.0.
1.5
3.0
4.5
6.0
7.5

varying of with Pr = 1.0, M = 0.1 and p = 1.0.
294
0.75
0.8
Skin friction , Cfx
0.60
Temperature,
0.6
0.45
0.15
0.00
0.0
0.4
=0.01
=0.05
=0.10
=0.20
0.30
0.5
1.0
1.5
2.0
0.2
2.5
0.0
0.0
3.0

varying of with Pr = 1.0, M = 0.1 and p =1.0.
3.0
4.5
6.0
7.5
0.75
0.60
Skin friction , Cfx
Heat transfer, Nu x
1.5

varying of p with Pr = 1.0, M = 0.1 and = 0.01.
0.4
0.3
0.45
0.2
=0.01
=0.05
=0.10
=0.20
0.1
0.0
0.0
0.5
1.0
1.5
2.0
p=1.00
p=1.25
p=1.50
p=1.75
0.30
0.15
2.5
0.00
0.0
3.0
Fig.9. Variation of rate of heat transfer against x for varying

of with Pr = 1.0, M = 0.1 and p =1.0.
0.5
1.0
1.5
2.0
2.5
3.0

varying of p with Pr = 1.0, M = 0.1 and = 0.01.
0.30
p = 1.00
p = 1.25
p = 1.50
p = 1.75
0.20
0.10
Heat transfer, Nu x
0.4
f'
Velocity,
p=1.00
p=1.25
p=1.50
p=1.75
p=1.00
p=1.25
p=1.50
p=1.75
0.3
0.2
0.1
0.00
0.0
1.5
3.0
4.5
6.0
0.0
0.0
7.5
Fig.10. Variation of velocity profiles against y for varying

of p with Pr = 1.0, M = 0.1 and = 0.01.
0.5
1.0
1.5
2.0
2.5
3.0
Fig.13. Variation of rate of heat transfer against x for

varying of Pr with = 0.01, M = 0.5 and p = 1.0.
295
6. CONCLUSION
MHD-conjugate free convection flow from horizontal
circular cylinder considering stress work effect is studied.
The effects of the magnetic parameter, Stress work
parameter and Conjugate conduction parameter are
analysed on the fluid flow with Prandtl number Pr = 1.0.
The velocity of the fluid within the boundary layer
decreases with increasing magnetic parameter, stress work
parameter and conjugate conduction parameter. The
temperature distribution increases for increasing magnetic
parameter while it decreases with increasing stress work
parameter and conjugate parameter. The skin friction
coefficient along the surface decreases for all three
parameters however the rate of heat transfer increases for
increasing stress work parameter while it decreases for
increasing magnetic parameter and conjugate conduction
parameter.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
Merkin, J. H., Free convection boundary layer

on an isothermal horizontal cylinder, ASME,
paper no.76-HT-16, (1976).
Merkin, J. H. and Pop, I., A note on the free
convection boundary layer on a horizontal
circular cylinder with constant heat flux,
Wrme und Stoffbertragung, Vol. 22, pp.7981 (1988).
Kuehn, T. H. and Goldstein, R. J., Numerical
solution to the Navier-Stokes equations for
laminar natural convection about a horizontal
isothermal circular cylinder, Int. J. heat Mass
Transfer, Vol.23, pp.971-979 (1980).
Wang, P., Kahawita, R. and Nguyen, T. H.,
Numerical computation of natural convection
flow about a horizontal cylinder Using
Splines, Numerical Heat Transfer, Part-A,
Vol.17, pp.191-215(1990).
Gdalevich, L. B. and Fertman, V. E.,
Conjugate problems of natural convection,
Inzh-Fiz. Zh. 33, pp.539547(1977).
Miyamoto, M., Sumikawa, J., Akiyoshi, T.
and Nakamura,T., Effects of axial heat
conduction in a vertical flat plate on free
convection heat transfer, International Journal
of Heat and Mass Transfer 23, pp.15451553
(1980).
Pozzi, A. and Lupo, M., The coupling of
conduction with laminar natural convection
along a flat plate, International Journal of
Heat and Mass Transfer 31, pp.18071814
(1988).
Kimura, S. and Pop, I., Conjugate natural
convection from a horizontal circular
cylinder, Numerical heat transfer, Part A,
Vol.25, pp.347-361 (1994).
Wilks, G., Magnetohydrodynamics free
convection about a semi-infinite vertical plate
in a strong cross field, ZAMP, Vol.27,
pp.621-631 (1976).
Aldoss, T. K., Ali, Y. D. and Al-Nimr, M. A.,
[11]
[12]
[13]
MHD mixed convection from a horizontal

circular cylinder, Numerical Heat Transfer,
Part A, Vol.30, pp.379-396 (1996).
El-Amin, M. F., Combined effect of viscous
dissipation and Joule heating on MHD forced
convection over a non-isothermal horizontal
cylinder embedded in a fluid saturated porous
medium, Journal of Magnetism and Magnetic
materials, Vol.263, pp.337-343 (2003).
Gebhart, B., Effects of viscous dissipation in
natural convection, Journal of Fluid
Mechanics, Vol. 14, pp.225232, (1962).
Pantokratoras, A., Effect of viscous
dissipation and pressure stress work in natural
convection along a vertical isothermal plate.
New results, International Journal of Heat and
Mass Transfer , Vol. 46 pp. 49794983 (2003).
[14]
Nazar, R., Amin, N. and Pop, I., Free convection

boundary layer on an isothernal horizontal circular
cylinder in a micropolar fluid Heat transfer,
Proceeding of the 12th int. conference (2002).
[15]
Keller, H. B., Numerical methods in the

boundary layer theory, Annual Reviews of
Fluid Mechanics 10, pp.417433 (1978).
Cebeci, T. and Bradshaw, P., Physical and
computational aspects of convective heat
transfer, Springer, New York (1984).
[16]
NOMENCLATURE
296
Symbol
Meaning
Unit
Radius of the outer cylinder

Thickness of the cylinder
Applied magnetic field
Skin friction coefficient
Specific heat
Acceleration due to gravity
(cm)
(cm)
(N)
(J/Kg.K)
(cm/s2)
M
Nux
p
Pr
Tb
Tf
Magnetic parameter
Local Nusselt number
Conjugate conduction parameter
Prandtl number
Temperature of the inner cylinder
Temp. at the boundary layer region
(K)
(K)
Ts
T
Temp. of the solid of the cylinder

Temperature of the ambient fluid
Velocity components
(K)
(K)
(cm/s)
Dimensionless velocity components
Cartesian coordinates
(cm)
a
b
B0
Cfx
cp
f
u, v
u,v
x ,y
x, y
Greek symbols
Dimensionless Cartesian coordinates
Meaning
Unit
Kf
Co-efficient of thermal expansion
(K-1)
Stress work parameter

Density of the fluid inside
Kinematic viscosity
Viscosity of the fluid
Dimensionless temperature
Electrical conductivity
Thermal conductivity of the fluid
Ks
Thermal conductivity of the solid
(Kg/m3)
(m2/s)
(N.s/m2)
J/msK
(kW/mK
)
(kW/mK
)

SHIP RECYCLING PROSPECTS IN BANGLADESH

Kh. Akhter Hossain, K. Shahriar Iqbal and N. M. Golam Zakaria
Bangladesh University of Engineering and Technology, Dhaka, Bangladesh
E-mail: ksiqbal@yahoo.com
ABSTRACT
This paper deals with the recent status of ship-recycling industry in Bangladesh which has received
considerable attention during last two decades. The social and environmental impacts of the ship recycling in
Bangladesh have also been covered. Considering its positive economical contribution as well as some negative
effect like lack of occupational health and safety standard, the search has made to address whether Bangladesh
should continue supporting this business on their soil. Some viable recommendations are made at the
conclusion.
Keywords: Ship breaking, ship dismantling, industrial safety, safety hazards, environmental hazards.
1. INTRODUCTION
Ship recycling is the process of dismantling old
ship to recover steel scrap and other materials. This
industry had begun in Bangladesh in 1960, when a
cargo vessel named M D Alpine of Greek was
affected by violent storm and beached near sea shore
of Fauzdarhat at Chittagong. The ship had to remain
there for a long time. In 1965, the vessel was
scrapped by Chittagong Steel House. In 1974, Al
Abbas, a salvaged Pakistan Navy vessel, which was
sunk during liberation war, was scrapped by
Karnafully Metal Works and since then commercial
shipbreaking began in Bangladesh. Finally in the year
1984, Bangladesh appears as a major ship recycling
nation in the world [1].
Until 1960s, ship breaking/recycling activity was
considered as a highly mechanized operation that was
concentrated in industrialized countries- mainly USA,
UK, Germany and Italy. The UK was accounted for
50% of the industry-Scotland ran the largest ship
breaking operation in the world. During the 1960s
and 70s, ship breaking activities migrated to semiindustrialized countries like Spain, Turkey and
Taiwan, mainly for the availability of cheap labour
and the existence of re- rolling steel market [2]. But
from early 1980s, ship breaking is no more costeffective in those countries and so to maximize profits
ship owners sent their vessels to the scrap yards of
India, China, Pakistan, Bangladesh, the Philippines
and Vietnam, where health and safety standards are
minimal and workers are desperate for work.

Bangladeshi industrialists also took the opportunities
of this lucrative business which resulted importing
more and more ship to Bangladesh. Thus, within a
short period, Bangladesh established herself as a
leading ship recycling nation in the international
market.
Ship-recycling industries have been located along
the coastal belt of Bangladesh. The industry is
providing the countrys main source of steel and
generates large amounts of revenues for various Govt.
authorities. There are more than 50,000 people
directly involved in the industry. This industry is
currently supplying 60% of the raw materials for the
local steel industry. Table 1 shows the steel market
scenario of Bangladesh related to the ship recycling
industry [3]. It has also a significant contribution to
the local shipbuilding industry. On the other hand,
frequent accident due to explosion, inhuman working
condition, death, fatal injuries and permanent
disabilities also brought this industry to the focus of
attention. That is why we have to consider all
economic benefit from the industry and at the same
time the social and environmental costs [4] should
also be taken into account.
This paper highlighted some of the relevant aspects
of current ship-recycling industry in Bangladesh. At
first an overview of the present ship breaking industry
in Bangladesh has been depicted through studying its
working condition and economic contribution. Based
on data collected from the field survey, contribution
297
of each component from dismantled ship has been

identified. Social and economical costs have been
described. Some recommendation regarding safety
aspect and OSH have also been put forward in line
with the study.
Table 1: Yearly statistics of steel market in
Bangladesh
Steel consumption
Steel production
Scrap steel from ship recycling
yard
Scrap steel contribution to steel
production as raw material
Number of re-rolling mills
Scrap yards
Estimated no. of direct workers
5 million MT
2.5 m MT
Up to 1.5 m MT
(30% of total
consumption)
60%
About 350
120 (Active 80)
50000
2. DATA COLLECTION AND ANALYSIS

Data collection through field survey was not easy
as these industries are very restricted for common
people. However, the authors managed to convince
some of the owners of ship breaking yards to extend
their help for this study. Visiting some renowned
ship breaking yards, the working conditions were
observed, physical data was collected and analyzed to
understand the real life scenario. Since a good
database on such industry was not available, data was
also collected from various print media, NGO reports,
private and government organizations.
3. SHIP-RECYCLING INDUSTRY IN
BANGLADESH
Ship-recycling industries have been located along
the coastal belt of Bangladesh, from Bahatiary to
Barwalia at Sitakund. There are more than 100
registered ship-recycling yards located in this area.
The industry is providing the countrys main source
of steel and generates large amounts of revenues for
various Govt. authorities. There are more than 50,000
people directly involved in this industry, and more
than 100,000 people are involved indirectly. It is
currently supplying substantial portions of the raw
materials for steel production. It is also contributing
to the local shipbuilding through supplying used
machineries and materials.
The beaching method widely applied in the local
ship breaking is unique of this kind. This method,
with flat muddy land and huge labour force, replaced
the demand of expensive and heavy infrastructure,
which is required for pulling and docking the ship. In
this beaching method, the ship is sailed with its
maximum speed using its own power during the high

tide and forced to be beached over the flat muddy
land where it is dismantled to small parts using semiskilled and unskilled labour during low tide. Then the
dismantled parts are pulled to the dry shore area using
electric winch and labour force.
Gas cutting is widely used to make relatively small
pieces from the steel structure. Usually no study is
carried out to follow any order of which parts should
be separated first and which should be the next.
Minimum knowledge of safety is used in this process.
So, sometimes this method of inhuman working
condition leads to frequent accidents due to
explosion, death, fatal injuries and permanent
disabilities. That is why, apart from economic benefit
from this industry, the social and environmental costs
demand huge attention for further development of this
industry.
It is not a miracle that makes Bangladesh a paradise
of ship recycling industry. There are few distinct
reasons for the development of ship recycling
industry in Bangladesh:
Availability of long beach made of soft
sand & muddy land
Perfect low slope for beaching
Large tidal difference of water depth
Stable weather conditions
Abundance, availability and low cost of
labor
Low level of environmental awareness
and moderate enforcement of laws
Possibility of reusing of almost 100%
scrap materials.
High demand of goods recovered from
scrap ships
Again there are a few additional definite reasons
for choosing the present location at Chittagong
(Fauzdarhat to Kumira), for ship recycling industry in
Bangladesh, which are,
Availability of a locality to support a
heavy industry
Availability of linkage industries within
a reachable distance
Having direct road communication with
linkage industries
Non-existence of any sensitive area
around
4. PERFORMANCE OF SHIP RECYCLING

INDUSTRY IN BANGLADESH
Though the ship breaking in Bangladesh started in
sixties, commercially it started in late seventies. The
past data of ship recycling showed that Bangladesh
played a significant role [1] in the ship recycling
world, particularly during 2004 to 2009. In Figure 1,
the share of Bangladesh to the world total of ship
298
recycling in LDT is shown. In Figure 2, it is seen that

Bangladesh led the world ship recycling during 2004
to 2008 and India was just behind it. Before 2004,
India was the number one in ship recycling for long
period. However, currently India has again retained
their leading position in the ship recycling world
leaving China and Bangladesh as their follower. The
trend of Bangladesh showed ups and downs in early
of this decade, but sharply increased in the last three
years. In 2009 Bangladesh recycled around two
million ton scrap. So Bangladesh is now being
considered as one of the global leaders in ship
breaking.
A comparison of percentage cost breakdown [4] of
breaking a 14800 LDT tanker in three Asian countries
has been shown in Figure 3. Figure 4 shows estimated
cost and profit related to the same tanker in the three
countries. From the figure it is seen that cost profit
ratio is 2 times higher than Indian ship braking yards
and six times higher than breaking yard of Pakistan.
Bangladeshi ship breakers usually offer highest price
while purchasing scrap ships, but due to cheap labour
and other benefit compared to neighbouring countries,
cost profit ratio is high. In 2009 Bangladesh recycled
around two million ton scrap ship and thus playing a
vital role in the global ship recycling market [5].
Figure 2: Comparison among major ship breaking

nations ( Year 2002-2009)
Figure 3: Comparison of cost breakdown of sample

Tanker
Statistics of Shipbreaking in Bangladesh
L D T in m illio n
9
8
7
6
5
4
3
2
1
0
BD LDT
Total LDT
Year Year Year Year Year Year Year Year

2002 2003 2004 2005 2006 2007 2008 2009
Figure 4: Cost and profit related to recycling the

sample tanker
Figure 1: Statistics of Ship breaking in Bangladesh
5. A CASE STUDY
( Year 2002-2009)
L D T in m illio n
Comparision with other major shipbreaking

nation
3.5
3
2.5
2
1.5
1
0.5
0
China LDT
India LDT
BD LDT
Pak LDT
Year Year Year Year Year Year Year Year

2002 2003 2004 2005 2006 2007 2008 2009
Ship breaking is an important activity in context of

Bangladesh. Though Bangladesh is playing a leading
role as a ship breaking nation for a long time, but the
data related to these industries are very rare as these
yards are highly restricted to the commoner. In this
study, efforts have been made to assess the social and
environmental impact of ship recycling on the basis of
field data collected from existing recycling yards. A
sample 5000 LDT multipurpose container ship data
has been collected for this purpose. Table 2 shows
where the materials and components of the sample
ship were recycled in Bangladesh. Table 3 shows the
contribution of materials and components to the local
industries as recycling materials and the amount of
pollutants from this ship disposed to the environment.
299
From the Table 2, it is seen that ship braking

industry is providing enormous contribution to the
national economics through supporting inland
shipbuilding industry, construction industry, rerolling mills, steel mills, oxygen plants, and also
cable, ceramic, and furniture factories. It provides
about 35,000 tons of processed wood & furniture
annually and henceforth preventing de-forestation.
Its contribution to the national economy is about US$
1 billion.
On the other hand, ship destined for ship breaking
may contain significant quantities of hazardous
materials and that is why ship breaking operations
involve high risks and harms. In Bangladesh, ship
breaking takes place on sandy beaches following very
rudimentary ways. Because of socio-economical
conditions, labours are very cheap and they do not
bother to take any risky job. They do not have even

basic knowledge of occupational safety and hazard.
As a result, many fatal incidents occur in these yards.
Presence of inflammable substances or gas pockets
leads to break out fire accompanied by explosions
specially during cutting operation. Exposure of toxic
fumes during cutting by torch is also very common.
The workers have no information regarding the
hazardous materials they are handling or the
minimum safety measures that they require.
Unprotected handling of the toxic substances may
cause wide range of complications. However it was
not possible to assess such damage carried out by the
sample ship, as no such information directly related to
the sample ship was available.
Table 2: Materials/Machineries Collected from 5000 LDT Multipurpose ships and their uses
Materials
a. Steel
b. Electric cable and cable sheathings
c. Navigational instrument such as compass,

navigation light, life boat & buoy, life raft, fog
horns, generator, battery, various maps, fire fighting
equipment, etc
d. Marine engine
e. Generator Pumps, Compressors,
Mechanical Equipments & Machineries
Other
Uses
1. Raw material for re-rolling mills.
2. Steel plate, frame, girder, stiffener, longitudinal, etc.
are used for construction of inland vessels
1. House hold and industry
2. Cable sheathing is used in rubber industry
3. Inland Shipbuilding industry
2. Other Industry
1. Export
1. Household use
3. Other industries and some are exported
f. Motor, Light, Fans, Fridge, Switch, Switchboard,

Various Electrical & Electronic Materials
1. Household use
3. Used in industries and some are exported
g. Steering Gear, Capstans, Windlass, Crane, David,

Derik, Anchor, Cable, Chain, Block, Pulley, Wear
rope, Bollard, Fairlead, Deck-eye, Hatch, Hatch
Coaming, Various fittings & fixture, etc
h. Furniture, utensils, bedding materials, bathroom

fittings, refrigerator, washing machine, etc.
1. Household use
3. Used in industries and some are exported
1. Transport industry
2. Inland Vessels
1. Brick field
1. Refrigerant Industry
1. Dying industry
1. Recycled in metal industries
2. Other Industry
3. Export
i. Fuel and lube oil

j. Burnt oil and oil sludge
k. Coolants
l. Dye
m. Heavy metals like copper, zinc, mercury, brass,
alloy metal
300
Table 3: Amount of materials/substances collected, recycled, and disposed from the sample ship
Name of the Items/Materials collect from the
ship
Amount
in Tons
Steel plate, Frames, Girder, Longitudinal,

Stiffeners, Doors, Hatches, Anchor, Bollards,
Fairleads, Hatch cover, Deck eyes, Bulkhead,
Deck plate, etc
Paint containing lead, cadmium, tins, arsenic,
zinc, chromium, silicon, chromates, mercury,
etc
Various hazardous wastes, rubber, Cables,
Batteries, Composite materials, sealants
containing PCBs, etc
Various types of asbestos (Used in very old
ship. Presently not used)
Engine oil, bilge oil, hydraulic and lubricants
oils and grease
Residual oil
Machineries, Equipments, Prime-movers,
Generators, Pumps, Compressors, Gear box,
Shaft, Propeller, Boats, Crane, David, Derik,
Fittings, Fixture, Tools, Spares, Electric
switch & others, TV, Fridge, AC Unit,
Washing machine, various household, etc,
Boats, Crane, David, Derik, Fittings, Fixture,
Tools, Spares, Electric switch & others, TV,
Fridge, AC Unit, Washing machine, various
household, etc
Percenta
ge
recycled
Percentage disposed
to environment
(Pollution)
4600
Percentage
of amount
of all
materials/
substances
92%
100%
Negligible
10
0.20%
01
0.02%
95% (as
remain
with
plate)
95%
5% (During cutting
dragging and
transporting of plate
and others fittings)
5%
2.5
0.05%
95%
0
386.5
0
7.73%
0
100%
5% (Again most of the

amount is collected by
local boat-men)
0
0
6. CONCLUSIONS
Ship breaking is an important activity in the
lifecycle of maritime transportation. It makes a
significant contribution to the global conservation of
energy and resources. From the above discussion and
case study, the following conclusions can be drawn:
In Bangladesh, almost 100% materials and
equipments collected from a scrap ship is
recycled. Through this process ship recycling has
become a major contributor to the development
of national economy.
Contribution of ship breaking to inland
shipbuilding in Bangladesh is enormous [6].
Occupational health and safety can be ensured by
strict monitoring of govt. regulatory bodies as
well as imparting training to the worker and also
providing personal protective clothing &
equipment to the workers.
Environmental pollution can be kept to a

minimum level by following standard
international procedure for disposing ships
materials. Development of detailed ship
recycling guidelines for local scrap yards is also
a necessity for improving the standard of this
industry.
In Bangladesh, ship breaking takes place on sandy

beaches without any containment or barriers to
prevent water and soil pollutions. But environmental
pollution by other industries like tannery, paint,
dying, brickfield, inland/coastal vessels and road
vehicles are not even less severe in Bangladesh.
Actually, level of environmental awareness among the
different industries is still very poor and ship
301
breaking is no exception. Hazardous waste and

materials, which cannot be recycled, are usually
dumped on the spot because of lacking of reception
facilities. So, pointing only to the ship recycling
industries for destroying our environment and human
health is not unbiased.
REFERENCES
[1] Shipbuilding Statistics, The shipbuilders
association of Japan, march 2010.
[2] Misra, H., Status of ship breaking industry in
India, The IUP journal of managerial
economics, Vol. VII, Nos. 3 & 4, 2009.
[3] Statistics of World Bank, Dhaka office,
September 2010.
[4] Abdullah, H.M., Mahboob, M. G. and Biruni,
A.A., Drastic expansion of ship breaking yard
in Bangladesh: a cancerous tumor to the coastal

environment, Proc. of international conference
on environmental aspects of Bangladesh
(ICEAB 10), Japan, Sept. 2010
[5] Cotzais Shipping Group (online), Accessed on
june 2010, available: www.cotzias.gr
[6] Hossain, K. A., Evaluation of potential
prospect and challenges of Bangladeshi
shipbuilding in the light of Global contest,
M.Sc. Engg. Thesis, submitted to Dept. of
Naval Architecture and Marine Engineering,
BUET, Dhaka, 2010.
302

CONCEPTUAL FRAMEWORK FOR THE EVALUATION OF CONTAINER

TERMINALS EXPANSION BY MARGINAL APPROACH
K.B.Loke1, A.S.A.Kader2 and A.M.Zamani3
1
Universiti Teknologi Malaysia (UTM),
Johor, Malaysia.
E-mail: gsk0053@yahoo.com
2
Johor, Malaysia.
E-mail: abdsaman@fkm.utm.my
3
Johor, Malaysia.
E-mail: zamani@fkm.utm.my
ABSTRACT
Most of the container terminals expansion models are focused on fulfillment of future throughput demand.
The purpose of expansion only described overall terminal expansion with respect to increase in demand. At this
moment, container terminals expansion models did not consider small changes in commercial viability with
small changes in expansion magnitude over time. Therefore, this study intended to look at the alternative ways
of container terminals expansion model. Critical reviews are presented to compare the existing approaches and
underlined the merits, drawbacks and specify area for that approaches. At the same time, an alternate approach
would be highlighted for further study. It expanded from existing approaches and draw attention to marginal
approach by using Net Present Value (NPV) to evaluate the increment requirement for future throughput
demand. The positive NPV represent the significant of increment of the expansion component and magnitude
respectively for each expansion period. The purpose of using marginal approach is to assure a sustainable and
economical effective expansion plan.
Key words: container terminal, expansion model, marginal approach, net present value, expansion component
and magnitude, suitable and economical effective.
1. INTRODUCTION
90% of cargoes were likely to be
containerised. In highly developed trades, it was
estimated that some 70% of containers move in FullContainer-Load (FCL) basis; and the remaining by
Less-Than-Container-Load (LCL) shipments (Branch,
1986). Likewise, more than 90% of international
trades move through seaports and 80% of sea cargoes
move in containers through major seaports (Won et
al., 1999). The study has proved that the worldwide
container throughput increases approximately 10.7%
annually. Hence, the development and expansion of
container terminal has become crucial in order to meet
the demand for container traffic.
To meet the container terminals development
and expansion requirement, physical port layout is
one of the important, such as seaside and landside. To
ensure a well coordination, reliability of operation, in
favor for profit and benefit, the port layout,
networking, etc also must be designed to fix to the
expected future demand (Chalid, 2009). To be success
to support the additional capacity throughput, features
of port expansion normally including extra shipping
berths, terminal land, depth of dredging area, road and
rail connection, additional facilities and etc.
As its importancy and needs, a lot of studies

and scientific methods have been proposed to solve
the port development and expansion problems. To be
clear and value on it, this paper has be underlined and
elaborated details on port development and expansion,
past and current design approaches in section 2.
To ensure a well manner, several of study has
been done. UNCTAD, 1985 used the planning chart
concept to leaded the different facilities or
infrastructure. The formula has been converted into
chart for immediate use. Frankel, 1987 employed
mathematical techniques to familiarize with the issues
and methods of port planning and development.
Thomas, 1999 specified on container handling system,
by given significant efficiency and competence to that
container terminal selected. Zamani, 2006 utilize
fuzzy methods to development planning model. He
tries to improve the lack of human modes in planning
approaches. Dekker, 2008 apply marginal approach to
determine when, size and interval expansion time in
such method. Figure 2.1 shows the existing and
current port development and expansion approaches
303
Initial
Development
UNCTAD (1985)
Empirical
Approach
Initial
Development
Frankel (1987)
Empirical
Approach
Zamani Model
M
(2006) are special forr initial port set
s
up. Niswaari Model (20005) and Dekkeer Model (20008)
are particu
ular for port exxpansion estim
mation. Howeveer,
Saman Model
M
(1997)) is used for
f
port initiial
developm
ment and expannsion planningg. But, it uniquue
for inlandd water way designs. On the other hannd,
Kendra Model
M
(1997) and
a Thomas Model
M
(1999) are
a
specific for environm
ment protectio
on during poort
ment and appproximation multi
m
containner
developm
decision making
m
respecttively.
1
1990
Kendra (1997)
Theoretical Approach
Environmental
Preservation in
Development
Saman (1997)
Costing Approach
h
Initial
Development
2.1
Thomas (1999)
Weightage Approach
Equipment
Selection
2.1.1 UN
NCTAD MOD
DEL, 1985
2000
Capacity
Expansion
Contaainer terminall developmentt and expansioon

planning need
n
to considder the containeer traffic demannd
with the terminal caapability. First, its need to
determinee the annual thrroughput folloowed by terminnal
productivvity, facility am
mount and sizze, and level of
service provided. Thhe main coonsideration of
UNCTAD
D model is to relate the term
minal capabiliity
with the level of servicee provided.
Niswari (20
005)
Empiricaal
Approach
h
Zamani (2006)
Fuzzy System
S
Approach
Initial
Developmentt
Deekker (2008)
Marg
ginal Approach
Capacity
Expansion
EM
MPIRICAL A
APPROACH
H
2010
Figure 2.1 Existiing and Current Port Dev

velopment and Expansio
on Approaches
D
between
b
Existin
ng and Currennt
Table 2.1 Discrepancy
Port Devellopment and Exxpansion Apprroaches
The developmen t and expan sion model of

UNCTAD
D is focused on five elem
ments. The firrst
element is to determinee the containerr traffic demannd
and is statted in TEUs (T
Twenty-Foot Equivalent
E
Unitts).
Containeer traffic dem
mand is the total
t
contain er
throughpuut handled by terminal. Thee other elemennts
are contaiiner park area,, container freiight station areea,
berth-day
y requirement,, and ships cost
c
at terminaal.
The size
s
of containner part area is determined by
b
total conttainer throughhput, average transit time of
container,, area requirrement per TEU,
T
containner
stacking height,
h
and saafety factor. Next,
N
the size of
container freight statiion area is determined by
b
container throughput, average stacking height of
general cargo, accesss area, and safety factoor.
Subsequent, the annuaal berth day requirement is
determineed by average ship operatin
ng hour per daay,
average number
n
of conttainer units perr hour per cranne,
number of
o crane per shhip, average nuumber of movves
per ship, and number of ships per day.
d
Finally, thhe
i determined by annual berrth
ships cosst at terminal is
day requirrement, numbeer of berth tariiff days per yeaar,
berth utiliization, and avverage daily shhips cost. All of
these elem
ments are convverted in the form
fo of planninng
chart.
Application: Particcular in seaport developmennt.
Meritts: To determinne the capacityy requirement to
support future
fu
demand based on fourr variables, theey
are contaiiner park area,, container freiight station areea,
berth-dayy requirement, and
a ships costt at terminal.
d
beetween existingg
Table 2..1 shows the discrepancy
and curreent port devvelopment annd expansionn.
UNCTAD Model (1985)), Frankel Moddel (1987), andd
Draw
wbacks: Lack oof control on un
ncertainly. Som
me
more, doesnt detail ouut container handling
h
systeem
and terminnal other areass.
304
2.1.2 FRANKEL MODEL, 1987

Frankel model taken on the same standard with
UNCTAD model, and agreed that the container
terminal layout and equipment is important in
container terminal planning and is depending on
container moves, terminal layout, equipment
performance, and operation strategy. Frankel Model
considers for four main elements, there are container
demand, area required, equipment selection and
systems cost. But, it provided details in calculation of
container part area and calculation of number of berth.
Application: Particular in seaport development.
Merits: Provide the details algorithm calculation in
container part area and number of berth requirement.
Drawbacks: Lack of control on uncertainly. Some
more, doesnt detail out container handling system
and terminal other areas.
2.1.3 NISWARI MODEL, 2005

This model related the operational capacity and
financial requirement with cost expansion. Its
considered that the expansion plan is a manner to
build up the operation capacity. It is assumed that the
container terminals capacity is determined by total
quay moves, based on the total of container handle
each year.
Application: Particular in container terminal
expansion.
Merits: Evaluate the operation and financial needs
in expansion.
Drawbacks: Its focus on berth and yard expansion
needs only.
2.2
THEORETICAL APPROACH
2.2.1 KENDRA, 1997

This study point out that most of the seaport
expansion is underlying the environmental
consideration. So, this model links the development of
dry cargo and container terminal with environmental
consideration. This theoretical model identifies the
different style of addressing coastal and terminal
conservation. It is tied into concepts of sustainability
and environmental control in terminal development
and expansion needs.
Application: Application
container terminal.
in
dry
cargo
and
Merits: Provide foundation understanding in

coastal management and preservation.
Drawbacks: Lack of association
theoretical and practical control.
2.3
between
COSTING APPROACH
2.3.1 SAMAN, 1997

Saman model highlighted cost estimate based on
physical component development or expansion. The

cost estimate model is useful for preliminary
evaluation of inland waterway terminals cost
investment judgment. This model focused on track
cost model, vessel cost model, and terminal cost
model. Track cost model direct with waterway and
land constructions cost; like dredging the channel,
construction of locks, bridges, levees, bank
strengthening, bank and bottom protection, road
construction, landscaping, plantation, and fencing.
Vessel cost model determined with traffic flow
projections, vessel capital requirement plus operating
cost. Terminal cost model can be classify in fixed and
variables costs based on terminal operating needs.
Application: Unique in inland waterway transport
systems
Merits: To understand in preliminary cost
requirement in development or expansion of inland
waterway terminal.
Drawbacks: Lack of control in profitability and
investment return.
2.4
WEIGHTAGE APPROACH
2.4.1 THOMAS MODEL, 1999

Thomas model highlighted that the selection of
container handling system is an important issues in
determine the efficiency and cost effective of a
container terminal. This model given summary of
container handling system based on their features; like
land utilization, terminal development costs,
equipment cost, equipment maintenance costs,
manning level, and operating factors. Thomas model
has proposed six handling system; like tractor-trailer
system, straddle carrier direct system, straddle carrier
relay system, yard gantry system, front-end loader
system, and combination system.
Application: Particular in seaport development.
Merits: Its has classified the handling system
performance based on their features.
Drawbacks: Lack of control on uncertainly and
future demand need.
2.5
FUZZY SYSTEM APPROACH
2.5.1 ZAMANI MODEL, 2006

Zamani model expanded from UNCTAD model,
Frankel model, and Thomas model. Its included four
of the elements mention by UNCTAD model and
Frankel model, 1 element from Thomas model, and
adds in 1 additional element. The six elements are
container part area, container freight station area,
ships cost at terminal, berth-day requirement,
container handling system, and terminal other areas.
Application: Specific in container terminal
development.
Merits: Zamani model used fuzzy expert system to
305
overcome the short of the conventional methods of

container terminal development planning. The
conventional methods used planning charts and
mathematics formula to determine the port develops
needs. Zamani model improved the conventional
methods by apply the natural modes of reasoning that
involves approximate, imprecise, linguistic and
subjective values.
Drawbacks: Lack of control on financial and
economical.
2.6
MARGINAL APPROACH
2.6.1 DEKKER MODEL, 2008

Dekker model important the issues of when a
terminal need to be expand, what size should be
expand, what is the interval capacity the terminal
should be expand. This model used marginal approach
to analysis the needs of expand. The optimal
expansion of a container terminal is determined by the
steady demand growth. Then, its related the
expansion needs of container terminal with TEU
forecasting. Where, a container terminal achieves a
certain TEU throughput that desires the needs of
expansion.
Application: Specific in container terminal
expansion.
Merits: Related the future demand need with
financial control.
Drawbacks: Lack of control on expansion
component and magnitude.
spent based on change in demand dQ, but some

expansion component could sustain dQ will others
will not, eg. Storage area may need to be expanded
will the number of quay crane can still be maintained.
This sustenance period will continue until dQ further
increase to a new level to justify for the next
expansion, eg. storage and quay crane. Container
terminal expansion will be more accurate if dQ and
periods of sustenance for each expansion component
could be identified so that the expansion of expansion
component is at correct magnitude and at the correct
time.
Therefore, this study intended to look at the
alternative ways of container terminals expansion
model. It expanded from existing approaches and
draw attention to marginal approach by using Net
Present Value (NPV) to evaluate the increment
requirement for future throughput demand. The
positive NPV represent the significant of increment of
the expansion component and magnitude respectively
for each expansion period. The purpose of using
marginal approach is to assure a sustainable and
economical effective expansion plan.
It is extent from Dekker (2008) model. But,
Dekker (2008) model lack of concentrated on small
changes in expansion component and magnitude. So,
this paper tries to overcome the limitation and take on
the expansion component variables into account. The
expansion component variables adopted from Zamani
model (2006). Zamani model (2006) is the recent and
more absolute version that using fuzzy system
approach with completed expansion variables.
4.
3. MAPING THE DIRECTION

Since 2000 until 2010, the container terminal
development and expansion approaches are more
concerned about financial and economic impact. It
shows that the future diagram is started giving their
attention and interest in this area.
Empirical Approach,
UNCTAD (1985)
The expansion cost for expansion component is
Empirical Approach,
Frankel (1987)
container park area,

container freight
station area, berth-day
requirement, and
ships cost at terminal
Frankel (1987) clarified that the objective of port

development and expansion is to maximum net profit
or minimum cost of expenses. For economic point of
view, a port authority should meet the port service
requirements with minimum cost expenses or
maximum profit. With limited resources and supplies,
the port authority should depend on availability of
resources allocation to plan for development and
expansion needs.
On the other hand, most of the container terminals
expansion approaches are focused on fulfillment of
future throughput demand. The purpose of expansion
only described overall terminal expansion with
respect to increase in demand. At this moment,
container terminals expansion models did not
consider small changes in commercial viability with
small changes in expansion component and magnitude
over time.
CONCEPTION FRAMEWORK
FORMULATION
container part area

and number of berth
Weightage Approach,
Thomas (1999)
container handling
system
Modify
Fuzzy System Approach,

Zamani (2006)
container park area, container freight

station area, berth-day requirement,
ships cost at terminal, container
handling system, and terminal other
area
Marginal Approach
Finance
Modify from fuzzy system

approachs expansion
component (container park area,
container freight station area,
berth capacity requirement,
container handling system, and
terminal other area
Marginal Approach,
Dekker (2008)
Container
Terminal
Expansion
Model
TEU/year (maximum capacity

extension)
Figure 4.1 Conception Framework Formulations for

Container Terminal Expansion Model
306
Most of the conventional models used the

empirical approach to describe container terminal
development or expansion (UNCTAD, 1985 &
Frankel, 1987). Thomass model (1999) focused on
container handling system. Zamanis model (2006)
modifies UNCTAD (1985), Frankel (1987), and
Thomas models (1999); while trying to overcome the
lack of human approximation style by using
linguistics terms in conventional models. It considers
on container part area, container freight station, berthday requirement, ship cost at terminal, container
handling system, and terminal other areas.
Dekker model (2008) highlighted the lack of
marginal cost in conventional models. This model
used marginal approach to control the optimal
expansion in between marginal investment costs and
marginal benefits.
For marginal approach definition, Roger (2004)
expressed marginal approach is the additional cost of
produce an extra unit of production. Its also can be
used for cost saved by producing one less unit. So,
marginal approach is the method of cost incurred or
cost saved in the produce of the marginal unit of
production.
5. PRELIMINARY MODELING
EVALUATION
Before any expansion plan taking action, the future
throughput demand needs to be determining first.
After that, current capacity for each expansion
component needs to be matching with the future
throughput demand. If the current capacity is biggest
than future demand, then that component not need to
expand. But, if the current capacity is smaller than
future demand, then that component needs to be
expanding. Figure 5.1 show the mapping of
preliminary modeling evaluation. The future
throughput demand, time, expansion size and
significant of expansion will discuss in next section.
Receive
Future
Demand
Matching
Current
Capacity
with Future
Demand
Identify
Capacity
Sustain the
Demand
Yes
Dont
Expand
No
(2006) and Dekkers model (208), the expansion

component variables are based on each infrastructure
component that need to analyze. It is focus on 5
variables, there are container part area, container
freight station, berth capacity, container handling
system, and terminal other areas.
,
Where,
cte
cpa
cfs
bcr
chs
toa
- container terminal expansion

- container part area
- container freight station
- berth capacity requirement
- container handling system
- terminal other areas
Insist of the unit for all infrastructure components

is in unit $, therefore berth-day requirement (in
Zamanis model, 2006) is changed to berth capacity
requirement. Ship cost at terminal using by Zamanis
model (2006) no take in account, because is the ship
tariff paid by shipping line by using port service. But,
it will calculate in next section, as one of the variables
in marginal approach formula. The further details of
calculate method is discuss on next section.
Current container terminal expansion model only
describe overall terminal expansion with respect to
increase in demand. Incremental expansion of
individual infrastructure taking into account, and its
ability to sustain increase in demand has been
neglected.
Currently,
expansion
cost
for
physical
infrastructure is spent based on change in demand but
some infrastructure could sustain while others not, eg.
container part area may need to be expanded while the
container freight station can still accredit the
expansion. This sustainance period will continue until
TEU increases to a new level to justify for the next
physical expansion. Container terminal expansion will
be more accurate if the quantity and periods of
sustainance for each infrastructure could be identified
so that expansion of infrastructure is of correct of
magnitude and at the correct time. Figure 5.2 show
the container terminal expansion component plan for
each variable respectively (cpa, cfs, bcr, chs, toa).
Expand
Calculate How
Much to Expand
(Capacity)
Capacity Sustainance
Calculate Time
To Expand
Expansion
Size
Calculate
Significant of
Expand
TEU0 TEU1
Display Result
Figure 5.1 Map of Preliminary Modeling Evaluation

Based on the recent study, Zamanis model
TEU2
TEU
Figure 5.2 Container terminal infrastructure expansion

plan for each variable respectively (cpa, cfs, bcr, chs,
and toa)
307
Therefore, each of the infrastructures is

independent to deal with future container throughput
respectively. The expansion terminal is the
combination of each expansion of infrastructure to
meet the future demand.
cpa
cfs
bcr
chs
toa
Where,
if t < h
if t > = h
Where,
Q
- Quantity Demand
- Number of Year
- Planning Horizon
cte
- container terminal expansion
Q0
- Quantity Demand at Time = 0
cpa
- expansion of container part area
Qt
- Quantity Demand at Time = t
cfs
- expansion of container freight station
Qh
- Quantity Demand at Time = h
bcr
- expansion of berth capacity requirement
- Annual Demand Growth Rate
chs
- expansion of container handling system
toa
- expansion of terminal other areas
Each of the expansion components will deal with

the expansion cost and investment return respectively.
However, the container terminal expansion model
would consider the combination result from all
expansion components.
After that, the container terminal expansion model
will use the concept of Net Present Value (NPV) to
evaluate increment requirement when the NPV = +ve;
where the +ve NPV = attractive; the significant of
increase the expansion component respectively for
each expansion period.
ve NPV
attractive
This study would look at the demand pattern of

each expansion component respectively. So all of the
expansion component would analysis and investigate
the future demand needs in that order.
6.2
SIGNIFICANT OF EXPANSION
Thereafter, Net Present Value (NPV) would be use

to determine the container terminal capacity
expansion. NPV define as rule holding that one should
invest if the present value of the expected future cash
flows from an investment is larger than the cost of the
investment (Robert et al, 2009). Once a company
decides to invest a new investment, the future cash
flows that the new investment generates would affect
the decision making. NPV use to calculate the return
from investment and compare with cost of investment.
If the output is positive, then the new investment can
carry on. Therefore, the formula of NPV is defined as
below:
Where,
+ve NPV
- Positive Net Present Value
- sum of income
Where,
- sum of principle of investment
NPV
- Net Present Value
ctri
- sum of container terminal return
from investment
- Principle of Investment
ctec
expansion cost
- Net Income
- Discount Rate (Inflation Rate)
- Number of Year
sum
of
container
terminal
Then, this part would estimate when t = j, the

period that the expansion component need to be
expand.
6. PRELIMINARY ALGORITHM
EVALUATION
6.1
FUTURE DEMAND PLANNING
The future demand planning is important in

determine a series of future capacity expansion in
order to support the future demand. Dekker Model
(2008) used steady linear growth to determine future
demand pattern. Its also used by Dekker (2005) in
determine the port capacity planning. To determine
the future demand at time t, Qt, the following equation
is using:
----- (1)
Where,
I - Income
D Dues Collected
C Cost of the Expanded Part of the Expansion
Component
308
Next, this division would discuss on the return

from investment and cost of the expansion for the
expansion components. Grigalunas et. al. (2001)
address that the income for the port should be the
revenue received in year t. However, the cost of the
expansion should in include operators investment
spend in year t, fees paid by the operator in time t,
operating costs in year t, maintenance costs in time t,
and mitigation costs in time t.
be determine in nex section. Each of the investment of

expansion component for the expansion period would
be decide by NPV. If the output value is positive, then
the expansion project is significant for expand.
6.3
C3
Capacity (TEU)
For return from investment, it determine by the

dues collect from revenue
TIME CONSTRAINT
----- (2)
Where,
Dt is total dues (revenues, tariff, etc) collected
after expansion of expansion component
Quantity Demand
Capacity Supply
Critical
Year, Tc2
ECQ
C2
Safety
Capacity x
Capacity Supply
Critical
Year, Tc1
Setup
Point, Ts1
C1
For cost of the expansion, this part more concern

the cost after initial principle investment
T0
Ts1
Tc1
Tc2
Time (Years)
Sustain Period Expand Period
Figure 6.1 Time Frames for the Demand and Install

Capacity
Figure 6.1 show the time frame for the demand
and install capacity. A new expansion component is
start to install at time Ts1, another at time Ts2, and etc.
By setup point and economic capacity quantity
approach, the expand size and time can be determine.
But, it strongly recommends that the second onward
setup point and economic capacity quantity can be
review from time to time. Because some uncertainly
may affect the output result; like politics, economic,
technology, and etc.
Economic Capacity Quantity (ECQ) is the
minimum or economical capacity expansion. Fawcett
et. al, ( 2007) proposed that
----- (3)
Where,
OI - Operators investment (e.g. training) spend
in year t
FP
time t
- Fees (e.g. salary) paid by the operator in
OC - Operating costs in year t

MC - Maintenance costs in time t
MIT - Mitigation (other costs) costs in time t
Insert equation 2 and 3 into equation 1.
ECQ =
----- (4)
After the total net income has obtained, insert
equation 4 into NPV values.
UNCTAD (1985) suggested that the development

of a port consists of a combination of medium term
and long term planning of new facilities plus. For the
existing port, short term actions also need to improve
the management and present facilities and their used.
Its recommended that planner needs to place 20 years
and draw a consistent picture of that entire he find at
the long term planning. So this study would set 20
year as the time horizon. Then, the sustainance and
expansion period in between the time horizon would
Where,
A
- Time Horizon Demand
S
- Cost per Setup
C
- Costs per Unit of Infrastructure
P
- Carrying Costs as a Percentage
Normally, the carrying costs are in between 20%
to 30% of the value of the product.
After determination of capacity of increment, the
second question is to determine sustain and expand
times for the expansion component. The critical point
in between the sustain time and expand time is call
setup point (SP).
First, calculate the daily demand.
Where,
DD
- Daily Demand
AD
- Annual Demand
309
BD
- Number of Days Open for Business

Second, calculate the setup point.
[6]
Where,
SP
- Setup Point
DD
- Daily Demand
LT
- Lead Time
[7]
7.
[8]
CLOSING
This study intends to develop a generic container

terminal expansion model for the entire expansion
component in a container terminal. It is also
determine the magnitude, sustain period, and expand
period for each expansion component in particular
time.
Therefore, this study would contribute an
expansion model for container terminal in developing
countries. The output of study would produce a multi
container decision making tool for port planner in
expansion estimation to meet the future throughput
demand
REFERENCES
[1] Branch A.E., 1986. Elements of Port Operation
and Management. Chapmen and Hall Ltd., New
York. 240 Pages.
[2] Chalid A.H., 2009. Model for Mixed Sea and
Inland Port Design. Doctor of Philosophy in
Mechanical Engineering. Universiti Teknologi
Malaysia. 203 Pages.
[3] Dekker Sander & Verhaeghe RJ, 2008.
Development of A Strategy for Port Expansion:
An Optimal Control Approach. Maritime
Economic and Logistics, Vol. 2008, No. 10. pp
258 274.
[4] Frankel Ernst G., 1987. Port Planning and
Development. John Wiley and Sons, New York.
795 pages
[5] Kendra James M, 1997. Seaport Development
Versus Envitonmental Preservation: The Case of
Sears Island, Maine, USA. Marine Policy, Vol.
21, No. 5, pp 409 424.
[9]
[10]
[11]
[12]
[13]
[14]
310
Niswari Astrini, 2005. Container Terminal

Expansion to Build Capacity: A Case Study.
Master of Science in Maritime Economics and
Logistics. Erasmus University Rotterdam. 73
pages.
Rogel L.C., 2004. Marginal Cost in the New
Economy Proposal for A Uniform Approach to
Policy Evaluations. M. E. Sharpe, Inc. New
York. 188 Pages.
Saman Kader, 1997. Cost Modelling for Inland
Waterway Transport Systems. Doctor of
Philosophy in Marine Engineering. Liverpool
John Moores University. 276 pages.
Thomas B.J., 1999. Improving Port Performance
Container Terminal Development, United
Nation Fonference on Trade and Development
and Swedish International Development
Authority, UNCTAD
UNCTAD, 1985. Port Development A
Handbook for Planners in Developing
Countries. UNITED NATIONS, New York. 227
pages.
Won Y.Y. and Yong S.C., 1999. A simulation
Model for Container-Terminal Operation
Analysis Using an Object-Oriented Approach.
International Journal of Production Economics,
Volume 59, Issues 1-3, March 1999. Pages 221230.
Zamani Ahmad, 2006. The Application of Fuzzy
Expert System to Preliminary Development
Planning of Medium Size Container Terminal.
Doctor of Engineering in Engineering Business
Management. Universiti Teknologi Malaysia.
355 pages.
Grigalunas et. al., 2001. Comprehensive
Framework for Sustainable Container Port
Development for the United States East Coast.
University of Rhode Island Transportation
Center. Department of Environmental and
Natural Resource Economics. 46 pages.
Fawcett et. al., 2007. Supply Chain Management
Form Visio to Implementation. Pearson
Education, Inc., New Jersey. 530 pages.
AN ALTERNATIVE STRUCTURAL DESIGN FOR TANKERS

Mohammad Ashickur Rahman Khan
School of Marine Science & Technology,
University of Newcastle upon Tyne, United Kingdom
E-mail: ashick23@yahoo.com
ABSTRACT
The aim of this paper is to initiate an alternative design for tankers. The target of this investigation is to
create a new structural arrangement for the double hull region of the tanker, using corrugated plates after
replacing all the secondary support members such as longitudinals, floors, intercostal girders, brackets,
stringers, stiffeners, frames, angle clips, bounding bars etc. This work has been mainly influenced to generate
a shape with easy accessibility, well visibility and less complication to fabrication, since this portion of vessel
is really congested for inspection and maintenance. However, this study is predominantly focused on the
calculations of thicknesses and section modulii of the proposed corrugated plates and comparison with the
required section modulii for a model vessel in the same region according to the classification rules. All the
calculations have been carried out according to the Common Structural Rules For Double Hull Oil
Tankers, adopted by International Association of Classification Societies (IACS), which was issued in July
2009. As a model vessel the Yakumosan (Hull no. 3223), owned by Star Express Inc. is considered.
Keywords: Tanker, Double hull, Corrugated plate, Section modulus, Web and Flange plate.
1. INTRODUCTION
A tanker is used to transport liquid cargoes in
bulk, the most common type being the oil tanker,
which is also known as petroleum tanker. A wide
range of other liquids are also carried in tankers
such as liquefied petroleum gas (LPG), liquefied
natural gas (LNG), chemicals (e.g., ammonia,
chlorine and styrene monomer), fresh water, wine
etc. Tankers ordered [6] after 1993 had to comply
with the MARPOL double hull regulation. This is
opposed to single hull tankers where one or more
cargo holds are bounded in part by the ship's shell
plating. In the double hull design the cargo tanks are
completely surrounded by wing and double bottom
tanks which can be used for ballast purposes.
Therefore, this portion of the vessel needs
continuous inspection, cleaning and maintenance.
But due to the construction technique in
conventional way, this region is crammed with a
variety of support members such as longitudinals,
floors, intercostal girders, brackets, stringers,
stiffeners, frames, angle clips, bounding bars etc. To
commence an alternative of all these members, this
work has been prepared to suggest a design with a
single cross sectional segment in the double hull
region which will execute all the requirements of the
IACS [1].
Figure 1. The naming and launching ceremonies [2]

for this 300,000 t class double-hull tanker were held
at the Kure Shipyard of IHI Marine United
Company Ltd. on November 28, 2008. MOL
(Mitsui O.S.K. Lines, Ltd.) Chairman Kunio Suzuki
named the ship and rope cutting honours went to
Mrs. Yukie Yoshida, the wife of Mr. Genichi
Yoshida, Executive Vice President of Mitsui &
Company Ltd. The Yakumosan left on its maiden
voyage to Saudi Arabia under the arch of a rainbow
celebrating its bright future. This tanker offers the
safest, most environment-friendly features such as
double-hull fuel tanks and VECS (Vapour Emission
Control Systems). Recently, it is transporting crude
oil from the Middle East to Asia. The general
particulars [5] of Yakumosan are shown in Table 1.
Table 1. General Particulars of the Model Vessel
Yakumosan
Owner
Star Express Inc.
Builder
IHI Marine United
Hull No
3223
Ship Type
VLCC
Length (OA)
333.00 m
Length (BP)
324.00 m
Beam
60.00 m
Depth
29.00 m
Draught
19.20 m
DWT
300,000 t
GT
160,300 t
2. MODEL VESSEL
A VLCC named Yakumosan (Hull no: 3223) is
considered as the model vessel, which is shown in
311
Proceeding
gs of MARTEC 20010
Main En
ngine
Maximuum Continuouss
Revolutiion (MCR)
Service Speed
Classificcation
Complettion
DU-Sulzeer 7RT-flex
84T-D dieesel x 1 unit
1. Longitudinally, onlly amidships is consideredd

(i.e.
162)
27,160 kW
W
b
2. Transvversely, for double bottom:
26, 0,
0 26
andd
for
doouble
28, 29, 30
0, 28, 29, 30
considdered.
15.8 kt
NK
Novembeer 28, 2008
side::
aree
3. Verticcally, for doublle bottom:

0, 1, 2 and forr
doublee side:
4, 1
19.2, 29 are connsidered.
5. CAL
LCULATION
NS
5.1 Assu
umptionss
Figure 1. The Modeel Vessel Yaku

umosan [2]
3. PROP
POSED STRUCTURAL DESIGN
D
The purppose of this project
p
is to deevelop a new
double huull configuratioon for tankers. Since, the
original miid-ship section
n of the model vessel
v
has not
been possible to collect; therefore it is assumed that
the vessel has been bu
uilt in a conv
ventional way
following the classificattion rules. In the proposed
design, all the supporting members aree replaced by
simple lonngitudinal corruugated plates. The breadths
of web andd flange of theese plates are taken
t
as 1100
mm and 1000 mm resspectively. Meeanwhile, the
corrugationn angle of thhese plates is taken as 65
degrees. Moreover,
M
theree are two side girders in the
double botttom region on
n each side of the
t centreline
at a distannce of 8748 mm
m and 17496 mm from the
centreline as well as a centreline girder.
g
In the
double sidde region theree are also twoo longitudinal
girders at a distance of 8748 mm and 17496 mm
b
of the wing
w
tank on each
e
side (i.e.,
from the bottom
port side and starboard side). Since the mid-ship
section of the proposed design
d
as show
wn in Figure 2
is mainly focused on thee double hull region of the
vessel, theerefore no strucctural memberrs in the deck
region are shown.
To sim
mplify the calcculations some assumptionss
are consid
dered. They aree as follows:
a) The cargo region off the tanker is parallel to thee
centree line of the vesssel.
b) The double
d
bottom tanks and douuble side tankss
(i.e., wing
w
tanks) aree used for water ballast.
c) Ballasst tanks are designed for ballast waterr
exchan
nge by flow-thhrough method.
d) The position
p
of thee aft bulkheadd is at
65
5
(i.e.
0.2 ) and tthe forward bulkhead
b
is att
290 (i.e. 0.9 ). Therefore, the cargo tankk
n is in between 0.2 to 0.9 .
region
e) The bilge
b
starts at
26 in port side andd
26 in starboarrd side.
f) Abovee
4, the loocal breadth at any draught iss
same as
a the breadth of the vessel.
g) Heigh
ht of air pipe orr overflow pipee is 1 m abovee
the freeeboard deck.
4. OVER
RVIEW OF THE
T
WORK
K
PROCES
SS
This sttudy is mainnly concentraated on the
calculationns of thicknessses and sectionn modulii for
the corruggated plates being
b
considerred. For this,
proper siggn conventions are used (ii.e. -axis is
positive to
owards forwarrd of the vesssel, -axis is
positive tow
wards port sidde of the vessell and -axis is
positive towards
t
upwaard side of the vessel).
Meanwhilee, the calculatiions are based on a specific
point of the vessel. Therefore,
T
som
me particular
points are considered. The co-ordinaates of these
points are as
a follows:
Figure 2.
2 Midship Sectiion of the Propoosed Structurall
Design
5.2 Calcculating Blocck Coefficien

nt
An emppirical formulaa developed by
b Ayre [4], iss
defined ass:
3
3.33
1.67
where:
rule length,, in m = 324 m;
m
servicee
mouldedd
speed, inn knots = 115.80 knots;
displacem
ment volume at the scantling draught,
d
in m3.
308,338 tonne
Hence:
312
Meanwhile, according to IACS (Section 4/1.1.9.1),

Block Coefficient at the scantling draught,
, is to
be taken as:
moulded breadth measured amidships
where:
at scantling draught waterline, in m = 60 m;
scantling draught, in m = 19.2 m.
,
Therefore:
0.826
Note: The numerical values of vessels

characteristics in this section are obtained from
Table 1.
5.5 Calculations of Corrugated Plate as a

Primary Member in Double Bottom Region
According to IACS (Section 8/2.5.6.4), the net
, of the web and flange of corrugated
thickness,
plates are to be taken as the greatest value calculated
for all applicable design load sets (i.e., each design
load set corresponds a particular load component at
a specific deign load combination, acceptance
criteria set and some parameters such as dynamic
load combination factors, metacentric height and
radius of gyration ), and given by:
5.3 Double Hull Arrangement

The double bottom and double side tanks and
spaces, protect the cargo tanks or spaces, and are not
to be used for the carriage of oil cargoes. According
to IACS (Section 5/3.2.1.1 & 5/3.3.1.1), the double
bottom depth and double side width are both taken
as 2.0 m.
P
= design pressure for the design load set
being considered, calculated at the point, in kN/m2
= breadth of plate:
According to IACS (Section 8/Table 8.2.1), the

thickness of plating in the cargo tank region is to
comply with the appropriate minimum thickness
requirements given in Table 2:
Scantling Location
Net Thickness (mm)

Formula
Value
Keel Plating
6.5 + 0.03L2
15.50
Bottom Shell
4.5 + 0.03L2
13.50
Bilge
4.5 + 0.03L2
13.50
Side Shell
4.5 + 0.03L2
13.50
Upper Deck
4.5 + 0.02L2
10.50
4.5 + 0.02L2
10.50
Hull internal
boundaries
tank
Non-tight bulkheads
and other plates
4.5 + 0.01L2
; (for flange plating, in mm)
; (for web plating, in mm)
= permissible bending stress coefficient:

= 0.75
; (for acceptance criteria set AC1)
= 0.90
= specified minimum yield stress of the

material, in N/mm2 = 235 N/mm2
Since, corrugated plates are used as primary support
members in the double bottom region, in the table
for Design Load Sets for Primary Support
Members of IACS (Section 8/Table 8.2.9), the
specifications for Double bottom floors and
girders are considered. The maximum values for
each design load set in this region are shown in
Table 3.
Table 3. Maximum Values Obtained from
Calculations for Each Design Load Set
Design
Load
Set
7.50
Load
Component
where:
rule length, L, but need not be taken
greater than 300 m (i.e.,
300 m).
Design
Load
Combi
nation
Acceptance
Criteria
Set
Value
S+D
AC2
197.83
AC1
193.06
S+D
AC2
296.40
AC1
278.40
Sea Pressures
2
According to IACS (Section 8/2.2.1.1), keel plating

has to be extended over the flat of bottom for the
complete length of the ship. The breadth,
, is not
800 5
mm.
to be less than:
800 5 300 2300 mm.
Therefore:
Since, the breadth of keel plating is not to be taken
less than 2300 mm; it is taken as 2500 mm.
mm
where:
5.4 Hull Envelop Plating
Table 2. Minimum Net Thickness for Plating in

the Cargo Tank Region
| |
0.0158
12
13
Net Cargo
Pressure-Sea
Pressure
Notes:
1. The first column specifies the Design load sets
which are considered for Double bottom floors
and girders according to IACS (Section 8/Table
8.2.9).
313
2. The second column specifies the pressure load

components of the design load set, which are
defined in Table 4, according to IACS (Section
7/Table 7.6.1).
3. The third column specifies which column in the
design load combination table of IACS (Section
7/Table 7.6.1) is to be applied for each design
load set, where S denotes the static design load
combination and S+D denotes the static plus
dynamic design load combination.
4. The forth column specifies the Acceptance
Criteria Set.
5. The fifth column gives the maximum values for
each Design Load Set obtained from calculations
according to corresponding IACS formulas.
design load combinations for each load component
are given in Table 4.
5.5.1 Flange Plate Thickness

In the proposed design, the breadth of flange plate
(i.e. ) is taken as 1000 mm. From Table 3 the
greatest value (i.e. P) calculated for all applicable
design load sets is 296.40 kN/m2. Since, this greatest
value is for design load set 12, therefore acceptance
criteria set AC2 is to be considered.
Hence, for flange plate:
0.0158
1000
Load Component
S+D
5.5.2 Web Plate Thickness

In the proposed design, the breadth of web plate
(i.e. ) is taken as 1100 mm. Hence, for web plate:
0.0158
1100
The greater of
a)
b)
| |
Notes:
= design sea pressure, in kN/m2.
2.
= design tank pressure, in kN/m2.
5.
20.57 mm
mm
where:
1.
4.
Meanwhile, according to IACS (Section 8/2.5.6.5),

where the corrugated plate is built with flange and
web plate of different thickness, then the thicker net
plating thickness,
, is to be taken as the
greatest value calculated for all applicable design
load sets and is given by:
Hull
envelope
3.
18.70 mm
Therefore, web thickness of the corrugated plate is

20.57 mm.
Design Load Combination
Ballast tanks
(BWE with
flow-through
method)
Therefore, flange thickness of the corrugated plate is

18.70 mm.
Table 4. Design Load Combinations
= net thickness of the thinner plating, either

flange or web, in mm;
= static sea pressure at considered

draught, in kN/m2.
= tank testing pressure, in kN/m2.
= breadth of thicker plate, either flange or

web, in mm
P
= static tank pressure in the case of

overfilling or filling during flow through
ballast water exchange, in kN/m2.

being considered, calculated at the point in
kN/m2
= permissible bending stress coefficient:
= added overpressure due to liquid flow

through air pipe or overhead pipe, in kN/m2.
= 0.75
= 0.90
7.
= dynamic wave pressure for a

considered dynamic load case, in kN/m2.

material = 235 N/mm2.
8.
= dynamic tank pressure for

considered dynamic load case, in kN/m2.
9.
denotes the static design load combination

and S+D denotes the static plus dynamic
design load combination.
6.
10. These values are obtained

corresponding formulas of IACS.
from
the
Since, flange is the thinner plating (i.e. 18.70 mm)

therefore, for acceptance criteria AC2:
|
.
.
18.70
22.32
Therefore, web thickness of the corrugated plates is

22.32 mm.
314
Hence, the thicknesses of flange and web of

corrugated plates are taken as 18.70 mm and 22.32
mm respectively. Meanwhile, according to IACS
(Section 6/Table 6.3.1), a local corrosion additions,
, is to be considered for all structural elements
in the cargo tank region and for the members which
are in and between ballast water tanks, the value
is 3.0 mm. Therefore, the thicknesses of flange and
web of corrugated plates are to be taken as 21.70
mm and 25.32 mm respectively.
All symbols are defined in Figure 3 and the values

are obtained from Figure 2.
Figure 3. Configuration of a Single Corrugation
5.5.3 Calculating Section Modulus

minimum net section modulus,
, is to be taken
as the greatest value calculated for all applicable
design load sets is given by:
| |
cm3
25.32
Therefore:
1100
34,604 cm3

being considered and calculated at the calculation
point
= 296.40 kN/m2 [Detail in Table 3]
= bending moment factor: for continuous
stiffeners and where end connections are fitted
consistent with idealisation of the stiffener as having
as fixed ends = 12 ; (for horizontal stiffeners)
Since, corrugated plates are used as primary support

members in the double side region, in the table for
Design Load Sets for Primary Support Members
of IACS (Section 8/Table 8.2.9), the specifications
for Side transverses are considered. The
maximum values for each design load set in this
region are shown in Table 5:
Table 5. Maximum Values Obtained from
Calculations for Each Design Load Set
Design
Load
Set
[Detail in Figure 2]
= stiffeners spacing, in mm
1458 mm [Detail in Figure 2]

material, in N/mm2 = 235 N/mm2
|
26,061 cm3
Hence, the section modulus,

corrugation unit is taken as:
, [3] of a single
3
where:
= thickness of the web plate
= 25.32 mm
Load
Combi
nation
Acceptance
Criteria
Set
Value
S+D
AC2
169.64
AC1
152.84
S+D
AC2
315.59
AC1
286.44
Sea Pressures
= permissible bending stress coefficient for

the design load set being considered, to be taken as
0.9 ; (since, from Table 3.2 the
maximum design pressure is for design load set 12,
therefore AC2 is to be considered).
Design
Load
Component
2916 mm
1000
which is greater than the required section modulus.

Therefore, the proposed structure is feasible.
= effective bending span, in m
= 2
5.6 Calculations of Corrugated Plate as a

Primary Member in Double Side Region
where:
= 8.748 m
1000
2
12
13
Note: All the particulars are similar to Table 3.
5.6.1 Flange Plate Thickness

In the proposed design, the breadth of flange plate
(i.e. ) is taken as 1000 mm. From Table 5, the
greatest value (i.e. P) calculated for all applicable
design load sets is 315.59 kN/m2. Since, this greatest
value is for design load set 3, therefore acceptance
criteria set AC2 is to be considered.
Hence, for flange plate:
0.0158
; (as calculated above)
= height of corrugation
= 1000 mm
= breadth of flange plate = 1000 mm

= breadth of web plate
Cargo
Pressures
1000
|
.
19.30 mm
Therefore, flange thickness of the corrugated plate is

19.30 mm.
= 1100 mm
315
5.6.2 Web Plate Thickness

In the proposed design, the breadth of web plate
(i.e. ) is taken as 1100 mm. Hence, for web plate:
0.0158
1100
21.23 mm
Therefore, web thickness of the corrugated plate is

21.23 mm.
Same as before, flange is the thinner plating (i.e.,
19.30mm) therefore, for acceptance criteria AC2:
|
19.30
23.03
Therefore, web thickness of the corrugated plates is

23.03 mm.
Hence, the thicknesses of flange and web of
corrugated plates are taken as 19.30 mm and 23.03
mm respectively.
Meanwhile, according to IACS (Section 6/Table
6.3.1), a local corrosion additions,
, is to be
considered for all structural elements in the cargo
tank region and for the members which are in and
between ballast water tanks, the value is 3.0 mm.
Therefore, the thicknesses of flange and web of
corrugated plates are to be taken as 22.30 mm and
26.03 mm respectively.
5.6.3 Calculating Section Modulus

Similar to Section 5.5.3, the minimum net section
modulus,
, is to be taken as the greatest value
(i.e., P = 315.59 kN/m2 ) calculated for all
applicable design load sets is given by:
|
.
.
And the section modulus

unit is taken as:
1
3
26.03
1000
27,748 cm3
of a single corrugation
3
1000
1100
35,574 cm3
which is greater than the required section modulus.
Therefore, the proposed structure is feasible.
6. RESULTS
The simple target of replacing all the support
members from this region by a similar crosssectional structure, has been successful by
considering a special specification of corrugated
plates (i.e., breadth of flange = 1000mm, breadth of
web = 1100mm and corrugation angle of these
plates = 65 degrees). In case of double bottom
region, at the end of calculations, the thicknesses of
the flange and web of the corrugated plates have
been achieved 21.70mm and 25.32mm respectively.
To evaluate these thicknesses, the structural strength

assessment has also been performed by calculating
the section modulus of this structure, and the value
is 34,604cm3, which is greater than 26,061cm3 as
required by IACS for stiffeners in the same region.
Similarly, in double side region the thicknesses of
the flange and web have been achieved 22.30 mm
and 26.03 mm respectively. The section modulus of
this structure is 35,574cm3, which is greater than
27,748cm3 as required by IACS for stiffeners in the
same region.
7. CONCLUSION
The key intention of this paper has been to
propose an alternative structural design for the
double hull region of tankers and to establish that
this new design meets the criteria of classification
rules and offers better strength. In addition it offers
a less complicated construction process due to the
requirement of less welding and a more inspection
friendly region which is easier to maintain as well.
However, due to the involvement of a greater
amount of plates, this structural will be expensive to
build.
Nevertheless, the total achievement is still
incomplete, since total section modulus of the midship section and the finite element strength
assessment of the structure is yet to be done.
Therefore, it is further recommended to perform
these analyses for the proposed structure and
validate these theoretical calculations.
REFERENCES
[1] International Association of Classification
Societies (IACS). Common Structural Rules for
Double Hull Oil Tankers (2009).
[2] Mitsui O.S.K. Lines. News Radar - New
Launchings. No.207 April 2009; URL: http://
www.mol.co.jp/opensea/0904_nnldo.html
[fulltext, free access, confirmed: 12/04/2010].
[3] Rahman, M.K. Optimization of Panel Forms for
Improvement in Ship Structures, Structural
Optimization, Vol. 11, pp 195-212 (1996).
[4] Schneekluth, H. & Bertram, V., Ship Design for
Efficiency & Economy. Second Edition ISBN 0
7506 4133 9. pp 3 (1998).
[5] Sea-Japan. Japan Ship Exporters Association.
No. 333 Feb. - pp 6, Mar. (2009). URL:
http://www.jsea.or.jp/VariableE/Sea333.pdf
[fulltext, free access, confirmed: 12/04/2010].
[6] Tupper, E. C. Introduction to Naval
Architecture. Forth Edition, ISBN 0 7506 6554
8. pp 344-345 (2004).
316

DEVELOPMENT OF G-CODE FROM CAD MODEL FOR ROTATIONAL PARTS

Md. Deloyer Jahan1
1
Department of Industrial and Production Engineering,

Bangladesh University of Engineering and Technology, Bangladesh
E-mail: mdjahan.mpe@aust.edu
ABSTRACT
Throughout the life cycle of engineering products, computers have a prominent, often central role. In the
process of product design and manufacturing and its interface; process plans, this role becoming increasingly
important as competitive pressure call for improvements in product performance and quality and for reduction
in development time-scales. Computer aided process planning engineers to improve the productivity with which
they carry out their work. Feature extraction and classification is considered as the bridge between ComputerAided Design (CAD) and Computer-Aided Process Planning (CAPP). This paper proposes a method that can
extract and classify for rotational parts taking a 2D data file as input. In addition, feature interactions are also
taken into consideration in this methodology. The proposed feature extraction and classification method consists
of three basic procedures. First, polyline of desired profile for certain object is drawn in certain manner and
saved in DXF format of AutoCAD. Second, feature is extracted from the 2D CAD DXF data file. Third, G-Code
compatible for CNC machine is generated using several logics. A sample application description is presented
for demonstration purposes. The system has been implemented in Visual Studio (Visual C++) on a PC-based
system.
Key words: CAD, CAPP, CAM, DXF, Process plans, Data Extraction, Feature Recognition, Tool Path Planning,
Part Program, G-Code.
1. INTRODUCTION
Process planning translates design information into
the process steps and instructions to efficiently and
effectively manufacture products. As the design
process is supported by many computer-aided tools,
computer-aided process planning (CAPP) has evolved
to simplify and improve process planning and achieve
more effective use of manufacturing resources.
Process planning encompasses the activities and
functions to prepare a detailed set of plans and
instructions to produce a part. The planning begins
with engineering drawings, specifications, parts or
material lists and a forecast of demand. A computeraided process planning system has various modules
such as for shape recognition, costing and process
database etc. Each module may require execution
several times in order to obtain the optimum process
plan. The input to the system will most probably be a
3-D wire frame or solid model from a CAD data base
or a 2-D model. The process plan after generation and
validation can then be routed directly to the
production planning system and production control
system.
2. FEATURE BASED CAM SYSTEM

An intelligent interface between CAD and CAPP
systems is imperative because the CAPP systems
depend on correct data obtained from CAD systems to
perform precise process planning. Feature recognition

techniques provide such a connection between CAD
and CAPP. CNC part programs are derived
automatically using automatic feature recognition
from 2D drawings in this system. The CAD model of
the component including half of the 2D upper profile
to be turned has to be designed in any CAD
environment and be converted to DXF data structure
to accomplish the feature recognition process of the
system. Part programs have been derived appropriate
to the Fanuc O-T control system. This system was
prepared using Delphi version 7. In addition, the
system is supported with material and cutting tool
database prepared according to the Sandvik Coromant
catalogue [1, 9]. The system is composed of three
important modules, file reading, feature recognition
and tool path planning. The General structure of the
system is shown in Figure 1.
2.1 File Reading System

After data input into the system using DXF format
a set of processes is done automatically according to a
specific hierarchy. Firstly, the features in the drawing
are found and defined to the system. In this stage, the
coordinates of the start and end points of arc features
which is not obtained directly from DXF database are
found using trigonometric equations and all of feature
coordinates are modified from right to left and from
317
below to above. These processes provide the precise

feasibility of the sequencing process to be done in the
following stages. It may be possible that the drawing
could have been constructed anywhere in the CAD
environment according to the Cartesian coordinate
system. In that case all of the feature coordinates of
the design are modified according to the quadrant II
(x<0,y>0) of the coordinate system because of the fact
that tool paths, according to the absolute dimension
system, are formed by generally taking the face of the
component in the manufacturing of cylindrical
components. After these processes, all coordinates of
the features are sequenced according to their start
points and transferred to the origin. Finally the shape
of the work piece is drawn with the program with its
sub symmetry [2, 10].
DXF data structure

File reading system
Feature recognition system
Interchange Format (DXF) is used for its rigidly

defined structure. It is also easy to write programs to
extract data from these files. [3, 11].
3.2 Structure of the DXF file

There are six sections in a DXF file. They are:
HEADER, CLASSES, BLOCKS, TABLES, OBJECT
and ENTITIES. Among them, only ENTITIES section
contains geometric information regarding any object
drawn in the design file. So, part feature information
is extracted from the section. Geometric information
regarding each entity is stored after different DXF
numbers or codes in the ENTITIES section. In actual
DXF file all data (DXF codes and respective values)
are in a single column one after another. DXF code 0
in the first row of the first column indicates that a new
section is started. Similarly name of an entity
(POLYLINE), handle number (BC), name of the
layer (0) are after DXF code 0, 5 and 8. X, Y,
and Z coordinate values of different points are
available after DXF code 10, 20 and 30 respectively. .
The ending of the current (ENTITIES) section is
identified by SEQEND. Similarly ENDSEC is for
ending of all six section of the DXF file. [7]. Detail
structure, DXF code and their description are
available in AutoCAD online help. [4, 12].
3.3 Part Data Extraction
Tool path planning

CNC Part program
Figure 1: Structure of the system
3. DATA EXTRACTION
3.1 Data Extraction from the CAD File
It is difficult for the end user to interpret or alter the
databases used by different CAD packages.
Commercial CAD package developers designed their
CAD system for design purpose only, not for
manufacturing systems. As stated earlier, one CAD,
CAM or CAPP system cannot completely understand
a design file created by different software and for this
reason, it is not possible for a third party to develop a
CAPP system that will completely compatible with
other CAD systems i.e. understand and extract all the
information kept in a design file. To resolve the
problem some standard formats like Initial Graphics
Exchange Specification (IGES), Data Interchange
Format (DXF) are being used for transferring design
information from one CAD system to another CAD,
CAM or CAPP system. In this research work, Data
In this work, a visual C/C++ program is used to

extract the information from the DXF file. As already
mentioned in a DXF file data is stored in a single
column. So, the program opens the DXF file and
search ENTITIES section line by line until a match
is found because data regarding the part geometry is
available in this section. As the profile of the part is
drawn by a polyline, when the program reaches the
ENTITIES section it searches for POLYLINE.
After this, X and Y coordinates of each VERTEX
which represents a point in the profile are collected
following DXF code 10 and 20 respectively from the
entire ENTITIES section. To know whether the line
between two points is curved or not, the program
search DXF code 40 after DXF code 10, 20, and 30
for individual vertex. DXF code 40 indicates that line
between the current point and the next point is
curved.The program continues until it reaches
ENDSEQ, i.e. the end of the entity section. [5, 13,
16]
3.4 Vertex Coordinate Extraction

The vertex coordinate extraction algorithm
utilizes headers and flags used in DXF files to identify
coordinate values. The VERTEX header, for instance,
indicates the beginning of vertex coordinates of edges.
All X and Y coordinates are placed under specific
flags following this header. The 10 flag precedes an X
coordinate value, whereas the 20 flag precedes a Y
318
coordinate value. The flow chart of the algorithm is

given in Figure 2. As illustrated in the figure, the
coordinates that represent the part are identified and
stored into the Vertex Coordinate Array (VCA). Other
attributes, such as the X and Y coordinates of fillet
beginning, sweeping and beginning angle of the fillet
and fillet radius, are extracted and stored into the
Fillet Array (FA). Both arrays are saved for further
processing.
Figure 2: Extraction algorithm for coordinates
4. PROCESS PLANS FOR ROTATIONAL

PARTS
Computer aided process planning is done in this
study by generating G-CODE for CNC machine for
rotational parts. A sample shape is studied here
featured with fillet. First of all it is needed to
understand what G-code is. G-Code, or preparatory
code or function, are functions in the Numerical
control programming language. The G-codes are the
codes that position the tool and do the actual work.
The programming language of Numerical Control
(NC) is sometimes informally called G-code. But in
actuality, G-codes are only a part of the NCprogramming language that controls NC and CNC
machine tools. Today, the main manufacturers of CNC
control systems are GE Fanuc Automation (joint
venture of General Electric and Fanuc), Siemens,
Mitsubishi, and Heidenhain, but there still exist many
smaller and/or older controller systems.
G-codes are also called preparatory codes, and are
any word in a CNC program that begins with the letter
'G'. Generally it is a code telling the machine tool
what type of action to perform, such as:
319
4.1 Profile With Fillet:

(10,24)
(26, 24)
(28, 22)
(38, 17)
(28, 20)
(40, 15)
(31, 17)
(10, 10)
(40, 10)
Figure 3: Profile of rotational parts
For this profile first part zero is determined. Here

part zero is at point (10, 10). Then for absolute
positioning other points are determined from the
vertex coordinate system. Points found from vertex
coordinate system are (40, 10), (40, 15), (38, 17), (31,
17), (28, 20), (28, 22), (26, 24), (10, 24) From those
points least value of X coordinate and least value of Y
coordinate is determined which is found to be (10, 10)
So by subtracting the value of X and Y from all the
coordinates conversion from global to local system is
made. Profile with fillet needs to consider extra
features. For those features extra calculation and idea
is generated. [6]
ENTITIES
0
LINE
5
89
330
1F
100
AcDbEntity
8
0
100
AcDbLine
10
40.0
20
10.0
30
0.0
11
40.0
21
15.0
31
0.0
0
LINE
5
8A
330
1F
100
AcDbEntity
80
100
AcDbLine
10
40.0
20
15.0
30
0.0
11
38.0
21
17.0
31
0.0
0
LINE
5
8B
330
1F
100
AcDbEntity
8
0
100
AcDbLine
10
38.0
2017.0
30
0.0
11
31.0
21
17.0
31
0.0
0
LINE
5
8C
330
1F
100
AcDbEntity
8
0
100
AcDbLine
10
28.0
20
Steps to generate G-Code from DXF file format:

1. Drawing is saved in DXF format.
2. From large list of data ENTITIES is to be
found. [8, 17, 18]
3. From ENTITIES several features are
identified
4. Every line is start with AcDbLine and ends
with a character 0
5. Xstart is shown by 10 and Ystart is by 20
6. Xend is shown by 11 and Yend is by 21
7. Lowest value of X and Y is determined and
is set to origin of the coordinate by
subtrac1ting them from all the coordinates.
8. Generation of G-Code is executed from the
highest value of x at which the value of Y is
no equal to the value of coordinate origin value
of Y there is such one point, in this case this
value is (40, 15)
(40, 15)
(40, 10)
20.0
30
0.011
28.0
2122.0
31
0.0
0
LINE
5
8D
330
1F
100
AcDbEntity
8
0
100
AcDbLine
10
26.0
20
24.0
300.0
11
10.0
21
24.0
31
Figure 4: DXF for the profile
320
0.0
0
ARC
5
8E
330
1F100
AcDbEntity
8
0
100
AcDbCircle
10
31.0
20
20.0
30
0.0
40
3.0
100
AcDbArc
50
180.051
270.0
0
ARC
5
8F
330
1F
100
AcDbEntity
8
0
10022.0
30
0.0
40
2.0
100
AcDbArc
50
0.0
51
90.0
0
ENDSEC
0
AcDbCircle
10
26.0
20
9.
As cutting is starting from some far from

the workpiece thus for rapid positioning
cutting tool is positioned in the position
at +2 in X axis and -2 in y axis from the
first point(in this case (40, 15)).
10. For this case starting point is (42, 13).
Angle 270
Angle 270
Angle 180
Angle 180
Angle 0
Angle 90
(42, 13)
Angle90
15. So for corresponding G-code for

clockwise circular interpolation is G02
and for counterclockwise circular
interpolation is G03.
16. For this sample problem for first fillet
which is clockwise and following G-code
is obtained for the input of DXF file.
11. For G-Code representation X axis of the

coordinate system is represented as Z and
Y axis of coordinate system is as X. The
Codes used is elaborated later in the
chapter. So for initial rapid positioning Gcode is G00. The formulae used to define
is as following
(X-Xorigin) = Xstart
G02 Z18 X20 R 3

17. For this sample problem for second fillet
which is counterclockwise and following
G-code is obtained for the input of DXF
file.
Represented as Z.
(Y-Yorigin) 2 = diameter
G03 Z16 X28 R 2
Represented as X
After linear interpolation when there is

some gap program understand it as fillet
and by the descending order of X
coordinate it arrange the G-code
command.
Similarly
circular
interpolation is done for all AcDbCircle
command.
So after rapid positioning first line

become
G 00 Z 32 X 6
12. Then linear interpolation is made with
respect to descending order of X axis
coordinate in the DXF file 10, 11
commands. So next line in G-code
become
5. IMPLEMENTATION
The intended use for this G-Code operation is
to enable specialist machine builders to use for
Motion Coordination. It provides solutions for
the control of X, Y axes where the motion
program source is from a CAD/CAM program
that outputs G-Code sequences. This G-Code
implementation is not intended to replace full
functioning CNC controllers on sophisticated
metal turning machines. G-Code processing has
been implemented on Motion Coordinator system
softwares. The system allows up to 2 axes to be
controlled by a sequence of G-Code commands
that are resident in one or more of the programs
can be saved to the Motion Coordinator. This will
be implemented to aid flexibility of CNC
operation. A typical application would require a
startup program, written in BASIC and the GCode dispatcher program which is also a BASIC
language program. [19] G-Code programs can be
loaded into the memory and can be selected for
running as required.
G 01 Z 28 X 14
13. Another type of feature include in the
profile is fillet. To represent fillet circular
interpolation
is
necessary.
When
AcDbCircle is found in the DXF file it
implies there is some circular shape.
14. For circular shape 10, 20 imply
coordinate of center of the circle and 40
implies radius of the circle. One
important measure is either the circle is
clockwise or anticlockwise. 50, 51 imply
the starting angle and ending angle. If 50
and 51 range from 0 to 179 degree then
the circular shape is clockwise and if
range of 50, 51 is 180 to 359 then circular
shape is counter clockwise.
321
6. CONCLUSION
Computer-Integrated Manufacturing has gained
recognition as a most effective tool in increasing
manufacturing competitiveness. In this work, the
work provides a rigorous method for the
understanding of process planning of rotational parts
and the development of effective and efficient
Computer-Aided Process Planning systems. The main
contribution of this paper is to develop CAPP module
based on G-Code that has been conceived and realized
for the manufacturing of data transmissions in an
integrated production system. Realizing the CAPP
module for data transmissions has led to an increase
of the flexibility and correctness of the technological
planning, having favorable influences on the
manufacturing costs.
[9]
AutoCAD Databook, Prentice Hall Inc (UK)

ltd
[10]
IJCSNS International Journal of Computer

Science and Network Security, VOL.10 No.2,
February 2010 vol. 27(4), pp. 148-154.
[11]
Alam, M. R., Lee, K. S., Rahman, M. and

Zhang, Y. F., (2000), Automated process
planning for the manufacture of sliders,
Computers in Industry, vol. 43, pp. 249-262.
Aslan, E., Seker, U and Alpdemir, N., (1999),
Data extraction from CAD model for
rotational parts to be machined at turning
centres, Journal of Engineering and
Environmental Science, vol. 23, pp. 339-347.
Atkinson, A., (1991), Manufacturing parts
features: CIMs technological common
denominator, Integrated Production Systems
Design, Planning, Control and Scheduling,
fourth edition, Institute of Industrial Engineers,
Norcross, Georgia.
Cay, F. and Chassapis, C., (1997), An IT view
on perspectives of computer aided process
planning research, Computers in Industry, vol.
34, pp. 307 337.
ElMaraghy, H. A., (1993), Evolution and
future perspectives of CAPP, Annals of CIRP,
vol. 42(2), pp. 1-13
int. j. Computer Integrated Manufacturing,
1997, vol. 10, no. 1-4, 266 280
Asmundsson, J., Rardin, R.L. and Uzsoy, R.,
Tractable non-linear capacity models for
aggregate production planning. Technical
Report IN 47907 1287, Purdue University,
School of Industrial Engineering, West
Lafayette, 2003
Han, J. H., Pratt, M. and Regli, W. C., (2000),
Manufacturing feature recognition from solid
models: a status report, IEEE Transactions on
Robotics and Automation, vol. 16, no. 6,
December. [18]. J. Rossignac and A. Requicha,
"Solid modeling", in the Encyclopedia of
Electrical and Electronics Engineering, Ed. J.
Webster, John Wiley & Sons, 1999.
Hayong Shin, Classification of geometric
model, IE752 geometric modeling class in
KAIST, 2002
http://www.triomotion.com/upload/AN208%20Using%20GCODE.pdf
[12]
ACKNOWLEDGMENT
This work would not have been possible without the
support and encouragement of my thesis supervisor,
Dr. Nafis Ahmad, under whose supervision I chose
this topic and began the thesis. I would also like to
thank the department of industrial and production
engineering, BUET for supporting me during my one
year of study and research on this topic by providing
full access of their capacity. I would like to give
special thanks to Md. Shihabuddin (B.Sc. In
Computer Science and Technology, BUET) for
helping me to make the program for my work. I
cannot end without thanking my family, on whose
constant encouragement and love I have relied
throughout my time at the academy.
[13]
[14]
[15]
[16]
REFERENCES
[17]
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Gindy, N. N. Z.,A hierarchical structure for

form features, International Journal of
Production Research, Vol. 27, No. 12, pp.
2089-2103
Jaekoo, J., et. Al., Efficient feature-based
process planning for sculpture pocket
machining,
Computers
and
Industrial
Engineering,
Li, R. K and Adiga, S.,Part feature recognition
system- A vital link in the integration of CAD
and CAM, Proceedings of the IXth ICPR, pp198-205
Auto CAD online help. Version 2000, 2004,
2007.
Tr. J. of Engineering and Environmental
Science 23 (1999) , 339 { 347. (c) TUBITAK
Ahmad, N., A dynamic model of Computer
Aided Process Planning for rotational
components. M.Engg. thesis, BUET, 2001
Online help AutoCAD 2007, Autodesk Inc.
Hill, A.E. and Pilkington, R.D.,A Completed
[18]
[19]
322

EFFECTS OF TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY AND

VISCOUS DISSIPATION ON CONJUGATE FREE CONVECTION FLOW ALONG A
VERTICAL FLAT PLATE
ATM. M. Rahman1, NHM. A. Azim 2, M. M. Rahman3 and M. K. Chowdhury3
1
Department of Computer Science and Engineering,
Dhaka International University,
Email: atmmrahman2005@gmail.com
3
Department of Mathematics, Bangladesh
University of Engineering and Technology,
Email: mkamal@math.buet.ac.bd
2
School of Business studies,
Southeast University,
Email: nhmarif1@gmail.com
ABSTRACT
The effect of the temperature dependent thermal conductivity (TDTC) on conjugate free convection flow along a
vertical flat plate with viscous dissipation has been investigated. The transformed non-linear ordinary partial
differential equations are solved using the implicit finite difference method with Kellerbox scheme. Numerical
results for the velocity and temperature profiles as well as local skin friction co-efficient and surface
temperature distributions for different values of thermal conductivity variation parameter, viscous dissipation
parameter and Prandtl number are presented graphically. The values of the Prandtl number are considered to
be 0.733, 1.099, 1.63 and 2.18 that corresponds to air, water, glycerin and sulfur dioxide, respectively.
Keywords: TDTC, Conjugate free convection, vertical flat plate, viscous dissipation, finite difference
method.
1. INTRODUCTION
Free convection heat transfer and fluid flow in

enclosures with various form and wall
conductions has been studied widely in current
year due to its wide range applications, such as
building thermal design, solar energy collector
etc. Aydin [1] studied conjugate heat transfer
through a double pane window. On the other hand
the thermal interaction between laminar film
condensation forced convection along a conducting
wall was investigated by Chen and Chang [2]. The
axial heat conduction effect in a vertical flat plate on a
free convection heat transfer was studied by
Miyamoto et al. [3]. Pozzi and Lupo [4] investigated
the coupling of conduction with laminar natural
convection along a flat plate. Merkin and Pop [5]
presented conjugate free convection on a vertical
surface. In all the aforementioned analyses the effects
of temperature dependent thermal conductivity has
not been considered. But, laminar free convection
flow from an isothermal sphere immerged in a fluid
with thermal conductivity proportional to linear
function of temperature has been studied by Molla et
al [6]. Gebeart [7] has shown that the viscous
dissipation effect plays an important role in natural
convection for various devices which are subjected to
large deceleration or which operate at high rotative
speeds and also in strong gravitational field processes
on large scales (on large planets) and geological
processes. Takhar and Soundalgekar [8] have studied
dissipation effects on MHD free convection flow past

semi-infinite vertical plate. Therefore the objective of
the present work is to investigate the effect of (TDTC)
on the free convection flow along a vertical flat plate
with heat conduction and viscous dissipation.
Let us consider a steady two-dimensional natural
convection flow of viscous and incompressible fluid
along a vertical flat plate of length l and thickness b
(See Fig.1).It is assumed that the temperature at the
outside surface of the plate is maintained at a constant
temperature Tb, where Tb>T, the ambient
temperature of the fluid.
323
insulated
interface
b
Tb
u
T ( x , 0)
g
v
Fig.1: Physical model and coordinate
, at y = 0, x > 0
y
u 0, 0 as y , x > 0
The x-axis is taken along the vertical flat plate in

the upward direction and the y-axis normal to the
plate.
The governing equations of such flow under the
usual boundary layer and the Boussinesq
approximations with temperature dependent thermal
conductivity are given below:
u v
+
=0
x y
u
u
2u
u
+v
= 2 + g (T f T )
x
y
y
T f
(1)
(2)
T f
1
u
(3)
u
+v
=
( f
) +
x
y
y
c p x
c p y
Here we will consider the form of the temperature
dependent thermal conductivity, which is proposed by
Charraudeau [9]
(4)
f = [1 + (T f T )]
2
where is the thermal conductivity of the ambient

fluid and is a constant.
The appropriate boundary conditions to be
satisfied by the above equations are
u = 0,
v =0
T f
, at y = 0, x > 0 (5)
= s (T f Tb )
T f = T ( x ,0),
y
b f
y ,
x=
x
y
ul
, y = Gr , u = Gr
l
l
Gr =
1
2
,v=
vl
Gr
1
4
, =
T f T
Tb T
1
5
f +
1
20
(11)
16 + 15 x
6 + 5x
f f
f 2
20 (1 + x )
10 (1 + x )
h +
Pr
Pr
16 + 15 x
+
20 (1 + x )
(12)
x 5 2
x 5
h
h h +
1
Pr
x
+
1 + x
1
f h
f h
5 (1 + x )
(13)
where prime denotes partial differentiation with

respect to . The boundary conditions as mentioned in
equation (10) then take the following form
f ( x,0) = f ( x,0) = 0,
x (1 + x) h ( x,0) 1
h( x,0) =
1
9
1
(1 + x) 4 + x 5 (1 + x) 20h( x,0)
f ( x, ) 0, h ( x, ) 0
1
5
(8)
2
u
1
2
+ v = (1+ ) 2 + + N
x y Pr
y Pr y
y
where Pr = c p / k is the Prandtl number,
f
h
+ N x f 2 = x f
h
x
x
conjugate
f
f
+ h = x f
f
x
x
g l 3 (Tb T )
u
u 2 u
+v
=
+
x
y y 2
the
= x (1 + x) h ( x,)
here is the similarity variable and is the nondimensional stream function which satisfies the
continuity equation and is related to the velocity
components in the usual way as u = y and
v = x . Moreover h (x,) represents the
dimensionless temperature. The momentum and
energy equations are transformed for the new coordinate system. Thus we get
, (6)
where l is the length of the plate, Gr is the Grashof

number, is the non-dimensional temperature.
By applying the dimensionless quantities defined
by equation (6) in to equations (1) to (3) we get the
following non -dimensional equations
u v
(7)
+
=0
x y
= x 5 (1 + x) 20 f ( x,), = y x 5 (1 + x)
x >0
is
(10)
conduction parameter. This coupling parameter

determines the significance of the conduction
resistance within the wall. In the present investigation
we have considered p = 1.
To solve the equations (8) and (9) subject to the
boundary
conditions
(10)
the
following
transformations are then introduced
The non-dimensional governing equations and

boundary conditions can be obtained from equations
(1) - (5) using the following non-dimensional
quantities
1
4
p = (k b ) / (k s l ) Gr 1 / 4
where
T f
u 0, T f T as
u = 0, v = 0, 1 = (1+ ) p
(9)
= (Tb T ) , is the dimensionless thermal

conductivity variation parameter and
N = 2 Gr l 2 c p (Tb T ) is the dimensionless
viscous dissipation parameter.

The corresponding boundary conditions (5) take the
following form
(14)
From the process of numerical computation, in

practical point of view, it is important to calculate the
values of the surface shear stress in terms of the skin
friction coefficient. This can be written in the nondimensional form as
(15)
C f = Gr 3 / 4 l 2 / ( ) w
where w [= ( u y ) y = 0 ] is the shearing stress. Using

the new variables described in equation (6), the local
skin friction co-efficient can be written as
(16)
C f x = x 2 / 5 (1 + x) 3 / 20 f ( x,0)
324
The numerical values of the surface temperature

distribution are obtained form the relation
(17)
( x,0) = x1 / 5 (1 + x) 1 / 5 h ( x,0)
3. COMPARISON WITH PREVIOUS WORK
AND PROGRAMME VALIDATION
A comparison of the surface temperature and local
skin friction coefficient obtained in the present work
1
5
with x = and = 0 and obtained by Merkin and

Pop (1996) and Pozzi and Lupo (1988) have been
shown in Table-1 and Table-2, respectively. It is
clearly seen that there is an excellent agreement
among the respective results.
Table-1: Comparison of the present numerical results
of surface temperature with Prandtl number Pr =
0.733 and = 0
Pozzi and
Lupo
(1988)
Merkin and
Pop (1996)
Present
work
0.7
0.651
0.651
0.651
0.8
0.684
0.686
0.687
0.9
0.708
0.715
0.716
1.0
0.717
0.741
0.741
1.1
0.699
0.762
0.763
1.2
0.640
0.781
0.781
x5 =
Pozzi and
Lupo
(1988)
Merkin and
Pop (1996)
Present
work
0.7
0.430
0.430
0.424
0.8
0.530
0.530
0.529
0.9
0.635
0.635
0.635
1.0
0.741
0.745
0.744
1.1
0.829
0.859
0.860
1.2
0.817
0.972
0.975
The effect of thermal conductivity variation

parameter on the velocity and the temperature
profiles within the boundary layer with N = 0.01 and
Pr = 0.733 are shown in Fig. 2 and Fig. 3,
respectively. It is seen from Fig. 2 and Fig. 3 that the
velocity and temperature increase within the boundary
layer with the increasing value of. It means that the
velocity boundary layer and the thermal boundary
layer thickness increase for increasing values of .
Fig. 6 and Fig. 7 illustrate the velocity and

temperature profiles for different values of Prandtl
number Pr with N = 0.01 and = 0.01. From Fig. 6, it
can be observed that the velocity decreases as well as
its position moves toward the interface with the
increasing Pr. From Fig. 7, it is seen that the
temperature profiles shift downward with the
increasing values of Pr. Since the viscosity is the
resistance to flow of a fluid, so the velocity decreases
for increasing value of Pr.
Table-2: Comparison of the present numerical results

of local skin friction coefficient with Prandtl number
Pr = 0.733 and = 0
Cfx
The present work is used to analyze the effect of

thermal conductivity variation due to temperature on
free convection flow along a vertical flat plate in
presence of heat conduction and viscous dissipation.
The values of the Prandtl number are considered to be
0.733, 1.099, 1.63 and 2.18 that corresponds to air,
water, glycerin and sulfur dioxide respectively.
The effect of viscous dissipation parameter on the

velocity and temperature within the boundary layer
with = 0.01 and Pr = 0.733 are shown in Fig. 4 and
Fig. 5, respectively. It is seen that from Fig. 4 and Fig.
5 that the velocity and temperature increase within the
boundary layer with the increasing value of N.
(x,0)
x5 =
Fig. 8 and Fig. 9 illustrate the effect of the thermal

conductivity variation parameter on the local skin
friction coefficient and surface temperature
distribution against x with N = 0.01 and Pr = 0.733. It
is also seen that the local skin friction coefficient
increases for the increasing. From Fig. 8, it is seen
that the surface temperature increases for the
increasing .This is to be expected because the higher
value for the thermal conductivity variation parameter
accelerates the fluid flow and increases the
temperature as mentioned in Fig. 2 and Fig. 3,
respectively.
The effect of viscous dissipation parameter on the
local skin friction coefficient and surface temperature
distribution against x with = 0.01 and Pr = 0.733 are
shown in Fig. 10 and Fig. 11, respectively. It is
observed from Fig. 10 that the local skin friction
coefficient increases for the increasing values of N.
From Fig. 11 it can be seen that the surface
temperature distribution increase for increasing values
of N.
325
0.8
0.8
= 0.01
= 0.24
= 0.48
= 0.72
0.5
f'
f'
0.5
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
0.3
0.0
0.0
0.3
2.0
4.0
0.0
0.0
6.0
Fig. 2: Variation of velocity profiles for different

values of with N = 0.01 and Pr = 0.733
2.0
4.0
6.0

values of Pr with N = 0.01 and = 0.01
2.0
= 0.01
= 0.24
= 0.48
= 0.72
1.0
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
1.5
1.0
1.5
0.5
0.5
0.0
0.0
1.5
3.0
4.5
6.0
0.0
0.0
Fig. 3: Variation of temperature profiles for different

1.5
3.0
4.5
6.0

0.8
= 0.01
= 0.10
= 0.20
= 0.30
fx
3.0
f'
0.5
4.0
2.0
= 0.01
= 0.24
= 0.48
= 0.72
0.3
1.0
0.0
0.0
2.0
4.0
6.0
0.0
0.0
9.0
12.0
15.0
Fig.8: Variation of skin friction for different values of

with N = 0.01 and Pr = 0.733
2.0
3.0
= 0.01
= 0.10
= 0.20
= 0.30
1.0
2.0
(x,0)
1.5
6.0

values of N with = 0.01 and Pr = 0.733
= 0.01
= 0.24
= 0.48
= 0.72
1.0
0.5
0.0
0.0
3.0
1.5
3.0
4.5
6.0
0.0
0.0
3.0
6.0
9.0
12.0
15.0

Fig. 9: Variation of surface temperature for different

326
8.0
= 0.01
= 0.10
= 0.20
= 0.30
fx
6.0
4.0
2.0
0.0
0.0
3.0
6.0
9.0
12.0
15.0
Fig. 10: Variation of skin friction for different values

of N with = 0.01 and Pr = 0.733
= 0.01
= 0.10
= 0.20
= 0.30
(x,0)
6.0
4.0
Fig. 12 and Fig. 13 deal with the effect of Prandtl

number on the local skin friction coefficient and
surface temperature distribution against x with N =
0.01 and = 0.01. It can be observed from Fig. 12 that
the local skin friction coefficient decreases for the
increasing Pr. From Fig. 13. It can be seen that the
surface temperature distribution decreases for the
increasing values of Pr. Since the skin friction
decreases for increasing value of Pr as a result the
temperature distribution within the boundary layer
decreases for increasing Pr and also surface
temperature decreases for increasing Pr.
5. CONCLUSIONS
In this paper the effect of thermal conductivity

variation due to temperature and conduction on free
convection flow along a vertical flat plate with
viscous dissipation has been studied. From the present
numerical investigation the following conclusions
may be drawn:
2.0
0.0
0.0
3.0
6.0
9.0
12.0
15.0

2.5
1.5
fx
2.0
0.5
0.0
0.0
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
1.0
3.0
6.0
9.0
12.0
15.0
The velocity within the boundary layer

increases for decreasing values of Prandtl
number and increasing values of the thermal
conductivity variation and viscous dissipation
parameters.
The temperature within the boundary layer
increases for the increasing values of thermal
conductivity variation parameter and viscous
dissipation parameter and decreasing values of
the Prandtl number.
The local skin friction co-efficient decreases for
the increasing values of Prandtl number and
decreasing values of the thermal conductivity
variation and viscous dissipation parameters.
An increase in the values of the thermal
conductivity variation parameter, viscous
dissipation parameter leads to an increase in the
surface temperature and decreases for the
increasing values of Prandtl number.
NOMENCLATURE
Fig. 12: Variation of skin friction for different values

of Pr with N = 0.01 and = 0.01
(x,0)
1.5
1.0
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
0.5
0.0
0.0
3.0
6.0
9.0
12.0
15.0
b
Cfx
Cp
f
g
Gr
l
N
P
Pr
T
Tb
Tf
T
u
v

327
Plate thickness
Local skin friction coefficient
Specific heat at constant pressure
Acceleration due to gravity
Grashof number
Length of the plate
Dimensionless viscous dissipation parameter
Conjugate conduction parameter
Prandtl number
Temperature of the interface
Temperature at outside surface of the plate
Temperature of the fluid
Temperature of the ambient fluid
Velocity component in x- direction
Velocity component in y- direction
u
v
x
y
x
y
Dimensionless velocity component in xdirection

Dimensionless velocity component in ydirection
Cartesian co-ordinate along the vertical flat plat
in the upward direction
Cartesian co-ordinate normal to the plate
Dimensionless Cartesian co-ordinate along the
vertical flat plat in the upward direction
Dimensionless Cartesian co-ordinate normal
to the plate
Greek Symbols
Co-efficient of thermal expansion

Thermal conductivity variation parameter
Vector differential operator
Similarity variable
Dimensionless temperature
(x,0) Surface temperature distribution
Thermal conductivity of the ambient fluid

s
Thermal conductivity of the solid
Thermal conductivity of the fluid

Viscosity of the fluid
Kinematic viscosity
Density of the fluid inside the boundary layer
Dimensionless shearing stress
REFERENCES
[1] Aydin O., Conjugate heat transfer analysis of

double pane window. Build, Environ, Vol. 40,
pp.109-116, 2006.
328
[2] Chen H. T and Chang S. M., The thermal

interaction between laminar film condensation
and forced convection along a conducting wall,
Acta Mechanica, Vol.118, pp.13-26, 1996.
[3] Miyamoto M., Sumikawa J., Akiyoshi T.,and
Nakamura.T., Effect of axial heat conduction in
a vertical flat plate on free convection heat
transfer, Int. J. Heat Mass Trans., Vol.23,
pp.1545-1553, 1980.
[4] Pozzi A. and Lupo M., The coupling of
conduction with laminar natural convection
along a flat plate, Int. J. Heat Mass Trans.,
Vol.31, No. 9, pp.1807-1814, 1988.
[5] Merkin J. H and Pop I., Conjugate free
convection on a vertical surface, Int. J. Heat
Mass Trans., Vol.39, pp.1527-1534, 1996.
[6] Md. Mamun Molla , Azad Rahman and Lineeya
Tanzin Rahman , Natural convection flow from
an isothermal sphere with temperature
dependent thermal conductivity. Journal of
Vol.2, pp.53-64, 2005.
[7] Gebhart. B., Effect of Pressure stress work and
viscous dissipation in some natural convection
flows, Int. J. Heat Mass Trans., Vol. 24, No. 10,
pp. 1577-1588, 1981.
[8] Takhar H. S and Soundalgekar V. M.,
Dissipation effects on MHD free convection
flow past a semi-infinite vertical plate. Appl.
Sci.Res.Vol.36, pp 163-171, 1980.
[9] Charraudeau J., Influence de gradients de
properties physiques en convection force
application au cas du tube. Int. J. Heat Mass
Trans., Vol.18, pp.87-95, 1975.

EFFECT OF TEMPERATURE DEPENDENT VISCOSITY INVERSELY

PROPORTIONAL TO LINEAR FUNCTION OF TEMPERATURE ON
MAGNETOHYDRODYNAMIC NATURAL CONVECTION FLOW ALONG A
VERTICAL WAVY SURFACE
Nazma Parveen and Md. Abdul Alim
Department of Mathematics, Bangladesh University of Engineering and Technology,
E-mail: nazma@math.buet.ac.bd
maalim@math.buet.ac.bd
ABSTRACT
In this paper, the effect of temperature dependent variable viscosity inversely proportional to linear function
of temperature on Magnetohydrodynamic (MHD) natural convection flow of viscous incompressible fluid along
a uniformly heated vertical wavy surface has been investigated. The governing boundary layer equations with
associated boundary conditions for phenomenon are converted to non-dimensional form using suitable
transformations. The resulting nonlinear system of partial differential equations are mapped into the domain of
flat vertical plate and then solved numerically employing the implicit finite difference method, known as Kellerbox scheme. The solutions are obtained in terms of the skin friction coefficient, the rate of heat transfer, the
streamlines and the isotherms over the whole boundary layer and are shown graphically for the effects of the
pertinent parameters, such as the viscosity parameter () and the magnetic parameter (M) for Prandtl number
Pr = 0.73 and the amplitude of the wavy surface = 0.3.
Key words: Magnetohydrodynamics, temperature dependent viscosity, natural convection, uniform surface
temperature, Keller-Box method and wavy surface.
1. INTRODUCTION
It is necessary to study the heat transfer from an
irregular surface because irregular surfaces are often
present in many applications. It is often encountered
in heat transfer devices to enhance heat transfer.
Laminar natural convection flow from irregular
surfaces can be used for transferring heat in several
heat transfer devices, for examples, flat-plate solar
collectors and flat-plate condensers in refrigerators
and heat exchanger. One common example of a heat
exchanger is the radiator used in car, in which the heat
generated from engine transferred to air flowing
through the radiator. Alam et al. [1] have studied the
problem of free convection from a wavy vertical
surface in presence of a transverse magnetic field. The
viscosity of the fluid to be proportional to a linear
function of temperature, two semi-empirical formulae
were proposed by Charraudeau [2]. The effect of
temperature dependent viscosity on natural
convection heat transfer from a horizontal isothermal
cylinder of elliptic cross section have been studied by
Cheng [3]. Hossain et al. [4] investigated the natural
convection flow past a permeable wedge for the fluid
having temperature dependent viscosity and thermal
conductivity. Molla et al. [6] studied natural
convection flow along a vertical wavy surface with

uniform surface temperature in presence of heat
generation/absorption. Natural convection flow from
an isothermal horizontal circular cylinder with
temperature dependent viscosity was also investigated
by Molla et al. [7]. Hossain et al. [8] investigated
natural convection of a viscous fluid with viscosity
inversely proportional to linear function of
temperature from a vertical wavy cone. Numerical
study on a vertical plate with variable viscosity and
thermal conductivity has been investigated by Palani
and Kim [9]. Nasrin et al. [10] investigated MHD free
thermal conductivity and viscosity depending on
temperature. The natural convection heat transfer
from an isothermal vertical wavy surface was first
studied by Yao [11, 12] and using an extended
Prandtls transposition theorem and a finite-difference
scheme. He proposed a simple transformation to study
the natural convection heat transfer from isothermal
vertical wavy surface. Yet the preceding literature
survey shows that in natural convection heat transfer
the variation of viscosity with temperature and
magnetic field along a vertical wavy surface has not
been well investigated.
329
The present study is to incorporate the idea on the

effects of temperature dependent viscosity inversely
proportional to linear function of temperature in
presence of strong magnetic field of electrically
conducting fluid with free convection along a vertical
wavy surface. However, it is known that viscosity
may change significantly with temperature. The
governing partial differential equations are reduced to
locally non-similar partial differential forms by
adopting some appropriate transformations. The
transformed boundary layer equations are solved
numerically using implicit finite difference scheme
together with the Keller box technique [5]. We have
focused our attention on the surface shear stress in
terms of local skin friction and the rate of heat transfer
in terms of local Nusselt number, the stream lines and
the isotherms for selected values of parameters
consisting of the magnetic parameter M and the
viscosity variation parameter .
The boundary layer analysis outlined below allows
( X ) being arbitrary, but our detailed numerical

work will assume that the surface exhibits sinusoidal
deformations. The wavy surface may be described by
(1)
where L is the characteristic length associated with

the wavy surface.
The geometry of the wavy surface and the twodimensional Cartesian coordinate system are shown in
Figure 1.
Momentum Equations
1 P 1
V
V
+ .(U )
+V
=
X
Y
X
2
+ g (T T ) 0 0 U
1 P 1
V
V
+V
=
+ .(V )
X
Y
Y
(3)
(4)
Energy Equation
T
T
k
+V
=
2T
X
Y C p
(5)
where (X, Y) are the dimensional coordinates along

and normal to the tangent of the surface and (U, V) are
the velocity components parallel to (X, Y), k(T) is the
thermal conductivity and (T) is the dynamic
viscosity of the fluid in the boundary layer region
depending on the fluid temperature.
The boundary conditions for the present problem
are
U = 0, V = 0, T = Tw
at Y = Yw = ( X )
U = 0, T = T , P = p
(2)
as Y
(6)
(7)
There are very few forms of viscosity variation

available in the literature. Among them we have
considered that one which is appropriate for liquid
introduced by Hossain et al. [8] as follows:
U V
+
=0
X Y
n X
Y w = ( X ) = sin
Under the usual Boussinesq approximation, we

consider the flow governed by the following boundary
layer equations:
Continuity Equation
1 + (T T )
*
(8)
where is the viscosity of the ambient fluid and * is

a constant evaluated at the film temperature of the
flow T f = 1 2 (Tw + T ) .
Following Yao [11], we now introduce the

following non-dimensional variables
Tw
x=
Y
u=
Figure 1. Physical model and coordinate system
330
1
X
Y
L2
, y =
Gr 4 , p = 2 Gr 1P
L
L
L 12
L 14
Gr U , v =
Gr (V xU ) ,
g (Tw T ) 3
d d ,
T T ,
Gr =
L
x =
=
dX
dx
2
T w T
Introducing the above dimensionless variables into

Equations (2)(5), the following dimensionless form
of the governing equations are obtained:
u v
+
=0
x y
u
(1+ x2 ) u
Mu +
(1+ )2 y y
1 p
u u
(1+ x2 ) 2u
+ v = Gr 4 + x
y
y (1 + ) y2
x
1 x
(1 + x2 )
3
1
f + ff + x xx2 f 2 +
(1 + )
4
2
1
1
+
+
x2
x
(10)
x (1+ x2 ) u
xxu2
(1+ )2 y y
1
2
+v
= (1 + x ) 2
u
x
y Pr
y
f ( x, o) = f ( x, o) = 0,
( x, o) = 1
f ( x, ) = 0,
( x, ) = 0
(12)
Equation (11) indicates that the pressure gradient

along the y-direction is O (Gr 4 ) , which implies that
lowest order pressure gradient along x-direction can
be determined from the inviscid flow solution.
1
Equation (11) further shows that Gr 4 p / y is

O(1) and is determined by the left-hand side of this
equation. Thus, the elimination of p / y from
equations (10) and (11) leads to
u
u (1 + x2 ) 2u x xx 2
+ v =
u
x
y (1 + ) y2 1 + x2
(1+ x2 ) u
M
1
u+
2
2
(1 + ) y y 1 + x
1+ x2
(13)
at y = 0
as y
(17)
(18)
The rate of heat transfer in terms of the local

Nusselt number, Nux and the local skin friction
coefficient, Cfx take the following forms:
Nu x (Gr / x)
C fx (Gr / x )
= 1 + x2 ( x, o)
/2 =
1 + x2
(1 + )
f ( x, o )
(19)
(20)
This paper concerns the natural convection flow of
viscous incompressible fluid along a uniformly heated
vertical wavy surface in presence of strong magnetic
field and variable viscosity inversely proportional to
linear function of temperature along a vertical wavy
surface using the very efficient implicit finite
difference method known as Keller box scheme
developed by Keller [5]. This method has been
extensively used recently by Hossain et al. [1, 4, 6, 7
and 8].
The corresponding boundary conditions for the

present problems then turn into
u = v = 0, = 1
u = = 0, p = 0
(1 + )
f
Mx
f
f = x f f
f
x
1 + x2
(1 + )
x
The boundary conditions (14) now take the

following form:
(11)
In the above equations, Pr, and M are, respectively

known as the Prandtl number, the dimensionless
viscosity parameter and dimensionless Magnetic
parameter, which are defined as
C p
2 L2
, = * (Tw -T) and M = 0 0
Pr =
1
k
Gr 2
(16)
2
x
2
1
3
f

(1 + x2 ) + f = x f

Pr
4
x
x
x u
(15)
where is the pseudo similarity variable and is the

stream function that satisfies the equation (9).
Introducing the transformations given in equation
(15) into equations (13) and (12) the momentum and
energy equations take the following forms,
(9)
1
u u
p
p (1+ x2 ) 2u
+ v = + Gr 4 x +
x y
x
y (1+ ) y2
= x 4 f ( x,), = yx 4 , = ( x,)
(14)
Now we introduce the following transformations to

reduce the governing equations to a convenient form:
The governing equations (16) and (17) with the

boundary conditions in equation (18) are solved
numerically employing a very efficient implicit finite
difference method together with Keller-box scheme.
Numerical values of the shear stress in terms of the
skin friction coefficients Cfx, the rate of heat transfer
in terms of the Nusselt number Nux, the streamlines
and the isotherms are presented graphically for
different values of the viscosity parameter = 0.0
(constant viscosity) to 1.0, the magnetic parameter M
331
(a)
= 0.0
= 0.5
= 1.0
0.8
Cfx
0.6
0.4
0.2
0
0
(b)
Nux
0.4
Pr = 0.73 , M = 0.5 , = 0.3
= 0.0
= 0.5
= 1.0
0.2
Pr = 0.73 , M = 0.5 , = 0.3

0
0
Figure 2. Variation of (a) skin-friction

coefficient (b) rate of heat transfer against x for
varying of with M = 0.5, = 0.3 and Pr = 0.73.
M = 0.0
M = 0.5
M = 1.0
M = 1.5
M = 2.0
(a)
1.2
1
Cfx
0.8
0.6
0.4
0.2
Pr = 0.73 , = 0.3 , = 0.5
0
0
0.4
(b)
Pr = 0.73 , = 0.3 , = 0.5
Nux
= 0.0 (non magnetic field) to 2.0 while Prandtl

number Pr = 0.73 and the amplitude of the wavy
surface = 0.3.
The influence of the parameter = (0.0, 0.5 and
1.0) on the surface shear stress in terms of the local
skin friction Cfx and the rate of heat transfer in terms
of the local Nusselt number Nux are illustrated
graphically in Figures 2(a) and 2(b) respectively when
the values of amplitude of wavy surface = 0.3,
magnetic parameter M = 0.5 and Prandtl number Pr =
0.73. Figure 2(a) indicates that increasing the values
of the viscosity-variation parameter , the skin friction
coefficient decreases slowly along the downward
direction of the plate. On the other hand it can be
shown from Figure 2(b) that the value of the rate of
heat transfer along the wavy surface increases for
increasing the values of the viscosity-variation
parameter . Here we conclude that for high viscous
fluid inversely proportional function of temperature
the skin-friction coefficient is slow and the
corresponding rate of heat transfer is large. In Figure
2(a), the maximum values of local skin friction
coefficient Cfx are 0.86640, 0.78213 and 0.77294 for
= 0.0, 0.5 and 1.0 respectively which occur at
different values of x and it is seen that the local skin
friction coefficient Cfx decreases by 10.79% as
increases from 0.0 to 1.0. Again Figure 2(b) shows
that the rate of heat transfer increases 7.82% due to
the increased value of . Increasing values of lead to
increase the amplitude of the Nux.
0.2
= 0.0
= 0.5
= 1.0
= 1.5
= 2.0
0
0
Figure 3. Variation of (a) skin-friction

coefficient (b) rate of heat transfer against x for
varying of M with = 0.5, = 0.3 and Pr = 0.73.
The effect of the magnetic parameter M = (0.0, 0.5,
1.0, 1.5, 2.0) on the surface shear stress in terms of the
local skin friction Cfx and the rate of heat transfer in
terms of the local Nusselt number Nux are depicted
graphically in Figures 3(a) and 3(b) respectively when
the values of amplitude of wavy surface = 0.3,
viscosity parameter = 0.5, and Prandtl number Pr =
0.73. As electrically conducting fluid in presence of
magnetic field generates electrical current, the
magnetic field is changed and the fluid motion is
moderated. As a result, the velocity gradient f"(x, 0)
decreases with the increase of the magnetic parameter
M. The same result is observed on the local rate of
heat transfer Nux due to that an increase in the
magnetic parameter M at different position of x. The
skin friction coefficient and the rate of heat transfer
coefficient decrease by 33.55% and 11.71%
respectively as M increases from 0.0 to 2.0.
Figures 4 and 5 illustrate the effect of the
temperature dependent viscosity variation parameter
on the development of streamlines and isotherms
respectively which are plotted for Prandtl number Pr =
0.73, amplitude of wavy surface = 0.3 and M = 0.5.
We find that for = 0.0 the value of max = 6.46, for
= 0.5 the value of

value of
max = 6.68 and for = 1.0 the
max = 6.78. From Figure 4, it is seen
that
the effect of viscosity parameter , the flow rate in the
332
(a)
15
6
0.0
0.13
0.0
0.31
0.69
0.94
00
20
0.06
00
0.1
6
10
(c)
0.0
7
15
10
0.43
0.07
0.20
0.13
0.4
7
0
0.4
0.93
15
1.87
00
1.69
20
10
(b)
10
(c)
2.19
3.13
3.75
0.3
6.78
6.43
00
20
2.44
5.57
5.14
2.65
10
(c)
10
2.24
1.63
1.22
1
0.8
04
2.
0.43
4
10
3.05
15
2.85
1.29
2.14
4.71
4.29
3.00
3.43
6.00
15
10
0.94
20
1.56
10
4.69
2.81
4.38
15
4.07
0.42
00
4.37
0.62
6.68 5
6.3
5.92
5.50
4.66
5.08
3 .8
3.39
2.54
10
5.62
5.00
10
(b)
3.12
10
9.37
6.25
8.74
8.12
9
7.4
6.87
15
20
(a)
2
20
00
0 .7 1
20
00
0.24
0.35
0.
24
0.3
10
8
10
x 6
Figure 5. Isotherms for (a) = 0.0, (b) = 0.5
and (c) = 1.0while M = 0.5, = 0.3 and Pr =
0.73.
6.46
6.03
5.60
5.17
4.74
4.31
3.88
3.45
2.58
1.72
1.29
10
6
0.0
10
00
15
15
0.4
(a)
(b)
20
0.25
10
5
are 9.37, 4.69 and 3.05 for magnetic parameter M =

0.0, 1.0 and 2.0 respectively. Finally we conclude that
for much roughness of the surface with the effect of
magnetic parameter the velocity of the fluid flow
decreases in the boundary layer.
20
0.06
boundary layer increases. From Figure 5, we observe

that owing to the effect of , the thermal state of the
fluid decreases causing the thermal boundary layer
decrease.
The effect of variation of the surface roughness on
the streamlines and isotherms for the values of M
equal to 0.0, 1.0 and 2.0 are depicted by the Figure 6
and Figure 7 respectively while Prandtl number Pr =
0.73, amplitude of wavy surface = 0.3 and viscosity
variation parameter = 0.5. We observe from Figure 7
that as the values of magnetic parameter M increases
the thermal boundary layer thickness becomes higher
gradually. Figure 6 depicts that the maximum values
of decreases steadily while the values of M
increases. The maximum values of , that is, max
20
0.
Figure 4. Streamlines for (a) = 0.0, (b) = 0.5

and (c) = 1.0 while the magnetic parameter M
= 0.5, amplitude of wavy surface = 0.3 and
Prandtl number Pr = 0.73.
00
2
4
8
10
x 6
Figure 6. Streamlines for (a) M = 0.0, (b) M =
1.0 and (c) M = 2.0 while = 0.5, = 0.3 and Pr
= 0.73.
333

20
REFERENCES
(a)
0.1
00
0.
0.25
0.1
0.7
4
0.0
10
15
19
0. 6
10
20
(b)
0.4
1
0.3
0.13
5
00
0.19
0.06
10
0.1
15
0.25
0.75
0.81
8
10
20
0.
06
(c)
00
0.3
0.25
38
0.
0.1
10
0 .3
15
0.6
4
0 .7
5
10
Figure 7. Isotherms for (a) M = 0.0, (b) M = 1.0

and (c) M = 2.0 while = 0.5, = 0.3 and Pr =
0.73.
5. CONCLUSION
The effect of temperature dependent variable
viscosity inversely proportional to linear function of
temperature on Magnetohydrodynamic (MHD)
natural convection flow of viscous incompressible
fluid along a uniformly heated vertical wavy surface
has been investigated. From the present investigation
the following conclusions may be drawn:
The effect of increasing viscosity parameter

results in decreasing the local skin friction
coefficient Cfx and increasing the local rate of
heat transfer Nux.
An increase in the values of M leads to decrease

the skin friction coefficient Cfx and the local rate
of heat transfer Nux.
The flow rate decreases and the thermal

boundary layer increases when the effect of
magnetic field is considered.
The streamlines increase and the thermal

boundary layer decrease when viscosity
parameter increases.
[1] Alam, K. C. A., Hossain, M. A. and Rees, D. A.

S., Magnetohydrodynamic free convection
along a vertical wavy surface, Applied
Mechanics and Engineering, Vol. 1, pp. 555566
(1996).
[2] Charraudeau, J., Influence de gradients de
properties physiques en convection force
application au cas du tube, International
Journal of Heat and Mass Transfer, Vol. 18, pp.
87-95 (1975).
[3] Cheng, C. Y., The effect of temperature
dependent viscosity on natural convection heat
transfer from a horizontal isothermal cylinder of
elliptic
cross
section,
International
Communications of Heat and Mass Transfer,
Vol. 33, pp. 10211028 (2006).
[4] Hossain, M. A., Munir, M. S. and Rees, D. A. S.,
Flow of viscous incompressible fluid with
temperature dependent viscosity and thermal
conductivity past a permeable wedge with
uniform surface heat flux, International
Journal of Thermal Science, Vol. 39, pp. 635644 (2000).
[5] Keller, H. B., Numerical methods in boundary
layer theory, Ann. Rev. Fluid Mech. Vol. 10, pp.
417433 (1978).
[6] Molla, M. M., Hossain, M. A. and Yao, L. S.,
Natural convection flow along a vertical wavy
surface with uniform surface temperature in
presence of heat generation/absorption,
International Journal of Thermal Science, Vol.
43, pp. 157163 (2004).
[7] Molla, M. M., Hossain, M. A. and Gorla, R. S.
R., Natural convection flow from an isothermal
horizontal circular cylinder with temperature
dependent viscosity, Heat and Mass Transfer,
Vol. 41, pp. 594-598 (2005).
[8] Hossain, M. A., Munir, M. S. and Pop, I.,
Natural convection of a viscous fluid with
viscosity inversely proportional to linear
function of temperature from a vertical wavy
cone, Int. J. Therm. Sci. Vol. 40, pp. 366371
(2001).
[9] Palani, G. and Kim, K.-Y., Numerical study on
a vertical plate with variable viscosity and
thermal conductivity, Arch Appl. Mech., Vol.
80, pp. 711-725 (2010).
[10] Nasrin, R. and Alim, M. A., MHD free
thermal conductivity and viscosity depending on
temperature, Journal of Naval Architecture and
Marine Engineering, Vol. 6, No. 2, pp. 72-83
(2009).
[11] Yao, L. S., Natural convection along a vertical
wavy surface, ASME Journal of Heat Transfer,
Vol. 105, pp. 465468 (1983).
[12] Yao, L. S., A note on Prandtls transposition
theorem, ASME J. of Heat Transfer, Vol. 110,
pp. 503507 (1988).
334

DEVELOPMENT OF AN INTELLIGENT SYSTEM FOR MONITORING THE

SHIP ENGINE ROOM USING MOBILE PHONE
BadrunNahar1, Md. Liakot Ali2, and Md. Sharfaraj Hossain3
1
Institute of Information and
Communication Technology
Bangladesh University of Engineering
and Technology
Dhaka-1000
Email: tanjina84@gmail.com
2
Technology
Dhaka-1000
E-mail: liakot@iict.buet.ac.bd
2
Bangladesh University of
Engineering and Technology
Dhaka-1000
ABSTRACT
In todays world, ensuring security for important places as well as hazardous environment is a burning issue.
Different surveillance methodologies such as alarm system, CCTV, PC based video system are developed to
ensure this security. But using all these systems, it is not possible for a person to monitor the security of his or
her desired location when they are outside. Now-a-days anybody can communicate with anyone at any time
around the globe with the help of mobile phone technology. By keeping the technological facility of mobile
phone in mind, a mobile phone based intelligent surveillance system has been described in this paper. This
paper will give a solution for the security of the ships engine room by accessing from the mobile phone when
the ship is within the range of radio network. In this system there are server and client (captain or engine
operator) end. Server will store images being captured by the webcam of the engine room .Then based on
clients request sending from their mobile phone to the server, clients will be able to view the images from their
mobile phone that is stored in the server. Clients are also able to move the webcam by sending control
instruction from their mobile phone to view the images of their desired position. The developed system has been
tested first using the GUI (emulator) designed by NetBeans IDE. It has also been tested using different mobile
phones to see the images in real time. This surveillance system can also be implemented for other handheld
devices like PDA.
Keywords: Java 2 micro edition (J2ME),Java Media Framework (JMF),NetBeans IDE, Java programming for
server and client software, usage of mobile phone as GUI.
1. INTRODUCTION
Security is a prime concern in our day-today life.
Everyone wants to be as much secure as possible. A
security guard can be a physical solution but if the
specific location like ships engine room, office or
parking area can be seen and monitored from a
remote place, it will be more secured. That is why
different monitoring systems like alarm system,
CCTV, PC based video system etc. have been
proposed [1-4]. But it's not always feasible to be
physically near to the system. So, to be in touch with
this sort of important systems by not being physically
close, we need some sort of remote solution. Todays
communication world ensures that anybody can
communicate to anyone anywhere anytime across the
globe with the help of mobile phone technology [5].
Various type of research works have been conducted
335
in different time by different authors such as a video

surveillance system based on Multimedia Messaging
Service (MMS) [6-7]. This system is desired to
monitor an area continuously and to capture the
hazardous momentary event and to send this video
image to the user as an MMS or as an e-mail
immediately. But the problem with this system is that
as the mobile operators do not encourage sending
multimedia message to mobile phone from e-mail in
fear of spam, the concept could not be verified with
the mobile phones. Another research work is
controlling remote system using mobile telephony
[13-14] which introduces the mechanism of the
mobile phone so that the ordinary services of the
mobile phones can be leveraged to communicate with
and control the remote systems. After reviewing the
past researches and exploiting the communication
facility of the mobile phone, a mobile phone based
Proceedinggs ofMARTEC 20110
surveillancce system haas been deveeloped in this

research foor giving a sollution about th
he insecurity of
ships engiine room whilee the captain iss far away from
m
the enginee room[13]. Using this system
s
captainn
(client) or engine operattor will be abble to watch or
m from anywhhere of the shipp.
monitor their engine room
Although different typees of cameraa with built-inn
f moving it in
i any directioon are availablee
hardware for
in the marrket [15], these are very exppensive. In thee
proposed system a hardware circuit using stepper
motor and
d other accesssories are developed for thee
movement of the webccam to providde a low cosst
solution. The
T proposed system will prrovide the truee
sense of reeal mobility an
nd security [166] by accessingg
the desiredd place from th
he mobile phonne. The aim of
this researrch is to deveelop a mobilee phone basedd
surveillancce system throu
ugh which enggine operator or
captain cann get snaps off their engine room anytimee
from anyw
where. To acchieve the abo
ove goal, this
research haas following ob
bjectives:
To develop a Java
J
Program for capturingg
image sequences of the importaant location.
To develop an interface throough which a
municate with the server andd
perrson can comm
acccess the captuured images to
t monitor thee
dessired location at
a any time.
To develop a hardware ciircuit for thee
moovement of the webcam.
Next sectiions of this paper

p
describee the hardwaree
design thatt is designed too move the weebcam based onn
clients insstruction. Then
n next portionn elaborates thee
system arcchitecture and
d its working procedure andd
then the experimental
e
reesults getting from emulator
(GUI) and real mobile phhone.
Parallel port
p is a simplee and inexpenssive medium for
f
building computer conttrolled system
m. The simpliciity
and ease of programminng makes paraallel port popullar
onic hobbyist world.
w
The paraallel port is ofteen
in electro
used in computer coontrolled robots, Atmel/PIIC
mers; home auttomations etc. The primary use
u
programm
of paralleel port is to connnect printers to computers annd
is specifically designedd for this purpoose. The pins in
DB25 co
onnector are diivided into thrree groups, theey
are 1) Daata pins 2) Conntrol pins) 3) Sttatus pins. In thhe
proposed system paarallel port is used for
f
communiicating betweenn PC and hardw
ware by sendinng
controllinng bits throughh data pins 2 to 5 as shown in
Figure 1aand 3.ULN20003 is a 7-bit 500V 500mA TTL
Linput NP
PN transistor. In this researcch it is used to
drive the stepper motorr. A stepper motor
m
is basicallly
an electro
omechanical deevice which coonverts electriccal
pulses innto discrete m
mechanical movements.
m
Thhe
rotation of
o the motor has
h direct relaationships to thhe
applied input
i
pulses. In this propoosed system thhe
pulses wiill be applied frrom the paralleel port.
3. SYST
TEM ARCH
HITECTURE
E
3.1 Opeerational Procedure of the System
S
The prooposed work consists of th
hree modules 1)
image caapture module 2) parallell port interface
module 3) surveillancce software module. Imagge
m
is respoonsible for cappturing an imagge
capture module
of the monitored
m
areaa and storing it in a speciffic
location of
o the server. Parallel port interface moduule
is responnsible for transsferring the co
ontrol bits to thhe
stepper motor
m
in order tto move the weebcam on cliennts
command
d. The softwarre part sends thhe images to thhe
concernedd person.
2. HARD
DWARE DESIGN
The ideaa of the hardw
ware design is to move thee
webcam based
b
on cliennts command sending from
m
their mobiile phone. Thee working invo
olves a paralleel
port (DB225), a main IC
C ULN2003 and
a a unipolaar
stepper motor.
m
The schematic
s
diaagram of thee
hardware circuit
c
is shown
n in Figure 1.
Figure 2. System architeecture.

h
the systeem
Figure 2 shown abovve describes how
works to monitor? In this research, there are clieent
and serveer end. A webccam will be connnected with thhe
server annd it will continnuously capturre images of thhe
desired lo
ocation and saave it to a specific location of
the serverr. On the otherr hand client will
w start fetchinng
images frrom the serverr by login the URL from theeir
mobile phone.
p
That iss why at firstt client need to
Figurre 1. Schematiic diagram of thhe hardware

ciircuit.
336
establish a connection with the serv

ver. Once thee
connectionn is establisheed client will start fetchingg
images fro
om the specifiic location of the local hosst
where imaages are being saved that is captured
c
by thee
webcam using HTTP prrotocol. Figuree 3 shows how
w
nstruction willl be transmittted from thee
control in
parallel poort to the ULN22003 on clientts demand andd
then it driives the steppeer motor whicch in turn willl
move the webcam.
w
Figure 4. Flow chart of the overall procedure.

Figuree 3. Hardware connection at server side.
3.2 How to
t Control thee Webcam?
Client haas also optionn to send conttrol instructionn
from the mobile
m
phone for horizontall movement of
the webcam
m. The proceduure is as follow
ws:
1) A parallel port will be conn
nected with thee
sttepper motor via ULN20
003 IC. Thee
w
webcam
will be
b placed ontto the stepper
m
motor
and it willl move the weebcam based onn
thhe instruction getting
g
from thee client.
2) A
After viewing im
mages client caan send controol
innstruction from
m their mobilee phone to thee
seerver for horrizontal moveement of thee
w
webcam
that is left
l and right movement.
m
3) A
After receiving
g the control instruction for
leeft movement of the webcaam, server willl
seend a combinnation bits thaat will hit thee
paarallel port firrst. For right movement thee
coombination bits will be differrent.
4) Thhen this com
mbination bit will
w move thee
sttepper motor as well as thhe webcam viaa
U
ULN2003
based on the instrruction gettingg
frrom server.
5) Fiinally clients are
a able to view the images
baased on their co
ontrol instructiion.
For better understandingg a flow chart of the overalll
procedure is given in Fig
gure 4.
337
4. GRA
APHICAL US
SER INTER
RFACE
(GUI) DESIGN
D
The Grraphical User Interface is a type of usser
interface item that alllows people to
t interact wiith
ways than tyyping such as
programss in more w
computerrs; hand-held devices
d
such as
a MP3 playerrs,
portable media
m
players or gaming devvices; househoold
appliancees and office equipment
e
witth images rathher
than text commands. A GUI offers graphical iconns,
and visuual indicators,, as opposed
d to text-baseed
interfacess, typed commaand labels or teext navigation to
fully reprresent the infoormation and actions
a
availabble
to a user. In this researcch GUI is usedd to help the usser
t URL by ppressing Loginn from the mennu
to enter the
option to
o view the capptured images as well as sennd
control in
nstruction from
m menu bar juust like the reeal
mobile ph
hone for testinng purpose beffore testing wiith
real mob
bile phone. A
After login when
w
the usser
confirms the address, tthe GUI startss displaying thhe
s
locatioon.
images itt fetched from the servers specific
From thee menu optionn the user cann also select thhe
control in
nstruction to be transmitted to
t the server for
f
controllinng the horizonttal movement of the webcam
m.
Figure 5 and 6 show
ws the snaps taken from thhe
emulator (GUI) while testing the proggram using loccal
host.
Figure 7. Viewing imaages by NOKIA

A 5300 and Sony
Ericsson W810i.
Pictures as
a shown in Fiigure 6 and 7 are
a same in booth
emulator (GUI) and reaal mobile phonnes. So it provves
that the proposed systtem functions successfully in
real time..
Figurre 5. a) Sendin
ng request for establishing
e
connection b) sending control instrucction to server.
After esttablishing connnection to thee server, acceess

time for various
v
mobilee phones is alsoo being observeed
during exxperiment whicch is also given
n in Table 1.
Table 1.Access
1
time for various mobile
m
phones
Figure 6. a) Left imagee b) right imagee of the desiredd

loocation.
Model
Time in
i sec (~)
Nokia N73
10
Nookia 6630
15
Sony Ericsson W810ii
25
NO
OKIA 5300
25-30
Siem
mens M75
30-35
Phhilips 960
35
6. CON
NCLUSION
5. EXPE
ERIMENTAL
L RESULTS
S
In order to test the efffectiveness off the proposedd
w have conduucted using various
v
mobilee
system, we
phones succh as NOKIA N-73,
N
Siemenss M75, NOKIA
A
6630, Sonny Ericsson W810i,
W
NOK
KIA 5300 andd
Philips 9600 and the resuult is successfuul for the abovee
devices. Buut time for fetcching images varies
v
as GPRS
S
class is not same in alll the devicess and the dataa
service pro
ovider supportting EDGE peerforms betterr.
NOKIA 53
300 can fetch 90 images onee after anotherr.
Fig.7 show
ws the images fetched
f
by NO
OKIA 5300 andd
Sony Ericssson W810i froom the server being capturedd
by webcam
m.
338
This reesearch is reelated to the mobile phonne

environm
ment as it is beccoming a most used technologgy
and it is growing too ffast. Exploiting
g the benefits of
the mobiile phone techhnology, a neew approach of
surveillan
nce system thhat can monittor any desireed
places fro
om a remote llocation has been proposed in
this research. A webcam integrateed with steppper
motor annd other hardw
ware accessoriees are connecteed
to a serv
ver placed in the
t desired loccation. User caan
access th
he server from any remote location
l
anytim
me
and view
w the images of his/her desired
d
locatioon
captured by the webcam
m. This work has been testeed
u
an emulator (GUI) whhich is designeed
first by using
using thee NetBeans ID
DE and then using differeent
Proceedings ofMARTEC 2010
types of mobile phones. The result is same in both

platforms and it proves the proper functionality of the
proposed system. But to implement the proposed
system for the security of the ship engine room, we
need to interface our existing system with two
different environments (radio and satellite network)
usually used in ship [17]. We have also need to
interface our system with the ships built-in
controlling system so that we can use the system in
both environments as well as can access the internet
and these will be done in next version of this
research. In order to increase the security, movement
of the webcam can be controlled not only horizontally
but also vertically too. This feature can also be
incorporated in the next version of this research.
REFERENCES
[1] http://www.securitypark.co.uk/security_news_9
_CCTV%20System.html, Last accessed on
05.01.2010.
[2] http://www.deskshare.com/Resources/Articles/
wcm_setup_PCvideosurveillancesystem.aspx;
Last accessed on 26.01.2010.
[3] http://www.securitypark.co.uk/security_article2
64278.html, Last accessed on 26.01.2010.
[4] Kanade, T., Collins, R. T. & Lipton, Advances in
cooperative multi-sensor video surveillance,
Proceedings of DARPA Image Understanding
Workshop. pp. 3-24, November 1998.
[5] http://www.buzzle.com/articles/benefits-ofusing-a-cell-phone.html; Last accessed on
26.01.2010.
[6] Chandramathi, S., Chandra, S. N., Adarsh,
M.G. A Novel Video Surveillance System
based on Multimedia Messaging Service,
Journal of Computer Science, Science
Publications 2005.
[7] Rita, C. Multimedia surveillance systems,
Proceedings of the 3rd ACM international
workshop on Video surveillance & sensor
networks, November 11-11, 2005, Hilton,
Singapore.
[8] Kissell, E. T. Stepper Motor Theory of
Operation, September 6, 2006.
[9] http://en.wikipedia.org/wiki/Stepper_motor ; Last
accessed on 17.02.2010.
[10] http://en.wikipedia.org/wiki/Parallel_port; Last
accessed on 17.02.2010.
339
[11]
Miller, P. F., Vandome, F. A. and McBrewster,

J. machine, Paperback - Oct. 11, 2009.
[12] http://www.java-forums.org/; Last accessed on
16.05.2010.
[13] Yoshiro, I., Makoto, O., Daisuke, Y., Osamu,
S., Yutaka, I., Shin'ichi M. Mobile PhoneEnhanced User Interface of Remote
Monitoring System, Proceedings of the
International
Conference
on
Mobile
Business, p .63-68, July 11-13, 2005.
[14] Shahriyar, R., Hoque, E., Naim, I. and Sohan,
S.M. Controlling Remote System using
Mobile Telephony, Undergraduate Thesis,
ICST (Institute for Computer Sciences, SocialInformatics
and
Telecommunications
Engineering), 2008.
[15] Michael, B., Andreas, D., Arnold, M.,
Bernhard, R., Helmut, S. Distributed
Embedded Smart Cameras for Surveillance
Applications, Computer, v.39 n.2, p.68,
February 2006.
[16] Stefania, B., Fabio, S. Improving the
effectiveness of monitoring and control
systems
exploiting
knowledge-based
approaches,
Personal
and
Ubiquitous
Computing, v.9 n.5, p.301-311, September
2005.
[17]http://www.brighthub.com/engineering/marine/art
icles/32355.aspx;Last accessed on 10.11.2010.
CROSS-SHORE SEDIMENT TRANSPORT DUE TO WAVE: A LABORATORY

STUDY
Badal Mahalder and Umme Kulsum Navera
Department of Water Resources Engineering,
Bangladesh University of Engineering and Technology, Dhaka-1000
E-mail:badal@wre.buet.ac.bd
ABSTRACT
A study was carried out in the Hydraulics and River Engineering Laboratory of Water Resources Engineering
Department, BUET to investigate the cross-shore sediment transport process by wave induced current. Wave is
an important parameter for near-shore sediment transport. Morphological changes in the nearshore coastal
area is governed by wave as it generates velocity, so in laboratory flume an artificial beach profile with a slope
of 1:15 was developed to observe the sediment transport process. In this set-up non-breaking wave was
generated by using wave maker in a glass sided laboratory flume (70 feet long and 30 inch width). In this flume
waves were generated for 1 hr 15 minutes with different water depth and wave period to measure the velocity at
different locations on the slope. A 3.0 cm thick sand bed was used on the slope with d50 value ranging from
0.11mm-0.2mm for each run. The final bed elevation was measured after completing each run at different
locations to understand the sediment transport process. Comparing to others (tide and wind), wave generated
velocity is the simplest one but is an important parameter for sediment transport. This study provides
information regarding the process of sediment transport due to wave in the cross-shore direction due to varying
water depth and wave parameters in the laboratory. The wave generated velocity depends on the wave height,
water depth, wave period etc; hence the change of bed elevations also depends upon these parameters. In this
paper the changing pattern of bed elevation (sediment transport) with different wave parameters is presented.
The results obtained from the laboratory experiment showed good agreement with the change of different
initial parameter like wave period and water depth. From the analysis it is observed that the changes of bed
level depend on water depth and wave period greatly as well as the wave height.
Key words:cross-shore, sediment transport, wave.
1. INTRODUCTION
Nearshore and beach formation problems often
require an understanding of sediment movement.
Wave is the predominant factor for the sediment
movement. Waves stir up the sediment with different
threshold conditions. The current on the other hand
carries the sediment to the long-shore direction. Such
flows generate shear stresses on the bed which
are modifications of the shear applied by either
flow
alone.
Bed shear stress under either
unidirectional or oscillatory flows is also another
important factor for the sediment to move. For
oscillatory flow, Bagnold (1963) reasoned that the
wave-induced oscillatory water motion causes the
sediment to be moved back and forth with a net
expenditure of energy. Although no net transport
results from the oscillatory flow, the energy
dissipation acts to support the sediment in an
amount proportional to the local energy dissipation

rate. The governing parameter for sediment transport
in the cross-shore direction is wave and its direction
with the shoreline. The types of wave may be both
breaking and non-breaking. This study has been
conducted in the laboratory of Hydraulics and River
Engineering Laboratory under the Department of
Water Resources Engineering, BUET by forming
non-breaking waves in the flume.
Wave-induced transport processes are related to
the velocities generated by high and low-frequency
wave phenomena. Net onshore transport is dominant
in non-breaking wave conditions, whereas net
offshore transport is dominant in breaking wave
conditions (Van Rijn, 1990). In the non-breaking zone
the sediment transport pattern is more or less simple,
but for breaking zones the process exhibits rather
341
complex and intricate pattern (Zhou, 2001). Few

attempts to relate the sediment transport/erosion
process in a controlled laboratory observation have
been made. Threshold conditions for homogenous
sands in combined flows have been examined for
co-linear conditions (George and Sleath, 1979;
Hammond and Collins, 1979) and for waves and
currents acting perpendicular to each other (LeeYoung and Sleath, 1988). Likewise, co-linear
interactions above a mixed sand bed have been
studied to examine erosion processes rather than
the associated flow parameters (Tomlinson, 1993).
The mechanism of coastal sediment transport can be
defined as: Wave stirs up sediment and current
transports the sediment. This study is confined in a
narrow scale as the only mode of sediment transport
process is due to wave and wave generated current,
which transports the sediment to the cross shore
direction.
When mean currents due to current are weak,
nearshore sediment transport is driven predominantly
by wave-orbital velocities (Fredsoe and Deigaard,
1992). As waves shoal, their shapes become skewed
(relatively sharp crests and broad, flat troughs) and
asymmetric (pitched forward, with steep front faces
and gently sloping rear faces) (Hsu et al., 2006).
During the passage of the steep front faces of
asymmetric waves, fluid is accelerated strongly as the
orbital velocity rapidly changes from maximum
offshore to maximum onshore (e.g., Elgar et al., 1988,
and Gallagher et al., 1998). The symmetrical wave
induced sediment transport is evaluated by energetic
procedure. Energetics process for evaluating sediment
transport process is driven with moments of time
series of fluid velocities measured above the bottom
boundary layer. It suggest that offshore bar migration
or the sediment transport, observed when incident
waves are energetic, is driven by strong offshoredirected mean currents that are maximum near the bar
crest (Thornton et al., 1996; Gallagher et al., 1998).

An energetic process explains that sediment transport
is associated with the strong flow accelerations under
steep wave faces (Hoefel and Elgar, 2003) which
suggests that onshore sediment transport is observed
when incident wave energy is moderate and mean
currents are relatively weak. This is also related to the
change in cross-shore gradients of beach profile. Hsu,
et al. (2006) suggest that when mean-current-induced
transport is small, wave-induced transport leads to the
observed onshore bar migration, or the sediment
transport. This experiment was carried out in the
laboratory, where symmetrical waves were generated
by wave generator with different wave period with
varying water depth.
2. EXPERIMENTAL SETUP
The experiment was carried out in a 70 ft (21.34
m) long, 30 inch (0.762 m) wide and 30 inch (0.762
m) deep rectangular tilting flume (Fig.1) in the
Hydraulics and River Engineering Laboratory under
the Department of Water Resources Engineering,
BUET. The side walls of the flume are made of glass
and the bed is painted by water resistance color to
avoid any unnecessary bed friction development. In
this flume wave generator was set at one end and at
the other end the defined 1:15 slope was placed. On
the slope a 3.0 cm sand layer with d50 value ranging
between 0.11 mm to 0.24 mm. The flume consists of a
reservoir and a stilling chamber which is located
behind the wave generator. Flume bed was kept
horizontal and it is supported on an elevated steel
truss. The height of the flume is limited to carry out
runs in large-scale experiment. After the wave
generator, several numbers of screens were set to
reduce the wave reflections made of coarse wire
mesh.
Wave generator
Wire mesh (screens)
3 cm sand bed
15
70 ft (21.34m)
Inlet tank
Length of slope (depends

upon the water depth)
Figure 1: Detail of experimental setup
342
30 inch
along the slope of artificial beach was taken as the

area of interested.
Here, total 9 segments were taken (Fig. 2) and
the length of each segment was 30 cm. The starting
of first segment was from zero water depth (still
water level) and it extends towards the deep water.
The measured wave run-up length was other
point of concentration. The bed profile of wave
run-up area (zero water depth to the end of wave
run-up) was plotted by the same way and total
volume of sand was calculated per unit width of the
flume.
3. METHODOLOGY
The experiment was conducted with different
test scenarios (Table1). The bed profile was
measured with the help of point gauge after
completing each run with different water level and
wave periods. The run period was taken as 1 hour
15 minutes. This period was chosen after
conducting several test runs. For different wave
periods, the setting procedure of the wave
generator was an essential task as the motor
requires different number of rotation with changing
wave period. Designated stepwise procedure was
followed for generating non-breaking wave in the
laboratory flume.

The sediment movement pattern or the change
of total sediment volume can be observed from
Fig.3 to Fig.5. From these figures it may be stated
that, along the cross-shore line (i.e. below the zero
water depth) the onshore movement of sediment is
prominent when the water depth is low, but for
higher water depth the sediment movement
characteristics is towards the off-shore area. This
analysis shows that for higher water depth and
wave period the sediment movement in the offshore region increases (Table 2).
Table 1: Experimental Scenario

Water Depth
(cm)
50
40
35
Wave Period
(sec)
1
2
1
2
1
2
Wave Height
(cm)
11.0
12.0
11.0
11.8
10.6
11.9
4. DATA ANALYSIS
The bed elevation after each run was measured
and the obtained data were plotted to calculate the
final volume of the cohesion less sediment. Data
were plotted in AutoCAD, where 270 cm length
Water Level
Total 9 Segments along the slope (Spacing 30 cm)
12 3 4 5 6 7 8 9
1:15 slope
Bottom of Flume
Figure 2: Schematic diagram of studied region
343
Figure 3: Change in sediment volume for 50 cm water depth
But for the same wave period (2 sec) the total

volume of sediment or bed material decreases form
the initial volume in case of lower water depth (35cm
and 40cm) though the change of overall sediment
volume is increased (Table 4). In the wave run-up

zone the rate of accumulation of bed material is
increased with the increase of wave period and water
depth. This is due to the increase of wave energy
344
which is active in the onshore area of the slope. In

this zone the effect of wave period plays a significant
driving force for sediment transport, as in this
experiment only the fine cohesionless sand is used so,

the effect of cohesive and coarser material on the
transport cannot be interpreted.
Table 2: Change in volume below zero water depth upto 270 cm along the slope
Water
Depth (cm)
Wave
Period
(sec)
Actual sand Volume

(cm3/ unit width)
Final sand volume

(cm3/ unit width)
Change in volume (cm3/

unit width)
810
930.05
120.05
810
906.90
96.90
50
1
810
893.77
73.77
810
734.75
-75.25
810
855.07
45.07
810
675.48
-134.53
40
35
From the data (Table 3) other important

parameter, wave run-up value, about sediment
transport in the nearshore zone can be identified. This
experiment suggests that with lower wave period the
wave run-up value is lower and vice-versa. Hence the
onshore movement of sediment for higher wave
period is more compare to lower wave period. This
can be described as, the wave period increases the
time of incident wave in the surf zone also increases
and the orbital velocity also increases. Hence it gets
more time for sediment mixing in the water and the
suspended sediment is carried by the wave induced
current towards the onshore zone. The experimental
results also support this phenomenon. But for the offshore zone there is a great deal of variation. As it is
known that the wave orbital velocity increases with
the increase of wave period but with the increase of
water depth the propagating wave gets less sediment
concentrated water. But when the waves move
towards the opposite direction, it carries some
sediment towards the deep water zone, so the
sediment volume decreases in the off-shore zone for
35 and 40 cm water depth in higher wave periods. On
the other hand for 50 cm water depth the change of
sediment volume is positive but it is lower than that of
1 sec wave period.
Table 3: Change in volume above zero water depth

Water
Depth
(cm)
Wave
Period
(sec)
Distance of wave RunActual sand

up form initial water Volume (cm3/ unit
level (cm)
width)
Final sand
volume (cm3/
unit width)
Change in volume
(cm3/ unit width)
50
150
156.05
6.046
180
540
730.00
190.00
64
192
224.56
32.56
145
435
603.67
168.67
60
180
202.80
22.80
130
390
539.41
149.41
50
40
35
Table 4: Total change of sediment volume

Water
Depth
(cm)
Wave
Period
(sec)
Actual sand Volume

(cm3/ unit width)
Final sand volume

(cm3/ unit width)
Change in volume (cm3/

unit width)
960
1086.10
126.10
1350
1636.90
286.90
1002
1078.32
96.32
1245
1318.41
73.41
990
1067.87
77.87
1200
1244.88
44.88
50
40
35
345
6. CONCLUSION
The objective of this study was to investigate the
amount of sediment movement under variety of wave
parameter and water depth. Here only the effect of
wave period on sediment transport is discussed
keeping the wave height nearly same (Table 1). This
experimental result suggests that the amount of
sediment movement in two different zones is different
for different water depth and wave period. The
movement of onshore sediment volume increases with

the increase of wave period, but for off-shore zone
this value decreases with a significant scale. Hence
wave period is an important parameter for sediment
movement in the formation of beach profile or other
onshore/off-shore bar as a whole the sediment load
movement in the coastal areas.
REFERENCES
[1]
Bagnold, R. A., 1963. Mechanics of marine

sedimentation, in The Sea, vol. 3, edited by
M. N. Hill, Interscience, New York.
[11]
Sekiguchi, T. and Sunamura, T., 2004. Effects

of bed perturbation and velocity asymmetry on
ripple, J. of Coastal Engineering, 59, 231239.
[2]
Elgar, S., Guza, R.T. and Freilich, M. H., 1988.

Eulerian
measurements
of
horizontal
accelerations in shoaling gravity waves. J.
Geophys. Res. 93, 92619269.
[12]
Thornton, E., Humiston, R. and Birkemeier,

W., 1996. Bar-trough generation on a natural
beach. J. Geophys. Res 101, 1209712110.
[3]
Fredshoe, J. and Deigaard, R., 1992.

Mechanics of Coastal Sediment Transport.
Advanced Series in Ocean Engineering,
Volume 3. World Scientific Publishing Co. Pvt.
Ltd. Singapore. 260-287.
[13]
Tomlinson, B. N., 1993. Erosion studies of

mixed sand beds under the combined action
of waves and currents.
[4]
Gallagher, E. L., Elgar, S. and Guza, R. T.,

1998. Observations of sand bar evolution on a
natural beach. J. Geophys. Res. 103, 3203
3215.
[5]
George, C. B. and Sleath, J. F. A., 1979.

Measurements of combined oscillatory and
steady flow over a rough bed. J. Hydraul.
Res., 17, 303-313.
[6]
Hammond, T. M. and Collins, M. B., 1979.

On the threshold of transport of sand-sized
sediments under the combined influence of
unidirectional
and
oscillatory
flow.
Sedimentology, 26, 795-812.
[7]
Hoefel, F. and Elgar, S., 2003. Wave-induced

sediment transport and sandbar migration.
Science 299, 18851887.
[8]
Hsu, T. J., Elgar, S. and Guza, R. T., 2006.

Wave-induced sediment transport and onshore
sandbar migration, J. of Coastal Engineering,
53, 817824.
[9]
Lee-Young, J. S. and Sleath, J. F. A., 1988.

Initial Motion in Combined Wave and
Current Flows. In: Proc. 21st Int. Conf. on
Coastal Eng., Malaga. ASCE, pp. 1140-l151.
[10]
Rijn, Leo C. Van, 1990. Principles of sediment

transport in rivers, estuaries and coastal seas,
Amsterdam: Aqua Publications - I11. ISBN 90800356-2-9 bound NUGI 816/831.
346
[14] Zhou, L., 2001, Sediment Transport, Aalborg

University.

AN OBJECT ORIENTED APPROACH TO BEAM ANALYSIS

M. S. Islam, G. M. Khalil and N. A. Nahian
Bangladesh University of Engineering and Technology, Dhaka-1000
E-mail: shahid113@gmail.com, nishat.wms@gmail.com
ABSTRACT
An Object-Oriented Programming (OOP) procedure is presented for solving structural mechanics beam
problems by using the deflection, bending moment and shear force equations. The program is developed for a
single span and for cantilever and simply supported beam. A beam is divided into one hundred segments, i.e.,
one hundred and one nodes and deflection, bending moment and shear force are calculated on each node for
applied loads and moments.
Key words: Beam analysis, C++, Object oriented programming, shear force, bending moment, deflection.
1. INTRODUCTION
Beam Analysis is a very important and
fundamental topic of the subject Mechanics of
Materials studied by engineering students. The
behaviors of a beam are represented by bending
moment, shear force and deflection curves. A beam
can have supports of fixed, hinged, simple or of
combined type and is classified accordingly. The topic
Beam Analysis informs us how a beam behaves
under different loading conditions and supports. The
main objective of this research is the development of
a reliable and user friendly program in the field of
beam analysis. By using this program, an engineer or
an engineering student alike can easily generate his
own answers to basic questions concerning the
behavior of beams. This program will help them
understand the subject better and verify/correct their
approach to the problems. The program is developed
in C++ language. The method uses object oriented
approach and superposition technique. The program
takes the value and direction of loads and moments
applied to the particular type of beam as an input. The
total length of the beam is divided into one hundred
parts by one hundred and one nodes. The program
will give the nodal values of deflection, shear force
and bending moment using corresponding equations
given in section 2 of this paper. These nodal values
can be used to draw accurate curves.
2. RELEVANT EQUATIONS FOR

DEVELOPING THE PROGRAM
The deflection equations1 for the different loading

conditions are listed below. The shear force and
bending equations are the second and third derivatives
respectively of the deflection equations.
The
derivations of these equations (deflection, shear force

and bending moment) are available in Pytel and
Singer (2005). Downward load is considered as
negative and upward load as positive. Clockwise
moment is considered negative while the anticlockwise moment is taken as positive. The
fundamental properties of a beam are E (modulus of
elasticity), I (moment of inertia), L (length of the
beam), W (point load) and w (uniformly distributed
load per unit length).
2.1 Cantilever beam

Loading condition 1: Point load at any point on the
beam (Fig.. 1)
Where a is the position of load from the fixed end,
b is the position of load from the free end and x is the
variable distance from the fixed end.
When the variable distance x is less than or equal to
a, the deflection equation is
y=
Wx 2 (3a x )
6 EI
(1)
The bending moment equation is the second

derivative of the deflection equation (1), which is
EI
d2y
= W (a x)
dx 2
(2)
Finally the shear force equation is the third

derivative of the deflection equation (1) and the first
derivative of the equation (2),
347
EI
d3y
= W
dx 3
(3)
When the variable distance x is greater than a, the

deflection equation is
y=
Wa (3 x a )
6 EI
y=
wx 2 6 L2 4 Lx + x 2 wa 4 wa 3 (x p )
24 EI
8EI
6 EI
(9)
(4)
w
Fig. 3: Uniformly distributed load
Fig. 1: Point load at any point on the beam
Loading condition 4: A gradually varying load (Fig. 4)
Loading condition 2: Uniformly distributed load (Fig.

2)
A gradually varying load of highest magnitude w is

subjected to a beam of length L where the variable
distance is x. The deflection equation is
A uniformly distributed load of magnitude w per

unit length is applied to a cantilever beam of length L
where the variable distance is x. The deflection
equation is
y=
wx 2 6 L2 4 Lx + x 2
24 EI
y=
w 10 L3 10 L2 x + 5Lx 2 x 3
120 EIL
(10)
(5)
w
L
Fig. 4: Gradually varying distributed load

L
Fig. 2: Partially uniformly distributed load
2.2 Simply supported beam:

Loading condition 3: Partially uniformly distributed
load (Fig. 3)
Where p is the distance of the starting position of
the uniformly distributed from the fixed end (in
Figure 2, p is zero), q is the distance of the uniformly
distributed load from the free end, a is equal to L q
and b is equal to L-p.
Loading condition 1: Point load at any point on the

beam (Fig. 5)
y=
wx 2 6a 2 4ax + x 2
24 EI
for
0< x<b
(11)
Pb L
y=
(x a )3 + L2 b 2 x x 3
6 EIL b
w
a
When p = 0 and x a, the deflection equation is
y=
Pbx 2
L x2 b2
6 EIL
(6)
for
a<x<L
When p = 0 and x > a, the deflection equation is

L
y=
pa
pa (x a )
+
8EI
6 EI
4
(7)
Fig. 5: Point load at any point on the beam
When q = 0 and x p , the deflection equation

y=
wx 2 6 L2 4 Lx + x 2 wx 2 6 p 2 4 px + x 2
24 EI
24 EI
When x > p and q = 0, the deflection equation
(8)
Loading condition 2: Uniformly distributed load over

the length of the beam
wx 3
y=
L 2 Lx 2 + x 3
(12)
24 EI
Loading condition 3: A gradually varying load (Fig. 6)
348
y=
wx
7 L4 10 L2 x 2 + 3x 4
360 EIL
(13)
Figure 6: A gradually varying load
3. THE PROGRAM
The programming language used is C++, an
extension of C. It was developed by Bjarne Stroustrup
in the early 1980s at Bell Laboratories (Deitel, 2005).
The reason for using this language is that it provides
capabilities for object-oriented programming.
As seen in section 2, for each of the two types of
beams (cantilever and simply supported), a number of
loading conditions are present. Procedural program
does not work very well for large and complex
problems. Though it can not be claimed that the
presented program is such a complicated one, but
object oriented programs are easier to understand,
correct and modify (Deitel, 2005). It results in
programs that are clearer, more reliable and more
easily maintained. Instead of viewing a program as a
series of steps to be carried out, it views it as a group
of objects that have certain properties and can take
certain actions.
The developed program is a multifile program. The
source file main_beam.cpp contains the function
main. This function makes an object beam of class
Beam. Its next task is to call a member function of
class Beam named add_node(). Then according to
the type of beam and presence of point load,
uniformly distributed load, partially distributed load,
gradually varying load and moment, it calls different
member
functions
of
class
Beam
like
add_pointload_cantibeam()to find out the
effect due to the presence of corresponding loads or
moments. Finally to add up all the effects on the
beam, ie, to apply the method of superposition, the
function summation_beam()is called.
In the header file named class_beam.h, all the
classes used in the program are created. These classes
are class
pointload,
class
unidl
(uniformly distributed load), class
gvdl
(gradually varying distributed load),
class
partidl (partially distributed load), class
moment, class Node and class Beam.
The first five classes mentioned above contain
member data about the positions and values of loads
and moments. The class Node has member data

about deflection shear force and bending moment due
to different types of loads and finally total deflection,
total shear force and total moment to sum up all the
individual effects. The class Beam has member
data as typebeam (value 1 means cantilever beam
and 2 means simply supported beam), E (Youngs
modulus of elasticity), I (moment of inertia) and
length. The beam is divided into 100 segments, i.e.,
contains 101 nodes. Then for applied loads (maximum
four point loads) and moments (maximum four),
appropriate member functions are created to evaluate
deflections, shear force and bending moments at each
node. Then a member function named void
summation_beam()is created to add up the
individual deflection, shear force and bending
moment components to give total values at each node.
In short, all the member functions are declared in
corresponding classes in class_beam.h file.
These
functions
are
defined
in
either
canti_beam.cpp or in simply_beam.cpp
and are called in main_beam.cpp.

To investigate the quality of the output generated by
the program, a problem (A cantilever beam of 3 m
length is carrying a load of 1 KN (downward) at the
free end , 2 KN (downward) at a distance 1 m from
the free end,3 KN(downward) at a distance 2 m from
the free end and finally a partially uniformly
distributed load of 2 N/mm as shown in Figure 7. The
modulus of elasticity of the beam is 200x103 N/mm2
and moment of inertia is 48.6x106 mm4) is selected.
Then the generated output is compared with the
outputs from a professional FEM software and hand
calculation.
349
3 KN
1m
2 KN
1m
1 KN
Figure 7: Model problem

Node
0
Bending moment
in (N-mm)
1.50E+07
Shear force
(N)
8000
Deflection
(mm)
0
1.48E+07
8000
-0.000690741
1.45E+07
8000
-2.75E-03
---
---
---
---
49
4.65E+06
5000
-1.24E+00
50
4.50E+06
5000
-1.28E+00
51
4.35E+06
5000
-1.32E+00
---
---
---
---
98
6.36E+04
1120
-3.59431
99
30900
1060
-3.64472
100
1000
-3.69513
are shown in Figures 8-10. Demo version of the

professional software BEAM (based on finite element
method), used for comparison (Table 2), gives only
bending moment and shear force data. For that reason
the space for deflection in the table of comparison is
kept empty. It is clear from the table that the
developed program gives accurate results.
Table1. Nodal values for the developed software
Though the problem discussed above has only point

loads but the program is capable of performing the
analysis of beams subjected to a combination of point
loads, uniformly distributed loads, gradually varying
loads, non-uniformly distributed loads and moments.
At present both the cantilever beam and simply
supported beam can be analyzed. The program is
capable of analyzing beam of only one span.
Figure 8: Deflection Curve
6. CONCLUSION
An Object-Oriented Program is developed for
solving structural mechanics beam problems by
using the deflection, bending moment and shear force
equations. The developed program generates output
without any error. At present no graphics is associated
with the program. The generated results are
transferred to Microsoft Excel to create the curves.
Future goal is to create graphical input and output
system.
Figure 9: Shear Force Curve
REFERENCES
1. A. Pytel and F. L. Singer (2005), Strength of
Materials, Fourth Edition, Harper Collins publishers.
2. H. M. Deitel (2005), C++, How to Program, Fifth
Edition, Prentice Hall.
Figure 10: Bending Moment Curve

5. COMPARISON
Developed software output is compared with the
finite element software and hand calculation below.
Finite
Hand
element
Developed
calculation
software
software
BM(max)
N-mm
SF(max)
N
DEF(max)
mm
1.50E+07
15000127
8000
8000.06
-3.69513
------
1.50E+07
8000
-3.69513
Table 2: Table of Comparison

The curves drawn from the results of the program
350

IMO BALLAST WATER CONVENTION AND ENFORCEABILITY OF G2

Oihane Cabezas-Basurko and Ehsan Mesbahi
School of Marine Science and Technology
Newcastle University,
Newcastle upon Tyne, NE1 7RU
Email: Ehsan.Mesbahi@ncl.ac.uk
ABSTRACT
IMOs G2 guideline (Guidelines for ballast water sampling) clearly states that ballast water samples are
required to be representative of the whole ballast water discharged. Enforcement and implementation of the
ballast water convention largely depend on defendable case based on evidence that a ship entering territorial
waters has not complied with regulations. Statistical representativeness of ballast water samples have been
discussed in this study with additional emphasis on practicality of the sampling procedure. Universally accepted
mathematical methodologies for determination of a representative number of samples from an unknown
population have been used to identify minimum number of samples which could be considered statistically
representative of the ships ballast water. Results clearly indicate that a large amount of ballast water must be
sampled to achieve certain level of confidence which could be translated as true representation of the ships
ballast water discharged and potentially used in any further legal actions by states or ship operators.
Key words: underwater acoustic communication, direct-sequence spread spectrum, delay-locked loop,
synchronization.
1. INTRODUCTION
Ballast water convention foresees that ships
calling at ports will be required to present adequate
documentations with respect to their ballast water
operations to prove that they are not suspected of any
violations. If there are clear basis for port authorities
to be suspicious of violation of the ballast water
regulations or there are clear grounds for potential
damage and pollution to the territorial waters or in
some cases on a random basis, ships could be
subjected to an onboard inspection. According to
the Convention, inspections will include sampling of
the ships ballast water. As a result and in particular
cases, port authorities may ask a suspected ship to
leave the territorial waters prior to completion of her
cargo and ballast water operations (Blair 2008). All
these cases could potentially lead to legal cases,
costing port authorities, port state controls, ship
operators, ship owners and cargo owners exceeding
millions of dollars.
Sampling is an essential part in any environmental
pollution control or prevention policy. According to
guidelines G2 (Guidelines for ballast water sampling),
ballast water samples should be representative of
the whole ballast water discharged (IMO 2008b).
Representativeness of ballast water samples has not
yet been clearly discussed and currently there are no
clear guidelines on how to achieve such true
representativeness.
Factors that certainly add high level of

uncertainties to any statistical analysis of ballast water
samples could be categorised as:
Sampling points and techniques (also
referred to as sampling regimes)
Sampling from tanks (sampling

location)
Sampling from discharge line and

during ballast water operations
Biological
content
and
homogenous
distribution of species inside a tank
o
Size and shape of the tank
Source of the sea water
Any sampling regime will have to consider a

minimum number of ballast water samples from the
total amount of ballast water to be discharged in order
to achieve a statistically representative result and
with high level of confidence. This will certainly lead
to legally defendable cases when there is a potential
for prosecution or arrest of the ship.
This research is funded and supported by UKs
Maritime and Coastguard Agency (MCA). The
statistical representativeness of ballast water sampling
is analysed using the methodology for determining the
sample size for a population. This will provide the
amount of ballast water that needs to be collected to
351
ensure that the ballast water samples are true

representative of the ballast water discharged. In order
to define samples proportion the analysis has been
focused on the non-compliance of D-2 standard of the
Convention, which is to have more than 10 viable
organisms that are 50 m in minimum dimension.
pollution. This makes ballast water management the

first case in the marine field to require representative
samples.
Sampling for environmental protection
Regulatory and guiding documents referring to

quantitative measures regarding ballast water
management are three: the regulation D-2 (Ballast
Water performance standard) of the Convention, and
Guidelines G8 (Guidelines for approval for Ballast
Water Management Systems) and G2 (Guidelines for
Ballast Water Sampling).
Any ballast water management systems will have
to be tested and type-approved by G8 which will be
subsequently challenged by G2 and during its real
operations. It is imperative to have relevance,
consistency and agreement between various parts of
the regulations and guidelines which refer to
quantitative measures for sampling for type
approval and sampling for compliance.
Compliance with standards is, indeed, an essential

condition for effective implementation of any
environmental regulation. To achieve this objective,
adequate inspection and monitoring is usually
required (RCEP 1998). There is abundant literature
about the usage of sampling in environmental
policies. Some of the Statutory Instruments of British
law in relation to environmental regulations, for
instance, imply sampling (SI 2000, SI 2007). There
are also guidelines emphasizing the need of having
sampling when pollution is introduced into the
environment, such as oil pollution after an accident
(Watterson et al 1999).
Moreover, there is a strong trend considering that
sampling for compliance should incorporate a
statistical analysis to ensure the results are statistically
representative. The Royal Commission on
Environmental Pollution has recognised the
importance of statistical analysis presented a
document to the British Parliament (RCEP, 1998)
highlighting that any numerical environmental
standard needs to be robust, must recognise scientific
assessment and should try to be a statistically
verifiable standard. It also pointed out that the
standard should embed the substances nature as well
as the statistical variation in the parameter to which it
relates to.
In the field of maritime environmental protection,
there are also few regulations including standards for
the environmental protection. Discharging oily water
with more than 15 ppm into the sea is prohibited by
MARPOL 73/78, therefore oily water separators are
continuously monitored (IMO 2006). Determination
of sulphur content in the bunker oil is another
example of sampling in the maritime regulations.
During bunkering, the fuel supplier provides a
certificate indicating that the content of the fuel are
complying with Annex VI of MARPOL requirements
(IMO 2008a). Ships are not generally equipped with
the facilities onboard to test and analyse the fuel
against the standards. Three samples are taken at the
beginning, middle and end of the supply and kept
onboard. In case of any dispute bunker suppliers will
be held responsible and may be prosecuted since it
was certified that the content of the fuel met the
requirements of Annex VI of MARPOL 73/78.
Nonetheless, unlike ballast water management, none
of the marine environmental protection regulations
requires a continuous sampling and representative
analysis as part of their regulatory process to control
2. QUANTITATIVE
MEASURES
IN
BALLAST WATER REGULATIONS
AND GUIDELINES
D2 Ballast Water performance standard

D2 is one of the two standards developed by the
Convention which numerically quantifies the quality
of the ballast water that ships are allowed to
discharge. This regulation also sets the performance
standard for ballast water treatment systems.
To enter into force, the Convention needs to be
ratified by at least 30 States which combined
merchant fleet is at least 35% of the gross tonnage of
the worlds merchant shipping. At the moment only
14 States have ratified it. Nevertheless, the
Convention is expected to be able to entry in force by
2010 (Matheickal 2008).
G8 Guidelines for approval for Ballast Water
Management Systems
The G8 defines the recommended procedures for
type approval of ballast water treatment systems
and ensures that the system under consideration meets
the standards set by regulation D2 of the Convention.
According to G8 and as one of the final stages of
the approval when the ballast water treatment system
is going to be tested onboard the ship and in real
operations, three replicate samples of treated ballast
water need to be collected during the beginning,
middle and the end of discharge (IMO 2005).
G2 Guidelines for Ballast Water Sampling
Guideline G2 has been under discussion for a
while and it was not until October 2008 that a
resolution was achieved at MEPC 173(58). G2
describes the sampling protocol for determining
whether a ship is in compliance with the Convention
(IMO 2008). According to G2, any sampling protocol
testing compliance with the Convention should
352
observe ten principles to help ensure consistency of

approval between Parties and to provide certainty to
the shipping industry.
The most relevant principles to the current study
are number 2, 5, 6 and 7, which are as follows:
-
the sampling protocol should result in

samples that are representative of the whole
discharge of ballast water from any single
tank or any combination of tanks being
discharged
- the quantity and quality of samples taken
should be sufficient to demonstrate whether
the ballast water being discharged meets with
relevant standards
- sampling should be undertaken in a safe and
practical manner
- sampling should be concentrated to a
manageable size
G2 does not provide any further guidelines or
indication on what a representative sample is but it
clearly mentions that representativeness will depend
on the statistical significance and is required.
calculating a proportion from a finite population

(Desu & Raghavarao 1990, Thompson 2002;
Stattucino Applet 2008). This equation will provide
the number of samples that are required to be taken in
order to ensure that the sampled proportion is
representative of the population. In other words, the
following equation will estimate the amount of ballast
water (in m3) that needs to be sampled for ensuring
that the sampled ballast water is representative of the
ballast water discharged:
n=
A population in statistic terms could be defined

as the group of items from which samples are
taken. If the number of samples and the size of
each sample are increased, this will lead to
increase in accuracy and precision of the results,
but at the expense of higher cost (Barnett 2004)
and time.
The sample size is the number of population

members selected when a sample is taken from a
population.
A proportion refers to the targeted variable

from the population that is subjected to study;
it usually refers to the relation of the targeted
variable over the sample size.
The main mathematical formula used for

determining a sample size for a proportion is the
following which is universally accepted and used for
(N 1) e 2 + z2 2 P(1 P )
where
n=
sample size or the number of 1m3 ballast

water samples that needs to be taken in
order to have a true representation of the
ballast water discharged.
N=
population size or total amount of ballast

water discharged, in m3.
P=
probability of success or the estimated

proportion of an attribute that is present in
the population. In this context it refers to
the probability of having 10 viable
organisms/m3 of ballast water sampled.
(1-P) =
is the probability of failure (Q) that

represents the probability of having < 10
viable organisms/m3 of ballast water
sampled.
Z/2 =
confidence coefficient for a given

confidence interval (for a confidence level
of 95%, Z/2=1.96)
e=
standard sampling error: error that is

assumed while the sampling is carried out.
A value of 1-5% is considered as normal.
3. STATISTICAL EPRESENTATIVENESS
OF BALLAST WATER SAMPLES
The concept of representative samples implies that
the collected sample reproduces the same
characteristics of the environment that was originally
taken from; and that every individual has an equal
chance of appearing in the sample, this is usually
achieved by simple random sampling (Barnett
2004). Statistical representativeness in relation to
ballast water management, however, is a concept that
has not yet been discussed.
Before the methodology is presented some
terminologies need to be clarified:
- A sample is a subset of the population sampled,
in other words, it is a small proportion of the
population.
N z2 2 P(1 P )
In large statistical analysis, it is usually

recommended to collect some samples before
determining the sample size. This pre-sample
analysis should indicate the amount of viable
organisms present in the water (i.e. the proportion)
and will help defining the P (probability of success).
Since in D-2, G8 and G2 we will have to consider the
whole discharged ballast water, pre-sample of ballast
water in this case is not possible; hence, the proposed
analysis has been carried out without priori
knowledge of the content of ballast water onboard the
ship. This will lead to assumption for the maximum
variance (P=0.5; Q=0.5) meaning that 50% of the
water that is sampled meets the criteria of P (ballast
water contains 10 viable organisms per m3), and
another 50% meet the criteria of (1-P) (ballast water
contains < 10 viable organisms per m3).
To discuss the statistical representativeness of
ballast water samples collected onboard a ship our
353
analysis has considered the definition of noncompliance according to D2, which is to have more
than 10 viable organisms that are 50 m in
minimum dimension in one m3 sample of the ballast
water.
Assumptions
Organisms are heterogeneously distributed inside
ballast tanks (Murphy et al 2002). Additionally,
source of ballast water in each tank could vary from
tank to tank and is dependant of ships mission profile
and ballasting patterns. Sampling equipments and
sampling points present extreme limitations and
uncertainties during sampling and limit accessibility
to all corners of the tank. In order to structure and
simplify the real scenarios, the following assumptions
have been made:
- Population is the whole volume of ballast
water to be discharged. (to comply with G2)
-
One m3 of ballast water is a member of the

population. (to comply with D2)
One sample contains 1m3 of ballast water.
Each member of a population contains

homogeneously distributed organisms.
Ship has 8 ballast tanks
A Bio-Section is a part of ballast water

inside a ballast tank which could be sampled
inside the tank or during discharge. (to
consider G8)
A Bio-Section has homogeneously

distributed biological content.
Any 1 m3 sample of ballast water taken from

a Bio-Section will be a true representative
of that Bio-Section.
Hypotheses
Various hypotheses were examined to make sure
that all aspect of sampling procedure and in particular
practicality of sampling was considered. Bearing this
as an starting point and in order to see whether
biological representativeness of a ballast tank or
discharging period can be truly defined in a three
layer approach, five hypotheses were set out.
Hypothesis 1: Full Ballast Capacity
The population (N) is equivalent to the total
amount of ballast water that may be discharged. This
hypothesis assumes that the origin of the ballast water
onboard ship is the same, and there are neither BioSections nor the ship has ballast tanks or
subdivisions. Hence, the 8 ballast tanks were
considered as one large tank. In this case, if the
Master decides to discharge 1,000 m3 of ballast water,
then the population will be 1,000.
Hypothesis 2: Ship has 8 ballast tanks and each tank

contains 27 Bio-Sections
This hypothesis follows the same idea as
Hypothesis 2 but considers that each ballast tank can
be divided into 27 Bio-Sections as shown in figure
1(b). Statistical population for this hypothesis is
216.
contains three Bio-Sections
In this hypothesis it is considered that a ship has 8
ballast tanks and each tank is divided into 3 BioSections as shown in figure 1(a). The population of
ballast water onboard would be 24.
a)
b)
Figure 1 Ballast tank containing Bio-Sections of

ballast water: a) tanks with 3 Bio-Sections, b) tanks
with 27 Bio-Sections
is a Bio-Section, each m3 of a tank is a true
representative of the tank
Each tank is a Bio-Section and the volume of
water in each tank is homogeneously distributed. Each
tank may contain ballast water of different source;
hence, one m3 of ballast water taken from a tank is
biologically a true representative of the particular tank
only. Statistical population in this case will be 8.
Hypothesis 5: Ship has 8 ballast tanks and same
Bio-Sections for all ballast tanks
Similar to hypothesis 3, there are 3 Bio-Sections
in each ballast tank. But in this hypothesis we
consider that all ballast tanks contain the same ballast
water (same source) and the characteristics of the
three Bio-Sections are also the same in the 8 ballast
tanks. Total population in this case will be 3.
Determination of number samples to collect
In these hypotheses, the populations of ballast
water to discharge are considered to be 3, 8, 24, 216
and 1,000. Table 1 shows the amount of ballast water
that requires to collect as sample if small volumes of
ballast water are to be discharged and for a confidence
level of 95%, maximum variance and sampling error
of 1%.
354
Table 1 Sample sizes for the populations considered in the hypotheses and for a sampling error of 1%,
Confidence level of 95% and maximum variance (P=50%, Q=50%)
Population of ballast water (m3), N
% of Population of ballast water required
for sampling
Sample size (tonnes)
Number of 1 m3 samples (m3), n
24
216
1,000
100%
100%
100%
97.7%
90.6%
3.00
7.9
23.9
211.2
905.7
10000.00
9000.00
8000.00
7000.00
6000.00
5000.00
4000.00
3000.00
2000.00
1000.00
0.00
0
20000 40000 60000
80000 100000 120000 140000 160000
Total ballast water discharged (tonnes)
Figure 2 Sample size (n) for a confidence level of 95% and sampling error of 1%
With a maximum variance, when the discharged
volume is small, usually the total amount of water
discharged will need to be sampled (e.g. for
population size of 24, all the 24 m3 need to be
sampled).
These populations are unrealistic as ships carry
and discharge larger volumes of ballast water; but it
provides a preliminary result that when discharging
small volumes of ballast water, almost all of the
discharged amount need to be sampled. In addition, in
the case that the ballast tanks were divided into BioSections
or
what
G8
suggests
as
top/middle/bottom
sections
it
requires
homogeneous distribution of organisms in that
section, which in reality is not true. Hypothesis 1 is
the one which is closest to a real discharge scenario.
In addition, it is worthwhile to mention that following
the findings the number of samples to collect for
type approval purposes (G8) significantly differs
from the statistical representativeness analysis and
sampling for compliance purposes.
If a large volume of ballast water is to be
discharged, and assuming that no Bio-Sections are
considered (Hypothesis 1), then the amount of ballast
water to collect as sample is not as high as the volume
discharged (e.g. for a ballast water discharged volume
of 50,000 m3, 8056 m3 of water needs to be sampled).
Despite the reduction, the volume to collect as sample

still remains comparatively very high.
If the sampling error is increased the amount of
ballast water discharged that needs to be sampled
decreases. On the other hand, the amount of ballast
water that needs to be sampled increases when the
confidence level is increased. Moreover and as
mentioned previously, the percentage of ballast water
discharged that needs to be collected as sample
decreases when the total ballast water discharge is a
large amount. The most conservative scenario for
having a representative condition and meeting the
specifications of the Convention and guidelines is
given by the case of Figure 2 (for a sampling error of
1%). In this specific case, the minimum amount of
ballast water required for sampling would be around
9% of total ballast water discharged. This value
corresponds to discharging 100,000 m3; still the
amount to sample remains very high (9000 m3).
The graphs of figure 3 illustrate the trend of the
percentages of the ballast water discharged that needs
to be collected as sample (% BW to sample) to
ensure that the samples collected are statistically
representative of the total ballast water discharged.
The graphs show the results for a maximum variance,
confidence levels from 90 to 95% and sampling errors
of 1% and 5%.
355
Figure 3 Representation of the amount of ballast water required to sample to ensure statistical representativeness
for the case of maximum variance sampling error of 1% (right) and 5% (left)
4. CONCLUSIONS
Statistics play an important role in the
development of any environmental policy. Statistical
analysis which will lead to definition of
representativeness of ballast water sampling has not
been clearly defined or discussed within IMO
guidelines and ballast water Convention.
IMOs G2 requires that sampling regimes or
protocols should result in samples that are
representative of the whole discharge of ballast water.
Neither biological contents of ballast tanks are
homogenous, nor are sources of the ballast water in
the tanks the same. This will lead to the observation
that statistics can provide the definition of what a
representative sample is.
In order to asses the representativeness of the
sampling results a thorough statistical analysis has
been carried out. Various hypotheses and assumptions
were examined to make sure that all aspect of
sampling procedure and in particular practicality of
sampling has been considered.
Results of the statistical analysis clearly indicate
that to achieve 95% confidence that samples provide a
representative of the whole discharge of ballast water
a large volume of ballast water need to be sampled
and analysed. As an example, for a ship to discharge
216, 5,000, 10,000 and 50,000 m3 of ballast water, it
is required to sample 211, 3,288, 4,899 and 8000 m3
of her ballast water respectively. This is certainly not
practical and manageable due to high cost, time and
undue delay to ships operation.
In addition to biological representativeness of
samples, implementation and enforcement of the
Convention should also ensure that the samples taken
and the ballast water standards are statistically
representative of the ballast water discharged. D-2,
G2, G8 all include quantitative based standards and
guidelines for compliance and type approval
purposes. Currently, the G8 does not agree with
statistical
representativeness
and
compliance
requirements for G2. In addition, verifiable statistical

procedures to provide adequate confidence in
representativeness of ballast water samples have not
been discussed although 16 States have currently
agreed to ratify the Convention and its associated
regulations as it is.
ACKNOWLEDGMENTS
This research was funded by UKs Maritime and
Coastguard Agency (MCA) under the contract The
control of spread of non-indigenous species through
ballast water. Part B: Ballast water sampling:
methods, analysis, representativeness and legal
issues. Results and opinions expressed in this paper
do not present MCAs views.
REFERENCES
[1]
Barnett V., 2004. Environmental statistics

methods and applications. Wiley Series in
probability and statistics. John Wiley & Sons,
Ltd., Chichester, UK.
[2]
Blair M., 2008. Ballast Water Management in

USA. 4th International Conference on Ballast
Water Management (ICBWM08), 16-17
October 2008, Singapore.
[3]
Desu M.M. & Raghavarao D., 1990. Sample

Size Methodology. Academic Press, Inc.,
London.
[4]
IMO, 2004. International Convention for the

control and management of ships ballast
water and sediments. BWM/CONF/36, 16 Feb
2004, London, UK.
[5]
IMO, 2005. Report of the Marine Environment

Protection Committee, Annex 3: Resolution
MEPC. 125(53) Guidelines for approval of
ballast water management systems (G8).
MEPC
53/24,
International
Maritime
356
Research, 24(7):729-734.
Organization, London.
[6]
IMO, 2006. MARPOL consolidated edition

2006. International Maritime Organization,
London, UK.
[11]
RCEP (Royal Commission on Environmental

Pollution), 1998. Setting environmental
standards. 21st report, HMSO, London.
[7]
IMO, 2008a. Guidelines for the sampling of

fuel oil for determination of compliance with
MARPOL
Annex
VI.
Resolution
MEPC.96(47)) in MEPC 58/23.
[12]
SI (Statutory Instrument, UK), 2000. The

Water Supply (Water Quality) Regulations
2000, No 3184. HMSO, London, UK.
[13]
[8]
IMO, 2008b. Report of the Marine

Environment Protection Committee, Annex 3:
Resolution MEPC. 173(58) Guidelines for
ballast water sampling (G2). MEPC 58/23,
International Maritime Organization, London.
SI (Statutory Instrument, UK), 2007. Air

quality standard regulation 2007, No 64.
HMSO, London, UK.
[13]
Stattucino
Applet
http://www.berrie.dds.nl/calcss.htm,
visited on the 18/12/08
[14]
Thomson S.K., 2002. Sampling. 2nd ed. New

York. John Wiley & Sons, Inc.
[15]
Watterson J., Halliwell C., Meharg A.,

Campbell G., 1999. Environmental sampling
after a chemical accident. Department of the
Environment, Transport and the Regions,
HMSO, London, UK.
[9]
[10]
Matheickal J., 2008. Marine Biosecurity: Need

for an integrated Approach and Global
Partnership. 4th International Conference on
Ballast Water Management (ICBWM08), 1617 October 2008, Singapore.
Murphy K., Ritz D., Hewitt, C.L. 2002.
Heterogeneous zooplankton distribution in a
ship's ballast tanks. Journal of Plankton
357
2008.
website

THREE DIMENSIONAL MHD FLOW WITH HEAT AND MASS TRANSFER

THROUGH A POROUS MEDIUM WITH PERIODIC PERMEABILITY AND
CHEMICAL REACTION
N. Ananda Reddy 1, S. Vijaya Kumar Varma 2 and M. C. Raju3
1
Department of Mathematics, A.I.T.S.,
Affiliated to Jawaharlal Nehru
Technological University (A), Rajampet,
Kadapa - 516126. A.P. India. Email:anandareddyn@gmail.com.
2
Department of Mathematics, S.V.
University, Tirupati. A.P. India
E-mail: svijayakumar@yahoo.co.in
3
Principal, APGCCS,
Rajampet, Kadapa. A.P. India.
E-mail: mcrmaths@yahoo.co.in
ABSTRACT
A three - dimensional MHD flow with heat and mass transfer of a viscous, incompressible, conducting fluid,
through a semi-infinite porous medium, in the presence of viscous dissipative heat and chemical reaction, is
considered. The porous medium is bounded by an infinite porous flat surface, kept at a uniform temperature.
The cross flow (suction) velocity at the surface is assumed to be constant. A uniform magnetic field is applied
normal to the bounding surface. The expressions for the velocity, temperature and concentration are obtained
using the perturbation method, subject to the relevant boundary conditions. Velocity of the fluid is increasing
with velocity ratio parameter, Hartman number and Reynolds number whereas decreases with permeability
parameter. Increase in Prandtl number and Reynolds number led to increase in temperature of the fluid. Fluid
concentration decreased with increase in Chemical parameter and Schmidt number. The numerical results are
shown graphically for different values of the parameters entering into the problem.
Keywords: MHD, periodic permeability, chemical effect, Heat and mass transfer.
1. NOMENCLATURE
T
Tw
: temperature of the fluid

: temperature of the fluid at the surface
: temperature of the free stream
: molar species concentration of the fluid

: molar species concentration of the fluid at the
Cw
surface
C : molar species concentration of the free stream
Cp : specific heat of the fluid at constant pressure
D : chemical molecular diffusivity
kT : thermal conductivity of the fluid
Ec : Eckert number
K : chemical reaction parameter
chemical reaction rate constant
K|
:
k : permeability of the porous medium
k : permeability parameter
: wave length of the permeability distribution
l
p : fluid pressure
Re : Reynoldss number
Pr : Prandtl number
Sc : Schmidt number
M : Hartmann number
Bo : uniform magnetic field
u, v, w : velocity components in the x, y , z directions
u,v,w : velocity components in the x, y, z directions
U : free stream velocity
V : suction velocity at the surface
x, y, z : Cartesian co-ordinates
x, y, z : dimensionless Cartesian co-ordinates
Greek Symbols
: kinematic viscosity of the fluid
: dimensionless temperature
o ,1 : dimensionless temperatures of rest and far
away plates
: dimensionless concentration
o , 1
: dimensionless concentrations at rest and far
away plates
: fluid density
: electrical conductivity of the fluid

: velocity ratio parameter
359
2. INTRODUCTION
In recent years, the possible use of MHD is to
affect a flow stream of an electrically conducting fluid
for the purpose of thermal protection, braking,
propulsion and control. Model studies on the effects
of magnetic field on flows through porous medium
have been made by several investigators. Heat and
mass transfer on flow past a vertical plate have been
studied by several authors; viz. Soundalgekar and
Ganesan [11] and Lin and Wu [5] in numerous ways to
include various physical aspects. In the above
mentioned studies the chemical effect is ignored.
Chemical reaction can be codified as either
heterogeneous or homogeneous processes. This
depends on whether they occur at an interface or as a
single phase volume reaction. In many chemical
engineering processes, there does occur the chemical
reaction between a foreign mass and the fluid in
which the plate is moving. These processes take place
in numerous industrial applications, e.g., polymer
production, manufacturing of ceramics or glassware
and food processing. Das et al. [3] studied mass
transfer effects on moving isothermal vertical plate in
the
presence
of
chemical
reaction.
Muthucumaraswamy et al. [7] studied the effect of
chemical reaction on unsteady MHD flow through an
impulsively started semi-infinite vertical plate. The
effect of chemical reaction on an unsteady hydro
magnetic free convection and mass transfer flow past
an infinite inclined porous plate is studied numerically
by Alam et al. [1]. Muthucumaraswamy and Ganesan
[6] derived numerical solution of the natural
convection flow of an incompressible viscous fluid
past an impulsively started semi-infinite isothermal
vertical plate with chemical reaction of first order.
Ananda Reddy et al. [2] studied effects of thermo
diffusion and chemical reaction with simultaneous
thermal and mass diffusion in MHD mixed
convection flow with ohmic heating.
Singh and Sharma [8], Gupta and Johari [4], Singh
and Sharma [9] and Singh et al. [10] have studied the
problems of three dimensional flows by considering
various flow parameters that entering into the
problems. But, no attention has been paid to the
problem of three dimensional hydro magnetic flows
with heat and mass transfer in presence of chemical
reaction. Hence, the objective of this paper is to study
the three dimensional MHD flow with heat and mass
transfer through a porous medium with periodic
permeability and chemical reaction.
3. BASIC EQUATIONS
A three dimensional steady flow with heat and
mass transfer of a viscous, incompressible
conduction fluid through a semi-infinite porous
medium, bounded by an infinite porous surface is
considered. The surface lying, horizontally on the xz plane is subjected to a constant suction. The x-axis
is taken along the infinite surface being the direction
of flow and the y-axis is taken normal to the surface
directed into the fluid flowing laminarily with free
stream velocity. A uniform magnetic field (Bo) is
applied normal to the fluid flow i.e., in the direction
of y-axis. Let u , v and w be the velocity
components in x , y a n d z directions respectively.
The flow is considered under the following
assumptions:
i. The viscous dissipation is considered.
ii. The induced electrical field is neglected
iii. The permeability of the porous medium is
periodic.
iv. The permeability of the porous surface is
assumed to be of the form
k (z ) =
k
1 + cos z / l
The problem becomes three dimensional due to such a

permeability variation.
Under the above assumptions, the flow in the semiinfinite porous medium is governed by the following
equations.
Continuity Equation:
w
+
=0
y z
(1)
Momentum equations:
u
u
1 p
2u 2u
+w
=
+( 2 + 2 )
y
z
y
z
x
(2)
2
B0
(
+
)u
k (z )
1 p
+w
=
y
y
z
(3)
2
2
)
v
+( 2 +
k (z )
y
z 2
w
w
1 p
+w
=
z
y
z
(4)
B0 2
2w 2w
)w
+( 2 + 2 ) (
+
y
z
k (z )
Energy equation:
T
T
k 2T 2T
+w
= T ( 2 + 2 )+
C p
y
z C p y
z
2 w 2
= 2
+
y z
Where
u 2 w 2 u 2
+ +
+
+
y y z z
Mass Transfer
360
equation:
(5)
C
C
2C 2C
+w
= D( 2 + 2 ) K| (C C )
y
z
y
z
y = 0: u = 0; =; w = 0; = 1; = 1
y : u =1; p = p; w = 0; = 0; = 0
(6)
(15)
For free stream, the momentum equation become
1 dp
B0 2
)U = 0
+(
+
dx k ( z )
The corresponding boundary conditions are:
y = 0: u = 0; =V; w = 0; T = Tw; C = Cw
y : u = U; = 0; w = 0; p = p ;
T = T; C = C
(7)
Where f denotes u, v, w, p, and . When = 0,

the problem reduces to a two -dimensional flow and is
governed by the following equations obtained from
(9) to (14) using (16).
The following dimensionless quantities
y
z
u
w
;z = ; u = ; = ; w = ;
l
l
U
U
U
C p
p
Ul
;Re =
; Pr =
; Sc = ;
p=
2
U
kT
D
(T T )
(C C )
=
; = ; =
;
(Tw T )
(C w C )
K
B0 2l 2
U2
M =
; K = | 2 ; Ec =
C p (Tw T )
V
0
We note that the amplitude of the injection

velocity is very small and hence, using Perturbation
technique, the solution to this flow problem may be
assumed to be of the form:
(16)
f ( y, z) = f0 ( y) + f1( y, z) + ......
y=
(8)
u u 1 u u Re M
+ w = 2 + 2
+ (u 1)
y
z Re y z k(z) Re
(10)
p 1 Re
+w = + 2 + 2
v
y
z
y Re y z k(z)
(11)
w
w
p 1 2 w 2 w
+w
= + 2 + 2
y
z
z Re y
z
1 2 2 Ec
+w =
+
+
y
z Pr Re y2 z2 Re
(12)
(19)
(20)
y = 0; u0 = 0; 0 = 1; 0 = 1
y ; u0 = 1; 0 = 0; 0 = 0
(21)
The solutions for the equations (17) - (20) w.r.t the

corresponding boundary conditions (21) are
0 = c1 e
+ (1 c 0 ) e
0 = e t y ;
t1 y
(22)
Pr R e y
with
(13)
2 w 2
= 2
+
+
y z
u 2 w 2 u 2
+
+
+

y y z z
1 2 2
+w =
+
Re K
y
z Sc Re y2 z2
u 0 = 1 e t1 y ;
where
(18)
Using (16) in (15), the corresponding boundary

conditions are
Re M
+ w
k ( z) Re
d 2u0
du0
Re2
Re2
+
+
=
Re
M
u
M
dy2
dy
k
k
d 20
d
+ Sc Re 0 Sc Re2 K0 = 0
2
dy
dy
(9)
2
(17)
du
d 20
d
+ Pr Re 0 = Ec Pr 0
dy 2
dy
dy
are introduced into the equations (1) (6). Then the

governing equations become
w
+ =0
y z
d 0
=0
dy
w0 ( y ) = 0; 0 ( y) = ;
(23)
(24)
p0 ( y ) = p ;
When 0, substituting the expressions (16) into

the equations (9) to (14) respectively and equating the
coefficients of like powers of on both sides and
neglecting those of 2 , 3 etc., the following
equations are equations are obtained.
(14)
The relevant boundary conditions obtained from (7)

using the dimensionless quantities (8) are as under:
361
1 w1
+
=0
y
y
u
1 2 u1 2 u1
u
v1 0 1 =
(
)
+
Re y 2
y
y
z 2
Re
M
( u1 + ( u 0 + 1) cos z )
u1
k
Re
(25)
(26)
1
p
1 2 1
2 1
= 1+
( 2 +
)
y
y Re y
z 2
Re
( v1 cos z )
k
(27)
w1
p
1 w1 w1
= 1+
(
+
)
y
z Re y 2
z 2
(28)
Re M
(
+
) w1
k
Re
2
21
1
1 0 1 =
( 21 +
)
y
y
z 2
Pr Re y
(29)
2 Ec du 0 u1
+
Re dy y
2
(33) and substituting in equation (32), the solutions

for v1 and w1 are obtained as follows:
1 = ( c2 e t y + c3e t y + c4 ) cos z ;
3
w1 =
0
1 21 21
(
) Re K1 (30)
1 =
+
y
y Sc Re y2 z2
(et3 y et4 y )sin z
4.2 Solutions of the main flow, temperature

and species concentration fields:
The equations governing the main flow, the
temperature and the species concentration fields are
given by (26), (29) and (30) respectively. The
solutions to these equations may be assumed as:
u 1 ( y , z ) = u 1 1 ( y ) c o s z
1 ( y , z ) = 1 1 ( y ) c o s z
1 ( y , z ) = 1 1 ( y ) c o s z
The following boundary conditions are obtained using

the above substitutions in (31),
y = 0 : u 1 1 = 0, 1 1 = 0, 1 1 = 0
y : u 1 1 = 0, 1 1 = 0, 1 1 = 0
The corresponding boundary conditions are
y = 0 : u1 = 0,1 = 0, w1 = 0,1 = 0, 1 = 0
y : u1 = 0, 1 = 0, p1 = 0, w1 = 0, (31)
1 = 0, 1 = 0
c5
and 1 in
equations (26), (29) and (30), the following equations
for velocity, temperature and concentration of main
flow are obtained.
Putting the above substitutions for u1,
4.1 Solution for the Secondary flow:

The equations governing the secondary flow are
given by (25), (27) and (28). The solutions for
1 (y, z) , w 1 (y, z) and p1 (y, z) are assumed to be of
the following form.
u||11 ( y) + Re u|11 ( y) (M + 2 +
1 ( y , z ) = 11 ( y ) cos z
Here
1
w1 ( y , z ) = |11 ( y ) sin z
p1 ( y , z ) = p11 ( y ) cos z
are
the
derivatives
of
A4 e Pr Re y + A5 e t1 y + A6 e 2 t1 y +
A7 e t5 y + A8 e t6 y + A9 e t8 y + A10 e t9 y +
A11e t10 y + A12 e t11 y + A13e t12 y
Here
||11 ( y ) , |11 ( y ) are
the
derivatives
of
11( y) with respect to y.
||11 ( y ) + Sc R e |11 ( y ) (R e 2 K Sc +
(33)
2 )11 ( y ) = A14 e t y + A15 e t
+ A16 e t15 y
Here ||11 ( y ) , |11 ( y ) are the derivatives of 11 ( y )
2
Putting the above substitutions (32) for v1, w1 and

p1 in equations (25), (27) and (28), the equation for
velocity components v1, w1 and pressure p1 in terms of
11(y) is
Re2 ||
) 11 ( y )
k
Re 2
Re2
Re 2 |11 ( y ) + 2 ( 2 +
) 11 ( y ) =
k
k
||||11 ( y) + Re |||11 ( y) ( M + 2 2 +
||||11( y), |||11( y), ||11( y), |11( y)
u ||11 ( y ) , u |11 ( y )
||11 ( y ) + Pr Re |11 ( y ) 211 ( y ) =
Using the above substitutions (32) in (31), the

following boundary conditions are obtained.
Here
Re2
)u11 ( y)
k
u 1 1 ( y ) with respect to y.
(32)
11(y) with respect to y.
y = 0 : v11 = 0, v = 0
y : v11 = 0, v = 0
t5 y
= Ae
+ A2et6 y + A3et1 y
1
The expressions for 1 ( y, z ) and w 1 ( y, z ) have

been chosen so that the continuity equation (25) is
satisfied. Here |11( y) denotes the differentiation of
|
11
|
11
(34)
14
with respect to y.
Solving these equations subject to the boundary
conditions (34), the following solutions are obtained:
u1 = (c6et1y + c7et5 y + c8et6 y + c9et7 y )cos z
1 = (c10e PrRe y + c11et y + c12e2t y + c13et y

1
are the
derivatives of 11(y) with respect to y. Solving the

above equation subject to the boundary conditions
362
+ c14et6 y + c15et8 y + c16et9 y + c17et10 y

+ c18et11y + c19et12 y + c20et13y ) cos z
1 = ( c 21e t y + c 22 e t
2
+ c 23 e t15 y
+ c 24 e t16 y ) cos z
14 y
1.1
1
0.9
Here t1, t2,.t16, c1,c2,c3,.c23, c24 are the

constants obtained, but not mentioned because of
brevity.
Substituting
the
values
of
0.8
0.7
0.6
u0 , 0 , w0 , 0 , 0 ,
0.5
u1 , 1 , w1 , 1 and 1 in equation (16), we get the
u 0.4
solutions of equations governing the flow.
0.3
0.2
Re = 1, k = 3
Re = 2, k = 3
Re = 1, k = 6
0.1
5. DISCUSSION
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
In the present paper, a three dimensional MHD

flow with heat and mass transfer through a porous
medium with periodic permeability and chemical
reaction is studied. The Prandtl number for air at 2980
K and 1 atm is given by Pr = 0.71 while = 0.01 and
z = 0.3. It is observed that the velocity of the fluid u
increases as the velocity ratio parameter and
magnetic parameter M increases as shown in Figure 1.
From Figure 2, an increase in Reynolds number Re
lead to an increase in u. But an increase in
permeability parameter k decreased u. The effects of
Pr and Re on temperature of the fluid are shown in
Figure 3. An increase in Pr and Re resulted in
decrease of . This is due to the fact that there would
be a decrease of thermal boundary layer thickness
with the increase of Prandtl number. Finally, the
effects of K and Sc on concentration of the fluid are
also studied. is decreasing with increments in K and
Sc as shown in Figure 4.
Figure 2. Velocity variations for Re and k
1
Pr = 0.71, Re = 1
Pr = 1.0, Re = 1
Pr = 0.71, Re = 2
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Figure 3. Temperature variations for Pr and Re
1.1
0.9
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.5
u 0.4
0.3
0.3
= 0.5, = 1
= 1.5, = 1
= 0.5, = 2
0.2
0.1
0
0.4
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
0.2
K = 0.02, Sc = 0.22
K = 0.05, Sc = 0.22
K = 0.02, Sc = 0.78
0.1
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Figure 4. Concentration variations for K and Sc
Figure 1. Velocity variations for and M
363
REFERENCES
[1] Alam M. S., Rahman M. M. and Sattar M. A.,
Effects of Thermophoresis and chemical
reaction on unsteady hydro magnetic free
convection and mass transfer flow past an
impulsively started infinite inclined porous plate
in the presence of heat generation/absorption,
Thammasat International Journal of Science and
Technology, Vol. 12(3), pp. 44-53, (2007).
[2] Ananda Reddy N., Raju M. C. and Varma S. V.
K., Thermo diffusion and chemical effects with
simultaneous thermal and mass diffusion in MHD
mixed convection flow with ohmic heating,
Journal of Naval Architecture and Marine
Engineering, Vol. 6(2), pp. 84-93, (2009).
[3] Das U.N., Drka R.K., and Soundalgekar V.M.,
Effects of mass transfer on flow past an
impulsively started infinite vertical plate with
chemical reaction, The Bulletin of GUMA, Vol.
5, pp. 13-20, (1999).
[4] Gupta G.D. and Johari Rajesh, MHD three
dimensional flow past a porous plate, Indian
Journal of Pure and Applied Mathematics, Vol.
32(3), pp. 377-386, (2001).
[5] Lin H. T. and Wu C. M., Combined heat and
mass transfer by laminar natural convection from
a vertical plate, Heat and Mass transfer, Vol.
30(6), pp. 369-376, (1995).
[6] Muthucumaraswamy R. and Ganesan P., Natural

convection on a moving isothermal vertical plate
with chemical reaction, Journal of Engineering
Physics and Thermophysics, Vol. 75(1), ( 2002).
[7] Muthucumaraswamy R., Maheswari J. and
Pandurangan J., Unsteady MHD flow past an
impulsively started semi-infinite vertical plate in
the presence of chemical reaction, International
Review of Pure and Applied Mathematics, Vol.
4(1), p.119, (2008).
[8] Singh K.D. and Sharma R., Three dimensional
couette flow through a porous medium with heat
transfer, Indian Journal Pure and Applied
Mathematics, Vol. 32(12), pp.1819-1929, (2001).
[9] Singh K.D. and Sharma R., Three dimensional
free convective flow and heat transfer through a
porous medium with periodic permeability,
Indian Journal of Pure and Applied Mathematics,
Vol. 33(6), pp.941-949, (2002).
[10] Singh P., Sharma V.P. and Misra U.M., Threedimensional free convection flow and heat transfer
along porous vertical plate, App. Sci. Res., Vol.
34, pp.105-115, (1978).
[11] Soundalgekar V. M. and Ganesan P., Finite
difference analysis of transient free convection
with mass transfer as an isothermal flat plate,
International Journal of Engineering Science, Vol.
19(6), pp. 757-770, (1981).
364

ESTIMATION OF SWAY, HEAVE AND ROLL OF A MOORED FLOATING

BREAKWATER DUE TO INTERACTION WITH WAVES
Md. Ataur Rahman1 and Silwati Al Womera2
1
Associate Professor
Department of Water Resources Engineering, BUET
E-mail: mataur@wre.buet.ac.bd
2
Lecturer
Department of Water Resources Engineering, BUET
E-mail: silwati@wre.buet.ac.bd
ABSTRACT
Floating breakwaters offer an alternative to conventional fixed breakwaters and usually preferred in
relatively low wave energy environments or where water depth or foundation considerations preclude the use of
a bottom-founded structure. In this study a two-dimensional numerical model has been developed that simulates
the dynamic displacements, i.e., sway, heave and roll of a moored floating body under wave action. The model
is based on the coupling of SOLA-VOF (Volume Of Fluid) method and porous body model. A pontoon type
submerged floating breakwater of rectangular shape supported by mooring chain is considered in the model.
The SOLA scheme is employed to calculate the pressure and velocities in each time step and the added
dissipation zone method is adopted to treat the open boundary. The analysis is done for both the vertical and
inclined mooring lines alignments. It is assumed the weight of the floating body is much less compared to the
buoyancy force acting on it, so that there does not occur any slack state in the mooring lines during the course
of wave interaction. Considering this assumption, in case of vertical mooring lines alignment, the roll motions
of the body can not occur and only sway and heave displacements are seen. On the other hand, all three
displacements develop when the floating body is moored by the inclined mooring lines and the above
assumption is still considered. The model simulations of the dynamic displacements of the floating body are
compared with the data measured through laboratory experiment. The very good agreement between the
simulated and measured data demands the successful performance of the developed numerical model.
Keywords : VOF method, dynamic displacements, floating breakwater.

1.
INTRODUCTION
Floating breakwater can be considered as an

alternative solution to conventional fixed breakwaters
in coastal areas with mild wave conditions. In the last
years, an evolution of the floating breakwater was
seen both regarding the largest structures protecting
big harbors and the smaller ones defending craft
harbors or marinas. They have been adopted at
number of sites where water depth or other
constraints render the rubble mound and caisson
structures too costly. They are environmentally
advantageous because they produce minimal
interference on water circulation, sediment transport,
and fish migration. They are more adaptable to the
water level changes that occur at harbors that are built
on reservoirs and in coastal areas having a large tidal
range. As a result of all these positive effects, many
studies have been carried out to investigate the
hydrodynamic performance of these floating
breakwaters. Several types of floating breakwaters
have been developed, however, the most commonly
used type of floating breakwaters is the one that
consists of rectangular pontoons connected to each
other and moored to the sea bottom with cables or

chains. Many studies have been involved in floating
breakwaters to investigate their performance, mooring
forces, and motion responses. Adee (1975) developed
a two-dimensional linear theoretical model to predict
the performance of catamaran type floating
breakwaters in deep water and compared the results
with measurements in a model tank and from a
prototype installation in the field. Yamamoto et al.
(1980) solved the problems of wave transformation
and motions of elastically moored floating objects by
the direct use of Greens identity formula and
validated their solutions with experimental
investigations. They found that if the mooring system
is properly arranged, the wave attenuation by a small
draft breakwater can be improved moored. Isaacson
and Byres (1988) reported the development of a
numerical model, based on linear diffraction theory,
to investigate floating breakwater motions,
transmission coefficients, and mooring forces, in
obliquely incident waves. Sen (1993) developed a
numerical method to simulate the motions of twodimensional floating bodies. Sannasiraj et al. (1998)
365
adopted a two-dimensional finite element model to

study the behavior of pontoon-type floating
breakwaters in beam waves. Williams et al. (2000)
investigated the hydrodynamic properties of a pair of
long floating pontoon breakwaters of rectangular
section. Lee and Cho (2003) developed a numerical
analysis using the element free Galerkin method and
mainly concerning the influence of mooring line
condition on the performance of floating breakwater.
Seah and Yeung (2003) have studied the sway and
roll hydrodynamics of cylindrical sections. In this
study, a two-dimensional numerical model is
proposed that combines the VOF method and the
porous body model to simulate the nonlinear wave
interaction with the floating body. A rectangular
shaped pontoon type submerged floating breakwater
supported by mooring chain is considered in the
model. The alignment of mooring chains is
considered for both vertical and inclined directions.
The dynamics of the body is calculated considering
its time-marching finite displacements. The timemarching boundary conditions enable it to consider
the non-linear dynamic interaction between the waves
and the floating body.
with q* as the source strength which is only located at

x = x S and t is the time.
2. NUMERICAL MODELING
As suggested by Brorsen and Larsen (1987), the flux

density of wave generation source q is gradually
intensified for initial three wave periods from the start
of computation for stable wave generation. The mesh
sizes are used as x = 2 cm and z = 1 cm. An
added dissipation zone method is used to treat the
open boundaries. The pressure in the full cell can be
calculated by means of the SOLA scheme. It is
considered that the floating body is of very light
weight compared to the buoyancy forces acting on it
vertically. This assumption results no slack state in
the mooring lines that causes no impulsive force on it.
The dynamics of the floating body due to wave action
are calculated as shown in Fig. 1. The horizontal
displacement (sway), vertical displacement (heave)
and the rotational movement (roll) of the body are
calculated with respect to its center of gravity. The
wave forces acting on the body H3, H4, H5, H6, V3,
V4, V5, and V6 are calculated by integrating forces
acting on the respective surface of the body which are
estimated from the pressure of the respective cells.
Applying the Newtons second law, the horizontal
motion equation of the body can be written as below:
The numerical model consists of the continuity

equation,
the
Navier-Stokes
equation
for
incompressible fluid and the advection equation that
represents the behavior of the free surface. The twodimensional numerical domain is divided into
staggered meshes in both horizontal (x-axis) and
vertical (z-axis) directions. As there occurs the
dynamic displacements (sway, heave and roll) of the
breakwater due to wave action, the numerical cells
can be classified into five types; a full cell filled with
fluid, an empty cell occupied by air, a surface cell
containing both fluid and air, an obstacle cell that
represents the structure and the porous cell containing
the fluid, the structure and/or air. So the continuity
equation, Navier Stokes equations and the VOF
function equation should be modified as below
considering the effects of x, z and v, where x and z
represent the ratio of the permeable length to the cell
length in vertical and horizontal directions
respectively and v represents the ratio of the
permeability volume in a cell.
The continuity equation is,
( xu) ( z w)
+
= q( x, z, t )
x
z
The Navier-Stokes equations,
p
u
u
u
+ xu
+ zw
= v
+
x
t
x
z
u
x 2 +
x x z
u w

z +
z x
p
w
w
w
v + xu + zw = v +
z
t
x
z
u w w
x + + z 2
x z x z z
1 q
+ vg
3 z
(3)
(4)
where p is the pressure, is the kinematic viscosity,

is the fluid density and g is the gravitational
acceleration.
The advection equation of VOF function F,
F ( x uF ) ( z wF )
+
= Fq
+
t
x
z
FX
(1)
(5)
= m.a x
or , H 3 + H 5 H 4 H 6 2.T 3 cos 3 +
q ( z , t )....................x = x S
q ( x, z , t ) =
(2)
0...............................: x x S
where u and w are the flow velocity of x and z
direction respectively, q is the wave generation source
2.T 4 cos 4 = ma x
(6)
where, ax denotes the acceleration of the body in

horizontal direction and T3 and T4 represent the
offshoreside and onshoreside mooring force
respectively
366
M Cg . X + M cg.Z + M cg.T = I
where,
L
L
Z
Z2
Z1
X1
L2=L
Z1 = L ( X 1 + X ) 2
(9)
(LB X 1 X B cos )2 + (Z1 + B sin )2 = L2
X2
M Cg . X , M Cg .Z , M Cg .T represent
the summation of moment governed by the horizontal

wave forces, vertical wave forces and the mooring
lines forces respectively, I is the mass moment of
inertia of the body and is the roll displacement of
the body. Also, considering no slack condition, from
the geometry of the Fig. 1, we write Eq. (9) and (10).
B
B
(8)
(10)
LB
V4
H5
H4
W
H3
By solving above five simultaneous equations (Eq.

(6) to Eq. (10)), the unknown parameters ax, az, , T3
and T4 are calculated. Finally the sway and heave
displacements are estimated using ax and az values.
Mizutani et el. (2004) developed a VOF simulation
model to study the dynamics of the floating
breakwater considering finite displacement. They
solved the dynamics equations explicitly and the
model considers only vertical moorings. In the
present model, we consider both vertical and inclined
mooring alignments and the simultaneous equations
are solved implicitly.
V6
V5
V3
H6
4
T4
3
T3
Fig. 1 Calculation of floating body dynamics
3.
LABORATORY EXPERIMENT
A laboratory experiment is done in a two-
The similar equation for vertical motion of the body

can be written as below:
FZ
= m.a z
or , V 4 + V 6 + W V 3 V 5 + 2T 3 sin 3 +
2T 4 sin 4 = m.a z
(7 )
where, az denotes the acceleration of the body in
vertical direction and W is the weight of the body.
Again, the rotation of the body is governed by the

moments acting on the body. Taking the moment
acting of the center of gravity of the body in anticlockwise direction, following equation is found.
dimensional wave tank of 30 meters long, 0.7 meter

wide and 0.9 meter of height. The floating body is
made of acrylic material with vacuum inside. The
body is 40 cm long, 68 cm wide and 18 cm of height.
The body is anchored to the bottom of the tank by
chain with three different inclinations - vertical (=
90) and two inclined (= 60 and = 45). Photo
view of the all three types connections are shown in
Fig.2. The floating body is always anchored so that
its top surface was at the height of 62 cm from the
bottom of the wave tank. Water depth in the wave
tank was varied as 62 cm, 65 cm and 68 cm. Regular
waves are generated from the wave generator and the
wave steepness (H/L) is varied as 0.01,0.02 and 0.03.
Wave
(a)
Wave
(b)
(c)
Fig. 2 Photo view of the floating body dynamics at the wave tank during experimental run (a) vertical mooring,
=90 (b) inclined mooring, =60 (c) inclined mooring, =45.
367
The experiments are conducted for the waves of ten

different wave periods, i.e., T=0.8 sec,0.9 sec, 1.0
sec,1.1 sec, 1.2 sec,1.3 sec, 1.4 sec, 1.6 sec,1.8sec
and 2.0 seconds. During the experiment, sway (x),
heave (z) and the rolling () of the body occurs due
to the wave action. These displacements of the
floating body are recorded with the laser system.
Two vertical laser rays and two horizontal laser rays
are focused on a white paper box that is set up at the
top surface of the floating body. When the floating
body moves due to the wave action, the box also
moves and the corresponding displacements are
recorded with the laser system. The rolling (), sway
(x) and heave (z) of the floating body are
calculated using the laser system data.
observed that the wave breaks over the top surface of

the inclined floating body during its anticlockwise
rotational motion. On the other hand, during the
clockwise rotational motion of the floating body, its
top surface becomes sloped to the onshore side, but
at the same time its offshore side vertical surface
becomes sloped that also encourages the wave
breaking resulting wave energy dissipation.
B / 2H
x / H
z / H
4.2 Comparison between measured and simulated

motion responses of the floating body
The developed numerical model can simulate well
the motion responses of the floating breakwater due
to its interaction with wave. The time series numerical
results of the displacements of the breakwater are
verified with measured results from the recorded laser
displacement sensors data during the experiment,
4.1 Experimentally measured displacements of which are presented in Fig. 4. In case of the mooring
lines anchored vertically (=90) (Fig. 4a), it is
the floating body
observed that there occurs no roll displacement of the
Due to wave action, there occur three
body and only the sway and heave values are shown
displacements of the body - heave (z), sway (x)
in the figure. For inclined mooring with =45 (Fig.
and rolling () for inclined moored condition. These
4b), sway, heave and roll are observed and shown in
displacements are shown in Fig. 3 for the condition
the figure. The positive sway magnitudes represent
of floating body anchored with bottom of the tank
the horizontal displacements of the floating body
along the onshoreward, i.e., along the direction of
positive x-axis. On the other hand, the negative sway
H/L=0.01
0.14
H/L=0.02
magnitudes represent the horizontal displacements of
0.12
H/L=0.03
0.10
the floating body along the offshoreward which is
0.08
0.06
along the direction of negative x-axis. The positive
0.04
heave values represent its displacement along the
0.02
0.00
downward direction and the negative heave values
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
3.0
represent its displacement along the upward direction.
2.5
2.0
The anticlockwise rotation of the body is assigned as
1.5
positive roll and clockwise rotation is defined as
1.0
0.5
negative roll in the figure. The comparison shows
0.0
good agreement of the numerical simulation results
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
0.8
compared to measured data.
0.6
0.4
5. CONCLUSIONS
0.2
0.0
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L
Relative
displacements
(c)
h=68cm
( d=6cm )of the
Fig. 3
floating body (=60, h=68cm)
with 60 inclination and 6 cm submergence depth

(h=68 cm). The half-filled symbols in the figure
represent the wave breaking over or behind the
floating body. It is seen that wave breaking occurs up
to the value of B/L=0.22 for wave steepness
H/L=0.03 and up to the less value of B/L for wave
steepness 0.01 and 0.02. The effects of wave
steepness (H/L) on these displacements are also
shown in the figure. This rolling motion of the body
makes it tilted and causes its top surface inclined that
result it acts as a wave absorber of the incoming
waves. Moreover, during the experiment, it is
A two-dimensional numerical estimation method

for calculating dynamics of a pontoon type moored
floating breakwater under wave action is developed.
The model couples the VOF method with porous
body model. SOLA scheme with the wave generation
source method and the added dissipation zone method
has been applied. The model can simulate the floating
body dynamics for both vertical and inclined
alignments of the mooring lines. The model results
are varied with the experimentally measured data. The
good agreement between the numerical and
experimental results regarding the dynamic
displacements of the floating body (sway, heave and
roll) confirms the validity of the developed numerical
model.
368
sway(cm)
10
8
6
4
2
0
-2
-4
-6
-8
-10
Numerical
REFERENCES
Experimental
[1]
0.8
10 11 12 13 14 15 16 17 18
Numerical
Experimental
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
4
10 11 12 13 14 15 16 17 18
t (sec)
sway (cm)
(a)
5
4
3
2
1
0
-1
-2
-3
-4
-5
Numerical
5.5
6.5
Experimental
7.5
Numerical
8.5
8.5
Experimental
heave (cm)
2
1
0
-1
-2
-3
5.5
6.5
12
10
8
6
4
2
0
-2
-4
-6
-8
-10
-12
7.5
Numerical
Experimental
Roll (degree)
heave (cm)
0.6
5.5
6.5
7.5
Adee, B. H., 1975, Analysis of Floating

Breakwater Performance, Proceedings of
Symposium on Modeling Techniques, ASCE,
pp. 15851602.
[2] Brorsen, M. and Larsen, J. (1987). Source
generation of nonlinear gravity waves with
boundary integral equation method. Coastal
Engineering, 11, Elsevier, Amsterdam, pp. 93
113.
[3] Isaacson, M., and Byres, R., 1988, Floating
Breakwater Response to Wave Action,
Proceedings of the 21st Coastal Engineering
Conference, Vol. 3, pp. 21892200.
[4] Lee, J., and Cho, W., 2003, Hydrodynamic
Analysis of Wave Interactions With a Moored
Floating Breakwater Using the Element
Galerkin Method, Can. J. Civ. Eng., 30, pp.
720733.
[5] Mizutani, N., Rahman, M.A., Hur, D.S. and
Shimabukuro, H., 2004, VOF Simulation for
Dynamic Behavior of Submerged Floating
Breakwater and Wave Deformation Considering
Finite Displacement Annual Journal of the
Coastal Engineering, JSCE, Vol. 51, pp. 701705.
[6] Sannasiraj, S. A., Sundar, V., and
Sundarravadivelu, R., 1998, Mooring Forces
and Motion Responses of Pontoon-Type
Floating Breakwaters, Ocean Eng., Vol. 25 (1),
pp. 2748.
[7] Seah, R.K.M. and Yeung, R.W., 2003, Sway
and roll hydrodynamics of cylindrical sections,
International Journal of Offshore and Polar
Engineering, ISPOE, Vol. 13, No. 4, pp. 241248.
[8] Sen, D., 1993, Numerical Simulation of
Motions of Two-Dimensional Floating Bodies,
Journal of Ship Research, Vol. 37, No.4, pp.
307-330.
[9] Williams, A. N., Lee, H. S., and Huang, Z.,
2000, Floating Pontoon Breakwaters, Ocean
Eng., 27, pp. 221240.
[10] Yamamoto, T., Yoshida, A., and Ijima, T.,
1980, Dynamics of Elastically Moored
Floating Objects, Appl. Ocean. Res., 2, pp. 85
92.
8.5
t (sec)
(b)
Fig. 4 Comparison of numerical and experimental results

of the time series values of displacements of floating
body (a) H=7.3 cm, T=1.3 sec, h=65cm, =90 (b) H=3.8
cm, T=0.9 sec, h=65 cm, =45
369

SAFETY STRATEGY FOR RO-RO PASSENGER FERRIES OPERATING ON

INDONESIAN WATERS
Sunaryo1, Antoni Arif Priadi2 and Tri Tjahjono3
1
Naval Architecture Study Program,
Mechanical Engineering Department,
University of Indonesia
Kampus UI, Depok, Indonesia
naryo@eng.ui.ac.id
2
Transportation Study Program, Civil
Engineering Department, University of
Indonesia
Kampus UI, Depok, Indonesia 16424
Antoni.bmmc@gmail.com
3
Transportation Study Program, Civil
Engineering Department, University
of Indonesia
Kampus UI, Depok, Indonesia 16424
tjahjono@eng.ui.ac.id
ABSTRACT
Indonesia is the worlds largest archipelago, 2/3 of the country is covered by sea, therefore sea crossing
transportation modes namely Ro-Ro ferries play very important role for the flow and movement of goods and
passengers from and to the islands within Indonesia and the neighbouring countries. But due to many factors a
lot of ferry accidents occurred every year, and claiming a large number of casualties. Efforts have been done to
improve the safety of domestic sea transportation, but since most of the Ro-Ro ferries are categorized as Nonconvention ships, as the result these ferries are not enforced to be fully compliance to the SOLAS regulations,
worsen by the varying sea and cargo characteristics, and low educated passengers, they are very vulnerable to
accidents. Research is being carried out to explore the most appropriate strategy for ensuring the safety of RoRo ferries operating on Indonesian waters using various methods, where four most influencing aspects i.e.
Software, Hardware, Environment, and Live-ware are put into considerations. The existing conditions and
experience learned from the past, backgrounds and sources of accident/incident are investigated, based on the
findings feasible strategy is proposed, which will include the design requirements; qualification of personnel,
refined rules, and operational efficiency.
Keywords: safety, strategy, Ro-Ro ferry, Indonesia.
ferries accidents occurred between 2006 and 2010,

which consisted:
1. INTRODUCTION
As the world largest archipelago Indonesia has more
than 17,000 islands, and 95,181 km length of coastal
line, 2/3 of the country is covered by sea water.
Considering this geographical situation therefore
cargo and passenger sea crossing transportation
modes namely roll on roll off (Ro-Ro) ferries play
very important role for efficient and economical
nation-wide flow and movement of goods and
passengers from and to the islands within Indonesia
territory and its neighbouring countries. But due to
many factors such as age of the ferries, sufficient
safety knowledge of the human resources, safety
awareness of the passengers and other service users,
sea and weather conditions, facilities and equipments
condition etc.
A lot of ferry accidents occurred every year on
Indonesian waters, and claiming a large number of
casualties (according to the official recorded data in
the period between 2006 and 2010 there were more
than 1000 deaths resulting from ferry accidents which
concerned everyone) [3]. According to the National
Transportation Safety Committee there are 19 major
Table 1. Ferry accident 2006 2010
Collission
Sinking
Fire
Grounding
Occurance
3
4
5
7
Casualty
n.a
1,116
29
N.A
As has been identified by Muchlis et al. [5] that

from these accidents 48% are due to human factors,
31% due to environment including weather, and 21%
due to technical aspects. According to Jinca [2] the
high casualties are mostly due to passengers were not
given enough information on emergency access and
how to use the safety appliances, sub-standard safety
equipment, and improper manning of the ships.
In response to this many efforts have been done by
the Government and other stakeholders to improve the
safety of domestic sea transportation, such as
ratification of the International Maritime Organization
(IMO) regulations and codes, establishing National
371
Safety Regulations, establishing Shipping Laws,

improving ports facilities, increasing the knowledge
of human resources, continuously monitoring the sea
and weather conditions and broadcaststing the
forecast to all stakeholders etc. But since most of the
Ro-Ro ferries operating on Indonesian waters are
categorized as Non-convention ships as regulated
by the International Maritime Organization, because
they are operating on the domestic waters, and most
of them are under 500GT, as the result these ferries
are not enforced to be fully compliance to the Safety
of Life at Sea (SOLAS) regulations, worsen by the
varying sea and cargo characteristics, and low
educated passengers, they are very vulnerable to all
sorts of accidents.
There are around 250 ferries operating on
Indonesian waters, consisted of various size, design,
building materials, and about 30% of them are
considerably old ranging between 15 - 30 years old,
more than a half of them are operated by State owned
company called Indonesia Ferries, and the rest are
owned by private companies. The routes and the
number of fleets in every route are regulated by
government under the Directorate General of land
transport based on the needs of the crossing locations.
The routes are divided into 3 main categories i.e.
commercial route, pioneer route, and subsidized route
[1].
In order to minimise ferry accident including the
number of casualties, research is being carried out to
set up safety strategy that can be applied to most of
the ferries operating in Indonesia.
2. RESEARCH METHODOLOGY
The research is conducted using both qualitative as
well as quantitative approaches. The qualitative
approach is used to review rules and regulations, and
research had been done on related subjects, and the
quantitative approach is used to explore and analyse
information gathered based on the field surveys.
Some modelling systems are used to explore the
most suitable approach in the strategy being
proposed, these include Generic Management System
Model; Modified SHEL (Software, Hardware,
Environment, Live ware) method [6]; and Safety
Assessment method, where four most influencing and
interrelated aspects on Ro-Ro ferries safety i.e. ship
design aspect, ship operational environment aspect,
human behaviour aspect, and rules and regulations
aspect are put into consideration.
Since the crossing environment conditions, the
cargo and passengers characteristics, and the ferries
are vary quite significantly from one location to
another, in order to get real operational information 4
busiest crossing locations have been selected to be
investigated in the research by direct interview and
distributing quetionnaires, namely Merak
bakaheuni in Sunda strait between Java and Sumatra
islands, Ujung Kamal in Madura Strait between
Surabaya (East Java) and Madura island, Ketapang

Gilimanuk in Bali Strait between Java and Bali
islands, and Padang Bai Lembar in Lombok Strait
Between Bali and Lombok islands.
The aims of this investigation is to gather
information the conditions of the ferries, the
qualifications of the crews, the bahaviour of the
passengers, the condition of the ports, and the sea
environment of the crossing routes.
3. SOURCES OF ACCIDENTS
The sources of accidents can be divided into internal
and external sources. The internal sources are the
sources of accidents that come from the ship itself and
everything in it, which include the operational of the
ship, the cargo being shipped, human influences
(intentionally or unintentionally), design and
construction of the ship, the arrangement of cargo and
passengers; and the external sources are the sources
that come from outside of the ship, which include the
environment such as the weather, wave and current,
collision (hit by other ship), acts of criminals, etc.
The potential accident hazards in the same way can
be grouped into four different aspects i.e. software,
hardware, environment, and live-ware (SHEL).
Software is consisted of organisation, management,
rules and regulation, operating system etc, Hardware
are consisted of the ship and all its machinery and
equipment, land based supporting facilities etc.
Environment is consisted of weather, wave, wind, and
current. And the live ware is everything relates to
human involvement both onboard and onshore.
Since the potential hazards are very much
determined by these four aspects, therefore the safety
strategy is also developed based on these four aspects.
4. THE SAFETY STRATEGY

The safety strategy is developed based on the four
aspects of potential hazards sources namely
Software, Hardware, Environment, and Live ware.
Each of aspect is then elaborated further into several
variables. Here, the identified variables are then
assessed by putting the potential hazards into safety
risk categories. Generic Management System model
is used for the assessment. The outline is described
in Figure 1.
Figure 1. Safety Assessment Model
372
The four aspects of potential hazards sources are

elaborated as follows:
4.1. Software
Included in the Software are: the organisation,
management, rules and regulation, and operating
system.
4.1.1. The Organisation
The organisation includes the organisation of the
ferry company, the organisation of the port
authority, and the organisation of the ship. The
structure of each of these organisations will
determine the effectiveness of the bureaucracy and
how the decisions are being made, this include the
manning system and the hierarchical system of the
organisation, how the organisation is carried out,
how each member of the organisation is interacting
one to another, which will directly or indirectly
affect the response to the safety requirements.
comfort of the ship and harm the people on board.

They are issued by governing authorities both local
as well as international usually based on the
experience learned from past accidents. Rules and
regulations are minimum requirements that are
applied to common ships and operational conditions,
but higher measures should be applied if the
situation is beyond the normal conditions, for this
purpose Generic Management System Circuit
approach that has been developed for addressing any
entity that is non absolute, which combines
management system circuit and process scheme [4]
may be implemented.
There are various rules and regulations to be
implemented for the operation of Ro-Ro ferries, that
are not applied only for the ships systems, but also
for the people working on board the ship, such as
SOLAS, ILO regulations, MARPOL, ISM Codes,
ISPS Codes etc.
4.1.4. Operating Systems

4.1.2. The Management System
Management system is a framework for managing
and continually improving organisations policies,
procedures, and processes that will transform the
input into desired output. The management system
usually involves 5 following steps loop:
Operating systems are instructions that should be

followed by all parties concerned in order to have an
effective operational condition, without good
operating systems the operation of the ship might
end up in chaotic situation and will create hazardous
condition, therefore the operating system should be
made simple, easy to be understood, systematic, and
accessible to the concerning parties.
4.2. Hardware
Included in the hardware are the design and the
structure of the ship including the machinery and
equipment of the ship, and all land based supporting
facilities.
4.2.1. Ship Design and Structure

Figure 2. Management system loop
Step 1 - DEFINE : The goal and policies
Step 2 - ORGANISE : Resources; plan the work and
assign responsibilities to different people
Step 3 - IMPLEMENT : In practice the policies in
order to achieve the goal
Step 4 - MEASURE : Performance by active
methods
Step 5 - REVIEW : The results obtained and
experience gained.
Step 5 will then to be used as feedback for
refining the whole process from the beginning. This
management system should be applied into all
aspects of the ship management in order to obtain
cost effective and safe operation of the ship.
4.1.3. Rules and Regulations

Rules and regulations are deterministic guidance
and instructions that are meant to prevent any
accident which might disturb the operational
The design and structure of the ship including all

its systems and sub-systems should be assessed
using Generic Ship Model, and Topographical
Reliability Block Diagram.
The Generic Ship Model describes how all the ship
functions are built up from subsystems and systems,
which is developed by utilising the model and
colected data. The topographical Reliability Block
Diagram gives a basis for the assessment/design
showing lack of system redundancies and effect of
fire and flooding situations aboard on top of the actual
ship arrangement. Results of the assessment will be
recommended for the improvement of the ship safety.
4.2.2. Land Based Supporting Facilities

The land based supporting facilities will
determine the safety of loading and unloading
both the ships cargo and the passengers. The
condition and efficiency of the facilities should
also be assessed using available data and
simulation tools.
373
4.3. Environment
Included in the Environment are the conditions of
the weather, wind, wave, current, depth of the sea,
traffic condition, and port description. First of all the
characteristics of the environment should be
investigated using available data and information
from the reliable sources, these data are known as
the normal environment characteristics. The normal
characteristics of the environment should be
matched with the safety requirements which are
related to the hardware aspects of the safety
strategy. Beside the normal environment
characteristics the day to day condition of the
environment should also be monitored and
evaluated by the authorised parties both on board as
well as on shore to determine whether the ship is fit
for operation.
Tolerable - These are the risks that might affect the

operation of the ship, or bring inconvenience to the
passengers, but will not cause major loss.
Ignorable These are the risks that have no effect
both to the operation of the ship and neither the
passengers.
5. APPLICATION OF THE STRATEGY

In order to apply the strategy into every aspect of
Ro-Ro safety the aspect being considered should be
focussed and then every component of the SHEL
potential hazard sources are listed into the strategy
matrix as follows:
Table 2. Safety Strategy Matrix
4.4. Liveware
Hardware
requirements
Environment
readiness
Live-ware
requirements
ACTIONS TO BE TAKEN
The potential hazard of the Ro Ro ferries

operation can be categorized as follows:
Grounding
Collision
Allision
Flooding
Fire
Sinking
Personnel Accident
The next step is put the potential hazard into risk
level categories from 0 to 5 using hazards matrix.
Where 0 means no risk and 5 means very high risk.
Example of the assessment is shown in figure 3.
Personnel
accident
POTENTIAL HAZARDS
Sinking
Untolerable These are the risks that should be

made as minimum as possible, because they might
threaten human lives, and cause major damage to
the ship as well as the environment.
Software
requirements
Fire
All the above aspects are integrated and evaluated

to minimise potential hazards as low as possible for
any determined conditions of the Ro-Ro passenger
ferries operating on Indonesia waters in a cost
affective manner based on the following risks
considerations:
Live-ware
potential
hazards
Flooding
4.4.3. People on Shore

Ferry safety are not only determined by those who
are on board, but at certain extents are also
determined by the people on shore, these include
port authorities, harbour workers, and land based
companys employees, aspecially those who have
influence on the operational of the ferries.
Environment
potential
hazards
Allision
4.4.2. Passengers
Passengers behaviour is determined by the
cultural, educational, ocupational, gender, and ages.
Since characteristics of the passengers are usually
very different from one place to another, therefore
special surveys are needed before any safety
procedures are introduced in to the Ro-Ro ferry
operation in relation to the passengers being
transported.
Hardware
potential
hazards
Collision
4.4.1. Ships Crews

Qualifications of the crews to be assign on board
should be complianced to the requiment standards,
this includes their knowledge, physical and mental
condition, and welfare.
Software
potential
hazards
Grounding
Included in the live-ware are the ships crews, the

passengers, and the people on shore that have
influence to the safety of the ferry.
SAFETY ASPECT TO BE CONSIDERED
Weather
Wind
Wave
Swell
Current
Sea depth
Traffict
Port
ENVIRONMENT
ASPECTS
374
Figure 3. Example of Safety Assessment

From the hazard category of the Ro-Ro ferries as
result of safety assessment then strategy matrix is
created for this aspect and then the actions to be
taken are integrated into an action plan as follows:
ACTIONS TO BE
SAFETY ASPECT 1
TAKEN
FOR
ACTIONS TO BE
SAFETY ASPECT 2
TAKEN
FOR
ACTIONS TO BE
SAFETY ASPECT 3
TAKEN
FOR
..........................................
..........................................
ACTIONS TO BE
SAFETY ASPECT n
TAKEN
FOR
A
C
T
I
O
N
P
L
A
N
Figure 4. Safety strategy action plan diagram
6. CONCLUSIONS
Four major sources for potential safety hazards are
identified namely Software, Hardware, Environment,
and Live-ware (SHEL), in order to obtain optimum
safety strategy for Ro-Ro ferries the influence of
SHEL sources should be first identified, and
requirements for anticipating the risks caused by
these sources should be established, and based on
these requirements the action plan is proposed.
REFERENCES
[1] Abubakar I., Introduction to Crossing
Transportation, Directorate General of Land
transportation, 2010.
[2] Jinca M Y., Keselamatan Transportasi Laut dan
Penyeberangan
(Sea
and
Crossing
Transportation
Safety),
RAKORNAS
Masyarakat Transportasi Indonesia, 2007.
[3] Komisi Nasional Keselamatan Transportation
(National Transportation Safety Committee),
Kajian analisis trend kecelakaan transportasi
laut tahun 2003-2008, 2009(Trent Analyses of
Sea TransportationmAccidents 2003 -2006,
2009), Page 46.
[4] Kuo.C, Contribution of Safety Management to
Ship Stability, Key note opening address, STAB
2009.
[5] Muchlis R. et al, Human Error Penyebab Utama
Kecelakaan Laut, http://www.republika.co.id,
Oktober 2009.
[6]
375
Sunaryo, Escape Route Geometry untuk

Memprediksi Arus Evakuasi Penumpang Kapal
Ferry Saat Terjadi Bahaya (Escape Route
Geometryfor Predicting the Evacuation Flows of
Ferry Passengersin Emergency), SENTA 2009.

EFFECTS OF TEMPERATURE DEPENDENT THERMAL CONDUCTIVITY ON

MHD FREE CONVECTION FLOW ALONG A VERTICAL FLAT PLATE WITH
HEAT GENERATION AND JOULE HEATING
A.K.M. Safiqul Islam1, M. A. Alim2 and ATM. M. R. Sharker3
1
Department of Mathematics,
Govt. Nazimuddin College, Madaripur.
Email: Safiqis@yahoo.com
2
Department of Mathematics, Bangladesh
University of Engineering and Technology,
Dhaka-1000, Bangladesh.
E-mail: maalim@math.buet.ac.bd
3
Department of Computer Science and
Engineering,
Dhaka International University, Banani,
Dhaka-1213, Bangladesh.
E-mail: mrrahmandiu@yahoo.com
ABSTRACT
Magnetohydrodynamic (MHD) free convection flow of an electrically conducting fluid along a vertical flat
plate with temperature dependent thermal conductivity, heat generation and joule heating effects is analyzed
here .The governing equations with associated boundary conditions for this phenomenon are converted to
dimensionless forms using a suitable transformation. The transformed non-linear equations are then solved
using the implicit finite difference method with Keller-box scheme. Numerical results of the velocity and
temperature profiles, skin friction coefficient and surface temperature for different values of the magnetic
parameter, thermal conductivity variation parameter, Prandtl number, heat generation and joule heating
parameters are presented graphically. Detailed discussion is given for the effects of the aforementioned
parameters.
Key words: magnetohydrodynamics, free convection, thermal conductivity variation, heat generation, joule
heating, finite difference method.
1. INTRODUCTION
Electrically conducting fluid flow in presence of
magnetic field and the effect of temperature
dependent conductivity on MHD flow and heat
conduction problems are important from the technical
point of view and such types of problems have
received much attention by many researchers.
Magnetohydrodynamics is that branch of science,
which deals with the motion of highly conducting
ionized (electric conductor) fluid in presence of
magnetic field. The motion of the conducting fluid
across the magnetic field generates electric currents
which change the magnetic field and the action of the
magnetic field on these currents give rise to
mechanical forces, which modify the fluid. It is
possible to attain equilibrium in a conducting fluid if
the current is parallel to the magnetic field. Then the
magnetic forces vanish and the equilibrium of the gas
is the same as in the absence of magnetic fields. In the
case when the conductor is either a liquid or a gas,
electromagnetic forces will be generated which may
be of the same order of magnitude as the
hydrodynamical and inertial forces. Thus the equation
of motion as well as the other forces will have to take
these electromagnetic forces into account.
Convection is the mode of energy transfer between

a solid surface and the adjacent liquid or gas that is in
motion, and it involves the combined effects of
conduction and fluid motion. The faster the fluid
motion, the greater the convection heat transfer.
Convection is the transfer of heat by the actual
movement of the warmed matter. It is also the transfer
of heat energy in a gas or liquid by movement of
currents. Considerable convection is responsible for
making macaroni rise and fall in a pot of heated water.
The warmer portions of the water are less dense and
therefore, they rise. Mean while, the cooler portions
of the waterfall because they are denser. Conduction
is most effective in solids but it can happen in fluids
also.
In electronics in particular and in physics broadly,
Joule heating is the heating effect of conductors
carrying currents. It refers to the increase in
temperature of a conductor as a result of resistance to
an electrical current flowing through it. At an atomic
level, Joule heating is the result of moving electrons
colliding with atoms in a conductor, whereupon
momentum is transferred to the atom, increasing its
kinetic or vibrational energy. Joule heating is caused
by interactions between the moving particals that
form current and the atomic ions that make up the
377
body of the conductor. It is the process by which the

passage of an electric current through a conductor
releases heat. Joules first law is also known as Joule
effect. It states that heat generation by a constant
current through a resistive conductor for a time whose
unit is joule. Heat generation is the ability to emit
greater-than-normal heat from the body. Joule heating
is the predominant heat mechanism for heat
generation in integrated circuits and is an undesired
effect.
Model studies of the free and mixed convection
flows have earned reputations because of their
applications in geophysical, geothermal and nuclear
engineering problems. Hossain [1] analyzed the
viscous and Joule heating effects on MHD free
convection flow with variable plate temperature.
Rahman et al. [2] investigated the effects of
temperature dependent thermal conductivity on MHD
free convection flow along a vertical flat plate with
heat conduction.Rahman and Alim [3] analyzed
numerical study of MHD free convective heat transfer
flow along a vertical flat plate with temperature
dependent thermal conductivity. Nasrin and Alim [4]
studied the combined effects of viscous dissipation
and temperature dependent thermal conductivity on
MHD free convection flow with conduction and joule
heating along a vertical flat plate. Alim et.al. [5]
analyzed the combined effect of viscous dissipation &
joule heating on the coupling of conduction & free
convection along a vertical flat plate. Alim et al. [6]
investigated Joule heating effect on the coupling of
conduction with MHD free convection flow from a
vertical flat plate.
The present study is to incorporate the idea of the
effects of temperature dependent thermal conductivity
on MHD free convection boundary layer flow along a
vertical flat plat with heat generation and joule
heating. The governing boundary layer equations are
transformed into a non-dimentional form and the
resulting non-linear system of partial differential
equations is reduced to local non-similar partial
differential
forms
by adopting
appropriate
transformations. The transformed boundary layer
equations are then solved numerically. Numerical
results of the velocity, temperature, local skin friction
coefficient and surface temperature distribution for
the magnetic parameter, thermal conductivity
variation parameter, Prandtl number, heat generation
and joule heating parameters are presented
graphically.
conducting, viscous and incompressible fluid along a

vertical flat plate of length l and thickness b (Figure1). It is assumed that the temperature at the outer
surface of the plate is maintained at a constant
temperature Tb, where Tb > T, the ambient
temperature of the fluid. A uniform magnetic field of
strength H0 is imposed along the y -axis i.e.normal
direction to the surface and x -axis is taken along the
flat plate. The coordinate system and the configaration
are shown in Figure-1.
The governing equations of such laminer flow
with heat generation joule heating and also thermal
conductivity variation along a vertical flat plate under
the Boussinesq approximations for the present
problem for continuity, momentum and energy take
the following form
u v
+
=0
x y
(1)
H 02 u
u
u
2u
+v
= 2 + g (T f T )
x
y
y
Tf
(2)
2
Tf
Tf
Q
H2 u
1
( f
) + 0 (Tf T) + 0 (3)
u
+v
=
Cp
x
y cp y
y Cp
x
Insulated
interface
H0
T ( x , 0)
Tb
-b
Figure 1: Physical model and coordinate system

Here is coefficient of volume expansion. The
temperature dependent thermal conductivity, which is
used
by
Rahman
[2]
as
follows
(4)
f = [1 + (T f T )]
where is the thermal conductivity of the ambient
fluid and is a constant, defined as
2. MATHEMATICAL FORMULATION OF
THE PROBLEM
At first we consider a steady two-dimensional

laminar natural convection flow of an electrically
f T
. The appropriate boundary

f
condition to be satisfied by the above equations are
378
u = 0, v = 0
Tf s
on y = 0, x > 0
Tf = T(x,0),
=
(Tf Tb)
y b f
u 0, Tf T as y , x > 0
= x 4 / 5 (1 + x) 1/ 20 f ( x, )
(5)
Here is the similarity variable and is the nondimensional stream function which satisfies the
continuity equation and is related to the velocity
components in the usual way as
x
y
ul
vl
Gr 2 , v =
Gr 4 ,
, y = Gr 4 , u =
l
l
T f T
g l 3 (Tb T )
=
, Gr =
Tb T
2
(6)
u=
where l is the length of the plate, Gr is the

Grashof number, is the dimensionless temperature.
Now from equations (1)-(3), we get using the
following dimensionless equations
u v
+
=0
x y
u
f +
M=
H0 l
Gr1 2
is the dimensionless magnetic
the
thermal
variation
conductivity
J=
H 0 Gr
Ql 2
is the non C p Gr1/ 2
on y = 0, x > 0
y
u 0, 0 as y , x > 0
here p = b
sl
1
Gr 4
1
1
h( x, 0) 1
(1 + x )
(14)
20h ( x , 0 )
For the numerical computation, it is important to

calculate the values of the surface shear stress in
terms of the local skin frication coefficient Cfx. This
can be written in the dimensionless form as Molla et
al. [8]
Cf =
Gr
3
4l2
(15)
w [ = ( u y ) y = 0 ]
is the shearing
stress. Using the new variables described in (6), the

local skin friction co-efficient can be written as
2
5
C f x = x (1 + x )
is the conjugate conduction
parameter. In the present work we considered p = 1.

To solve the equations (8) and (9) subject to the
boundary
conditions
(10)
the
following
transformations are proposed by Merkin & Pop [7]
(1 + x )
(1 + x ) 4 + x
f ( x , ) 0 , h ( x , ) 0
where
(10)
h ( x , 0 ) =
parameter,
dimensional
heat
generation
parameter.The
corresponding boundary conditions (5) then take the
following form
u = 0, v = 0, 1 = (1 + ) p
(13)
where prime denotes partial differentiation with

respect to . The boundary conditions in equation (10)
then take the form
non-dimensional
is the dimensionless joule
heating parameter and Q =
f
16 +15x
1
h
+
fh
f h = f h
x
20(1+ x)
5(1+ x)
x
1/ 2
C p (Tb T )
(12)
f ( x , 0 ) = f ( x , 0 ) = 0
parameter, =(Tb T) is
2
f
f
f + h = x f f
x
x
1
x 5
x 5 2
h +
hh +
h
Pr
Pr 1+ x
Pr 1+ x
22
16 + 15x
6 + 5x 2
ff
f
20(1 + x)
10(1 + x)
2
1
5
10
Mx (1 + x)
(8)
1
2
u + v = (1+ ) 2 + + Ju2 + Q (9)
x
y Pr
y Pr y
cp
where
is
the
Prandtl
number,
Pr =
and v =
.
y
x
Moreover, h (x,) represents the dimensionless

temperature. The momentum and energy equations are
transformed for the new coordinate system as
(7)
u
u
2u
+v
+ Mu = 2 +
x
y
y
(11)
= x1/ 5 (1 + x) 1/ 5 h( x, )
The non-dimensional governing equations and

boundary conditions can be obtained from equations
(1)- (3) using the following dimensionless quantities
x=
= yx 1/ 5 (1 + x) 1/ 20
3
20
f ( x,0)
(16)
The numerical values of the surface temperature

distribution (x, 0) are obtained form the relation
1
( x,0) = x 5 (1 + x)
379
1
5
h ( x,0)
(17)
This paper investigates the effect of the
temperature dependent thermal conductivity on
electrically conducting fluid in free convection flow
along a vertical flat plate with heat generation and
joule heating for strong magnetic field. Along with the
boundary conditions (14), the numerical solutions of
the parabolic non-linear partial differential equations
(12) and (13) will be found by using very efficient
implicit finite difference method together with Kellerbox elimination technique [9] which is well
documented by Cebeci and Bradshaw [10].
4. RESULT AND DISCUTION

The main objective of the present work is to
analyze the effect of thermal conductivity variation
due to temperature on MHD free convection flow
along a vertical flat plate in presence of heat
generation and joule heating. In the simulation the
values of the Prandtl number Pr are considered to be
0.733, 1.099, 1.63 and 2.18 that correspond to
hydrogen, water, glycerin and sulfur dioxide
respectively.
The velocity, the temperature, the local skin
friction cofficient and the surface temperature
distribution profiles obtained from the solutions of
equations (12) and (13) are depicted in Figures 2 to
11.Numerical computations are carried out for a range
of magnetic parameter M = 0.01, 0.15, 0.30, 0.45,
thermal conductivity variation parameter = 0.01,
0.25, 0.50, 0.75, heat generation parameter Q = 0.01,
0.05, 0.10, 0.15 and joule heating parameter J = 0.01,
0.20, 0.40, 0.60.
The interaction of the magnetic field and moving
electric charge carried by the flowing fluid induces a
force, which tends to oppose the fluid motion.
In Fig. 2(a), it is shown that the magnetic field action
along the horizontal direction retards the fluid
velocity with = 0.01, Q = 0.01, J = 0.01, and Pr =
0.733. Here position of peak velocity moves toward
the interface with the increasing M. From fig. 2(b), it
can be observed that the temperature within the
boundary layer increases for increasing values of M.
The effect of thermal conductivity variation
parameter on the velocity and temperature profiles
against within the boundary layer with M = 0.01, Q
= 0.01, J = 0.01 and Pr = 0.733 are shown in fig. 3(a)
and 3(b), respectively. It is seen from fig. 3(a) and
3(b) that the velocity and temperature increase within
the boundary layer with the increasing value of . It
means that both the velocity boundary layer and the
thermal boundary layer thickness increase for large
values of .
Figure 4(a) and 4(b) illustrate the velocity and

temperature profiles against for different values of
Prandtl number Pr with M = 0.01, Q = 0.01, J = 0.01
and = 0.01. From fig. 4(a), it can be observed that
the velocity decreases as well as its position moves
toward the interface with the increasing Pr. From fig.
4(b), it is seen that the temperature profiles shift
downword with the increasing Pr.
Figure 5(a) and 5(b) describe the velocity and
temperature profiles against for different values of
heat generation parameter Q with M = 0.01, Pr =
0.733, J = 0.01 and = 0.01. From Figure 5(a), it can
be observed that the velocity increases as well as its
position moves toward the interface with the
increasing Q. From Figure 5(b), it is seen that the
temperature profiles behave the same as increasing
within the boundary layer. It means that the velocity
boundary layer and the thermal boundary layer
thickness increase for large values of Q.
The effect of joule heating parameter J on the
velocity and temperature profiles against within the
boundary layer with M = 0.01, Pr = 0.733, Q = 0.01
and = 0.01 are shown in Figure 6(a) and 6(b)
respectively. The velocity and temperature increase
within the boundary layer with the increasing values
of J.
The variation of the local skin friction coefficient
and surface temperature distribution against x for
different values of M with Pr = 0.733, Q = 0.01, J =
0.01 and = 0.01 at different positions are illustrated
in Figure 7(a) and 7(b), respectively. It is observed
from Figure. 7(a) that the increased value of the
Magnetic parameter M leads to a decrease the skin
friction factor. Again Figure 7(b) shows that the
surface temperature increases due to the increased
value of the magnetic parameter M. It can also be seen
that the surface temperature increases along the
upward direction of the plate for a particular M. The
magnetic field acts against the flow and reduces the
skin friction and produces the temperature at the
interface.
Figure 8(a) and 8(b) illustrate the effect of the
thermal conductivity variation parameter on the skin
friction and surface temperature against x with M =
0.01, Q = 0.01, J = 0.01 and Pr = 0.733. It is seen that
the skin friction increases monotonically along the
upward direction of the plate for a particular values of
. It is also seen that that the skin friction increases for
the increasing . The same result is observed for the
surface temperature from Figure 8(b). This is to be
expected because the higher value for the thermal
conductivity variation parameter accelerates the fluid
flow and increases the temperature as mentioned in
Figure 3(a) and 3(b), respectively.
380
x with M = 0.01, J = 0.01 and = 0.01 and Pr = 0.733.

It can be noted that the skin friction and surface
temperature profile increase for the increasing values
of Q.
Figure 9(a) and 9(b) deal with the effect of Prandtl

number Pr on the skin friction and surface
temperature against x with M = 0.01, J = 0.01, =
0.01 and Q = 0.01.It can be observed from fig. 9(a)
that the skin frication increases monotonically for a
particular value of Pr. It can also be noted that the skin
friction coefficient decreases for the increasing values
of Pr. From Figure 9(b), it can be seen that the surface
temperature decreases due to the Pr increases along
the positive x direction for a particular Pr.
Figure 11(a) and 11(b) deal with the effect of J on

the skin friction and surface temperature against x
with M = 0.01, Q = 0.01,and = 0.01 and Pr = 0.733.
It can be noted that the skin friction and surface
temperature profiles increase for the increasing values
of J.
Figure 10(a) and 10(b) deal with the effect of Q

on the skin friction and surface temperature against
1.80
M = 0.01
M = 0.15
M = 0.30
M = 0.45
0.60
0.40
0.20
0.00
(a)
Temperature Profiles
Velocity Profiles
0.80
1.20
0.90
0.60
(b)
0.30
0.00
M = 0.01
M = 0.15
M = 0.30
M = 0.45
1.50
Figure 2(a) Velocity and (b) Temperature profiles against for varying of M with Pr = 0.733, Q = 0.01, =
0.01 and J = 0.01.
= 0.01
= 0.25
= 0.50
= 0.75
0.60
0.40
(a)
0.20
0.00
Velocity Profiles
0.80
1.20
0.80
0.40
0.00
= 0.01
= 0.25
= 0.50
= 0.75
1.60
(b)
0
Figure 3(a) Velocity and (b) Temperature profiles against for varying of with Pr = 0.733, Q = 0.01, M =
0.01 and J = 0.01.
381
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
0.60
0.40
0.20
(a)
0.00
1.60
Velocity Profiles
0.80
1.20
0.80
(b)
0.40
0.00
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
Figure 4(a) Velocity and (b) Temperature profiles against for varying of Pr with M = 0.01, Q = 0.01, = 0.01
and J = 0.01.
2
Q = 0.01
Q = 0.05
Q = 0.10
Q = 0.15
0.6
0.4
0.2
(a)
0
Velocity Profiles
0.8
1.5
(b)
0.5
Q = 0.01
Q = 0.05
Q = 0.10
Q = 0.15
Figure 5(a) Velocity and (b) Temperature profiles against for varying of Q with M = 0.01, Pr = 0.733, =
0.01 and J = 0.01.
2.00
J = 0.01
J = 0.20
J = 0.40
J = 0.60
0.60
0.40
(a)
0.20
0.00
Velocity Profiles
0.80
1.50
1.00
(b)
0.50
0.00
J = 0.01
J = 0.20
J = 0.40
J = 0.60
Figure 6(a) Velocity and (b) Temperature profiles against for varying of J with M = 0.01, Q = 0.01, = 0.01
and Pr = 0.733.
382

2
M = 0.01
M = 0.15
M = 0.30
M = 0.45
2.5
Surface temperature
1.5
(a)
0.5
1.5
0.5
(b)
0
M = 0.01
M = 0.15
M = 0.30
M = 0.45
0.5
1.5
2.5
Figure 7(a) Local skin friction coefficient Cfx and (b) Surface temperature distribution (x, 0) against x for
varying of M with Pr = 0.733, Q = 0.01, J = 0.01 and = 0.01.
3
2.5
Surface temperature
(a)
2
1.5
= 0.01
= 0.25
= 0.50
= 0.75
0.5
2.5
1.5
= 0.01
= 0.25
= 0.50
= 0.75
0.5
(b)
0.2
0.4
0.6
0.8
varying of with Pr = 0.733, Q = 0.01, J = 0.01 and M = 0.01.
2
(b)
(a)
Surface temperature
1.5
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
0.5
1.5
0.5
Pr = 0.733
Pr = 1.099
Pr = 1.63
Pr = 2.18
0
varying of Pr with M = 0.01, Q = 0.01, J = 0.01 and = 0.01.
383

3
Q = 0.01
Q = 0.05
Q = 0.10
Q = 0.15
2.5
1.5
(a)
0.5
Q = 0.01
Q = 0.05
Q = 0.10
Q = 0.15
2.5
Surface temperature
1.5
(b)
0.5
J = 0.01
J = 0.20
J = 0.40
J = 0.60
(a)
0
J = 0.01
J = 0.20
J = 0.40
J = 0.60
4.0
Surface temperature
varying of Q with Pr = 0.733, J = 0.01, M = 0.01 and = 0.01.
3.0
2.0
1.0
(b)
0.0
varying of J with Pr = 0.733, Q = 0.01, M = 0.01 and = 0.01.
REFERENCES
5. CONCLUSION
The effects of temperature dependent thermal
conductivity on MHD free convection flow along a
vertical flat plate with heat generation and joule
heating have been studied numerically and
graphically. From this investigation the following
conclusions may be drawn
i) The velocity profile within the boundary layer
increases for decreasing values of M, Pr and
increasing values of , Q and J.
ii) The temperature profile within the boundary layer
increases for increasing values of M, , Q and J and
decreasing values of Pr.
iii) The local skin friction coefficient decreases for the
increasing values of M, Pr and increases increasing
values of , Q and J.
iv) An increase in the values of M, , Q and J leads to
an increase in surface temperature distribution.
Moreover, this decreases for increasing values of Pr.
384
[1]
Hossain, M. A. The viscous and Joule heating

effects on MHD free convection flow with
variable plate temperature, Int. J. Heat transfer
Vol. 35, No. 12, pp.3485-3487, (1992).
[2]
Rahaman, M.M, Mamun A.A, Azim M.A and

Alim, M. A. Effects of temperature dependent
thermal
conductivity
on
magnetohydrodynamic (MHD) free convection
flow along a vertical flat plate with heat
conduction, Nonlinear Analysis: Modeling and
Control, Vol.13, No. 4,pp.513-524, (2008).
[3]
Rahaman, M.M and Alim, M. A. Numerical

study of magnetohydrodynamic (MHD) free
convective heat transfer flow along a vertical
flat plate with temperature dependent thermal
conductivity, Journal of Naval Architecture
and Marine Engineering, JNAME, Vol.6,
No.1, pp.16-29, (June, 2009).
[4]
[5]
[6]
Nasrin, R. and Alim, M. A., Combined effects

of viscous dissipation and temperature
dependent
thermal
conductivity
on
magnetohydrodynamic (MHD) free convection
flow with conduction and joule heating along a
vertical flat plate, Journal of Naval
Architecture and Marine Engineering,
JNAME, Vol.6, No.1, pp.30-40, (June, 2009).
Alim, M. A., Alam, Md. M., Mamun A.A. and
Hossain, Md. B., Combined effect of viscous
dissipation & joule heating on the coupling of
conduction & free convection along a vertical
flat plate, Int. Commun. In Heat & Mass
Transfer, Vol. 35,No.3, pp. 338-346,(2008).
Alim, M. A., Alam, M.and Mamun, A. A.,
Joule heating effect on the coupling of
conduction with MHD free convection flow
from a vertical flat plate, Nonlinear Analysis:
385
Modeling and Control, Vol.12, No. 3,pp.307316, (2007).

[7]
Merkin J.H and Pop I., Conjugate free

convection on a vertical surface, Int. J. Heat
Mass Transfer, Vol. 39, pp.1527- 1534,
(1996).
[8]
Molla M. Md, Rahman A.and Rahman T. L.,

Natural convection flow from an isothermal
sphere with temperature dependent thermal
conductivity. J. Architecture and Marine
Engineering, Vol.2, pp.53-64, (2005).
[9]
Keller, H. B. Numerical methods in boundary

layer theory, Annual Rev. Fluid Mech. Vol.
10, pp. 417-433, (1978).
[10]
Cebeci, T and Bradshaw, P. Physical and

Computational Aspects of Convective Heat
Transfer, Springer, New York, (1984)

SELECTING OF DIESEL ENGINE FOR HEAVY WEATHER SERVICE

IN INTER ISLANDS TRADE IN INDONESIA SEAWATER
Yusuf Siahaya
Energy Conversion System, Mechanical Engineering
Hasanuddin University, Makassar, Indonesia.
e-mail : yusufsiahaya@yahoo.com Phone/Fax : +62 411 588 400
ABSTRACT
One of the most important systems of the ship is the interaction between ship, machinery and propeller. Another
aspect which influences the hydrodynamics design of ship is the need to ensure not only good smooth water
performance but also that under average service conditions at the sea the ship shall not suffer from excessive
motions, wetness on deck, losing more speed than necessary in bad weather. In rough weather the ships resistance
will be changed by the actions of the waves and the wind, which result in change of the load on the propeller and
usually a reduction of the propeller efficiency. The marine engineering study made here is in relation to the
multipurpose inter insular cargo vessel of 3000 DWT, CARAKA JAYA III design concept. For the purpose of the
design layout of the power plant for CARAKA JAYA III, it is necessary to make route plan for this ship. In this case
two routes are made, one for West Indonesia, and one for East Indonesia. The results of the calculation of the
added resistance for CARAKA JAYA III is 0.3 % to 18 % higher than resistance at design condition for sea state
with Beaufort 1 to 4 and 38 % to 68 % higher than resistance at design condition for sea state with Beaufort 5 to 6.
The propeller load diagram for CARAKA JAYA III shows that with the wind a head of Beaufort 5, the resistance
increase by the factor fRS = 1.35 and maximum ship speed VS = 13 kn, the maximum propeller revolution nPSX = 120
RPM, and the maximum propeller power PPSX = 1526 kW. With a head wind of Beaufort 6 the, resistance increase
by the factor fRS = 1.66 at a maximum speed VS = 12.4 kn., the maximum propeller revolution nPSX = 115 RPM and
the maximum propeller power PPSX = 1477 kW. Therefore the selecting power of engine should have PPMC = 1800
kW, engine revolution nMCR = 600 RPM, and a maximum engine revolution nMAX = 636 RPM. In this paper a
method will be shown to pre-calculate the propeller load and the engine load in heavy weather conditions. With
these values, the design conditions for the propeller and main engine can be made clearer for Caraka Jaya III ship.
Keywords: resistance, propeller and engine power in heavy weather
1. INTRODUCTION
The power increase in heavy weather is caused
by an increase of resistance. The resistance increase
in waves has been the subject of the number of
interesting analytical and experimental studies.
Theoretical studies of the problem have been carried
out by Haverlock [4], Kreitner [5], Maruo [6], Jan
Olaf [7], Lai Trung Tho [ 8], Zhi Shu et al [13] and a
number others. However, those calculation methods
which gave full consideration to the physics of this
phenomenon have not been widely used in technical
practice, because of the great mathematical
difficulties involved. That is why a number of
practical techniques have been developed, which use
certain simplifying assumptions, thus facilitating the
determination of the value of the added resistance
due to the waves. Among such practical techniques,
suggestions have been made by Gerritsma and Beakelman

[1], Grossmann [2] and also by Xiaobo Yan [11].
Voyage reports show that with a headwind of Beaufort
8 the overall resistance of fully-loaded dry cargo
containership increases by the factor fRS = 3.1 and for a
refer containership with almost no containers on deck the
resistance goes up by the factor fRS = 2.35, Grosmaann [3].
The recorded values of power increase factor fPS = 4.45 on
a stormy trip against wind force Beaufort 9 for 24 hour of
containership Norasia Samantha, Grossmann [2].
2. EQUATION OF ADDED RESISTANCE
The added resistance according to the measured data
ship voyage at open ship (Pacific, Atlantic and India
Ocean) of 13 difference ship by Xiabo Yan [10], based on
387
the result of this investigation the added resistance

RAW for open sea equation is derived :
4. PROPELLER CHARACTERISTIC OF CJ III

AT DESIGN CONDITION
R AW open sea C 0 BL0.5VW 2
Using the Wageningen B-Series [10] computer program

at Institut fur Shiffskraftanlagen Technische Universitat
Berlin (TUB) the diagram of the KT factor as a function
of the P/D and advance number J can be determined, and
the diagram shown the maximum propeller open water
will be obtained when KTO/JO2 is tangent to
efficiency,
curve, therefore the values of KTO, P/D, and JO for CJ III
can be found from the following diagram,
(1)
The result of regression analysis by Xiabo Yan the

value of Co is 0.116.
The equation of the added resistance in Indonesia
Sea, the sea state Indonesia's water is restricted water
with low significant wave height at the same wind
speed for open sea. The equation of added resistance
for Indonesia Sea becomes:
R AW restricted sea 0.043 BL
0.5
VW
2.154
(2)
JO = 0.6590 KTO = 0.137

P/D = 0.95
= 0.6600
KTO/JO2 = 0.3994
Equation (2) is general equation which can be used to

calculate the added resistance in irregular waves of a
ship in any kind of restricted water in Indonesia Sea,
Yusuf Siahaya [12]. The equation of the added
resistance in irregular wave base on an extended
Schifrin Method [9] as follows:
.
AW
The equation (3) can be used to calculated the added

resistance in all sea are as, where the spectrum is
known
3. CASE STUDIES
To calculate the added resistance in irregular
waves in Indonesia Sea, CARAKA JAYA III (CJ III)
is chosen as an example for investigation. CJ III is a
multipurpose inter insular for semi container ship of
115 TEU. The principal particulars of CJ III as
follows:
Length Overall (m), LOA
Length between perpendicular (m), LPP
Breadth moulded (m) ,B
Draft moulded (m) , T
Dead weight (tonne)
= 98
= 92
=16.5
=5
=3242
All calculation for CJ III is made for a fixed

propeller design condition:
Ship speed
VSO = 12.5 kn
Resistance
RSO = 111 kN
Effective power PE = 713 kW
Figure 1: KT factor versus P/D and J

The propeller characteristic CJ III in open water can be
expressed as polynomials with the advance coefficient J.
KTS = 0.3960-0.2020J-0.2125J2 + 0.0378J3
(4)
KQS = 0.2578-0.2497J-.0.1666J2 -0.0423J3
(5)
5.
THE ADDED RESISTANCE OF CJ III

ACCORDING TO EQUATION OF RESTRICTED
WATER IN INDONESIA SEA
A method will be shown pre-calculate the propeller

load and the engine load due to sea and wind condition.
Calculation of the resistance increase fRS and the power
increase fPS for added resistance due to heavy weather in
Indonesia's seawater.
Total resistance due to weather condition:
R SS R AW R SOX
388
(6)
EC 2010
Where, RAW
= the addeed resistance due
d to weatherr
A
condition, RSOX = resistan
nce at design condition.
c
Resistancee increase facto
or due to weathher:
fRS = RSS / RSOX =1+RAW / RSOX
(7)
=[KTS/JJS2] / [ KTO/ JO2]
= fESS x DO
Bf
VW (Kn)
RAW
(KN)
1
2.0
2
5.00
3
8.5
4
13.55
5
19.0
6
244.5
0.33
2.44
7.5
20.22
42.2
733.0
The tootal resistance increase due to weather RSSX can

be determ
mined if the added resistaance RAW annd the
resistance at design condition RSO are
a known. Taable 3
shows the value of RSSX at the variationn of ship speedd VS
Propulsionn Efficiency du
ue to weather coondition:
DS
T added resisstance RAW of CJ

C III
Table 2: The
(8)
Where, fESS propulsive increase factor, DO propulsivee

efficiency at design cond
dition.
fES =
DS
DO
= [KTS/KQS
JS/JO]
Q ]/[KQO/KTO]/J
(9)
(10)
fPS = fRS/fES
] [VSOX/VSSX]
E = [PSSX/PSOX]/
Using the added resistan

nce factor fRS, the correlationn
between sllip factor s' and
d the fRS for CJJ III as follows::
4
s' = 5.55857 fRS-.05 - 4.5857
(11)
When the slip factor forr heavy weatheer condition iss

known thaan the new ad
dvance numbeer J for heavyy
weather coondition can bee determined:
JS = JO ( 1 sS ) = s' JO
Where slipp factor s' = ( 1 sS )
(13)
Resistancee at design condition of CJ

C III can bee
expressed as:
RO = aRO
R . VS
Figure 2: Tootal resistance of

o CJ III
(12)
(14)
aRO = RO / VS = 0.71
104 kN/ (kn)
The effecttive power PE is proportionnal to the shipp

speed VS 3 :
PE = aPEE . VS3
(15)
T
resistancee RSSX
Table 3: Total
Bf
1
2
R SSX R AW R SOX
45.33
47.1
10
71.33
73.0
52.5
65.0
87.0
1188.0
78.5
91.0
113.0
1444.0
12
1022.33
104.0
109.5
122.00
144.0
1755.0
14
139.33
141.0
146.5
159.00
181.0
2122.0
(KN )
aPE = PE / VS 3 = 0.365 kW/ (kn)3

- Results of calculation of CJ III at deesign conditionn
are shown in Table 1.
Table 1: Design
D
condition
n of CJ III
R0
n0
KN
rpm
m
111
1055
J0
KT00
KQ0
6590
0.173
36
0.267
T0
Q0
KN
KNm
141
86
0
0.65900
The addded resistancee RAW of CJ IIII in Indonesia'ss

seawater with
w
difference wind speedd are given inn
Table 2.
Figure 3: P
Propeller diagrram of CJ III at
a different Beeaufort
Number.
389
EC 2010
6. THE ADDED RE
ESISTANCE OF CJ IIII
ACCO
ORDING
TO
T
EQUA
ATION
OF
F
SCHIF
FRIN METHO
OD
The addded resistancee of CJ III cann be calculatedd
from equaation (3). Using the compputer program
m
(program zwide bestimu
ung des zusaatzwiderstande))
create by Xiaobo
X
Yan [11] in TUB Beerlin, the addedd
resistance RAW accordin
ng to Schifrinn equation aree
calculated and shown in Table
T
4.
Table 4: Thhe added resisttance of CJ III
g
a greater added resistannce for
for Schifriin Method to give
Bf 5 to 7.
7 It can be saaid that the calculation
c
resuult by
Schifrin Method
M
and restricted water equation (Indonesia
sea waterr) are the saame. Thereforre the equation of
restricted water and Schifrin
S
Methood can be ussed to
calculate the added resistance
r
in irregular waave in
Indonesia seawater. How
wever, the restrricted water eqquation
has advanntages both wiith in theory and
a in practicaal use,
and it is therefore
t
propposed that the added resistaance is
calculated by restrictedd equation (22) can be ussed to
calculate the
t power increeased in heavyy weather in irrregular
waves.
Bf
Vs
4
5
3
RAW
(Schifrin)
W
7. EVAL
LUATION OF
O
SAILIN
NG ROUTE
E
INDON
NESIA REST
TRICTED SEA
A WATER
0.0033
0.24
2.65
5
17.20
400.0
59.0
10
0.0033
0.26
2.72
2
19.00
477.0
71.0
12
0.0033
0.27
2.75
5
20.40
533.0
83.0
14
0.0033
0.28
2.75
5
21.50
600.0
96.0
For thee purpose of design

d
layout of
o the power pllant of
CJ III, it iss necessary to make two routte plans for thee ship.
In this casse two routes are made, onee for West Indonesia
sea waterr, and one foor East Indonnesia sea watter. In
Indonesiann sea water the direction of current and wind
w
is
differ for West Indonesiia and East Inddonesia. Thereefore a
sailing rouute plan is connsisting of twoo sailing periods. At
each part of the route itt is assumed that
t
the ship will
w be
porting at several ports of call in the morning and sailing
s
again in thhe evening. Fiigure 5, showss the western sailing
s
route for CJ
C III, where the
t heavy weather from Deccember
to March.
The values
v
of the added
a
resistancce in irregularr
wave RAW ( Schifrin Method)
M
is show
wn in Figure 4
for differennt ship speed CJ
C III.
Table.2 annd Table 4 are shown the addded resistancee
evaluation of CJ accordiing to equationn (2) the addedd
resistance RAW Indonesiaa sea water (resstricted water),
equation (3)
(
the added
d resistance RAW (Schifrinn
Method). The
T differencee between the two resistancee
increases is small for Bf 1 to 4, but there is a tendencyy
IN
The ship has to saill from Singapoore to Balikpappan in

1616 Km in
i December thhe route is diviided into 3 partts.
S1 = 520 Km
K Singaporee to Jakarta porrt
S2 = 749 Km
K Jakarta to Makassar portt
S3 = 302 K
Km Makassarr to Balikpapann port.
Total distaance = 1616 Km
m
Figure 4: The
T added resisstance RAW of CJ
C III
(Schhifrin Method)
Figure 5: The
T western saailing route CJ III, in Decembber
390
EC 2010
The ressults evaluationn of CJ III voyage in westerrn and

eastern Inndonesia sea water
w
can be compared
c
the results
r
investigatiion of Grossm
mann [2] when Noratian Sam
mantha
on trip against Bf 6 foor 24 hours. The
T voyage reccord
show the behavior
b
of thee real system propeller andd shipconforms very
v
nice to this theoretical.
9. CONCL
LUSION
The aim
m of this invesstigation is to match
m
ship, proopeller
and mainn engine for Caraka Jayya III to suiit the
requiremennts of the ship's service and to
t avoid overlooading
of the mainn engine in heaavy weather coonditions.
REFERENCES
Figure 6: The
T eastern sailling route CJ IIII, in August.
When the whole vo
oyage has to be finished inn
time t = 1666 hours, the whole
w
voyage sppeed should bee
VSM = 1616/166 = 9.73 Km/h.
K
The partt of the route iss
against Bff 5 and Bf 6 and
a the first leeg is against a
current of VCU = - 0.6 kn.
k In Table 5 the factors thee
added resiistance fRS, pro
opeller efficienncy factor fES,
propeller power
p
factor fPS for all westtern routes aree
evaluated.
Table 5: E
Evaluation off route factorss fRS, fES, andd
fPS.
Part Leength
Wind
d
fRS
fES
fPS
Km
Bff
S1
5200
1.28
0
0.95
1.36
S2
7499
1.63
0
0.91
1.76
S3
3022
2.10
0
0.86
2.37
8. PROPE
ELLER LOAD
D AT HEAVY
Y
WEATH
HER CONDIT
TIONS
If the CJ III is saiiling against Bf
B 5 the mainn
engine shoould be constaant speed nMSSX = 598 rpm.
This will give the prop
peller a slip of
o SS = 0.10,
increase the
From thee
t
resistancee factor 1.63.
propeller diagram
d
for CJJ III we get, JS = 0.6050, KTSS
= 0.1980, KQS = 0.0370
0, O = 0.62, and
a the powerr
factor, fPS = 1.76 the shiip speed CJ III is reduced att
same time to VSSX = 12 kn = 0.90 VMAX
y
M , the heavy
weather poower demand of
o the engine is
i PMSX = 17900
kW, and engine
e
torque in this condiition is QMS =
28.60 kNm
m. With this load
l
conditionns the selectedd
engine shoould be, PMCR = 1800 kW, nMCR
= 600 rpm,
M
nMAX = 636 rpm. One off the suitable engines
e
for thiss
purpose is the MaK M 453.
[1] Gerriitsma, J., Beeukelman, N.-V., Propulsiion in

regullar and irreggular waves, Internationall ship
buildding progress (IISP) Netherlannd, (1961).
[2] Grosssmann, G,. D
Design of dieseel engines for heavy
weathher, Institut fur
f schiffskraft
ftanlagen Technnische
Univversitat Berlin (TUB),
(
(1989)..
[3] Grossmann, G,. Simple method
m
of route
optim
mization for sinngle screw mootorship. Institut fur
schifffskraftanlagenn Technische Universitat Berlin
(TUB
B), ( 1989).
[4] Haveerlock, T.H,. The
resistancce of a ship among

a
wavees. Proceedinng of the Rooyal Society (PRS),
(
Uniteed Kingdom (11937).
[5] Jan Olaf Traung, Loss
of speedd in heavy Weaather.

Fishiing boat of the world [2009].
[6] Kreitnner , J., Heavee, pitch and reesistance in seaaway,
Transactions Instituution of Navall Architects (T
TINA),
(19455).
[7] Lai Trung Tho, Weather routing typpe of
recom
mmendations and
a advisories. (2009).
[8] Maruo, H., The exxces resistancee of a ship in rought
r
seas. Internationaal Ship Buildiing Progress (ISB),
(19755).
[9] Schiffrin, Naherunngformeln furr di brechnunng des
zusattzwiderstandess der schiffees in regelmassingen
seegaang. Leningraat 1973.
[10] Wageeningen B-Seriies Computer Program
P
[11] Xiaobo Yan, Beitrag zuur ermittlungg des
zusattzwiderstandess von schiff
ffen im seeggang.
Diserrtation Doctorr Ing (1992).
[12] Yusuuf Siahaya,. Prinsiples of
o designing main
propuulsion plant for
fo motorships for heavy weather
w
service in interr island tradde in Indonnesian
waterr.Disertation Doctor
D
(1992)..
[13] Zhi Shu.,
S
Torgein Mosn., Effect of avoidannce of
heavyy weather on the wave induuced load of ships.
s
Centeer for ships and ocean structur,
s
Norw
wegian
Univversity of sciencce and technology , May 20008
391

A PRACTICAL APPROACH TO TRIANGULATION OF PLANAR SURFACE

Md. Shahidul Islam, Naem-Ibne-Rahman and Md. Mezbah Uddin
Bangladesh University of Engineering and Technology, Dhaka.
E-mail: shahid113@gmail.com, naem_nir@hotmail.com
ABSTRACT
A triangular mesh generation algorithm is presented in this paper. Based on this procedure a fully automatic
object oriented program in C++ language is developed for any arbitrary 2D geometry. The program generates
unstructured triangular surface mesh which can be used for finite element analysis. The program gives mesh
output in a script file format for viewing in AutoCAD. Importance is given to the quality of the triangles
generated so that better results can be achieved by the finite element analysis. A very practical approach has
been taken in this regard. A number of example meshes are presented to show the effectiveness of the algorithm
and program.
Key words: Triangular mesh generation, C++, Object oriented programming, Finite Element Method,
Unstructured surface mesh.
1. INTRODUCTION
Finite element method (FEM) is a powerful tool for
the numerical solution for a wide range of engineering
problems. Applications range from deformation and
stress analysis of automotive, aircraft, building and
bridge structures to field analysis of heat flux, fluid
flow, magnetic flux, seepage and other flow problems.
FEM allows detailed visualization of where structures
bend or twist, and indicates the distribution of stresses
and displacements. FEM software provides a wide
range of simulation options for controlling the
complexity of both modeling and analysis of a
system. FEM allows entire designs to be constructed,
refined, and optimized before the design is
manufactured.
In this method of analysis a complex region
defining a continuum is discretized into simple
geometric shapes called finite elements. The material
properties and the governing relationships are
considered over these elements. An assembly process,
duly considering the loading and constraints, results in
a set of equations. Solution of these equations gives
the approximate behavior of the continuum.
The current research basically develops a program
to discretize a problem domain into triangular shaped
finite elements. Element generation is commonly
known as mesh generation. Triangles have been
chosen as the mode of discretization simply because it
is the most elementary geometrical shape when it
comes to mesh generation. Again, if the number of
total triangles is even, the mesh can be converted in to

quadrilateral mesh.
Automatic unstructured mesh generation is a
relatively new field. Within its short life span it has
tremendous advancement in many diverse fields.
There are a number of triangulation techniques. Of
these the Delaunay method [1] is one of the most
popular techniques of triangular mesh generation. The
Delaunay criterion in itself is not an algorithm for
generating a mesh. It merely provides the criteria for
which to connect a set of existing points in space. As
such it is necessary to provide a method for
generating node locations within the geometry. A
typical approach is to first mesh the boundary of the
geometry to provide an initial set of nodes/points. The
domain bounded by boundary nodes is then
triangulated according to the Delaunay criterion
which involves Voroni diagram [2]. In two dimensions
the Voroni diagram of a group of vertices divides twodimensional space into regions bound by straight line
segments. Each region or Voroni cell belongs to one
vertex. Although the Delaunay criterion has been
known for many years, it was not until the work of
Charles Lawson [3] and Dave Watson [4] that the
criterion was utilized for developing algorithms to
triangulate a set of vertices. The criterion was later
used in developing meshing algorithms by Timothy
Baker [5] at Princeton, Nigel Weatherill [6] at
Swansea, Paul-Louis George [7] at INRIA among
others.
Another very popular triangular mesh generation
algorithms is the advancing front, or moving front
393
method. It was first presented by Lo in 1985 [8]. He

presented a new mesh generation algorithm that
initially defines the boundary of the domain by a set
of line segments, the initial advancing front. All nodes
are generated within the domain bounded by the
initial advancing front and valid triangular elements
are then formed from the line segments and interior
nodes. Each element is generated by joining the end
nodes of a line segment to a third node with the
condition that it does not intersect the advancing
front. The front is then updated and the element
creation process goes on as long as the front is not
empty. Peraire et al [9] presented a modified version
of advancing front method in 1987. Most of the
subsequent advancing front method research has been
based upon Peraires algorithm.
As the triangulation algorithms discussed above are
very complicated, a very simple approach is presented
in this paper which can generate good quality
triangles. The developed program based on this
method requires the boundary nodes of the problem
domain in counter clockwise direction as input. The
main operation then includes selection of the
protruding point, generation of triangles from the
boundary nodes, insertion of internal points in
primary triangles satisfying the rules of triangulation,
generation of triangles from the inserted points, edge
swapping and finally the resultant triangles are
smoothed using Laplacian smoothing [10] technique.
After completing these operations the final resulting
mesh is presented in a script file format for viewing in
AUTOCAD software. Nodes and points represent
same entity in this study.
2. THE PROPOSED TRIANGULATION

PROCEDURE
2.1 Properties of Triangulation:
Irrespective of the technique followed for the
triangular mesh generation, the resulting triangulation
must follow some rules for it to be a correct one.
These are rules derived from simple geometry which
gives the relation between number of vertices, edges
and triangles generated. If the number of triangles,
edges, internal edges, boundary points and internal
points are represented by |T| , |E|, |Ei|, |Pb| and |Pi|
respectively, then these properties are:
domain are given in the anti clockwise direction. Thus

by joining these boundary nodes the geometry of the
problem domain will be established.
Since the geometry is defined by boundary points
and lines joining those points, the resultant domain
can be of two types: (1) convex region and (2)
concave region (Figure 1). A convex region is one in
which all the interior angles are not greater than 180.
On the other hand a region is said to be concave if any
of its interior angles are greater than 180. This
program is capable of triangulating both the convex
and concave regions. For our current analysis the
problem domain does not contain any hole.
(a) Convex region
(b) concave region
Figure 1. Problem domain
2.3 Steps of Triangulation

Our presented triangulation has five steps and these
are as follows:
Initial triangle generation
Internal point insertion
Further triangulation
Edge-swapping
Laplacian smoothing
All these steps are described below.
2.3.1. Initial Triangle Generation

In this step, initial triangles are formed from the
given boundary nodes. If there is n number of
boundary nodes then the number of initial
triangles will be n-2. In order to understand the
initial triangle generation process first it is important
to be familiar with the concept of the Proper triangle
and the Protruding Point.
Proper Triangle:
The vertices of any initial triangle will be three of
the input boundary nodes. A proper triangle is defined
as one that does not contain any other boundary nodes
except for its three vertices. The concept is illustrated
in Figure 2.
|T| = 2 |Pi| + |Pb| - 2

|E| = 3 |Pi| + 2 |Pb| - 3
|Ei| = 3 |Pi| + |Pb| - 3
2.2 The Problem Domain

The triangulation process starts with the definition
of the problem domain. The boundary nodes of the
394
(a) Improper triangle
(b) Proper triangles
Figure 2. Proper and improper triangles
It is worth mentioning that improper triangles can

only be produced in concave domains.
Protruding Point:
Protruding points are defined as the points that
produce proper triangles within the limits of a defined
domain. If Pi is a protruding point and Pi-1 and Pi +1 are
points before and after Pi respectively, then the
triangle produced by Pi is Ti ={ Pi, Pi-1, Pi+1 }. The
condition for a point Pi to be a protruding point is:
The interior angle of the problem domain at Pi must
be less than 180.
Ti must not contain any other boundary node of the
problem domain other than Pi , Pi-1 and Pi+1.
The above mentioned conditions determine whether
a point is a protruding point or not. Using the proper
triangle and protruding point concept the initial
triangulation is done. Figure 4 shows the resultant
mesh of the region bounded by points A, B, C, D and
E.
2.3.2 Internal Point Insertion

The next step in the triangulation process is the
internal point insertion. Internal points are important
because the initial triangles may be too large or may
have too small an angle to work with. Points are
generated at the centre of each initial triangle (Figure
3).
of good quality. A quality triangle is one that has

edges of almost equal dimension and has internal
angles of as close to 60 as possible. The triangles
generated so far thus have some room for
improvement. For some of the initial triangles it is
possible to increase the minimum internal angles by
the edge-swapping operation. So far the prominent
geometrical shape in this paper has been triangles. But
to be able to apply this technique the quadrilateral
shape needs to be considered. The technique is
explained below.
Quadrilateral Generation:
For edge swapping it is necessary to generate
quadrilaterals from the triangles that have been
generated so far. This is not a very tough job since
every two adjacent triangles form a quadrilateral
where the common edge is a diagonal of that
quadrilateral.
Even though every two triangle will form a
quadrilateral, not all the quadrilaterals present in the
domain will need edge swapping. In fact edge
swapping will not be possible for quadrilaterals
having an internal angle greater than or equal to 180.
Thus only the quadrilaterals having all four internal
angles less than 180 will be suited for edge
swapping. The unsuitable ones will have to be left
unchanged.
(a) A suitable quadrilateral.
(b) Not suitable for edge swapping
Figure 5. Suitability of quadrilaterals for edge

swapping
Figure 3. Internal point insertion for the domain

shown in Figure 2(b).
2.3.3 Further Triangulation

In this step, the initial triangles are further
triangulated. The internal point in each of the initial
triangles will divide the parent triangle into three new
separate triangles by joining the three vertices to the
internal point (Figure 4).
Figure 4. Further triangulation of the domain shown

in Figure 2(b)
2.3.4 Edge Swapping

Once triangulation for the second time is
done, there will be no more division of triangles. But
the triangles generated at this point may not look to be
In Figure 5(a), the quadrilateral ABCD is

constructed by the triangles ABC and ADC with AC
as the common edge. Here all the angles A, B, C and
D are less than 180. Thus ABCD is suitable for the
edge swapping operation. In case of Figure 5(b)
triangles EFH and FGH makes quadrilateral EFGH.
But here the angle H is greater than 180. So EFGH
will be skipped from the edge swapping operation.
The procedure of edge swapping is shown by using
the quadrilateral in Figure 5(a).
As discussed above, when a quadrilateral is
constructed from two adjacent triangles, the common
edge acts as a diagonal of that quadrilateral. The
Figure 6 (a) shows the six internal angles of the
triangles ADC and ABC. It is seen that the smallest
angle in this case (X1) is equal to 31. If the diagonal
of the quadrilateral is swapped as in Figure 6(b), the
smallest angle X2 becomes 20. The condition for
edge swap is X2 must be greater than X1. The whole
idea behind the edge swapping is to increase the
395
quality of the triangles by increasing the minimum

internal angles in the quadrilaterals. Thus edge
swapping is not suitable for quadrilateral ABCD.
(a)
Option 1
consideration. The neighboring points of a certain

point will be the points that share the same edge with
that certain point. Each internal node (not the
boundary ones) is moved to a new position given by
the average of the neighboring nodes.
Figure 8 (a) shows a domain just after triangulation
and Figure 8 (b) shows the same domain after the
application of smoothing process.
(b) Option2
Figure 6. Angle checking for edge swapping

Another important thing that should be kept in
mind during programming is that edge-swapping is an
iterative process. Each time a diagonal is swapped, the
process should be reset for the whole domain again.
Because swapping of a single edge changes a
considerable number of internal angles in the domain.
By iteration process one point will come when none
of the edges in the domain will need swapping. That
will be the end point of the edge swapping process.
For the domain in Figure 2(b), the result after edgeswapping will be as Figure 7.
(a) Before smoothing
(b) After smoothing
Figure 8. Laplacian smoothing
3. THE PROGRAM
As stated before, the triangulation is done in this
study using object oriented programming. Object
oriented programming approaches the problem in a
very organized manner. As a result it is very easy to
construct, maintain and modify for future
developments. It is also easier to understand
compared to other programming approaches. The
programming language used here is C++ as it is a very
well known object oriented programming language
for its relative simplicity and compactness in nature.
Some attributes of the program are given below:
(a) Before swapping.
(b) After swapping.
Figure 7. Example of edge swapping
2.3.5 Laplacian Smoothing

After finishing the edge swapping process, the
problem domain is now discretized into a finite
number of triangles. Thus the domain is ready for
finite element analysis. But the quality of the resulting
triangles may not still be satisfactory enough and it
has been discussed before that, poor quality elements
can give poor accuracy in results. Hence one final step
is required to complete the triangulation. Mesh
smoothing and refinement is a well known method to
improve the quality of a mesh. These methods adjust
the positions of the points in the mesh while
preserving its topology. Laplacian smoothing [10] is
the most popular method for node-based mesh
smoothing. In an iterative manner, it repositions the
points of the mesh by moving each interior node to
the geometric centre of its neighbors. It is often used
because it is computationally inexpensive and easy to
implement.
In this smoothing process the position of a point
and its neighboring points are taken into
Source files:
i.
Triangulation.cpp which contains the
function main()
ii.
Point.cpp
iii.
Line.cpp
iv.
Triangle.cpp
v.
Function.cpp
Header files:
i. Point.h: This contains the class Point. The
functions of this class mainly define a point by
the values of x and y coordinates.
ii. Line.h: This header file contains the class
Line. The functions define lines by the points it
contain. There are also functions that can
determine the middle point of a certain line and
the length of the line etc. There is also a bool
function that can compare between two lines to
say if the two lines are infact the same line or not.
iii. Triangle.h: This contains the class
Triangle. It has functions that define a
triangle by its three vertices and the three edges.
It also contain function to determine the centre of
the triangle and bool function that can
396
determine the common edge shared by two

adjacent triangles.
iv. function.h:
This header file includes
functions that are needed for other general
purposes of the program. The most important of
these is a function that determines whether an
input domain is a convex or a concave one.
Another important function determines whether a
certain point lies within a triangle or outside it.
First the input data are typed in the text file
input.txt. The program stores these data in a vector
named point_vector. Then initial triangles are
generated and the protruding points are stored in
a_point_vector. Upon completion of the initial
triangle generation process all the initial triangles are
stored in triangle_vector. Then interior points
are
inserted
in
each
member
of
the
triangle_vector and the internal points are
inserted in updated_point_vector. Next
further triangulation is done and the resulting triangles
are kept in updated_triangle_vector. Then
edge-swapping and smoothing operation are
conducted
in
updated_triangle_vector
consequently updating the triangles in it. Finally the
results are scripted in the file output.scr in a
format that enables the resulting triangulation to be
viewed in AutoCAD.
(a) Problem domain
(b) Meshed domain
Figure 11. Model 3

Model 4, 5, 6 and 7 represent convex regions with
variations so that the capability of the program can be
tested for different cases.
(a) Problem domain
(b) Meshed domain
Figure 12. Model 4
4. RESULTS
In this section some example mesh generated from
the given boundary nodes are presented. Model 1, 2
and 3 shows meshing of concave regions. The
variations of domain also have variation in point
distances, especially in model 3. The resulting mesh
shows the effectiveness of the program for these types
of domains,
(a) Problem domain
(b) Meshed domain
Figure 13. Model 5
(a) Problem domain
(b) Meshed domain
Figure 14. Model 6

(a) Problem domain
(b) Meshed domain
Figure 9. Model 1
(a) Problem domain
(b) Meshed domain
Figure 15. Model 7

(a) Problem domain
(b) Meshed domain
Figure 10. Model 2
397
5. CONCLUSION
At present, mesh generation is a very widely
studied topic. Lot of researches is taking place on this
subject. There are numerous methods and modes of
mesh generation being developed recently. This
research is just simply one of those efforts. The main
aim of this research was to develop a simple mesh
generation algorithm in a way that would be easy to
understand to everyone who has simply the
knowledge of elementary geometry.
This program is a work-in-progress. The aim from
the beginning has been to develop our very own FEM
software. Even in the triangulation, there is still a lot
of room for improvement. For example, after the
initial triangulation, if the triangles are first divided
into quadrilaterals and then again the quadrilaterals
are split into triangles and at last using smoothing
may be an alternate procedure. Anyway, a lot of
examples are presented in the result section and the
quality of the mesh is reasonably good. The future
plan is to develop a program that can surface mesh a
three dimensional body which later can be discretized
into hexahedrons [11]. The solid analysis will then be
possible.
[7] George, P. L., Hecht, F. and Saltel, E., "Automatic

Mesh Generator with Specified Boundary",
Computer Methods in Applied Mechanics and
Engineering, North-Holland, vol 92, pp.269-288
(1991).
[8] Lo, S.H.,A new mesh generation scheme for
arbitrary planar domains, International Journal
for Numerical Methods in Engineering, 21(8):
1403-1426, 1985.
[9] Peraire, J., Peiro, J, Formaggia, L., Morgan, K.,
Zienkiewicz, O.C., Adaptive remeshing for
compressive flow computations, Journal of
Computational Physics, 72(2): 449-466,1987.
[10] Hansbo, P., Generalized Laplacian smoothing of
unstructured
grids,
Communications
in
numerical methods in engineering, vol. 11, 455464 (1995).
[11] Islam, Md. Shahidul, Whisker weaving based
plastering algorithm-A new approach to
hexahedral mesh generation, Doctor of
Engineering Dissertation, Yokohama National
University (2005).
REFERENCES
[1] Delaunay, Boris, N., Sur la Sphere Vide.
Izvestia Akademia Nauk SSSR, VII Seria,
Otdelenie Matematicheskii iEstestvennyka Nauk,
vol 7, pp.793-800 (1934) .
[2] Voroni, G., Nouvelles applications des parameres
continues a la theorie des formes quadratiques.
Recherches sur les parallelloedres primitives,
Journal Reine angew.Math, 134, (1908).
[3] Lawson, C. L. ,Software for C1 Surface
Interpolation, Mathematical Software III,
pp.161-194 (1977).
[4] Watson, David F., Computing the Delaunay
Tesselation with Application to Voronoi
Polytopes, The Computer Journal, vol 24(2)
pp.167-172 (1981)
[5] Baker, Timothy J., Automatic Mesh Generation
for Complex Three-Dimensional Regions Using a
Constrained
Delaunay
Triangulation,
Engineering with Computers, vol 5, pp.161-175
(1989).
[6] Weatherill, N. P. and Hassan, O., Efficient Threedimensional Delaunay Triangulation with
Automatic Point Creation and Imposed Boundary
Constraints, International Journal for Numerical
Methods in Engineering, vol 37, pp.20052039(1994).
398

NUMERICAL-EXPERIMENTAL MODELING OF A PLANNING VESSELS

DYNAMICS IN REGULAR WAVES
A. Najafi1, S. Hamzeh2, S. Kalantari3 and M. S.Seif4
1
M.Sc.Student of Mechanical Engineering, Sharif University of
Technology,Tehran,iran;najafi.sharif@yahoo.com
2
M.Sc student of structure of at Amirkabir University of technology,
Tehran,iran;sadegh_hamzeh@aut.ac.ir
3
M.Sc.Student of Civil Engineering, Sharif University of
technology,Tehran,iran;saeed_kalantari@yahoo.com
4
professor of Mechanical Engineering, Sharif University of
technology, Tehran, Iran; seif@sharif.edu
ABSTRACT
In this paper, movements of a high speed craft with flight body have been studied both experimentally and
numerically. The aim of these studies are to compare the results obtained from experimental models and test
results from the numerical equation solving by couple vertical motion and longitudinal rolling using Savitsky
method. The most important issue in numerical modeling of ship flight is determining the conditions of constant
motion and the ships trim. Comparing the results showed that despite the uncertainty of moving conditions, the
corresponding results in the range are not different from each other. Using advanced numerical methods would
obtain to better estimation of movement of the vessels, especially in higher speeds.
Key words: Ship motion - High speed craft - time domain
1. INTRODUCTION
Prediction of vessel motion in waves resulted in
plan operations and implementation of ship research
is always considered. Therefore vessels movements
investigations are considered in design process.
Movement of a ship at sea with regard to issues
such as maximum speed of the ship in the waves,
voluntary slowdown due to increased forces on the
float and involuntary slowdown due to added
resistance, ships route optimization in order to reduce
transport time, fuel consumption and overall cost and
increase safety of vessels based on existing criteria
(calculated acceleration, the occurrence slamming,
stuff conditions on deck etc), is very important. The
first work in this field, based on Strip theory lay out,
to obtain coefficients of the equations of ships
motion.
Each of these methods, mentioned above, have
their limitations and conditions. Numerical methods,
although depended on the experimental results, but
with acceptable accuracy and their low cost, are their
benefits. Research in this area based on two general
potential theories, ([1] Strip theory and [2] boundary
element methods) and complete form of equations
Navier Stokes equations (such as RANS (Reynolds
Averaged Navier-Stokes Equations ([2]).
The main topic in numerical modeling of the ship
is determination of its status (heave and pitch motion)
during the test procedure. This position can be
identified with trim and depression level of ship in
water. It should be noted that changing ships speed

leads to change into these two variables.
In this research Savitsky method and relationships
used to calculate movements and resistance of flying
vessels in linear mode, and resistance movements
float is calculated.
The experimental model conducted with the
dimensions shown in table (2), and a constant wave
velocity and amplitude of various wavelengths is
examined. The aim of this study is investigation on
ship motions affected by input wave profile. Tensile
tests have been conducted in Basin of Sharif
University of Technology.
Both experimental and numerical methods are in
good agreement ships movements prediction.
Experimental methods, used of testing a similar
model, in smaller size.
Therefore, here numerical and experimental study
of a ship with high speed movements and its
interaction with waves has been studied, however this
categorize High Speed Crafts.
Movement of vessels at sea can be clearly
estimated in three categories. First, estimation of sea
environment (wave spectrum). Second, calculating
ships response (ships motion spectrum) and last
ships movement criteria in the sea and their
efficiency. Accuracy in modeling of environment
where ship interferes with is necessary. Appropriate
assumptions and the efficiency of ship (three) motion
based on the criteria, have significant effect on the
results.
399
2. EXPERIMENTAL MODELING
1.6
Tensile test were conducted using regular waves in

the Marine Engineering Research Center, Sharif
University of Technology.
This laboratory is
equipped with a traction system and the wave
generator produced regular, which is able to produce
both regular and random waves. Basin is 2.5 meters
wide, 24 meters long and 1.5 meters height.
For this study a laboratory model of a High Speed
Craft of 70 centimeters long is used. Profile test
model is shown in Table (2).
1.4
1.2
Rao Pitch
1
0.8
0.6
0.4
0.2
0
2.1 Model experiment with waves
0.5
1.5
/L
2.1.1 The effects of wave lengths on ship

movement
To investigate the effect of wave length on ships
movement, ship moved with Froude number equal to
2.250 and constant amplitude 1cm used with range of
wavelengths equal to 1.7 meters until 2 meters.
1.4
Figure 2: Effect of wavelength on the ships pitch

movement, Froude number is equal to 2.250
2.1.2 Effects of waves amplitude on ships

movement
To study effects of wave amplitude on ships
movement, ship moved with constant velocity of 2
meters per second, wavelength varied between 1till 2
centimeters used.
1.2
1.6
Heave amplitude(cm)
Rao Heave
1.4
0.8
0.6
0.4
0.2
1.2
1
0.8
0.6
0.4
0.2
0
0.5
/l
1.5
1.2
1.4
1.6
1.8
wave Apmlitude(cm)
Figure 1: Experimental study on the effects of

wavelength on ships heave motion, Froude number is
equal to 2.250
Experiments showed that heave motion increased
as, wavelength increases. In small wavelengths heave
tends toward zero. Obviously that is due to reduction
in wavelength, the water wills moves slowly. In calm
water heave motion tends toward zero. Waves with
larger wavelengths, range of heave motion tends
towards one, it is resulted because of the alignment of
the ships and wave moves.
In experimental study, showed that pitch movement
increases as wavelength increases. In small
wavelengths pitch tends to zero.

wavelengths on ships heave, Froude number is equal
to 2.25
5
4.5
Pitch Amplitude(degree)
3.5
3
2.5
2
1.5
1
0.5
0
1
1.2
1.4
1.6
Wave Amplitude(cm)
1.8

wavelengths on ships pitch, Froude number is equal
to 2.25
400
3. EQUATIONS OF MOTION
First it is necessary to declare coordinate (x, y, z)
which do not fluctuate with the ship. This coordinate
systemic moving with the same speed as vessel is
moving. When the ship is located in a stable position,
the coordinate origin is considered on center of
gravity of the ship. Z axis is vertical and its positive
direction is upward. X axis is horizontal and the
positive direction is toward the heel. Movement is
time dependent and is shown with k , for example
3 means the vertical movement of centroid and 5 is
longitudinal rolling in terms of radians. Angles of
trim and pitch are positive, if the prow comes up.
Longitudinal and transverse centers of gravity
respectively are showed by lcg and vcg.
Linear couple equations of ships heave and pitch
motions is considered:
d 23
d
d 2
d
+B33 3 +C333 +A35 25 +B35 5 +C355 =F3
2
dt
dt
dt
dt
(1)
d23 d3
d2
d
+B53 +C533 +(I55 +A55) 25 +B55 5 +C555 =F5
2
dt
dt
dt
dt
(2)
(M +A33)
A53
3.1 Moment and restoring force

Linear restoring coefficients in heave and pitch
motions calculated by using the following equation:
F j c
k
F L0 = lift force at angle

of dead rise zero
F L = lift force
w = Average wetted
ratio of length to width
deg = angle of trim in
planning level in term of
degrees
= angle of trim in
planning level in terms of
radians
Lc = wetted length of
Chine
L p = center of
pressure
U = Ship velocity
L = length of ship
vcg , lcg = position of
the center of gravity
Fnv = Volumetric
Froude number
(without speed)
U = Ship velocity
B = width of planning
level
It should be mentioned, that in the term of Froude
number of the width parameter is used instead of
length of the ship. The reason is the constant ratio of
width to instead length the ship. For example, wetted
length of keel before solving equilibrium equations
related the vertical force and trims moment at a
specified speed, is unknown.
Equation 4 in the range of 2o deg 15o and
w 4 is correct and usable. In figure 5 geometry of
M is mass of ship and I 55 is moment of inertia in

pitch movement, considering defined coordinate
system. (Ajk) is added mass. (Bjk) and (Cjk) are
Damping and restoring coefficient respectively. They
all will be provided, in following part of descriptions.
Damping force is due to body lift and restoring
coefficients are due to changes in buoyancy force.
C jk =
C L = lift force
coefficient
the ships body and angles and are specified.

Average ratio of wetted length to width ( w ) is equal
to 0.5 (L K + L C ) / B . LK and LC, are wetted length of
keel and Chine, respectively. In Savitsky approach
body form is assumed to be prismatic, also deadrise
angle is assumed to be constant through the length of
the ship.
According to the Savitsky equation, when the trim
angle approaches to zero, the value of the lift force
tends to zero. Implication of trim angle here is the
same as role of the rake angle in hydrofoil. It also
revealed that the lift force decreases with increasing in
deadrise angle.
(3)
j , k = 3,5
0
Here, zero means the static equilibrium position in

this case 3 = 5 = 0 .
Savitsky approach is used to define center of
pressure, lift force and drag. [4]
CL = CL 0 + 0.0065 CL 0 0.6
(4)
180 1.1 1.1

1.1
CL 0 = deg
0.012W0.5 = (
) 0.012W0.5
(5)
Figure 5: Coordinate (x, y, z) and parameters used in
calculations related to a planning, charter vessel
Savitsky 1964
Where:
C L0 = Coefficient of lift
force with angle of dead
rise at zero
Lk = wetted length of
keel
401
The L P is the distance along the keel of the ship

till heel (transom) to the center of pressure due to
hydrodynamic forces. In equation (5) and (6) force
and moment caused by hydrodynamic lift. Also in
this equation the effect of hydrostatic forces and the
production of surface waves are also considered. [4]
Lp
w B
= 0.75
5.21(
Fn B
w = 0 / 5 (LK + LC ) / B is calculated at
Value of
any time by using equations (8) and (9).

Using equations (5) and (6) to the amount of vertical
static force to, the results are as follows:
C
C33
C
= B L = B L 0 1 0.0039CL 00.4
2 2
3 0
3 0
0.5U B
(6)
) 2 + 2.39
(10)
C
C35
C
=B L =B L0 10.0039CL00.4
2 2
0.5U B
5 0
5 0
Horizontal and vertical distance between the

highest local pressure point applied and ship keel are
respectively equal to Vt + Z max and C = (Vt + Z max) /
tan . The fact that Vt = x would lead to a function
of x axis, thus it is demonstrated that flow separation
from the chine starts at x = x S = L K - L C which
would be calculated by the following equation:
Z
Z
x
B
tV
= (1 + max )
= (1 + max ) s
2
Vt tan
Vt tan
(7)
(11)
Pitch moment around COG can be expressed as

follows:
F5 c
0.5 U 2 B
=(
Lp
B
L cg
B
)C L
(12)
The L P value is achieved by using the equation 6.

Therefore:
Lcg
C L
1 L P
C 53
L
)B
=
B
C L + ( P
0.5 U 2 B 2
3 0
B
B
B 3
Table 1: The amount calculated parameters of Z max
(13)
Lcg
C L (14)
1 L P
C 55
L
)B
=
B
C L + ( P
0.5 U 2 B 3
5 0
B
B
B 5
Vt
for a wedge with constant vertical velocity V

impacted to the water surface [7].
3.2 Added mass in heave and pitch motions

Z max
Vt
0.5695
0.5623
7.5
0.5556
10
0.5361
15
0.5087
20
0. 4709
25
Also it is necessary to declare how
Calculations related to the added mass are based

on the Strip theory. Therefore, added mass coefficient
in the case of two-dimensional heave, 33 for the
wedge, is an important parameter for High Speed
Craft prismatic body. Analytical methods for
calculating 33 have been presented by many
researchers. Faltinsen approach is one of them: [5]
changes
a33 = d 2 K =
with 3 and 5 :
Vcg
Z +
wl 3
Lk = Lcg +
tan( +5 ) sin( +5 )
(8)
Using equation 7 and replacement for , + 5
(1.5 )
d 2
tan( ) sin( ) 2 (1 )(0.5 + )
(15)
Here d is the mug equal to 0.5 b tan , and b is

the wedge width. is gamma function.
For part of the body that xs, since there chine is
wetted, based on topic presented before equation (7)
the amount mug is written as follows:
there is:
Lc = L k x s = L k
0.5 tan
Z
(1 + max )( + 5 )
Vt
(9)
d = C (t ) ta n ( ) = (1 +
Z m ax
)x
Vt
(16)
Total amount of heave added mass is calculated as

followings:
402
xs
A331 = a331dx = K (1+

0
Z max 2 2 x s 2
) x dx
0
Vt
3.3 Damping in heave and pitch
(17)
K according to equation (15) and x s according to

equation 7 is defined as follows:
xs =
(18)
B tan( )
2 (1 + Z max )
Vt
Value of 1+ tan for different deadrise angles

Vt
from Table 1 is derived. After integration these

results achieved:
In this part damping induced by body lift and

heave velocity are considered. Calculations are based
on empirical equation of Savitsky. This equation
includes the hydrostatic part is, and only need to lift
part of the hydrodynamic force and moment. For this
purpose the equations (5) and (6) presented is using C
L0 and L P.
Because of the speed heave, attack angle (trim) will
be changed:
= =
F3 =
Added mass coupled heave and pitch is calculated

as:
xs
xs
A351 = A531 = (x xG )a331dx =K d 2(x xG )dx
xs
xs
U 2 B 2C L
(23)
The velocity acts heave also cause pitch-moment

around COG.Value of this moment by using equation
(6) when the considered Fn B , is calculated.
That will be result in:
(20)
F5 = F3 (LP Lcg ) = F3 (0.75W B Lcg )
(24)
Here, F3 vertical force at a time when Fn B is

considered. When this moment transfers to the left
side of equations of motion, coefficient damping, B53,
is expressed as follows:
That added mass pitch is calculated as followings:
When force term moves to left side of the equations,

damping coefficient, B33 is expressed as follows:
x G = L K Lcg ,
A 3 51
A 5 31
A 3 31 x G
K
ta n 4 ( )
=
=
4
4
3
B
B
B B
6 4 (1 + Z m ax ) 2
Vt
(22)
Angle of attack with lift force is:
(19)
A331 K tan 3 ( )
=
B 3 24 (1 + Z max )
Vt
d 3
V
=
/U
U
dt
A551 = (x xG )2a331dx =K d 2(x xG )2dx
(21)
Lcg
B 53
= 0.75W
B 33 B
B
A 55(1)
X
A (1)
tan 5
tan 4
K
K XG
=
+ ( G ) 2 33 3
B 5 160 (1 + Z max ) 3 3 32 B (1 + Z max ) 2 2
B
B
Vt
Vt
Therefore, using simplified analytical calculations

following results rose:
B) For the part of the wetted chine body d equals to:
B 55 = Ux T2 a33 (x T )
B
d =
tan
2
In mass added phrase integration from x

consequently:
(25)
(26)
B 35 = UA33 Ux T a33 (x T )
s
till LK,
A33(2)
L
= C1 C
8B
B 3
(2)
A55(2) A53(2)
X
L
X A
=
= C 1 ( K )2 ( S )2 + G 33 3
4
4
B B B
B
B
16 B
A55(2)
X
X X
A(2)
L
X L
=C1 ( K )3 ( S )3 C1 ( G ) ( K )2 ( S )2 + ( G )2 33 3
5
24 B
8 B B
B
B
B B B
3.4 Wave-induced force

To evaluate and analyze wave induced force, ship
movement out of water surface is neglected;
otherwise it is necessary to consider slamming forces.
Computation of slamming force of the ship body at
any given moment would be very difficult and
complicated.
3.4.1 Froude-Krilof force

When motions due to waves are considered, it is
necessary to apply wave force on the right side of
403
equation (1) and (2). And

A left side of the equationns
ve steepness iis
remains unnchanged. Asssume that wav
small thuus linear wavve theory can
n be appliedd.
Therefore, the wavelen
ngth is coonsidered greaat
T
calculatioon is assumed to be easier too
enough. This
lead. Colllision wave heeight at any giiven moment iis
expressed as
a followings:
=
s i n ( e t k x )
(27)
Here e = 0 + kU is collision freqquency, 0 iis

wave freqquency in cooordinate systtem which iis
2
connected to the earth
h, and k = 2 = w 0 is wavve
g
number.
Consequenntly vertical Frroude-Krilof foorce is equal too:
F3FK = C 33
3 a sin e t + C 35 k a cos e t
(28)
B35D = B35 + UA33

B 55D = B 35 + UA35
After replacement of coefficiennts and inpput

n and solve thhe
parameterrs in the equaation of motion
basic equ
uation in time t, domain motions of shipps
heave andd pitch comees by adding ships
s
mug terrm
and rannge of heave motion. New
w mug will be
b
developedd. The same prrocedure is appplied to deriveed
new trim term.
Using new mug and trim as input for the prograam
c
in coeffficients, as weell as changes in
leads to change
the stablee force of motiion. Solving th
his new equatioon
in time t + dt leads to new
n
range of heave
h
and pitcch.
t
value to the previous mug and trim
m,
Adding this
would leaad to new mugg and trim. Thee latter comes as
input for next time; thhis procedure continues doinng
ws
until convvergence of thhe answers. Figure (6) show
model ouutput of the tesst model. Inpuut parameters of
the prograam are the folloowing:
And Froud
de-Krilof pitch moment is equual to:
F5 FK = C 53
5 a sin e t + C 55 k a cos e t
Table 2: Model inputts for the test program

p
in the
experrimental tests
(29)
3.4.2 Difffraction forcce

In calcculations relatted to Froude-Krilof forcee,
effect of teerms of velocityy and waves acceleration
a
didd
not consid
dered. Thesee effects are considered inn
diffraction force calculaations. In thee first step thhe
otential functioon should be estimated withh
velocity po
the velocitty potential function
fu
of inpput waves, thhe
flow arounnd the body is
i zero. Thenn pressure and
force on thhe body is calcu
ulated.
F3 = F3s a sin e t + F3c a cos e t
F5 = F5s a sin et + F5c a cos et
(33)
(30)
(31)
ship sppeed (m / s)
ship width
w
(m)
0.15
L
Lcg
1
0.20
Vcg
M
ship length
l
(m)
Longituddinal center of
gravity position
p
(m)
Center of grravity height (m
m)
Ship mass (kg)
Deaddrise angle
20
Amplitude (m)
0.01
Periiod (sec)
2.250
Angle of w
wave approach to
t
thhe ship
180
Here Fjc and Fjs are ex

xpressed in teerms of addingg
) damping coefficients (Bjk) and restoringg
mass (Ajk),
coefficientts (Cjk):
F3 s = C 333 A 33 0 e + B 35D 0 k
F3c = C 355 k A 35 0 e k + B 33D 0
F5s = C 53 A 530e B 55D 0 k

F5c = C 55 k A 550e k + B 53D 0
D
= B33 and
Baased on linear theory B33
D
B53
= B53 . Also:
Figure 6: Model outputt results with characteristics

c
o
of
T
Table 2
404
4.
COMPARISON BETWEEN
EXPERIMENTAL AND NUMERICAL
RESULTS:
6
NumericalResults
Heave Amplitude(cm)
experientalResults
1.2
1.4
1.6
1.8
Wave Aplitude(cm)
Figure 8: Comparison between the results of ships

pitch motion with the conditions mentioned in Table 2
1.4
Numericalresult
1.2
ExperimentalResults
1
0.8
0.6
0.4
0.2
0
0
0.5
/l
1.5

heave motion with the conditions mentioned in
Table 2
1.6
NumericalResults
1.4
ExperimentalResults
Rao Pitch
1.2
1.6
In this section, numerical and experimental

results for the model mentioned in the case compared
with each other. Here the ships velocity and input
wavelength are constant and amplitude of the waves is
varied. Comparing the results of ships movement,
heave and pitch respectively give figures 9 and 10.
Consequently difference between numerical and
experimental results decreases as wave length goes
up. This difference decrement shows up in heave
motion (Figure 9) but in pitch motion, with increment
in wave length differences between numerical and
experimental results show up (Figure 10). Before
wavelength before reaches 1.5m experimental results
and numerical in pitch movement show a good
agreement.
numericalresults
4.2 Comparison between the results with

variable wavelength
1.8
ExperimentalResults
Rao Heave
In this section, numerical and experimental results

for the model mentioned in the case compared with
each other. Here the ships velocity and input
wavelength are constant and amplitude of the waves is
varied. Comparing the results of ships movement,
heave and pitch respectively give figures 7 and 8. It
shows that with increment range of input waves
amplitude, ships movement mode changes from
linear one, nonlinear. Indeed the numerical solution
assumed in linear mode. Consequently difference
between numerical and experimental results increases
as it goes up. This difference increment shows up in
heave and pitch motion (Figure 7) when they reach
1.5 in the range. As the range value goes up, this
difference moves up either. Before data range reach
1.5 there is a good agreement between experimental
and numerical results.
Pitch amplitude(deg)
4.1 Comparison between the results of with

variable amplitude
1
0.8
0.6
0.4
0.2
1.4
1.2
0.5
0.8
0.6
1
/l
1.5

pitch motion with the conditions mentioned
0.4
0.2
0
1
1.2
1.4
1.6
wave amplitude(cm)
1.8
5. CONCLUSION

heave motion with the conditions mentioned in
Table 2
Comparisons of experimental and numerical

results show that numerical solution are in acceptable
range and better results can be obtained by using more
405
accurate models and equations. Therefore, this issue

is important in the ships design procedure.
Obviously, this point should be noted that numerical
results for solving linear mode are acceptable
(numerical assumption). Increasing speed would lead
to nonlinear circumstances such as slamming, wet
deck surface and etc. Increasing the scope of wave
lengths, it is not recommended to use this numerical
method.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
Journee,
J.,
Massie,
W.,
"Offshore
hydromechanics", 1st Ed., Delft University of
technology, 2001.
Worbel, L., Aliabadi, M., "The boundary
element method", John Wiley & Sons, 2002.
OM Faltinsen, Hydrodynamics of High-Speed
Marine Vehicles. Cambridge University Press,
2005. ISBN 0-521- 84568-8.
Savitsky, D. (1964), "Hydrodynamic Design
of Planning Hulls", Marine Technology,
Vol.1, pp.71-95
OM Faltinsen,''Water impact in ship and and
ocean engineering'', keynote Lecture. In: Goda
Y, Ikehata M, Suzuki K (Eds) Proceedings of
the 4 th International Conference on
hydrodynamics (ICHD
2000).ICHD 2000
Local Organization Committee, Yokahama, pp
17-36
N. Salvesen, EO Tuck, OM Faltinsen,''Ship
Motions and sea loads'',
1970.
Transactions SNAME 78, 250-287.
R. Zhao, Odd M Faltinsen,''water emtry of twodimensional bodies'', J. Fluid Mech., 246,593612, 1993.
406

STUDY ON SOME COMPETITIVE PARAMETERS FOR SHIPBUILDING

INDUSTRY IN BANGLADESH
N. M. Golam Zakaria, M.M. Rahaman and Kh. Akhter Hossain
gzakaria@name.buet.ac.bd
ABSTRACT
Despite Bangladesh has glorious history & heritage as a shipbuilding nation since ancient time, shipbuilding
has been in focus only from the last couple of years. The optimism about this industry arose from the success
attained by a number of local entrepreneurs who brought the name and fame to Bangladesh as a country with
great potentials in shipbuilding by building and handing over some ocean-going vessels to overseas buyers.
Since then, the shipbuilding in Bangladesh did not have to look back and now new opportunities are knocking
at the door to flourish this industry further. However, due to global nature of this industry, an assessment of
suitability to modern shipbuilding in terms of global standard is of prime importance. This paper focuses on
studying some crucial competitive factors like labour skill, labour availability, labour man-hour, labour cost &
productivity for local shipbuilding which are the inherent part for expansion of this industry. The analysis also
focuses on the existing access to the resources like materials, knowledge and capital for shipbuilding.
Comparison of these parameters with other nations have been made qualitatively and quantitatively to find the
level of our shipbuilding. On the basis of existing condition and future vision, some recommendations have been
made which might be useful for accelerating the development of this sector.
Keywords: Shipbuilding, Competitive factor, Productivity, R &D, Labour cost.
1. INTRODUCTION
Due to congenial geographical location, the history
of indigenous shipbuilding in Bangladesh dates back
to a long ago. Chittagong port was the best centre of
building ocean going vessels in the middle of the 15th
century and even in the 17th century, the entire fleet of
ships of the Sultan of Turkey was built at Chittagong.
Also in 1805, British Navy built ships at Chittagong
for the famous battle of Trafalgar. However, with the
passage of time, glory of shipbuilding in this region
began to fade. Recently fame of shipbuilding revives
again when local shipbuilders exported some ocean
going vessel to foreign countries. The country now
foresees a huge economic progress into the arena of
shipbuilding industry by utilizing the economic
globalization, free international trade & shipping
boom of the recent time. Entrepreneurs also confident
in the good prospect of flourishing shipbuilding
industry because of its long history of maritime
activity, favorable geographical location and
availability of cheap work forces.
Shipbuilding is truly a global industry. It is also

technological intensive, labor intensive and capital
intensive industry. So, for the development and
expansion of this industry, numerous parameters are
to be considered. Because ultimate goal is to become
competitive relative to other neighboring countries
considering of all the related parameters.
As a labor intensive industry, labor input, labor
availability, labor cost, labor hour etc. are very crucial
to assess the present status and future development of
shipbuilding industry. Although cheap labor is big
advantage for Bangladeshi shipbuilding, however
international competitiveness and market shares shall
not depend on cheaper labour alone as lower
productivity and efficiency may nullify this
advantage. As a technology-driven sector, constant
improvement of engineering skill & adaptation of
new technology will keep one ahead of others in the
competition. Furthermore, the requirement of raw
materials, components as well as appliances for ships
by the industry also play important role for the
sustainability of this sector. On the other hand, the
development of the industry needs large investment
on facilities and infrastructures, which will absolutely
benefit other sectors. Political support is one of main
influencing factors in shipbuilding which is certainly
413
an international business. The downfall of the British

and Swedish shipbuilding industries and the successes
of Japanese and Korean shipbuilding were the direct
outcome of lose or gain political support.
In this paper, an effort has been made to assess
competitive parameters as mentioned above for local
shipbuilding in terms of global standard. For
qualitative and quantitative measurement of the
parameters, methodology described in the next section
has been applied. Based on the analysis of the above
parameters, some recommendations have been made.
2. METHODOLOGY
The present study provides an aggregate overview
and qualification profile of the shipbuilding sector in
Bangladesh. The following general methodology has
been adopted for data collection:
i)
Data were collected through extensive

review of literature, field visits, survey,
consultations
and
meetings
with
Government
authorities,
shipbuilding
associations, existing and upcoming
shipyards and its related stakeholders.
ii)
Private and public shipyards around the

country were visited to collect primary data
[1] and information about the local
shipbuilding practice through interacting
with structured, unstructured and open
ended questionnaires.
iii)
Data
about
shipbuilding
process,
organizational set up, working environment,
management practices and future vision of
local shipyards had been collected to get an
insight in context of work standards and the
requirements of shipyard compared to other
shipbuilding nations.
iv)
Qualitative and quantitative data about

labor, labor hour and labor cost were
collected to assess the labor productivity of
local shipbuilding industries.
v)
Professional participations were conducted

through information, questionnaire and
interviews. Information on different aspects
related to shipbuilding industries from the
stakeholders,
different
government
supporting
organizations,
educational
institutions including various published
sources as needed were collected.
Buriganga, Shitalakha and Meghna. About 20%

shipyards of Chittagong division are located along the
side of Karnapuli river and 6% are located along the
bank of Poshur river in Khulna division and the
remaining 4% are located in Barishal division.
Field data shows that most of these yards have
indigenous method of construction and are not
equipped with the man, machineries and equipment
required to build standard ships. These yards are
mainly engaged in building and repairing of inland
and coastal vessels, up to 3,500 DWT. Recently few
of these yards specially Ananda Shipyard and
Slipways Limited [2], Dhaka and Western Marine
Shipyards Limited [3], Chittagong have attained the
capability to manufacture ships of 10,000 DWT and
they are expanding their facilities to upgrade them up
to 25,000 DWT.
About 7% shipyards have the capability to entertain
quality shipbuilding right now. Another 19%
shipyards will be ready for construction of ships
(small & medium category) of international standard
with some renovation & expansion program. Some
new shipyards are also surfacing to build export
oriented ship.
A number of diversified types of vessels such as:
multipurpose vessel, fast patrol boat, container vessel,
cargo vessel, tanker, dredging barge, ro-ro ferry,
passenger vessel, landing craft, tourist ship, tug,
supply barge, deck loading barge, pleasure
craft/yatch, crane boat, speed boat, deep sea trawler,
self propelled barge, inspection vessel, cargo coaster,
troops carrying vessel, double decker passenger
vessel, hydrographic survey boat, pilot boat, hospital
ship, water taxi, pontoon, etc. are built in various
shipyards in Bangladesh.
The order [4] so far received from overseas
countries includes multipurpose cargo ship 80%,
container ship 7% and other types such as ferry,
landing craft, water taxi, floating reception vessel etc
13%.
STUDY OF SOME COMPETITIVE
PARAMETERS FOR LOCAL SHIPBUILDING
Analysing the competitiveness of shipbuilding from
the viewpoint of access to main resources or
production factors are crucial and in this paper, the
following resources are primarily highlighted:
3. INDUSTRY STRUCTURE
There are more than one hundred shipyards and
workshops in Bangladesh, most of them are privately
owned shipyards. Out of these shipyards,
approximately 70% are located in and around Dhaka
and Narayangong along the side of the river bank of
Manpower input, skill & employment

Wage costs
Raw materials -steel & Components
Knowledge
Capital
4.1 Manpower input, skill & employment

The availability of (skilled) manpower is an
important factor in determining the competitiveness
of the shipbuilding industry [5]. Fortunately, the
414
presence of shipyards in particular for long time with

large pool of workforce is the main strength of
Bangladesh shipbuilding industry. Bangladeshi
workforce is acclaimed to be disciplined, diligent,
hard working, obedient and quick learner and
comparatively cheaper than any other country in the
world [6]. Also availability of white color skilled
manpower in maritime field is an added advantage.
Management professionals, electronic, electrical,
information technology and communication engineers
of the country are capable of performing the
assignment in relation with shipbuilding. Also,
presence of technical institutes with 35,512 intake
capacity of producing skilled manpower with regard
to shipbuilding industry is very useful to support this
industry [4]. In addition to this capacity, presently
Khulna Shipyard Limited is producing 100 skilled
persons per year and Western Marine Shipyard Ltd. is
conducting training course in order to produce
international standard welders, painters, foremen, etc.
every year for quality ship production.
When taking a look at the skill levels in different
shipyards, similar patterns exist although there are
clear differences among different countries. The total
workforce in Bangladeshi shipbuilding can be broken
down in management & administration (7-10%),
graduate engineers (3-6%), technical & skilled
workers including diploma engineers, foreman,
certified welders (13-18%) and remaining portion
belongs to semi-skilled, unskilled workers at different
categories.
Although Bangladesh has a strong base for the
production of workforce, but the alarming picture is
that skilled workforce drain out every year and the
rate is very high as shown in Table 1. For the case of
naval architects, the situation is even worse. As a
result, shipyards are facing difficulties to carry out job
of ship production taking the engineers from other
backgrounds. The only positive side of drain out is
that every year about 10, 000 skilled shipbuilding
workers are contributing to countrys economy by
sending remittance. Shipbuilding related remittance is
about 2-4% of total remittance sending from abroad to
Bangladesh [4].
The ageing of population, demographic change and
negative growth of population is an issue [5] for the
shipbuilding countries like Europe, Japan, Korea,
Singapore etc. and there is a high demand of high
skill labours in coming years. Bangladesh with young
and energetic population has a huge opportunity to fill
the gap of those countries in coming years. Thus, to
capture market share of the growing manpower
industry particularly shipbuilding, Bangladesh needs
to invest more in human capital through education
and training. Training has become a necessity with
the structural shift towards activities demanding
higher skilled and emergence of automation.
The scenario of age distribution as well as education
qualification and experiences based on the data

collection among promising shipyards in the country
has been shown in Table 1.
4.2 Wage costs
Wages have a major impact on the competitiveness
of a shipyard: labor accounts for a large part of the
costs of a ship [7]. The labor share in total production
cost strongly depends on the wage levels and the
labor intensity of the production process. Figures on
this vary between 40-50%, while other sources state a
percentage between 21-23% in Europe and Japan and
19% in Korea. Indian shipbuilding labor costs are
estimated at 8-10% of the total shipbuilding costs [7].
As a result of labor cost competition, shipyards in
high labor cost countries are trying to reduce the manhours required to build the ship. This can be done by
improving facilities, systems and labor productivity.
Automation is important, but improved organization,
systems and product development may all play a part.
The competitive price advantage gained from
plentiful cheap labor played a significant role in the
lifecycle development of shipbuilding industry ( as
shown in Fig. 1) and it will likely to continue to do so
for the foreseeable future. Figure 2 shows the hourly
labour charge for local shipbuilding in Bangladesh
and Figure 3 & 4 show the comparison of labour cost
and productivity with other competing nations around
the world.
Figure 1: Shipbuilding lifecycle around the world
4.3 Access to raw materials: steel & components

For the shipbuilding sector, access to cheap raw
materials is very important in order to keep down
input costs of vessels. The main raw material one
that determines the price of a vessel to a large extent
is steel. Unfortunately, entire amount of steel for
class building has to purchase from outside country
although few local companies are producing limited
class steel plates in the recent time.
On the other hand, despite Bangladesh has a long
experience in shipbuilding, the component and
415
Productivity
Bangladesh
India
United
States
European
Countries
12
10
8
6
4
2
0
Figure 4: Shipbuilding productivity of different countries

around the world
Cost Distribution of Export Ship
25
Percentage of cost
20
15
Local Contribution
Foreign Contribution
10
5
Mark up
Labour and
Manpower
Financing
Cost,
Overhead and
Others
Machinaries,
Equipment
Steel
Propulsion
Plant
0
Design,
Drawing &
Classification
service suppliers/manufacturers for shipbuilding have

not grown in an organized manner and not up to the
quality required for class vessels except for few cases.
For export oriented ships, the proportion for
components/services required for classed vessel is
very insignificant [8] and at present it is only 5-10%.
The concerned shipyards, some local steel companies
together with some local marine component
manufacturers are now entertaining the small orders
that have been placed in Bangladesh. Because of non
availability of organized component and service
supply manufacturer for export oriented ship, they
have either procured from abroad (as shown in Figure
5) or have been engaged themselves for
manufacturing of the same.
On the other hand,
almost entire amount of the components /services as
well as steel plates, structural items required for
construction
of
inland/coastal
vessels
are
manufactured & supplied by various SME & ship
breaking yards.
In the recent year, Govt. is showing special interest
for export oriented shipbuilding, by waving custom
duties on import of raw materials and components
used in shipbuilding, green channel status etc.
Japan
Items
Figure 5: Percentage of Local and foreign Contribution of

Cost of a Sample Export Quality Combine Cargo Ship
Manufactured by Bangladeshi Shipyard
Hourly labour rate for local and class shipbuilding
H ourly R ate
$1.50
$1.00
4.4 Knowledge: R&D and Innovation
Local Ship
Classed Ship
$0.50
C asual Labour
Fitter
H elper
Machine Operator
Welder
Gas C utter
Specialized Welder and

Gas C utter
Specialized Machine
Operator
Foreman
Technician
$0.00
Labour Type
Figure 2: Hourly Labour Charge (in USD) for shipbuilding

industry in Bangladesh
Ita
ly
G
er
m
an
y
U
K
F
ra
nc
e
U
S
A
In
di
a
Ja
pa
n
B
an
gl
C
hi
na
S
K
or
ea
$35.00
$30.00
$25.00
$20.00
$15.00
$10.00
$5.00
$0.00
ad
es
h
Labour Charge in US$
Hourly Labour Charge for Different

Shipbuilding Nations
Name of the Countries
Figure 3: Average Hourly Labour Charges (in USD) for

Different Shipbuilding Nations around the World
Knowledge, R&D and innovation are of strategic

importance for the competitive position of
shipbuilding sector in the long run. Bangladesh
shipbuilding industry is one of few industrial sectors
that have gained international reputation and the
industry experienced a process from domestic
oriented industry to international-based industry in
last few years. Todays shipbuilding industry has been
gradually transformed into technology driven industry
from labor intensive one due to the growing needs to
meet environmental regulations on ship operators,
increased technological requirements of the clients
and extremely enlarged sized of the vessel. The global
market becomes severely competitive to late comers
armed with cheap labor cost and expanded facilities.
Under these circumstances, shipbuilding companies
are seeking effective ways to strengthen
competitiveness in quality, cost and delivery thus
ultimately benefiting the customers. In this context,
not only resources on high value-added ships, but also
much R & D efforts on production automation and
optimization of logistics are being conducted in order
to derive high level functionality, improve
performance, reduced cost and easy maintenance in
ship operation.
Bangladesh has received sizeable amount of
foreign orders in the recent time. To fortify the
position in the global market, shipbuilding yards have
416
to make hard effort to improve productivity and

quality through technological research and
development. Because of the unique nature of shipbuilding like short series production, sales of ship on
the basis of concept design, R & D is an inherent
characteristic of the individual end product.
Absolute R&D expenditures of most of the largest
shipbuilding countries have increased between 2000
and 2006 [5]. In absolute terms, South Korea has been
the country with the highest R&D expenditures (240
million euro in 2006) which is almost 1% of
production value and China takes the second place in
the absolute ranking (140 million euro) while Japan
is placed in the third (92 million euro in 2006).
But unfortunately, shipbuilding industry in
Bangladesh is still in its rudimentary stage because of
the negligence of R& D. Still Bangladesh is lacking
very basic facilities like Towing Tank, Cavitational
Tunnel, Structural laboratory etc. As a result, despite
having full potential of making basic design, we still
did not able to attain in-house design capability. And
also there is no in-house competence of doing testing
and checking.
4.5 Access to capital ship financing & Govt.
support
The demand for ships is also depending on the
availability of ship financing. If financial means are
difficult to obtain, the demand for ships will be low.
The financing of ships differs from other asset based
industries as ships are not just seen as a means of
transport, but also as a speculation and hence ship
finance is generally regarded as a specialist business
[5].
Shipbuilding financing cost in Bangladesh is very
high. Lack of adequate working capital loan, high rate
of interest on industrial and working capital loan
shipbuilding, high bank guarantees and bank
commissions, all lead to a very harsh environment for
the future nourishment of this industry. That is why
Govt. is presently taking a plan to formulate the
following policy regarding shipbuilding;
a)
Provide shipbuilding
advantages terms
with
loans
b) Exemption of export taxes and land rent

c)
Covers up to 50% of working capital
on
d) Incorporate shipbuilding expansion plan as

strategic development program of the
country
5. CONCLUSION
Bangladesh can turn itself into a hub of
shipbuilding in the global context for small and
medium-sized vessels [4]. Bangladeshi shipbuilding is
competitive in terms of low labor cost. Also easily
trainable workers can provide this country with a
decisive edge over other prospective countries.
However, in terms of design, R & D and supportive
industry, present condition is quite poor. Furthermore,
Bangladeshi shipbuilding productivity is rather low
due to inefficient management, which somehow
diminishes the benefits of low labor cost. However,
the major difficulty at the moment is the lack of
working capital.
To become more competitive than other
neighboring countries like India, China and Vietnam
and to attract and motivate new entrepreneurs to try
out their hands in this growingly lucrative sector, the
following recommendations should be met
immediately:
1) Apart from opening up a major new source
of export earnings, the growth of backward
linkages of this industry, opportunities for
employment expansion, higher revenue
earnings,
etc.,
deserve
appropriate
government support.
2) Bangladeshi shipbuilding needs a favorable
regulatory
framework
to
support
shipbuilding and maritime industry.
3) For sustainable development of this industry
immediate steps must be incorporated to
upgrade the level of R & D and to promote
the technological advancement. Basic
infrastructure like towing tank facility,
cavitation
tunnel, modern laboratories
should be set up as a priority basis for
supporting this industry.
4) Low-productivity persists because of low
technology in shipbuilding sector in
Bangladesh. Imparting training to the
workers, technological enhancement through
automation and modern management
practice make a huge improvement in terms
of productivity.
417

REFERENCES
[1] Churchill,
G.A.,
Marketing
Research
Methodological Foundation, Sixth edition, 1995.
[2] Brochure of Ananda Shipyard and Slipways Ltd,
Published by Anada, 2009.
[3] Booklet of Western Marine Shipyard Ltd,
Published by Western Marine, 2009.
[4] Hossain, Kh. A., Evaluation of potential,
prospect and challenge of Bangladesh
Shipbuilding in light of global context, M.Sc.
Engg. Thesis, submitted to Dept. of Naval
Architecture and Marine Engineering, April 10,
2010.
[5] Study on competitiveness of the European
Shipbuilding Industry, final report submitted by
ECORYS within the framework contract of

sectoral competitiveness studies- ENTR/06/054,
October, 2009.
[6] Zakaria, N.M.G., Hossain, K.A. and Islam, M.S.,
SWOT analysis of shipbuilding industries in
Bangladesh and its challenges to become
potential ship-exporting nation, Journal of Ship
Technology, Vol. , No.2, July 2010.
[7] Shenoi, R. A., Ship Production Technology,
University of Southampton, 2007, UK
[8] Zakaria, N.M.G. and Hossain, K.A., Service
providers and supporting industries of
shipbuilding sector in Bangladesh and its impact
on overall development of shipbuilding,
Proceeding of the 13th Annual Paper Meet, Paper
no. MED-38, September 25, 2010, Dhaka.
Table 1: Workforce Statistics in quality shipbuilding yards of Bangladesh
No.
Category of
worker
Age limit in
years
Academic
Education
Trade
Course
Experience
(Yrs)
% of workers
leaving
Bangladesh
1.
Unskilled
15-22
Class V
NA
NA
2.
Semi-skilled
18-25
Class V-VIII
Yes
2-3
10
3.
Skilled
22-30
Class V- SSC
Yes
5-10
15
Item
4.
Specialist
28-40
SSC
Yes
About 10
20
5.
Machine Operator
28-40
Class VIII- SSC
Yes
5-10
10
6.
Electrician
25-35
SSC
Yes
3-8
20
7.
Fitter
30-45
Class VIII- SSC
Yes
10+
15
8.
Welder
18-32
Class VI- SSC
Yes
5+
10
9.
Cutter
18-35
Class VI- SSC
Yes
5+
10
10.
Helper
15-25
Class V
NA
NA
11.
Painter
20-40
Class V-VIII
NA
3+
12.
Carpenter
20-35
Class V-VIII
Yes
3+
13.
TIG & MIG

Welder
32-45
Class VI- SSC
Yes
8+
20
14.
Crane Driver
28-40
Class VI- SSC
Yes
5+
10
15.
Drawing & Draft

man
28-45
SSC & Diploma
Yes
3+
10
16
Diploma Engineer
25-45
Diploma
NA
5+
20
17
Mechanical Engr
25-46
B Sc Engr
3+
25
18
Electrical Engr
25-45
B Sc Engr
3+
20
19
Naval Architect
25-40
B Sc Engr
2+
50
20
Admin Offr
30-50
M Sc
8+
21
Marine Engr
35-55
Class ll or l
10+
50
418

SUSTAINABLE CONTAINER SHIP FOR INLAND SHIPPING IN BANGLADESH

K. Shahriar Iqbal, N. M. Golam Zakaria, M. Imran Khaled and M. Milon Khan
Iqbal.shahriar@gmail.com
ABSTRACT
A project will be sustainable if it shows attractive economic performance, less detrimental to the environment,
and good service quality. This paper was aimed to select a sustainable design of a container ship from four
alternative designs for inland shipping in Bangladesh. In this selection process economic performances of these
ships were estimated as well as their environmental impacts were assessed. Before making the final selection, their
service qualities were also discussed.
Keywords: Economic performance, environmental impact assessment, ship design, global warming
1. INTRODUCTION
2. METHODOLOGY
A design, which is only commercially attractive, will

not sustain for long as it may face newly adopted
regulation to protect the global as well as local
environment. Such regulations often increase the cost of
operation. So its always wise to take these possibilities
into account while planning to launch a new project.
Bangladesh is going to add a new fleet of container
ships in its inland waterways to ease the mobility of
containers from and to the countrys sea ports. To do so
inland container ports and the container handling
facilities are being developed. New container ships are
being owned by some public and private entrepreneurs.
In most of the cases commercial attractiveness of the
ships are being considered.
This paper focused on a methodology to select a
design considering not only the commercial
attractiveness but also the environmental burden
imposed by the ships. The quality of service is also
discussed in this selection process.
Hasegawa and Iqbal [2], Iqbal and Hasegawa [3] and
Iqbal and Shill [4] adopted metholodogy to compare
inland water and road transportation systems to find the
best option for carrying cargo and passenger. This
method included comparison of economical benefit,
environmental burden imposed and service quality
rendered by the transportation systems. In this paper
similar methodology was used to compare alternative
design of container ships. Here the environmental
impact assessment was carried more in depth using the
software SimaPro [5].
Considering the total price of the ships and operating

costs throughout its 30 years life time, net present
values (NPV) [1] of the investments are estimated for a
number of alternative design of container ships suitable
for inland waterways of Bangladesh and aimed to be
operated in Dhaka-Chittagong route. The minimum
freight charge required to carry a container (TEU)
between Dhaka and Chittagong, that is, required freight
rate (RFR) [1], are also estimated and compared to find
the commercial attractiveness. In all the cases 12% rate
of return (RR) on the investment was taken into
account.
Environmental impact assessment (EIA) is the
technical qualitative and quantitative characterization
and assessment of the consequences of absorption and
emission of various materials and substances from and
to the environment. The impact analysis addresses
ecological and human health consequences and resource
depletion and could be divided into three sub-phases:
Classification: sorting of parameters into
environmental effect categories.
Characterization: calculation of the potential
contribution of the environmental loading to each effect
category.
Valuation: assessment of the total environmental
impact of the product life cycle.
In this study SimaPro, an EIA database software, was
used to assess the consequences on the environment
carried by the alternative container ships in their whole
life cycle. The effect categories considered here
included carcinogens, respiratory organics/inorganics,
427
climate change, radiation, ozone layer depletion, aquatic

acidification/eutrophication,
land
use,
mineral
extraction, fossil fuel extraction, etc. These effect
categories have impacts on resources, human health,
global warming, habitat alteration, biological diversity
and other hazards. Assessing the potential contribution
of the ships to all these impact types, total damages
were assessed and compared to find which one will
impose minimum burden to our environment.
Finally discussing the service quality available from
the alternative ships, the best alternative was proposed
for the inland shipping in Bangladesh.
3. MODEL CONSIDERED
Four model designs of container ships were
considered in this analysis. The particulars of the
models are shown in Table 1. Table 2 contains the
amount of some of the materials and energy used in
various phases of the ships life. The figures shown in
the tables were the major input data in this analysis.
4. ECONOMIC PERFORMANCE OF THE

MODEL SHIPS
attain 12% rate of return on investment as first cost of

ship and operating cost, was estimated for carrying 1
TEU container through a distance of 307 km between
Dhaka and Chittagong. Fifteen off-hire days per annum
for maintenance was taken into account in this
estimation. Net present value (NPV), another parameter
for comparison of economic performance, was also
estimated here to find economically the most attractive
ship model. For this estimation 8000 Tk./TEU, which is
the current rate charged by railway, was considered.
Both the parameters were calculated according to I. L.
Buxton [1]. The Results are shown in Figure 1 and
Figure 2.
The ship with hatch-cover showed the best
performance, that is, the minimum required freight rate
of Tk. 6035/TEU and the highest net present value of
Tk. 365,149,000. All with 12 % rate of return on
investment. The ship without hatch-cover was the next.
The worst among the four was the ship with deck
loading and 13.15m breadth. This is because the ship
with hatch-cover had the capacity of 108 TEUs and
comparatively low construction cost. The construction
costs of the deck-loading type ships were comparatively
higher due to its heavy deck construction to withstand
the containers weight.
Considering 30 years life time, required freight rate

(RFR), that is, minimum freight charge required to
Table 1: Particulars of Container Ships Considered
Ship Type
Deckloading
Deckloading
With hatchcover
Without hatchcover
Capacity
(TEUs)
100
80
108
108
Length
(M)
75.4
75.1
75.5
75.4
Breadth
(M)
15.66
13.15
13.15
13.15
Depth
(M)
4.2
4.2
6.2
8.0
Speed
(KN)
10
10
10
10
Engine Power
(KW)
1140
910
1130
1130
Estimated
Price (Tk)*
303,769,114
287,160,688
301,113,375
308,379,139
*1 USD= 69 Tk.
Table 2: Materials and Energy Consumption during Construction, Maintenance and Operation of four selected
container ships
Construction and Maintenance Phase
Operation Phase
Material
Ship Type
Deckloading
Hull
Machinery
Outfitting
Energy
Low alloy
Steel (kg)
Ferrochromium High
Carbon Steel (kg)
Brass (kg)
Electricity
(MJ)
Diesel
(kg)
Lub Oil
(kg)
8.74E5
6.95E4
7.5E3
1.6E6
1.48E7
2.95E5
Deckloading
8.25E5
6.95E4
7.5E3
1.52E6
1.26E7
2.49E5
With hatchcover
8.33E5
6.95E4
7.5E3
1.53E6
1.47E7
2.92E5
Without hatchcover
8.94E5
6.95E4
7.5E3
1.63E6
1.47E7
2.92E5
428
ProceeedingsofMARTEEC2010
RFRat1
12%RRinTaka
7418
6086
6035
WITHOUT
WITH
HATCHCOVER HATCHCOVER
WITHDEC
CK
LOADING&
M
B=13.15M
664
42
WITHDEECK
LOADING
G&
B=15.66
6M
Figure 1. Required freiight rate for carrrying 1 TEU between

b
Dhakaa and Chittagoong by alternatiive container shhips
NPV
Vat12%R
RRinThousandTaka
355584
365149
23371
15
80183
3
WITHDEC
WITHOUT
WITH
CK
R HATCHCOVEER LOADING&
&
HATCHCOVER
B=13.15M
M
WITHDEC
CK
LOADING
G&
B=15.66M
M
Figure 2. Net present vaalue of the cashh flow for carrrying containerrs between Dhaaka and Chittaggong by alternaative
conntainer ships
threee ship modeels. Then, ussing SimaProo, the
envirronmental burrden imposed by the ships were
assesssed. The results were shownn in Figure 4 as the
damaage assessmennt and Figure 5 as the singlee score
of thhe consequencces of environnmental burdeen. In
thesee results the im
mpacts on humaan health, ecossystem
qualiity, climate chhange and the use
u of resourcces are
show
wn. The best opption from the environmentall point
of viiew was the shhip with deck--loading and 13.15m
breaddth and the woorst was the shiip with deck-looading
and 15.66m breaddth. The impaccts imposed by
b two
otherr ships are verry close in maggnitude, with only
o
a
very little better ressult for ship wiith hatch-coverr.
5. ENVIIRONMENT
TAL IMPAC
CT
ASSE
ESSMENT
Environmental impactt assessment of
o the modelss
he software Sim
maPro. Amountt
were carrieed out using th
of some major
m
materiaals and energgy required too
construct and
a operate th
he model shipss were used ass
input to thhe system. Thee percentage of the materialss
which werre considered to
t be recycled was also usedd
as input. SimaPro
S
develo
oped life cyclee model of thee
ship with the inputs an
nd outputs froom and to thee
environmeent. The life cy
ycle model off the containerr
ship with deck-loading
d
and
a 15.66m breeadth is shownn
in Figure 3.
3 Similar mod
dels were geneerated for otherr
429
1 p
LIFE C Y C LE O F
SH IP WIT H
DEC K LO A DING
100%
1 p
SH IP WIT H
DEC K LO A DING
WIT H B=1 5 .6 6
4 .1 1 E7 kg
LA NDFILL
SENA RIO FO R
SH IP
291%
1 .0 2 E6 kg
C O NST RUC T IO
N WIT H
B=1 5 .6 6 M
4 E7 kg
O P ERA T IO N
8 .5 9 %
8 .7 E5 kg
SH IP S H ULL
4%
8 .4 8 E5 kg
Steel, electr ic,
un- and
low-alloyed, at
3 .9 6 %
-1 9 1 %
282%
6 .9 5 E4 kg
MA C H INA RY
P A RT S
1 .2 E5 kg
O UT FIT T INGS
4 .1 7 %
0 .6 2 7 %
5 .8 E4 kg
Fer r ochr omium,
high-car bon,
6 8 % C r , at
1 E4 kg
Fer r ochr omium,
high-car bon,
6 8 % C r , at
3 .4 5 %
0 .5 9 7 %
4 .2 E7 kg
Diesel,
low-sulphur , at
r efiner y/C H S
281%
1 .0 5 E5 kg
M A INT ENA NC E
WIT H
B=1 5 .6 6 M
4 .1 1 E7 kg
Recycling EC C S
steel B2 5 0
0 .5 1 1 %
-1 9 1 %
6 .2 E4 kg
Lubr icating oil,
at plant/RER S
0 .7 1 %
-3 .7 E7 kg
EC C S steel
sheet
-3 8 3 %
7 .5 E3 kg
Br ass, at
plant/C H S
0 .5 1 4 %
1 .5 6 E6 MJ
Electr icity,
medium
voltage,
1 .0 3 %
1 .6 4 E6 MJ
Electr icity, high
voltage,
pr oduction
1 .0 6 %
1 .6 6 E6 MJ
Electr icity,
pr oduction mix
UC T E/UC T E U
1 .0 5 %
Figure 3. Life cycle of container ship with deck loading and B=15.66M
430
120
110
100
90
80
70
60
50
40
30
20
%
10
0
-10
-20
-30
-40
-50
-60
-70
-80
-90
-100
-110
Human health
Ecosystem quality
LIFE
LIFE
LIFE
LIFE
CYCLE
CYCLE
CYCLE
CYCLE
OF
OF
OF
OF
SHIP
SHIP
SHIP
SHIP
Climate change
Resources
WIT H DECK LOADING & B=15.66M

WIT H DECKLOADING
WIT H HAT CHCOVER
WIT HOUT HAT CHCOVER
Comparing 1 p 'LIFE CYCLE OF SHIP WIT H DECK LOADING & B=15.66M', 1 p 'LIFE CYCLE OF SHIP WIT H DECKLOADING
Figure 4. Damage assessment of four model ships life cycle
4.2
4
3.8
3.6
3.4
3.2
3
2.8
2.6
2.4
2.2
2
1.8
kPt
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
-1.8
LIFE CYCLE OF
SHIP WIT H DECK
Human health
LIFE CYCLE OF
SHIP WITH DECKL
Ecosystem quality
LIFE CYCLE OF
SHIP WIT H HAT C
Climate change
LIFE CYCLE OF
SHIP WITHOUT
Resources
Comparing 1 p 'LIFE CYCLE OF SHIP WIT H DECK LOADING & B=15.66M', 1 p 'LIFE CYCLE OF SHIP WITH DECKLOADING
Figure 5. Single score of the consequences of environmental burden of four model ships
431
For loading and unloading of the containers, deckloading type ships will be favourable. On the other hand
ship with hatch-cover will protect the containers from
adverse weather condition. Among the models
considered here the ship with and without hatch-cover
have more capacity of carrying containers, which is 108
TEUs.
for comparison of different projects. The reasons behind

these uncertainties are usually due to,
uncertain data,
uncertainties on the correctness of the model,
uncertainties caused by incompleteness of the
model,
different opinion on weights of various impact
categories, etc.
7. CONCLUSION
REFERENCES
Analyzing three different criterion, that is, economic

benefit, environmental burden imposed and service
quality, the ship with hatch-cover would be the best
option among the models of the container ships
considered. The following reasons were in support of
this option,
it would require minimum freight rate to attain
specific rate of return on investment,
it would ensure maximum net present value of
the total cash flow in its life time,
though the ship with deck-loading and 13.15m
breadth showed minimum damage to the
environment, the ship with hatch-cover got
more capacity of carrying 108 TEUs
containers.
the loading - unloading facility is in favour of
deck-loading type container ship, but the ship
with hatch-cover would protect the containers
from damage occurred by bad weather.
1.
There are uncertainties in such analysis. One should

be aware of these uncertainties while using such model
5.
6. SERVICE QUALITY
2.
3.
4.
432
Buxton, I. L.; Engineering Economics and Ship

Design. Book published by British Ship Research
Association, UK, 1976.
Hasegawa K and Iqbal K S; Inland Transportation
System Planning by Life Cycle Impact Assessment:
A Case Study. Journal of Marine Science and
Technology, Springer-Verlag Tokyo, Vol. 5, pp. 18, 2000.
Iqbal K S and Hasegawa K; Inland Transportation
System Planning by Life Cycle Impact Assessment:
A Case Study, (second report: single comparison
index). Journal of Marine Science and Technology,
Springer-Verlag Tokyo, Vol. 6, pp 83-92,
December 2001.
Iqbal K S and Shill S K; A Comparative Study of
Two Passenger Transportation Systems between
Two Bangladeshi Cities. Jounral of Naval
Architecture and Marine Engineering, pp. 1-10, vol.
2, no. 2, 2005
PRe Consultants, SimaPro (Environmental Impact
Assessment Database Software), www.pre.nl
RECONSTRUCTION OF CAPSIZE TYPE OF ACCIDENTS BY FAULT TREE

ANALYSIS
M.T. Hossain1, Z.I. Awal2 and S. Das1
1
Undergraduate Student
Bangladesh University of Engineering and Technology (BUET)
2
Assistant Professor
Bangladesh University of Engineering and Technology (BUET)
ABSTRACT
This paper describes a research work on the reconstruction of capsize of marine vehicles using fault tree
analysis. Capsize of marine vehicles are observed to take place all around the world; however, such accidents
occur at an overwhelming rate in the inland waterways of Bangladesh. This paper, therefore, aims to illuminate
the chains of faults working behind the high frequency of capsizes with a view to execute groundwork for future
database development as well as accident analysis and reconstruction. Different factors triggering the accidents
are enumerated in a structured form after scrutinizing the accident data. During accident investigation, the
chain of faults squaring with the accidents can be easily identified with the help of an interactive system, being
directly linked with the accident factors featured in the fault trees. A case study has been shown where the
application of the developed tree is explained. Since the research has the limitation of being primly based on
secondary data sources, further research is recommended to refine it with the help of primary data. Finally
some steps to improve the inland water transportation sector are recommended in order to restrict the factors
contributing to marine accidents.
Key words: Fault tree analysis, hazard identification, accident database development, human factor.
NOMENCLATURE
ARI Accident Research Institute
BIWTA Bangladesh Inland Water Transport
Authority
DOS Department of Shipping
ERZ Economic Resource Zone
FTA Fault Tree Analysis
IWT Inland Waterway Transportation

MAIB Marine Accident Investigation Branch
MARS Marine Accident Reporting Scheme
MOS Ministry of Shipping
NSPAR National Strategy for Accelerated Poverty
Reduction
FAULT TREE ANALYSIS EVENT SYMBOLS

Primary Event Symbols
EXCLUSIVE OR - Output fault occurs if
exactly one of the input faults occurs.
BASIC EVENT - A basic initiating fault

which requires no further development.
UNDEVELOPED EVENT - An event which
is not further developed either because it is
of
insufficient consequence or because
information is unavailable.
Gate Symbols
Transfer Symbols
TRANSFER IN - Indicates that the tree is
developed further at the occurrence of the
corresponding TRANSFER OUT (e.g., on
another page).
TRANSFER OUT - Indicates that this
portion of the tree must be attached at the
corresponding TRANSFER IN.
AND - Output fault occurs if all of the input

faults occur.
OR - Output fault occurs if a least one of the
input faults occur.
1. INTRODUCTION
Bangladesh, a country having the area of 1,47,570
square kilometers, is inseparably entangled with a
circuitous network of around 700 rivers [1] covering a

length of 24,000 kilometers which is almost 7% of the
445
surface of the country. Bangladesh has about 9,000

square kilometers of territorial waters with a 720
kilometer long coast line and 20,000 square
kilometers of economic resource zone (ERZ) in the
sea [2]. She has a navigable waterway of 5,968
kilometers in length.
The major three river systems of the Ganges, the
Brahmaputra and the Meghna converge in Bangladesh
to form one of the worlds largest deltas. These three
rivers also generate one of the most complex river
systems in the world [1]. The complex water network
has spread in such a pattern that a huge portion of the
country especially the southern tip and the offshore
areas are hardly accessible by land transportation.
Therefore, the transportation infrastructure of the
country depends largely on waterways.
The IWT sector plays a vital role in the transportation
sector of Bangladesh. A substantial portion of rural
population (12.3%) only has a reasonable access to
the transportation system through IWT which is half
of all rural households who have access to river
transport [3]. Thus IWT makes a significant
contribution towards governments effort towards
growth and reduction of poverty under NSPAR [4].
With the advancement of industrialization, the
shipping industry of Bangladesh is also being
developed markedly. Being no longer confined to
indigenous pattern of country-boat and wooden ship,
Bangladesh observed a sheer expansion of shipping
business and also emerged as a ship exporter. To keep
the economy smooth-running, a safe and secured
marine transport system should be confirmed. But
disruption to marine transport system as well as to
growing commerce is mainly fetched by the high
frequency of marine accidents.
An investigation data showed that more than 9000
people have died or been reported missing in the past
25 years due to passenger ferry accidents [5]. These
were ascribed to various resources such as
newspapers, journals and personal communications to
the relevant offices (e.g., BIWTA and DOS,
Bangladesh).
Overloading and crowding at one side may be held
accountable for a great number of capsize events in
Bangladesh. Disaster lurks when the problem of
overloading is associated with structural faults. The
IMO criteria applied to the vessels showed that the
crowding of passengers on one side during a crisis
posed a greater risk than the beam wind. During
cataclysms, people run amok at one side and
simultaneously stability of the vessel gets harmed by
the storm; in most of the cases the vessel at this stage
gets overturned by the combined effect of crowding of
people at one side and storm [6].
Since the existing accident database of Bangladesh
has deficiencies, this paper focuses on using the FTA
as a tool of developing an improved database. For
accomplishing this, previous research papers provided

with accident factors and the data on previous
accidents from different authentic sources are carried
out. Keeping a categorized top-level system failure at
top position, the branches of its immediate earlier
factors are arranged in a descending order which end
up with the initial faults.
In Bangladesh, very few works had been done
regarding an in-depth analysis of marine accidents.
Out of them, most were limited to the identification of
major types of accidents, the stimulating factors and
their statistical distributions. The hazards involved
with the indigenous marine vehicles in Bangladesh
were featured extensively by Awal [7]. Another
investigation by Awal et al. revealed some interesting
findings portraying the major causes of passenger
vessel accidents [8]. Awal et al. in a similar kind of
research produced a statistical study of the primary
causes of accidents of various types of vessels [9]. An
investigation of accidents, executed by Zahanyar and
Haque, enumerated the factors triggering marine
disasters and proposed some actions for the
prevention of accidents [10]. Bangladesh Transport
Sector Study (BTSS) categorized the marine accidents
highlighting the major types of launch disasters [11].
A significant contribution was made by BIWTA,
conducting the safety and stability parameters of the
passenger vessels [12].
Furthermore, some significant works were done
regarding the improvement of stability to enhance the
probability of ship survival. Iqbal, et al. performed a
profound analysis on the structural and other faults
including the diminishing stability at lower angle of
inclination which contributes a lot to capsize. This
paper also featured a graphical representation of
significant heeling moment due to wind and crowding
of passengers [6]. Addressing the stability hazard
issue with restrictions like shallow draft and restricted
hull size, Podder suggested the use of sunken deck to
increase stability [13]. Khalil and Tarafder proposed
some adaption of ship design for the acquisition of
extra initial stability by upward shift of centre of
gravity to prevent the vessel from capsizing [14].
Iqbal, et al. devised an efficient way to increase the
restoring lever at large heeling angles by providing
additional reserve of buoyancy through the addition of
totally inflatable lifting bags under the rigid fender of
the vessel [15].
Aside from these, there are some other works being
equally applicable in Bangladeshi perspective.
Concluding the factors behind the loss of stability,
Kruger, et al. highlighted capsizing probability as a
function of metacentic height of the vessel and
recommended a demand for the introduction of
additional criteria to the IMO intact stability code,
accounting for dynamic phenomena related to the
behavior of ships in rough weather [16]. Addressing a
446
similar issue, Umeda and Ikeda proposed a more

physics oriented approach in stead of the classical
approach to use capsizing probability as a relative
measure for establishing stability criteria [17].
sources to develop a good concept on capsize

triggering factors in Bangladesh.
These works hashed out different features of capsize

in Bangladesh distinctly. No work has been done yet
to find the relationship among the attributes resulting
in a fiasco. As a result, a good compilation of events
triggering capsize is still elusive. To serve the purpose
of obtaining a scientific and practical accident
database system through comprehensive analysis of
accident events, FTA is a remarkable tool to start
with. In a book named Fault Tree Handbook with
Aerospace Applications, Vesely, et al. propounded
that FTA provides visual, logic model of the basic
causes and intermediate events leading to the top
event. The qualitativeinformation obtained from FTA
is of equal importance to the quantitative information
provided [18].
The accident database tool is developed with both the

primary and secondary data. The primary data are
collected through a structured observation method
[26], unstructured interview of the personnel at the
corresponding expertise and survey with a
questionnaire form for collecting data from the crews
and regular passengers of IWT sector.
Although no work has been done on Bangladeshi

perspective by using FTA, some valuable researches
were
accomplished
on
other
perspectives.
Amrozowicz, et al. applied FTA on grounding and
represented a probabilistic risk assessment [19]. Basu
and Bhattacharya applied FTA on the system failure
of floating structures particularly featuring structural
failure and compartment flooding [20]. Presenting a
fault tree on capsize issue, Webb performed risk
assessment and reviewed human factors related to the
risk of capsize during design and operation [21].
One of the paramount issues about capsize is human
factor. Rothblum suggested that the lions share of
accident factors is involved with human error [22].
Enumerating the human factors successfully Psaraftis,
et al. recommended putting more emphasis on human
factor than technological solution for reducing the risk
of marine accidents [23]. Talley, et al. concluded that
human mistakes result in higher number of fatalities
in passenger vessel accidents than environmental and
vessel related causes [24]. Calhoun addressed the
issue of shipboard operator fatigue and proposed
incorporation of human factor to ship design to
prevent fatigue as well as reduce human error to a
minimum level to ensure marine safety [25].
Based on the knowledge, developed from these
research works and the evaluation of accident data,
derived from various authentic sources, FTA is
applied on capsize, to assess the impact of an
intermediate event on the final failure.
3. DATA COLLECTION
The pivotal part of the research work is data
collection from legitimate sources and winnow the
pertinent issues with extensive accuracy. But, lack of
authentic data sources as well as shortage of extensive
analysis of capsizes makes the data processing
extremely difficult. So data are collected from various
3.1. Data Source
The prime source of secondary data is the research

paper published in various journals and conference
proceedings. Others are the accident data from
secondary sources, compiled at ARI and various
publications of the governing bodies such as BIWTA,
MOS and DOS. The data are also collected from the
reports of various newspapers. In order to conduct
case study, 3 accident investigation reports by MOS
as well as 26 capsize reports published by MAIB and
MARS were collected and studied to gain a proper
insight to the factors contributing to capsize.
3.2. Some Findings Regarding the Current Status
of Marine Safety in Bangladesh
Structured observation, unstructured interview and
literature study revealed some facts such as
overloading together with discrete and non-uniform
passenger flow, which enhances stability hazard.
During the main public holidays such as Eid and Puja,
a massive population is transferred from Dhaka to the
other parts of the country within a very short period.
This creates a huge pressure on the transportation
system as the means of transportation are limited. So
overloading is a prevalent issue here.
It is important to note that the vessels are kept
overloaded not only during this particular period but
also all over the year, even though there is no
waterway transportation crisis. This accounts for the
absence of prior ticketing system and insufficient
steps taken by the authority to prevent overloading.
Due to absence of prior ticketing system the actual
number of passengers onboard remains unknown
which allows continuing overloading more easily. In
addition, the incompetency between BIWTA and
DOS allows this phenomenon keep going easily. DOS
is directly responsible for ensuring ship safety by
preventing irregularities but it has insufficient man
power while BIWTA has it somehow with 22 officials
only but does has deficiency to enforce the orders.
The most alarming fact is the scarcity of trained,
skilled and competent crews for the safe operation of
the vessels. Deck Personnel Training Center (DPTC)
has been unable to provide sufficient number of
trained deck personnel to meet the current demand
which coerces the ship owners to rely on unskilled
and untrained crews. This allows the vessels to be
447
plied by inept masters and crews which often results

in a pitfall [5].
4. DOMINO THEORY OF ACCIDENT

CAUSATION
Domino theory was developed by Heinrich. He
devised a five-step accident sequence in which each
factor would actuate the next step just like a row of
falling dominoes [27]. The sequence of accident
factors were as follows:
1. Ancestry and social environment.
2. Worker fault.
3. Unsafe act together with mechanical and
physical hazard.
4. Accident.
5. Damage or injury.
Heinrich believed that removing a single domino from
the row would prevent the accident. According to
him, the key domino, to be removed from the
sequence, is number 3. This paper works on this step
by analyzing the accident causes and arranging them
in a chain of events leading to the top-level system
failure.
5. DEVELOPMENT OF FAULT TREE ON

CAPSIZE
Capsize, a top-level system failure, is intensely related
to buoyancy and stability of the vessel. These two
factors fluctuate as result of interaction with other
influences such as ship design, loading, ship steering
and so on [21]. Vessels capsize due to C1- loss of
stability and C2- no restoration of stability.
Stability loss is discussed into two parts: SS- static
stability loss and DS- dynamic stability loss. Static
stability of a vessel is lost due to SS1- reduced
buoyancy or SS2- reduced transverse stability.
Reduction in buoyancy is a function of FLD- flooding
or LoE- overloading, while transverse stability of a
vessel reduces as a result of SS21- GM/GZ reduction
or SS22- heeling moments. Improper opening and
poor drainage system keep a ship vulnerable to
flooding and the problem becomes severe during
cataclysms while the ships stability is harmed by
extreme beam wind. GM/GZ reduction is prior to
poor loading or overloading and any of the events
from FSE- free surface effect of liquid within the hull
or SS221- cargo shift or CRWD- rush of passengers
in one place results in producing heeling moments
vulnerable for the ship stability. During storms,
people stampede towards one place while the ship hull
is exposed to a severe beam wind. The concurrent
effect of severe beam wind and rush of people in one
side triggers heeling moment as well as reduction in
transverse stability and subsequent loss of static
stability.
Dynamic stability loss is initiated by DS1synchronicity of the vessel with waves or DS2- large
rolling angle or DS3- large acceleration. A large

rolling angle may lead to the total loss of the vessel
and a large acceleration causes massive cargo loss or
damage, severe damage to machinery or major safety
relevant systems and structural overload of safety
relevant members. It is seen that large accelerations
occur at high values of initial GM and there is no
particular criteria to limit the stability accordingly
(maximum GM limits) [16]. It is observed on the
vessels of Bangladesh that, the vessels have high
initial meta-centric height but GZ diminishes terribly
while the ship lists at larger angles [7]. Large rolling
angles occurs either at low values of initial GM or
during broaching situations [21]. Synchronicity with
waves occurs due to ROL- excessive rolling or Envresting on steep wave for a certain time. Excessive
rolling is ascribed to ROL1- insufficient roll damping
or DCE- critical resonance or DCE- very low initial
GM. Insufficient roll damping is liable to the selection
of unsuitable rolling reducer or the lack of sufficient
number of rolling reducers.
C22- insufficient or C21- delayed actions to restore
stability during high time may capsize the vessel.
Insufficient action is the result of Del- delayed action
or Err- erroneous action.
The whole capsize mechanism is illustrated in the
figure I.
6. APPLICATION OF THE DEVELOPED

FAULT TREE
In 27 November, 2009 at 10:45 PM a passenger ferry,
MV KOKO-4, capsized causing death to 81 people.
The sequence of the causes behind this accident is
demonstrated in the figure II.
From the investigation report [28], it was seen that the
vessel was extremely overloaded. Since the vessel
was not able to accommodate such a high amount of
people it is assumed that the load distribution was
very poor.
The vessel was rolling against the impact of the wave.
Although this rolling action was not responsible for
the capsize event, together with insufficient or
unsuitable rolling reducers, the ship might have a very
low GM which might have been exacerbated due to
poor load distribution raising the center of gravity.
This situation might enhance the loss of stability at a
lower angle of inclination during the ultimate failure
event. In addition, a lot of people gathered together in
the port side for buying ticket and disembarking
which created a great heeling moment and was proved
to be a disastrous factor later on.
The fore side of the vessel was attached to the ground
which, in combination with the previous crowding
factor, instigated imbalance to the ship and lowered
the aft portion. At that time aft maneuvering caused
the ship to be lower more and water started to ingress
from the starboard side. The people were so
frightened that they crowded together to the port side.
448
As a result, the vessel got imbalanced and capsized

instantly. Here the aft maneuvering by the master
suggested that an action was attempted to restore
stability which ended in a failure. The report defined
this action as erroneous one. The lowering of the aft
side which drew water was not reported to be a vital
threat as the water did not create any free surface
effect. It was the rush of the panicked people amok
one side of the vessel, under the startling
circumstances, which ensued capsize.
7. DISCUSSION AND LIMITATIONS

In Bangladesh, accident reporting and the subsequent
accident prevention procedures are still in an
unsophisticated and primitive state. No remedial
measure seems sufficient since it does not correspond
with the Domino or chain of events ending in a
disaster. Therefore, this paper accentuates the
importance of reporting accidents in a proper way by
uncovering the incidents behind a marine accident
elaborately.
Being primly based on secondary data sources is the
main limitation of this research work. The primary
data source includes observation and interviews which
are always questionable. Moreover, collecting data
from legitimate sources and filtering the relevant
issues with extensive precision results in a small
number of extensive analyses of accident events. This
is a big limitation of this work as such type work
demands a larger sampling. Certain events in the fault
trees are not further developed due to shortage of
genuine data sources. Furthermore, most of the
events, ended in a factor relating to the human
element, are not developed further.
8. RECOMMENDATIONS
electric onboard passenger counter may serve

this purpose.
9. CONCLUSION
This paper focuses on developing fault tree for
different types of accidents as a suitable tool for the
collection and analysis of accident data to identify
hazardous chain of events. The error reduction
technique as well as accident mitigation process
should be based on the facts discovered from the fault
tree.
REFERENCES:
[1]
[2]
[3]
[4]
[5]
[6]
Since the situation of ship safety has aggravated

abysmally in recent years, this is the high time some
proper and innovative insights are given to the
accident prevention scheme.
A rich and powerful database has to be
developed as soon as possible through proper
investigation. The fault tree may be a
quantum leap forward in this investigation
procedure.
The fault tree is recommended to be refined
and developed further by direct investigation
of accidents through an interactive system.
As soon as the database is developed, a
careful analysis of the data will result in the
discovery of certain chain of events
triggering a debacle. The accident mitigation
process should be based on this analysis.
Resolving the ambiguity between DOS and
BIWTA, the corresponding authority should
be given enough number of personnel and
power to stop vessels from overloading.
Furthermore, a prior ticketing system and an
[7]
[8]
[9]
449
Huq, N.A. and Dewan, A.M., Lauch Disaster in

Bangladesh: A Geographical Study, Geografia,
Vol.1, Issue 2, pp. 14-25 (2003).
Banglapedia, Water Transport, National
Encyclopaedia of Bangladesh, Web Page
Address:
http://www.banglapedia.net/HT/W_0034.HTM,2
005
Peoples Republic of Bangladesh: Revival of
Inland Water Transport: Options and Strategies,
World Bank (May, 2007).
Inland Water Transport Master Plan Study, Final
Report, Transport Sector Coordination Wing,
Planning Commission, Ministry of Shipping,
Government of Peoples Republic of
Bangladesh, (31 March 2009).
Iqbal, K.S., Bulian, G., Hasegawa, K., Karim,
M.M., Awal, Z.I., A Rational Analysis of Intact
Stability Hazards Involving Small Inland
Passenger Ferries in Bangladesh, J Mar Sci
Technol (2008).
Iqbal, K.S., Bulian, G., Hasegawa, K., Karim,
M.M., Awal, Z.I., Possible Remedies for Intact
Stability Hazards Involving Contemporary
Small Inland Passenger Ferries in Bangladesh,
J Mar Sci Technol (2008).
Awal, Z.I., A Study on Inland Water Transport
Accidents in Bangladesh: Experience of a
Decade (1995-2005), Proceedings of the
International Conference on Coastal Ships &
Inland Waterways, RINA, London, UK, pp. 6772, (2006).
Awal,Z.I.,
Islam,
M.R.
and
Hoque,
M.M.H.,2006, An Analysis of Passenger
Vessel Accidents in the Inland Waterways of
Bangladesh, Proceedings of the Marine
Technology
Conference,
University
of
Hasanuddin, Indonesia, pp. 211-219, (2006).
Awal, Z.I., Islam, M.R. and Hoque, M.M.,
Marine Vehicle Accident Characteristics in
Bangladesh: Study on Collision Type
Accidents, Proceedings of the International
Conference on Mechanical Engineering, BUET,
Dhaka, Bangladesh (2007).
[10] Zahanyar,
S.I.M.
and
Haque,
M.E.,
Investigation of Accidents, Damages and
Cargo Losses in Inland Shipping, B.Sc. Engg.
Thesis, Dept. of Naval Architecture and Marine
Engineering, BUET, Dhaka, Bangladesh (1989).
[11] Bangladesh Transport Sector Study (BTSS),
Volume I, Planning Commission of Bangladesh,
Dhaka, Bangladesh (1994).
[12] Safe, Efficient and Improved Water Transport
System: Problems and Prospects, Seminar
Report, Banaladesh Inland Water Transport
Authority (BIWTA), Dhaka, Bangladesh (2003).
[13] Podder, S., Comparison of Stability of Sunken
Deck and Weather Deck in Inland Passenger
Vessel, B.Sc. Engg. Thesis, Dept. of Naval
Architecture and Marine Engineering, BUET,
Dhaka, Bangladesh, (2004).
[14] Khalil, G.M. and Tarafder, M.S., Ferry Diasters
in the Rivers of Bangladesh: Some Remedial
Measures,
The
Journal
of
National
Oceanographic and Maritime Institution, Vol.
21, No. 2, pp. 33-53 (2004).
[15] Iqbal, K.S., Bulian, G., Hasegawa, K., Karim,
M.M., Awal, Z.I., Passenger Ferry Accidents in
Bangladesh: Design and Socio-economic
Aspects, Proceedings of the 10th International
Symposium on Practical Design of Ships and
Other Floating Structures (PRADS 2007), 30
September - 5 October 2007, Houston, Texas.
[16] Kruger, S., Hinrichs, Cramer, H., Performance
Based Approaches for the Evaluation of Intact
Stability Problems, 9th Symposium on Practical
Design of Ships and other Floating Structures,
Luebeck-Travemuende, Germany (2004).
[17] Umeda, N. and Ikeda, Y., Rational
Examination of Stability Criteria in the Light of
Capsizing
Probability,
(Web:
http://www.marine.osakafuu.ac.jp/~lab15/papersearch/papers/PDF/INTLA-64.pdf), STAB94 2.
[18] Vesely, W., Stamatelatos, M., Dugan, J.,
Fragola, J., Minarick, J., Railsback, J., Fault
Tree Handbook with Aerospace Applications,
Version 1.1, Prepared for NASA Office of
Safety and Mission Assurance, NASA
Headquarter, Washington DC, USA, (2002).
[19] Amrozowicz, M.D., Brown, A., Golay, M., A
Probabilistic Analysis of Tanker Grounding, 7th
International Offshore and Polar Engineering
Conference, Honolulu, Hawaii, (1997).
[20] Basu, R.I. and Bhattacharya, B., Challenges of
the Application of System Reliability Principles
to Floating Structures, Web Page Address:
http://www.nmri.go.jp/main/cooperation/ujnr/24
ujnr_paper_us/UJNR24-papers_us.htm
[21] Webb, R.D.G., Human Reliability and Ship
Stability, Report to NSSWG on Human
[22]
[23]
[24]
[25]
Web
[26]
[27]
[28]
450
Reliability and Capsize, on behalf of

Department of National Defense, represented by
Defense Research and Development, Canada
(2003).
Rothblum, A.M., Human Error and Marine
Safety, U.S. Coast Guard Research and
Development Center.
Psaraftis, H.N., Caridis, P., Desypris, N.,
Panagakos, G., Ventikos, N., The Human
Element as a Factor in Marine Accidents,
IMLA-10 Conference, St. Malo, France, (1998).
Talley, W.K., Jin, D., Kite-Powell, H.,
Determinants of the Severity of Passenger
Vessel Accidents, Maritime Policy and
Management 33(2), pp. 173 -186 (2006).
Calhoun, S.R., Human Factors in Ship Design:
Preventing and Reducing Shipboard Operator
Fatigue, U.S. Coast Guard Research Project,
Department of Naval Architecture and Marine
Engineering, University of Michigan (2006).
Page
Address:
http://www.manningaffordability.com/s&tweb/P
UBS/Fatigue/fatiguepaper.html
Kothari, C.R., Research Methodology:
Methods and Techniques, Second Revised
Edition, New Age International Publishers
(2004).
Heinrich,
H.W.,
Industrial
Accident
Prevention, McGraw-Hill, New York (1931).
Report on the Marine Accident caused by MV
KOKO-4 submitted to Ministry of Shipping,
Government of Peoples Republic of
Bangladesh (10 December, 2009)
Capsize
Figure I: Capsize FTA
Lossofstability
No restoration of stability
C1
C2
Static stability loss
Dynamic stability loss
No/delayed/wrong
detection of stability loss
Insufficient action
to restore stability
SS
DS
C21
Synchronicity
with waves
Large rolling angle
Large
acceleration
DS2
Reduced
buoyancy
DS3
DS1
Excessive rolling
Resting on steep wave for

a certain time
Reduced transverse
stability
SS1
Draft reduction
by grounding
C22
Manual
error
Sensor
error
Delayed
action
Erroneous
action
HE
MacE
Del
Err
SS2
Flooding
GM/GZ reduction
ROL
Heeling moments
SS22
SS21
FL
Env
Insufficient
roll damping
Critical
resonance
ROL1
DCE
Very low
initial GM
Leak
in hull
DCE
MatE
Improper
opening
DCE
Poor
drainage
DCE
Poor
loading
Weight
increase
(Overloading)
Free surface
effect of liquid
within the hull
Cargo shift
SS221
Rush of
passengers in
one place
LoE
CRWD
FSE
LoE
Inappropriate rolling reducer

DCE
Improper cargo seating
Insufficient rolling reducer

DCE
451
DCE
Improper cargo placing

DCE
Figure II: Application of the fault tree in case of the capsize of MV KOKO- 4
Figure I: Capsize FTA (Continued)
Capsize
Free surface
effect of liquid
within the hull
Delayed
action
Del
FS
Delayed
order
Delayed
response
HE
HE
Loss of stability
No restoration of stability
C1
C2
Unrestricted
liquid flow
Poor drainage
system
DCE
FSE1
Static stability is lost

SS
Loading
error
Insufficient
bulkheads (too much
length between them)
LoE
Liquid transfer
between bulkheads
C22
FSE1
Reduced
buoyancy
DCE
Heedlessness to
rules &
regulations
Unconsciousness
HE
Insufficient action
to restore stability
Reduced
transverse stability
SS1
Leak in
bulkhead
Non-liquid
tight bulkhead
MatE
DCE
SS2
Err
Draft
reduction
by
grounding
HE
Erroneous
action
GM/GZ
reduction
Heeling
moments
SS21
SS22
Poor
loading
Erroneous
response
Overloading
LoE
HE
LoE
HE
452
Erroneous
order
HE
Err
Erroneous
order
Erroneous
action
Rush of
passengers
in one place
CRWD
Erroneous
response
HE

A STUDY ON PULLOUT CAPACITY OF HORIZONTAL AND INCLINED PLATE

ANCHORS FOR OFFSHORE INSTALLATIONS
Baleshwar Singh1, Birjukumar Mistri2 and Ravi Patel2
1
Associate Professor
IIT Guwahati, Guwahati, Assam, India
Email: baleshwar@iitg.ernet.in
2
Post-Graduate Student
IIT Guwahati, Guwahati, Assam, India
Email: mistri@iitg.ernet.in; a.patel@iitg.ernet.in
ABSTRACT
Structures subjected to uplift tensile loading require anchoring systems to resist pullout loads. Anchors used
in offshore installations can be broadly classified as gravity anchors, anchor piles and plate anchors. Plate
anchors can be circular, square or strip in shape. The loading applied on plate anchors can be vertical or
inclined or horizontal depending on the anchor orientation. In this paper, various experimental, theoretical
and numerical studies for estimation of load capacity behavior of horizontal and inclined plate anchors have
been reviewed. A parametric study of pullout capacity has been carried out for plate anchors embedded in
sandy soils by varying embedment ratio for horizontal anchors and by varying inclination angle for inclined
plate anchors.
Keywords: Plate anchors, Vertical loading, Inclined loading, Pullout capacity.
1. INTRODUCTION
Embedded anchors are extensively used where the
foundations of structures are subjected to uplift
tensile forces. They are necessary when the uplift
load is greater than the self-weight of the structure.
They allow the transmission of the pullout load to
the soil at a greater depth and farther away from the
structure. Anchors of various types are now used for
the vertical uplift resistance of submerged pipelines
and transmission towers, and for the inclined or
horizontal tieback resistance of waterfront and
earth-retaining structures. The anchors are mainly
made of concrete or steel.
With the recent expansion of offshore exploration
and production in deeper waters, they have a
significant application based on technical and
economic considerations. As a result, a number of
embedded anchor systems are undergoing
development or have been developed. They are
commonly adopted in the mooring of various
compliant platform types, which are to be located in
deeper zones of the ocean. A proper understanding
of the anchor response to loading is essential to
evolve an acceptable design procedure.
Numerous researchers have proposed different
approaches to estimate the pullout load-deformation
response of plate anchors of various shapes in sandy
soils. In this paper, they are summarized in terms of
the procedures and relationships developed for
determining the ultimate pullout capacity. Most of
them are based on model and field tests, or on slipline limit equilibrium analysis approach, and very
few analyses have been based on rigorous numerical
analysis. The static ultimate pullout capacity (UPC)
for both horizontal and inclined plate anchors
embedded in sandy soils is worked out by using the
various correlations and techniques, and the
computed results are presented and compared.
2. HORIZONTAL PLATE ANCHORS

The ultimate pullout capacity Qu of a horizontal
plate anchor in cohesionless soil can be expressed as:
Qu = HANu
(1)
where = unit weight of soil, H = embedment

depth, A = Area of the plate, and Nu = breakout
factor. The magnitude of this dimensionless breakout
factor is influenced by the geometry of the failure
surface developed within the soil mass, and is
dependent on a number of factors including soil
parameters, relative depth and anchor dimensions.
2.1 Based on experimental approach

Based on laboratory and field tests in dense sand,
Balla (1961) found that for circular shallow anchors,
the failure surface extended to the ground surface as
circular arcs. The ultimate pullout capacity was
shown to comprise of two components: weight of soil
in failure zone, and the shearing resistance developed
along the curved failure surface. The UPC was
expressed as:
453
(2)
where h = diameter of plate. F1 and F3 are

functions dependent on the angle of internal friction
and embedment ratio H/h.
Andreadis & Harvey (1981) proposed an
expression for UPC which was dependent on
embedment ratio, anchor size and shape, and degree
of soil disturbance. It was concluded that the
proposed design procedure based on medium scale
laboratory tests had shortcomings mainly related to
scale effects present during repeated loading, and
field scale tests were necessary to confirm the
validity of the approach.
Murray & Geddes (1987) conducted several
laboratory tests to take into account the effect of
factors such as size and shape of plate, depth of
embedment, sand density and plate roughness. They
concluded that all methods overestimate the ultimate
pullout capacity in medium to loose sand. The
relation for UPC was proposed as:
For rectangular anchors,
1
]
(3)
For circular anchors,

1
2
2
3
tan 2
2
1
sin
(4)
where B = width of plate, and L = length of plate.

Frydman & Shaham (1989) performed pullout
tests on prototype slabs placed at various
inclinations and different depths in dense sand. A
semi-empirical expression was found to reasonably
predict the pullout capacity. Factors that accounted
for the shape and the inclination were then provided
for the estimation of the ultimate pullout capacity of
any slab anchor.
Hanna et al. (2007) presented an analytical model
to predict the pullout capacity and the load
displacement relationship for plate anchors in sand.
The model was developed based on the failure
mechanism deduced from laboratory testing, and the
limit equilibrium technique was used. Expression
was given to estimate the critical depth for a given
anchor/soil conditions, which separated deep from
shallow anchors. The radius of influence of a
individual anchor on the ground surface was
established, and accordingly the spacing between
anchors could be determined to avoid interactions
between anchors. The proposed theory compared
well with theoretical and experimental data
available in the literature.
2.2 Based on limit equilibrium approach

Meyerhof & Adams (1968) conducted a number
of model and full-scale uplift tests of footings with
special reference for transmission towers. They
presented an approximate general theory for uplift
capacity for a strip or continuous footing, by
considering forces acting on a curved failure surface
above the foundation. In sands, the geometry of the
failure surface was found to be fairly distinct but
varied in shape and extent depending on the
depth/width ratio of the foundation and on the
rigidity and relative density of the soil. Shape
factors were applied to the general expression to
account for the three-dimensional effect of
individual square or circular footings.
Vesic (1971) studied the problem of an explosive
point charge expanding a spherical cavity close to
the surface of a semi-infinite soil mass. If the
diameter of the cavity were made equal to a circular
plate size, there would be an ultimate pressure that
would shear away the soil located above the cavity.
On the basis of this concept, he proposed that the
ultimate pullout capacity of a horizontal plate
anchor comprised of the vertical component of the
force inside the cavity, effective self weight of the
soil, and the vertical component of the resultant of
internal forces.
Chattopadhyay & Pise (1986) proposed a
theoretical model for evaluating the ultimate
breakout resistance of horizontal plate anchors
embedded in sand, by assuming a curved
axisymmetric failure surface through the
surrounding soil. It indicated the existence of a
characteristic relative depth, beyond which breakout
factor approached a constant value. It was capable
of predicting the breakout factors for a wide range
of values of angle of shearing resistance of sand.
Saran et al. (1986) proposed an analysis to predict
the critical pullout load, the breakout load, and loaddeformation characteristics of anchors using
hyperbolic stress-strain curves of cohesive-frictional
soils as the constitutive law. The analysis
incorporated the effect of the shape of the anchor.
Strip, square, and circular anchors were analyzed.
Analytical data were compared with the available
experimental results, which showed good
agreement.
White et al. (2008) described a simple limit
equilibrium solution for predicting the uplift
resistance of plate anchors buried in sand. The
geometry of the solution was selected to match
model test observations. Simple charts were
presented for the prediction of peak uplift resistance
from the normalised burial depth, the critical state
friction angle and the relative density of the backfill. It
was shown that the solution for uplift resistance based
on the limit theorems of plasticity was generally very
unconservative due to an improbable uplift mechanism
not seen in model tests. It was shown that for a
454
Based on limit equilibrium analysis, Meyerhof

(1973) proposed the following equation for inclined
UPC in granular soils:
frictional material obeying normality, there is no

energy dissipation during shear, and hence the uplift
resistance is simply the weight of the soil.
2.3 Based on finite element approach
Hanna et al. (1988) developed an analytical

method based on limit equilibrium approach for the
estimation of UPC of strip anchors with inclination
angle varying from 0 to 60, and proposed that:
(7)
(5)
where
= soil friction angle in plane strain, and
= critical state friction angle in plane strain.
Kumar & Kouzer (2007) examined the vertical

uplift capacity of strip anchors embedded
horizontally at shallow depths in sand by using an
upper bound limit analysis in conjunction with finite
elements and linear programming. Even though the
analysis considered the development of plastic
strains within elements, it was noticed that the soil
mass lying above the anchor remained rigid, and a
planar rupture surface emanated from the anchor
edge making an angle
with the vertical. They
found that the influence of angle of internal friction
was higher at deep embedment depth.
3. INCLINED PLATE ANCHORS

In the pullout of inclined plate anchors, the force
is transmitted perpendicular to the anchor plane.
Inclination angle is defined as the angle between by
the pullout direction and the vertical axis. Harvey &
Burley (1973) proposed a method for inclined plate
anchors based on slip line approach with similar
assumptions considered by Balla (1961) for the case
of horizontal plate anchors.
Rowe and Booker (1979) proposed an analytical

technique for predicting the behaviour of an inclined
anchor of a general shape by dividing it into a series
of rectangular sub-regions. This approach was
applied to a square anchor and solutions were
presented in the form of influence charts for a range
of Poissons ratio, embedment ratio, anchor
inclinations and load inclinations. The solutions
were considered to be applicable for circular
anchors also.
Vermeer & Sutjiadi (1985) considered straight

rupture surfaces at an inclination to vertical equal to
the soil dilatancy angle, and proposed the following
simple relation for breakout factor:
Merifield et al. (2006) applied three-dimensional

numerical limit analysis and axi-symmetrical
displacement finite element analysis to evaluate the
effect of anchor shape on the pullout capacity of
horizontal plate anchors in sand. The anchor was
idealized as either square or circular in shape, and
rigorous solutions were presented for the ultimate
pullout capacity.
(6)
where Kb = earth pressure coefficient,

inclination of plate with vertical axis.
Rowe & Davis (1982) considered the effects of

soil dilatancy, initial stress state and anchor
roughness on the pullout capacity. The numerical
solutions presented were obtained from an elastoplastic finite element analysis. This approach
allowed the consideration of plastic failure within
the soil, anchor breakaway from the soil behind the
anchor, and shear failure at a frictional, dilatant soil
structure interface without the introduction of
special joint or interface elements. They found that
soil dilatancy had a significant effect on UPC,
whereas roughness and initial stress state had
negligible effects.
where H' = average depth of embedment, and Ks =

punching uplift coefficient.
Ghaly (1997) studied laboratory and field results
reported in the literature for shallow anchor plates of
various configurations and embedded in sands. This
data was incorporated in a generalized form to
predict the ultimate horizontal pullout resistance of
vertical anchor plates in terms of the influencing
parameters. The following expressions were
proposed for UPC:
For circular anchors,
5.5
(8)
For square and rectangular anchors,

3.3
(9)
Murray & Geddes (1989) presented laboratory

experimental results for the ultimate passive
resistance and corresponding displacements of
rectangular anchor plates pulled at inclination angles
through very dense sand. The results were compared
with theoretical solutions based on the upper and
lower bound limit theorems of soil plasticity.
Goel et al. (2005) worked out the breakout
resistance of inclined plate anchors in sand out using
limit equilibrium approach. The breakout resistance
was calculated for different soil friction angles with
varying relative depth ratio and anchor inclination.
It was found that the breakout factor increased
continuously with the inclination of the anchor. A
comparison of the predicted values of breakout
resistance from the proposed analysis with the
experimental values of the other researchers showed
455
reasonably good agreement.. The proposed breakout

factor was:
(10)
where K = coefficient of earth pressure, Ii =
coefficient of inclination, i = inclination angle of
plate anchor.
At the embedment ratio of 3, the inclined pullout

capacities in loose sand range from about 500 kN to
about 1000 kN for vertical orientation, with the
values of Goel et al. increasing exponentially
beyond 40 (Fig. 7). The same trend of the capacity
increasing with anchor orientation is also observed
for medium dense and dense sands (Figs. 8 & 9).
Balla(1961)
Andreadis&Harvey(1981)
Murray&Geddes(1987)
4. RESULTS & DISCUSSION

15000
UltimatePulloutCapacity
(kN)
To carry out a comparative study of the ultimate

pullout capacity of horizontal and inclined plate
anchors embedded in sand by using the above
empirical and theoretical correlations, calculations
have been made for a strip anchor of 2 m width and
unit length. If the correlation is applicable only for a
circular or square anchor, the equivalent area is
taken into consideration. The embedment ratio of
the horizontal strip anchor is varied from 2 to 10.
For the inclined strip anchor, the embedment ratio is
fixed at 3, and the inclination angle is varied from
0 to 90 from horizontal to vertical orientation. The
properties of sandy soils adopted in the
computations are presented in Table 1.
12500
10000
7500
5000
2500
0
0
4
6
8
10
EmbedmentRatio
Figure 1. Variation of UPC of horizontal plate
anchor in loose sand from experimental studies.
Balla(1961)
Andreadis&Harvey(1981)
Murray&Geddes(1987)
Table 1. Properties of sands.
Unit Weight (kN/m )

Friction angle ()
Modulus of elasticity
(kN/m2)
14
30
20,000
Medium
Dense
Sand
17
35
25,000
Poissons ratio
0.3
0.35
Dense
Sand
UltimatePulloutCapacity(kN)
Loose
Sand
15000
20
40
30,000
12500
10000
7500
0.4
The ultimate pullout capacities computed from

the experimental and theoretical predictions are
plotted against embedment ratio in Figs. 1 to 6 for
horizontal anchors and in Figs. 7 to 9 for inclined
anchors, respectively. From a comparison of the
anchor capacities for both dense sand and medium
dense sand (Figs. 1 & 2), it is observed that Ballas
correlation gives the highest values whereas that of
Andreadis & Harvey provides the lowest values. For
the loose sand (Fig. 3), Hanna et al.s relationship
yields the maximum values.
The vertical uplift capacities predicted from the
experimental studies can be compared with those
obtained from theoretical studies for the same
relative density of sand and the corresponding
embedment ratio. From the plots in Figs. 1 & 4 for
loose sand, at an embedment ratio of 3, the
experimental values range up to about 4000 kN and
are higher than the theoretical values which do not
exceed 2500 kN. However, at the same embedment
ratio, both the values are comparable for both
medium dense and dense sands, as observed from
Figs. 2 & 5 and Figs. 3 & 6, respectively.
5000
2500
0
4
6
8
10
EmbedmentRatio
Figure 2. Variation of UPC of horizontal plate anchor
in medium dense sand from experimental studies.
Balla (1961)
Andreadis & Harvey (1981)
Murray & Geddes (1987)
40000
Properties
35000
30000
25000
20000
15000
10000
5000
0
0
10
EmbedmentRatio
in dense sand from experimental studies.
456
Meyerhof(1973)
Meyerhof&Adams(1968)
Saranetal.(1986)
Rowe&Davis(1982)
Merifieldetal.(2006)
20000
15000
10000
5000
0
0
Saranetal.(1986)
Rowe&Davis(1982)
5000
0
6
25000
20000
15000
10000
5000
0
4
10
80
100
Goeletal.(2005)
2000
1500
1000
500
0
20
40
60
80
100
Rowe&Booker(1979)
Hannaetal.(1988)
4000
Goeletal.(2005)
3500
3000
2500
2000
1500
1000
500
0
0
EmbedmentRatio
in dense sand from theoretical studies.
60
Rowe&Booker(1979)
Meyerhof(1973)
Saranetal.(1986)
Rowe&Davis(1982)
30000
40
AnchorInclination()
Figure 8. Variation of UPC of inclined plate anchor
in medium dense sand.
35000
20
2500
10
EmbedmentRatio
in medium dense sand from theoretical studies.
Vesic(1971)
Chattopadhyay&Pise(1986)
Whiteetal.(2008)
Vermeer&Sutjiadi(1985)
Kumar&Kouzer(2007)
Hannaetal.(1988)
3000
10000
500
Meyerhof(1973)
15000
1000
AnchorInclination()
in loose sand.
20000
Goeletal.(2005)
1500
10
EmbedmentRatio
in loose sand from theoretical studies.
Vesic(1971)
Whiteetal.(2008)
Kumar&Kouzer(2007)
Rowe&Booker(1979)
Hannaetal.(1988)
2000
Vesic(1971)
Whiteetal.(2008)
Kumar&Kouzer(2007)
20
40
60
80
100
AnchorInclination()
in dense sand.
457
5. CONCLUSIONS
In deep water offshore installations, plate anchors
and their variants are being increasingly adopted. The
anchors are placed at orientations between the
horizontal and vertical depending on design
requirements of the application. Due to the nature of
loading on mooring systems used offshore, embedded
plate anchors will be subjected to a wide range of
sustained and repeated loads that will vary with the
tautness of the mooring lines. The degree of soil
disturbance adjacent to the anchor during installation
will also affect the pullout capacity.
Various techniques and procedures based on
experimental and theoretical studies found in the
literature have been reviewed, and the ultimate pullout
capacities of horizontal and inclined plate anchors
have been computed by varying the embedment ratio
and inclination angle. Comparisons have been made
among the predicted values. As the embedment ratio
increases, the breakout factor increases and tends to
reach a maximum value. The breakout load also
increases with anchor inclination for a deep anchor. It
is recommended that the anchors be installed deeply so
that inaccuracy in the embedded depth does not
substantially affect the designed ultimate capacity.
REFERNCES
[1] Andreadis, A. and Harvey, R.C., A design
procedure for embedded anchors, Applied
Ocean Research, Vol. 3, No. 4, pp. 177-182
(1981).
[2] Balla, A., The resistance to breaking-out of
mushroom foundations for pylons, Proceedings
of 5th Intl. Conf. on Soil mechanics &
Foundation Engineering, Paris, Vol. 1, pp. 569576 (1961).
[3] Chattopadhyay, B.C. and Pise, P.J., Breakout
resistance of horizontal anchors in sand, Soils
and Foundations, Vol. 26, No. 4, pp. 16-22
(1986).
[4] Frydman, S. & Shaham, I., Pullout capacity of
slab anchors in sand, Canadian Geotechnical
Journal, Vol. 114, No. 11, 1300-1317 (1989).
[5] Ghaly, A.M., Load-displacement prediction for
horizontally loaded vertical plates, Journal of
Geotechnical
and
Geoenvironmental
Engineering, Vol. 123, No. 1, pp. 74-76 (1997).
[6] Goel, S., Shalini and Patra, N.R., Break out
resistance of inclined anchors in sand,
Geotechnical and Geological Engineering, Vol.
24, pp. 15111525 (2005).
[7] Hanna, A.M., Das, B.M. and Foriero, A.,
Behavior of shallow inclined plate anchor in
sand, In Special Topics in Foundations,
Geotech. Special Technical Publication No. 16,
ASCE, pp. 54-72 (1988).
[8] Hanna, A., Ayadat, T. and Sabry, M., Pullout
resistance of single vertical shallow helical and
plate anchors in sand, Geotechnical and

Geological Engineering, Vol. 25, No. 4, pp. 559573 (2007).
[9] Harvey, R.C. and Burley, E., Behaviour of
shallow inclined anchorage in cohesionless
sand, Ground Engineering, Vol. 6, No. 5, pp.
48-55 (1973).
[10] Kumar, J. and Kouzer, K.M., Vertical uplift
capacity of horizontal anchors using upper bound
limit analysis and finite elements, Canadian
Geotechnical Journal, Vol. 44, pp. 698-704
(2007).
[11] Merifield, R.S., Lyamin, A.V. and Sloan, S.W.,
Three-dimensional lower-bound solutions for
the stability of plate anchors in sand,
Geotechnique, Vol. 56, No. 2, pp. 123132
(2006).
[12] Meyerhof, G.G. and Adams, J.I., The ultimate
uplift capacity of foundations, Canadian
Geotechnical Journal, Vol. 24, No. 4, pp. 589592 (1968).
[13] Meyerhof, G.G., Uplift resistance of inclined
anchors and piles, Proceedings of VIII
International Conference of Soil Mechanics and
Foundation Engineering, Moscow, Vol. 2, No. 1,
pp. 167-172 (1973).
[14] Murray, E.J. and Geddes, J.D., Uplift of anchor
plates in sand, Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 113, No. 3,
pp. 202-214 (1987).
[15] Murray, E.J. and Geddes, J.D., Passive inclined
anchorages in sand, Journal of Geotechnical
Engineering, Vol. 117, No. 5, pp. 810-814
(1989).
[16] Rowe, R.K. and Booker, J.R., The analysis of
inclined anchor plates, Third International
Conference on Numerical methods in
Geomechanics, Aachen, pp. 1227-1236 (1979).
[17] Rowe, R.K. and Davis, E.H., The behavior of
anchor plates in sand, Geotechnique, Vol. 32,
No. 1, pp. 25-41 (1982).
[18] Saran, S., Ranjan, G. & Nene, A.S., Soil
anchors and constitutive laws, Journal of
Geotechnical Engineering, Vol. 112, No. 12, pp.
1084-1100 (1986).
[19] Vermeer, P.A. and Sutjiadi, W., The uplift
resistance of shallow embedded anchors, Proc.
11th International conference on Soil Mechanics
and Foundation Engineering. San Francisco,
Vol. 3, pp. 1635-1638 (1985).
[20] Vesic, A.S., Breakout resistance of objects
embedded in ocean bottom, Journal of Soil
Mechanics and Foundations Div., ASCE, Vol.
97, No. 9, pp. 1183-1205, (1971).
[21] White, D.J., Cheuk, C.Y. and Bolton, M.D.,
The uplift resistance of pipes and plate anchors
buried in sand, Geotechnique, Vol. 58, No. 10,
pp. 771779 (2008).
458

EFFECT OF FGM COATING ON THERMOELASTIC

CHARACTERISTICS OF A ROTATING CIRCULAR DISK
A. M. Afsar1 and F. A. Sohag 2
Department of Mechanical Engineering,
E-mail:1mdafsarali1967@yahoo.com
fahmed.buet@gmail.com
ABSTRACT
This study focuses on the analysis of thermoelastic characteristics of a thin circular disc having a concentric
hole and a functionally graded material (FGM) coating at the outer surface. The disc is subjected to a
temperature gradient filed and an inertial force due to rotation of the disc. The coating region is assumed to
have an exponential variation of all the material properties except the Poissons ratio which is assumed to be
constant throughout the disc. The incompatible eigenstrain developed in the disc as a result of the temperature
gradient field and nonuniform coefficient of thermal expansion (CTE) is taken into account. Using the 2-D
thermoelastic theories, the problem is reduced to the solution of a second order differential equation which is
solved by a finite element model developed based on the variational approach and Ritz method. The finite
element model is demonstrated for an Al disc with an Al/Al2O3 FGM coating for the analysis of thermoelastic
characteristics corresponding to more practical boundary conditions of the disc. The numerical results reveal
that the FGM coating thickness is one of the important parameters to be considered in order to design a
grinding disc or a cutter with an FGM coating.
Key words: Functionally graded material, Coating, Circular disk, Thermoelasticity, Thermal load, Finite
element method.
1. INTRODUCTION
Functionally graded materials (FGMs) are a new
generation of engineered materials wherein the
microstructural details are spatially varied through
nonuniform distribution of the reinforcement phases.
The FGM concept was originated in Japan in 1984
during a space plane project, in the form of a
proposed thermal barrier material [1].Usually, these
materials are made of ceramics and metals to resist
severe environmental effects, such as wear, corrosion,
and large temperature gradient in one hand, and
ensure toughness and thermal conductivity on the
other hand. FGMs have been promising candidates for
various structural components such as FGM beams
[2-4], plates [5-6], and cylinders [7-8] which have
been studied under various thermal and mechanical
loading conditions. Afsar et al. [6] examined the

effects of nonhomogeneous parameter on the elastic
field in an FGM rectangular plate subjected to a
biaxial tensile load. FGM circular cylinders were
considered by Obata and Noda [7] and Liewet al. [8]
to analyze the thermal stresses. FGM circular
cylinders were also considered by Afsar and his coworkers [9, 10] for the analysis of brittle fracture
characteristics by taking into account the effect of
incompatibility of eigenstrain developed in the
cylinder due to nonuniform CTE as a result of
temperature change.
In the studies mentioned above, attention was focused
mainly on the structural performance of FGMs. To
investigate the functional performance of these
materials, FGMs are widely explored for thermal
barrier [11] and tribological [12] coatings. However,
the analysis of FGMs as cutting and grinding tools has
459
received only a little attention so far. Cho and Park

[13] investigated the thermoelastic characteristics of
functionally graded lathe cutting tools composed of
CrMo steel and ceramic tip. They demonstrated that
an added FGM layer between the steel shank and the
ceramic tip relaxes the thermo-mechanical stress
concentration. In our previous study [14], we have
demonstrated that a circular FGM cutter or grinding
disk can be designed with better thermoelastic
characteristics if certain parameters, namely,
temperature distribution, angular speed, radial
thickness, and outer surface temperature, are
controlled properly. For the purpose of analyzing the
thermoelastic characteristics, a finite element model
was developed. However, the FGM disk considered in
our previous study [14] had a material gradation
throughout the entire radial thickness, i.e., from the
inner surface to the outer surface of the disc. Such an
FGM disc is, in fact, not feasible due to
manufacturing limitations as well as cost. In an
attempt to get rid of this problem, an alternate
approach is sought to design the disc with a thin FGM
coating at the outer surface only instead of an FGM
disc having material gradation throughout the entire
radial thickness of the disc. To realize this goal, it is
an utmost necessity to understand and quantify the
thermoelastic characteristics of such a disc. The
present study is, therefore, aimed at analyzing the
thermoelastic characteristics of a thin circular disc
with a thin FGM coating at the outer surface of the
disc.
Fig. 1. Schematic diagram of a disc with an

FGM coating at the outer surface
The Poissons ratio is assumed to be constant
throughout the entire disc and the Youngs modulus
(E), CTE (), and density () is assumed to vary
exponentially as
(1a)
(1b)
(1c)
The constants of Eq. (1) are determined as:

(2a)
(2b)
(2c)
2. MATHEMATICAL MODEL OF THE

PROBLEM
The outer surface of a cutter or a grinding disc should
be made of such a material that it can withstand high
temperature and prevent wear. On the other hand, the
inner region of it should be thermally conductive to
facilitate cooling and tougher to absorb torsional
energy. Using a bimetallic disc can serve the purpose
but the unequal coefficient of thermal expansion
causes a miss-fit strain resulting in delamination and,
consequently, failure of the disc. To get rid of these
problems, the disc can be designed with an FGM
coating at the outer surface and a homogenous
material at the inner portion of the disc as shown in
Fig. 1.The disc has a concentric hole of radius a with
an outer radius b. The region (h-a) consists of
homogenous material B only. The coating region c=
(b-h) consists of A/B FGM where the distribution of
materials A and B varies continuously from r = h to b.
ln
(3a)
ln
(3b)
ln
(3c)
The parameters with subscripts A and B indicate the

properties of the respective constituent materials A
and B. Although Eqs. (1) - (3) express the properties
of the FGM coating region (hrb), the same
expressions can be utilized to determine the properties
of the homogeneous region (arh) by setting the
values of , , and to zero.

The disc considered in the present study is
subjected to a temperature gradient field and an inertia
force due to rotation of the disc. The inner surface is
assumed to be fixed to a shaft while the outer surface
460
is free from any mechanical loading. Thus, the

boundary conditions can be stated as
(i) r = a, ur= 0;
(ii) r = b, r= 0;
1
Due to nonuniform CTE and temperature gradient

field, an incompatible eigenstrain [15] is developed in
the disc which is given by
(4)
Here, T(r) is the change in temperature at any point r
of the disc. By making use of 2-D thermoelastic
theories and Eqs. (1) (4), one readily obtains [14]
3
(5)
Equation (9) is a system of algebraic equation, which
can be used to form a global system of algebraic
equation by satisfying the continuity condition
3
Solution of Eq. (5) gives the value of F which can be
used to determine different components of stress,
strain, and displacement from the following
expressions [14].
. Once values of is known at the global

node points, the components of stress, strain, and
displacement are obtained as:
(6a)
(6b)
(10a)
(7a)
(7b)
(10b)
(11a)
(11b)
(12)
(8)
Since the analytical solution of Eq. (5) is not realistic,
a finite element model is developed for the numerical
solution of the problem. As all the parameters are
function of r only making it a one dimensional
= (a, b) of the disc

e
is divided into N number of subdomains =
(re,re+1), where e = 1,2,,N. Then following the
problem, the radial domain
Variational approach and Ritz method Eq. (5)can be

reduced to
(9)
where
,
4. NUMERICAL
DISCUSSION
RESULTS
AND
The finite element model developed in the present

study is applied to an Al disc with an Al/Al2O3 FGM
coating at the outer surface. The materials A and B
mentioned earlier correspond to Al2O3 and Al,
respectively. The required properties are presented in
Table 1.
From the test of convergence, it is found that the
element size of 1 mm ensures the convergence of the
results. Therefore, all the numerical results present in
this section correspond to element size of 1 mm. To
observe the effect of temperature distribution profiles
on stress, strain, and displacement components, five
different temperature profiles as a function of
461
Table 1. Mechanical and thermal properties of Al

and Al2O3
normalized radial distance are considered as shown in

Fig. 2. For uniform temperature profile, we see that
the radial stress is compressive almost over the entire
region of the disc except some inner portion of it.
Also, for linear and parabolic2 distribution profiles of
temperature, some outer region including the FGM
coating experiences the compressive radial stress. For
other profiles of temperature distribution, radial stress
is positive over the entire region. For all profiles of
temperature distribution, radial stress is zero at the
outer surface which satisfies the boundary condition
of the problem.
Circumferential stress as shown in Fig. 4 has
different characteristics than the radial one. For all
profiles of temperature distribution, stress is
compressive at some part of the disc and tensile at
other portion of it. For the uniform profile of
temperature distribution, stress is negative whose
value remains almost constant over the inner portion
of the disc except the small outer region. Here, the
stress changes from negative to positive value and
continues to increase until the outer surface. For other
temperature profiles except the exponential one, stress
was positive in about 50% of the inner portion of the
disc and followed the trend of uniform one. On the
other hand, the exponential temperature distribution
yields stress whose magnitude is the minimum at
almost entire region of the disc. However from the
beginning of the coating, the stress changes to
compressive one whose value continues to increase
until the outer surface of the disc.
The effect of temperature distribution on the
displacement component is depicted in Fig. 5. As
seen, the displacement is zero at the inner surface
satisfying the boundary condition and gradually
increases to the maximum value at the outer surface
for all the temperature profiles. However, the gradient
of the displacement curves is smaller at the coating
region. The variation of displacement is the minimum
for exponential temperature profile.
To see the effect of angular speed on thermoelastic
characteristics of the disc, four different speeds are
considered and the results are presented in Figs. 6 to
8. As the inertial force is proportional to the square of
Fig. 2. Prescribed distribution of temperature

profiles
Fig 3. Effect of temperature distribution on radial

stress
Fig. 4. Effect of temperature distribution on

circumferential stress
the angular speed, and in general, the stress, strain,
and displacement are proportional to the force, the
magnitude of stress and displacement increases with
the increase of the speed as seen from the figures.
The radial stress exhibits a slightly different
characteristic in the coating region. Here, this stress
component does not vary so much with the speed as
well as with the radial distance.
462
Fig. 8. Effect of angular speed on displacement
Fig. 5. Effect of temperature distribution on

displacement
Fig. 9. Effect of coating thickness on radial stress

Fig. 6. Effect of angular speed on radial stress
Fig. 10. Effect of

circumferential stress
coating
thickness
on
Fig. 11. Effect

displacement
coating
thickness
on
Fig.7. Effect of angular speed on circumferential

stress
The effect of FGM coating thickness is illustrated
in Figs. 9 to 11. Fig.9 shows that over the major
portion of the disc, the radial stress is tensile and its
magnitude increases with the decrease of coating
thickness. But the coating region experiences
compressive stress whose magnitude and the span of
the region of compressive stress increase with the
increase of the thickness. Circumferential stress as
shown in Fig. 10 has the similar trend as that of the
radial one but the feature is different at the interface
463
of
of the coating and homogeneous part of the disc.

From Fig. 11, we see that the magnitude of the
displacement component decreases with the increase
of the coating thickness.
5. CONCLUSION
In the present study, a finite element model has
been developed for a rotating circular disc with an
FGM coating at the outer surface for the analysis of
thermoelastic characteristics of the disc. The Finite
element model developed in the study is applicable to
an FGM disc, disc with an FGM coating at the outer
surface, and also for a homogeneous disc by setting
the parameters = = = 0. From the study, it is
established that temperature distribution profile,
angular speed, and FGM coating thickness are the key
parameters to characterize the thermoelastic behavior
of the disc. The numerical results presented in the
study are helpful to design an actual cutter or grinding
disc with an FGM coating.
REFERENCES
[1] J. B. Holt, M. Koizumi, T. Hirai and Z. A.
Munir, Ceramic Transactions, Vol. 34:
Functionally Gradient Materials. The American
Ceramic Society, Westerville, Ohio, USA
(1993).
[2] H. J. Xiang and J. Yang, Free and forced
vibration of a laminated FGM Timoshenko beam
of variable thickness under heat conduction,
Composites: Part B 39, 292303 (2008).
[3] S.-R. Li, J.-H. Zhang and Y.-G. Zhao, Thermal
post-buckling of functionally graded material
Timoshenko beams, Appl. Math. Mech. 27, 803
810 (2006).
[4] A. K. Upadhyay and K. R. Y. Simha, Equivalent
homogeneous variable depth beams for cracked
FGM beams; compliance approach, Intl J.
Fracture 144, 209213 (2007).
[5] E. Feldman and J. Aboudi, Buckling analysis of
functionally graded plates subjected to uniaxial
loading, Compos. Struct. 38, 2936 (1997).
[6] A. M. Afsar, R. Rahman, Y. Q. Wang, Y. Shi and

J. Song, Finite difference solution of stress and
strain in an FGM plate, in: Proc. 16th Ann. Intl
Conf. Composites/Nano Engineering (ICCE-16),
Kunming, China (2008).
[7] Y. Obata and N. Noda, Steady thermal stresses in
a hollow circular cylinder and a hollow sphere of
a functionally graded material, J. Thermal
Stresses 17, 471487 (1994).
[8] K. M. Liew, S. Kitipornchai, X. Z. Zhang and C.
W. Lim, Analysis of the thermal stress behavior
of functionally graded hollow circular cylinders,
Intl J. Solids Struct. 40, 23552380 (2003).
[9] A. M. Afsar and H. Sekine, Optimum material
distributions for prescribed apparent fracture
toughness in thick-walled FGM circular pipes,
Intl J. Pressure Vessels and Piping 78, 471484
(2001).
[10] A. M. Afsar and M. Anisuzzaman, Stress
intensity factors of two diametrically opposed
edge cracks in a thick-walled functionally
graded material cylinder, EngngFract. Mech. 74,
16171636 (2007).
[11] J.-S. Xiao, B. Jiang, J. Liu and S.-Y. Huang,
Microstructures and mechanical properties of
ceramic/metal gradient thermal barrier coatings,
J. Wuhan Univ. Technol. 18, 10002413 (2003).
[12] L. Prchlik, A. Vaidya and S. Sampath, Damage
tolerant tribological coatings based on thermal
sprayed FGMs, in: Proc. 6th Intl Sympos.
Functionally Graded Materials, Estes Park,
Colorado, USA, September 1014 (2000).
[13] J. R. Cho and H. J. Park, High strength FGM
cutting tools: finite element analysis on
thermoelastic characteristics, J. Mater. Proc.
Technol. 130131, 351356 (2002).
[14] J. Go, A. M. Afsar and J. I. Song, Analysis of
thermoelastic characteristics of a rotating FGM
circular disk by finite element method, Adv.
Compos. Materials, Volume 19, Number
2(2010).
[15] T. Maru, Micromechanics of defects in
solids: mechanics of elastic and inelastic
solids. Kluwer Academic Publishers,
Dordrecht, The Netherlands (1987).
464

AN EXPERIMENTAL STUDY OF TWO-PHASE LOOPED THERMOSYPHON

A. K. Azad1, S. K. Sarkar2, C. M. Feroz3 and M. M. Alam4
3,4
Professor, Department of Mechanical Engineering
E-mail: mmalam@me.buet.ac.bd
1,2
Department of Mechanical Engineering

E-mail: azad_sgfl@yahoo.com
ABSTRACT
The present experimental work investigates the heat transfer performance of Miniature Looped Parallel
Thermosyphon (MLPT) which consists of two single tube thermosyphon connected by two U tubes of same
diameter at the top and bottom ends. For this purpose, the copper tube of 5.78 mm ID used with ethanol,
methanol, acetone and water as the working fluids. Heat transfer characteristics are determined experimentally,
based on the principle of phase change at different heat flux and different coolant flow rates. An analysis of the
experimental data gives that the axial wall temperature of both condenser and evaporator sections decreases
with the increase of coolant flow rate. The wall temperature increases with increase of heat flux. The thermal
resistance decreases with the increase of both coolant flow rate and thermal load. Overall heat transfer
coefficient increases with the increase of both coolant flow rate and heat flux.
Keywords: Heat flux; overall heat transfer coefficient; Working fluid; Thermosyphone; Phase change;
1. INTRODUCTION
Thermosyphon or heat pipe is a device of very
high thermal conductance. Among other cooling
technique heat pipe emerged as the most appropriate
technology and cost effective thermal design due to its
excellent heat transfer capacity, high efficiency and
structural simplicity [1-5]. The heat pipe can, even in
its simplest form, provide a unique medium for the
study of several aspects of fluid dynamics and heat
transfer, and it is growing in significance as a tool for
use by the practicing engineer or physicist in
applications ranging from heat recovery to precise
control of laboratory experiments [2, 6-9].
The idea of looped parallel thermosyphon (1) (2) is
developed to minimize several insufficient
performances of single tube thermosyphon such as
low maximum heat transfer rate and non-uniform wall
temperature in an evaporator section. In LPT, two
single type thermosyphons are joined by two U-tubes
at the top and bottom ends and thus total heat transfer
area is increased. A small quantity of fluid is placed
in the tube from which the air is evacuated and the
tube sealed. The lower end of the tube (evaporator
section) is heated causing the liquid to vaporize. After
that, the vapor moves to the cold end of the tube
(condenser section) where it is condensed. The
condensate is returned to the hot end by gravity. Since

the latent of evaporation is large, considerable
quantities of heat can be transported [9-11].
Thermal designers have widely accepted the miniature
looped parallel thermosyphon (MLPT) for their
thermal design solution and the area of application
increases day by day. Normally MLPTs have 3 to 6
mm diameter and less than 400 mm length. Most
preferable length is 150 mm. MLPT is relatively a
new technology; relevant data, information is quite
scarce (3). So, a thorough investigation of heat
transfer capability of MLPT is indispensable for
further
development
and
improvement
of
performance. The purposes of this study are to show
heat transfer characteristics in evaporator and
condenser sections in miniature looped parallel
thermosyphon (MLPT) and experimentally examined
the maximum heat transfer rate, wall temperature
profile and thermal resistance in the test sections
[7,11].
2. MINIATURE LOOPED THERMOSYPHON

The schematic diagram of experimental apparatus is
shown in the Fig. 2. The test loop consists of a pair of
evaporator, adiabatic, condenser and U tube sections
and is vertically oriented. Detailed dimensions of the
MLPT are shown in the Fig. 1. Two evaporator and
465
two condenser sections are located at bottom and top

of the loop, respectively. All sections are made of
5.78 mm ID copper tube. The length of the
evaporator, adiabatic, and condenser sections are 50
mm, 30 mm, and 70 mm, respectively. The both tubes
are connected by U tubes with the same inside
diameter at top and the bottom. The U tube is a half of
a circular ring with 31.72 mm inner radius.
Two single tube thermosyphon consists of three
sections as shown in Fig. 1.
a. Evaporator section
Water: Water is a colorless liquid. It boils at 373 K at

atmospheric pressure.
4. EXPERIMENTAL PROCEDURE
b. Adiabatic section
c. Condenser section
The detail dimensions of looped thermosyphon used
in the experiment are summarized in Table-1.
Evaporator section: It is bottom part of
thermosyphon. Heat is added to the thermosyphon
through evaporator section.
Condenser section: It is upper most part of the
thermosyphon. Heat is removed from the
thermosyphon through condenser section.
Adiabatic section: Adiabatic section is located
between the evaporator and condenser sections. This
section is actually kept with heat pipe to distinguish
evaporator section and condenser section. Adiabatic
section is thermally insulated.
Table 1: The detail dimensions of thermosyphon
Parameters
Acetone: Acetone is a flame-able, colorless liquid. It

is the simplest of the organic chemical called ketones.
It is completely soluble in water. It has a mild
pleasant odor. It boils at 329 K at atmospheric
pressure.
Dimension (mm)
Outside diameter of pipe, d0
6.48
Inside diameter of pipe, di
5.78
Length of evaporator, Le
50
Length of adiabatic section, La
30
Length of condenser, Lc
70
The Ni-Cr thermic wires are wound around the

wall of the evaporator at a constant interval of 1.5
mm. The heat added to the two evaporator sections of
MLPT is processed in the electrical method by using
the two separate DC power supply. The evaporator
sections are covered with glass fiber to minimize heat
loss. The adiabatic and the condenser sections are also
covered with the insulator.
The condenser sections are cooled by a constant
temperature water coolant, circulating in an annular
space between the copper tube and jacket. The water
coolant is supplied from an elevated water tank and
the flow is controlled by the flow meters. Nine
calibrated thermocouples of T type are attached at
each side at the wall of the MLPT to measure the wall
temperature. Five units in the each evaporator section,
two units at the each adiabatic section, and two units
are at the each condenser section. The inlet and outlet
coolant temperatures are also measured. The
thermocouples are attached at the wall surface using
adhesive. Temperatures are measured by the digital
thermometers.
The input power to the heater in the each evaporator
section is increased stepwise. The measurements are
made under a steady state condition at each input
power. Ethanol, methanol, acetone and water are used
as the working fluids.
In the present study, the performance of MLPT is
evaluated by measuring the thermal resistance, R
(0C/W), which is defined in Equation (1)
3. WORKING FLUIDS
The working fluids used in this study are
Methanol (CH3OH), Ethanol (C2H5OH), Acetone
(CH3COCH3) and Water (H2O). Some important
properties of these four working fluids are mentioned
below.
(1)
The overall heat transfer coefficient, Ut (kW/m2 0C) is
Methanol: Methanol or Methyl alcohol is a colorless,

flame-able liquid. Pure methanol boils at 327.85 K at
atmospheric pressure and molecular weight is 32.00.
Ethanol: Ethanol or Ethyl alcohol is a colorless,
flame-able liquid. It boils at 351.3 K at atmospheric
pressure
Te Tc
Q
obtained from Equation (2) as follows
Ut
466
Q
Ae(Te Tc)
(2)
4.1 SCHE
EMATIC DIIAGRAM
5. RESU
ULTS AND DISCUSSIO
ON
Forr better undeerstanding thee heat transffer
characteriistics of therm
mosyphon as well
w as the waall
temperatuure at differennt points of the two phaase
looped th
hermosyphon are measureed. From theese
measuredd data the tem
mperature proffiles are plotteed
against th
he axial distancces. By using equations
e
(1) annd
(2) therm
mal resistance, R and overaall heat transffer
coefficiennt, Ut are ddetermined. Each
E
of theese
propertiess are comparedd and explainedd in detail on thhe
followingg section with ggraphs.
140
(a) Acetone, mc=0.0 l/min (LHS)

2
q(kW/m )
120
2.94
3.92
4.9
5.88
7.35
Temperature( C)
100
(All dimenssions are in mm

m)
80
60
Fiig. 1: Dimensioons of thermossyphon

40
20
0
0
20
40
60
80
100
120
140
160
A
Axial
distance(mm)
Fig. 3: Temperature ddistribution forr various heat

fluxes
120
(i) Ethanol, mc=0.4 l/min (LHS)

2
q(kW/m )
100
2.94
3.92
4.9
5.88
7.35
Temperature ( C)
80
60
40
20
0
0
Fig. 2: Experim
mental apparattus
20
40
60
80
100
120
140
160
A
Axial
distance (mm)
1. U-Tube, 2. Water outt, 3. Condenserr, 4. Adiabatic

section, 5. Evaporator section,
s
6. Insuulator, 7. Mica
sheet, 8. Heater,
H
9. Insuulating tape, 10. Flow meterr,
11. Volt meter,
m
12. D.C. power
p
source, 13. T-joint, 144.
Volt sliderr, 15. A.C. pow
wer source, 16. Gate valve, 177.
Water tankk
T
d
distribution
for various heat
Fig. 4: Temperature
ffluxes
Figures 3 and 4 show the wall tempperature profilles
along thhe axial lenngth from thhe bottom of
thermosypphon having 55.78 mm ID for acetone annd
ethanol working fluiids respectiveely. The waall
with increasing heat flux. It also
temperatuure increases w
found in
n the figuress that both condenser annd
evaporatoor wall tempperature decreeases with thhe
increase inn coolant flow
w rate (i.e. 0.01--0.40 L/min).
467
120
100
(c) Methanol, mc= 0. 1 l/min (LHS)

100
70
60
Temperature ( C)
Temperature ( C)
80
2.94
3.92
4.9
5.88
7.35
80
(a) Ethanol, q=2.94 kW/m2 (LHS)
90
q(kW/m2)
60
40
mc(l/min)
50
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
40
30
20
20
10
0
0
20
40
60
80
100
120
140
160
Axial distance (mm)
20
40
60
80
100
120
140
160
Axial distance (mm)
Fig. 8: Temperature distribution for various coolant

flow rate
Fig.5: Temperature distribution for various heat
fluxs
120
2
(i) Methanol, q=7.35kW/m (LHS)

140
(a) Water, mc=0.0 l/min (LHS)
100
100
2.94
3.92
4.9
5.88
7.35
80
80
q(kW/m )
Temperature( C)
Temperature( C)
120
mc(l/min)
60
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
60
40
40
20
20
20
40
60
80
100
120
140
160
20
40
60
80
100
120
140
160
Axial distance (mm)
Axial distance(mm)

flow rate
Fig. 6: Temperature distribution for various heat fluxs
90
(a) Water, mc=0.0 l/min (LHS)

80
100
(a) Acetone, q =2.94 kW/m2, (LHS)

70
80
mc(l/min)
60
Temperature ( C)
Temperature ( C)
60
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
40
20
50
mc(l/min)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
40
30
20
10
0
0
20
40
60
80
100
120
140
Axial distance (mm)
0
0
20
40
60
80
100
120
140
160
Axial distance(mm)

flow rate

flow rate
Figures 5 to 7 show the wall temperature profiles
along the axial length from the bottom of
thermosyphon having 5.78 mm ID for methanol,
water and acetone working fluids. Figures indicate the
similar result as found for ethanol and acetone. The
wall temperature increases with increasing heat flux at
all elevation.
Figures 8 to 10 show the axial wall temperature

distribution of the thermosyphon at various coolant
flow rate for Ethanol, methanol and water working
fluids. At a lower coolant flow rate wall temperature
in the condenser section is found very irregular. This
irregularity decreases with the increase of coolant
flow rate. The evaporator wall temperature also
decreases with the increasing coolant flow rate. It also
468
160
shows the axial wall temperature distribution of the

thermosyphon at various coolant flow rate for water
working fluid. It is found in the figures that both
condenser and evaporator wall temperature decreases
with the increase in coolant flow rate.
CPU Using Micro Heat Pipe, Fujikura Co.,Technical

Note, Vol. 9, pp. 64~68, 1997.
[6] Mochizuki, M., Mashiko , K., Nguyen, T., Saito, Y.,

Goto, K., Cooling CPU Using Hinge Heat Pipe, Heat
Pipe Technology , Pergamon , pp. 218~229, 1997.
[7] Jian Ling, et. al. Experimental Investigations of

Radially Rotating Miniature High-Temperature Heat
Pipes, ASME Journal, 2001, pp. 512-518.
6. CONCLUSIONS
The results of the performance test for the
thermosyphon having 5.78 mm ID and length of 150
mm give the following conclusions:
[1] The axial wall temperature distribution of both
condenser and evaporator decreases with the
increase in coolant flow rate and it can be seen
that the wall temperature increases with increase
in heat flux. There is no significant variation
between left and right side heat pipes of
thermosyphon.
[2] The thermal resistance decreases with the
increase of coolant flow rate but there is no
significant change after a flow rate of 0.2 l/min.
[3] Thermal resistance of the thermosyphon also
decreases with the increase in thermal load.
[4] Overall heat transfer coefficient of the
thermosyphon increases with the increase of both
heat flux and coolant flow rate. At higher heat
flux condition of 7.35 kW/m2, the overall heat
transfer coefficient of methanol is 0.126 kW/m2
0
C, the overall heat transfer coefficient of acetone
is 0.2 kW/m2 0C, the overall heat transfer
coefficient of ethanol is 0.17 kW/m2 0C the
overall heat transfer coefficient of water is 0.16
kW/m2 0C. There is no significant variation of
thermal resistance and Overall heat transfer
coefficient between left and right side
thermosyphon.
[8] Lanchao Lin et. al High Performance Miniature

Heat Pipes, Thermal Management Research
Studies. Volume 1 (2006). Page-86.
[9] Amir Faghri et. al. Micro/ Miniature Heat pipe
AnalysisSorming Media, (2001)Page 175.
[10] Jun Zhuang et. al. Comparison of Heat Transfer
Performance Of Miniature Heat
Pipes,
pp.226~229,
proceedings
of
The
11th
International Heat Pipe Conference- Tokyo 1999.
[11] Thang Nguyen et. al. Prediction of long-term
performance of Miniature Heat
Pipe from
Accelerated Life tests, pp.230~233, proceedings
of The 11th International Heat Pipe ConferenceTokyo 1999.
NOMENCLATURE
Symbol
Meaning
Unit
do
Outside diameter of pipe
mm
di
Inside diameter of pipe
mm
Le
Length of evaporator
mm
Lc
Length of condenser
mm
La
Length of adiabatic section
mm
Te
Evaporator Temperature
Tc
Condenser Temperature
Fluid Flow rate
Ae
Area
REFERENCES
Thermal resistance
[1] Chowdhury, F., et al., Study on heat transfer
Ut
overall
heat
coefficient
characteristics of looped parallel thermosyphon, Proc.

4th European Thermal science Conf., S10-HPI-1, 2004.
[2] Kaminaga, F., et al., Heat transfer characteristics in an

evaporator section of a looped parallel thermosyphon,
Proc. 7th International Heat pipe Symposium, 2003, pp.
225-230.
[3] Kim, K. S., et. al. Cooling Characteristics of

Miniature Heat Pipes With Woven Wired Wick. pp
239 -244, proceedings of the 11th international Heat
Pipe Conference Tokyo 1999.
[4] Luca Rossi, Thermal Control of Electronic equipment

by
Heat
Pipes
and
Two-Phase
ThermosyphonsELBOMEC thermalloy , Via del
Tipografo, 440138 Bologna, Italy.
[5] Eguchi, K., Mochizuki, M., Mashiko, K., Goto,

K., Saito, Y., Takamiya , A.,Nguyen, T., Cooling of
469
L/min
m2
0
transfer
C/W
kW/m2
0
C

OPTIMIZATION OF SHIP HULL PARAMETER OF INLAND VESSEL WITH

RESPECT TO REGRESSION BASED RESISTANCE ANALYSIS.
Goutam Kumar Saha1 and Asim Kumar Sarker2
1
Associate Professor
Department of Naval Architecture & Marine Engineering
Bangladesh University of Engineering &Technology
Dhaka, Bangladesh
E-mail: goutamkumar@name.buet.ac.bd
2
Postgraduate Student
Department of Naval Architecture & Marine Engineering
Bangladesh University of Engineering & Technology
Dhaka, Bangladesh
E-mail: asimhitler@yahoo.com
ABSTRACT
This paper represents an integrated methodology for the preliminary optimum design parameters considering
the existing constraints provided by Bangladesh Inland waterways Authority (BIWTA). This research also
describes a parametric modeling optimization approach to the design of ship hull parameters to minimize the
ship resistance as well as allows creating and varying ship hull parameters quickly and efficiently within the
given constraints. The methodology of optimization process such as objective function, design variables and
used constraints are described. The hull resistance is chosen as the objective function and ship length, breadth,
draft and speed is selected for design variables. From the study it is shown that the methodology may be used
in the preliminary design stages for selecting hull parameter of inland vessel operating in Bangladesh
waterways.
Key words: Resistance, length, breadth, draft, speed, SQP
1. INTRODUCTION
Preliminary ship design is currently more art than
science, heavily dependent on highly experienced
naval architects [12]. In the early stages of conceptual
and preliminary design, it is necessary to develop a
consistent definition of a candidate design in terms of
its dimensions and other descriptive parameters such
as Length, Breath, Draught, Block coefficient, LCB,
etc. This description can then be optimized with
respect to some measures of merit. More detailed
design development involves significant time, effort.
It is important to be able reliably define and size the
vessel at parameter stage. Because these parameters
(length, breath, depth, draught, speed etc.) have
influences on resistance, capital cost, maneuverability,
longitudinal strength, hull volume, sea-keeping,
transverse stability, freeboard, etc.
Optimization means finding the best solution from
a limited or unlimited number of choices. Even if the
number of choices is finite, it is often so large that it is
impossible to evaluate each possible solution and then
determine the best choice. The target of optimization
is the objective function or criterion of the
optimization. It is subject to boundary conditions or
constraints. Constraints may be formulated as
equations or inequalities. All technical and
economical relationships to be considered in the
optimization model must be known and expressed as
functions. Some relationships will be exact and others
will only be approximate, such as all empirical
formulae, e.g. regarding resistance. Procedures must

be sufficiently precise, yet may not consume too much
time or require highly detailed inputs. Ideally all
variants should be evaluated with the same
procedures.
The result of the optimization model should be
compared against built ships. Consistent differences
may help to identify important factors so far neglected
in the model. A sensitivity analysis concerning the
underlying estimation formulae will give a bandwidth
of optimal solutions and any design within this
bandwidth must be considered as equivalent. If the
bandwidth is too large, the optimization is
insignificant.
A critical view on the results of optimization is
recommended. But properly used optimization may
guide us to better designs than merely reciprocating
traditional designs. The ship main dimensions should
be appropriately selected by a naval architect who
understands the relationships of various variables and
the pitfalls of optimization. An automatic optimization
does not absolve the designer and his responsibility. It
only supports him in his decisions. The result of the
optimization model should be compared against built
ships. Consistent differences may help to identify
important factors so far neglected in the model.
This paper will focus on parametric ship hull
development and design optimization of hull
parameters operating in inland waterways of
Bangladesh.
471
2. METHODOLOGY
This
methodology
describes
an
automatic
optimization procedure of hull parameter to minimize
the ship resistance within certain given constraints.
The methodology shall be valid for inland vessels are
performed as followed:
(a) Determination of initial ship hull parameters
(i.e. Length, Breath, draught, Speed etc)
(b) Other
hull
parameter
components
calculation.
(c) Resistance calculation with different
methods.
(d) Setup constrains for hull parameter
optimization.
(e) Using Algorithm for hull parameter
optimization.
2.1
Ship Hull Parameter
2.1.1 Length
A general consideration of hull resistance versus
length shows that frictional resistance increases with
length as the wetted surface increases faster than the
frictional resistance coefficient declines with
Reynolds number. The wave resistance, however,
decreases with length. The net effect is that resistance
as a function of ship length typically exhibits a board,
flat minimum. Since the hull cost increases with
length, an economic choice is usually a length at the
lower end of this minimum region where the
resistance begins to increase rapidly with further
length reduction. Below this length higher propulsion
requirements and higher operating cost will then
offset any further reduction in hull capital cost. A
number of approximate equation exits in the literature
for estimating vessel length from other ship
characteristics. In Posdunines formula,
VK
L m
C
/
VK 2
Where,
= Displacement (tones) and
VK = Speed (knots) and
C = a coefficient can be generalized from similar
vessels.
Another welknown formula for estimating vessel
length is Schneekluths [12] formula,
Lpp (m) = 0.3*V0.3*3.2*
Where:
LPP = length between perpendiculars [m]
= displacement [tonne]
V = speed (knot)
Fn = = Froude number, V/ .
The formula is applicable for ships with 1000 t
and 0.16 Fn 0.32.
Various non-dimensional ratios of hull dimensions can
be used to guide the selection of hull dimension or
alternatively used as a check on the dimension
selected based upon similar ships, functional

requirements etc. Each designer develops his or her
own preferences, but generally the length-breath ratio
L/B, and the breath-draught ratio B/D, prove to be
most useful.
2.1.2
Breadth
Where width can be chosen arbitrarily, the width

will be made just as large as the stability demand. For
ships with restricted dimensions (particularly
draught), the width required for stability is often
exceed. A lower limit for breadth comes from
requiring a minimum metacentric height and
indirectly, a maximum possible draught. The
metacentric height requirement formulation is an
inequality requiring values which are frequiently
obtained for some ships.The length-breath ratio can be
used to check independent choices of length and
breadth with initial length, a choice of L/B ratio can
be used to obtain an estimated breadth . The L/B ratio
has significant influence on hull resistance and
maneuverability (both the ability to turn and
directional stability). With the primary influence of
length on capital cost, there has been a trend toward
shorter wider hulls supported by design refinement to
ensure adequate inflow to the propeller. From Watson
and Gilfillan [10] recommendation,
L/B = 4.0
for L 30 m
L/B = 4.0 + 0.025 (L 30) for 30 L 130 m
L/B = 6.5
for 130 m L
2.1.3
Draught
The third most important non dimensional ratio is

the breadth-draught ratio B/T. The breath-draught
ratio is primarily important through its influence on
residuary resistance, transverse stability, and wetted
surface. In general, values range between
2.25B/T3.75. The breath-draught ratio correlates
strongly with residuary resistance, which increases for
large B/T. Recommendation for maximum breathdraught ratio by Roseman etal. [8] Is
(B/T) max = 9.625-7.5 CB
2.1.4 Other Hull Form Parameter

Other
hull
parameter
components
are calculated using following formulae:
2.1.4.1 Block Coefficient
Well-known Alexanders
coefficient, CB is given by
CB
Where, K=1.33
0.54
formula
0.5
for
block
VK
L
V
0.24 K
L
2.1.4.2 Midship coefficient (CM)

Mid ship coefficient CM is determined from
Schneekluth and Bertrams [10] formula
CM 1.006 0.0056CB .
472
CM
CB
Reynolds Number, Rn =
Where, = Viscosity of water
Detailed will be obtained from Holtrop [2, 3, 4]
2.1.4.3 Prismatic coefficient (CP )

Prismatic coefficient from empirical relation of
parameter
CB
CP
CM
2.1.4.4 Water plane coefficient (CWP )
Water plane coefficient from Schneekluths and
Bertrams [9] formula
CWP
CP
2.3.2 Hollenbach Method
CB
CM
3
And from Riddlesworths formula
1 2CB
CWP
3
2.1.4.5 Longitudinal Center of Buoyancy (LCB)
Longitudinal Center of Buoyancy from Schneekluth
and Bertrams [9] formula
CWP
LCB
LCB
2.2
8.80 38.9Fn
13.5 19.4CP
According to Hollenbachs [1] formula
B T
RR
CR
V
2
10
Where,
RR= Residual Resistance
CR= Coefficient of Residual Resistance
CR= CR, Standard * CR, Fnkrit * kL*(T/B)b1*(B/L)b2
*(Los/Lwl)b3*(Lwl/L)b4*(1+(TA- TF)/L)b5*(DP/TA)b6*
(1+NRudd)b7*(1+NBrac)b8*(1+NBoss)b9*(1+NThruster)b10
CR, Standard= c11+ c12Fn+ c13Fn2+CB*(c21+c22Fn+ c23Fn2)+
CB2*(c31+c32Fn+ c33Fn2)
CR, Fnkrit = max(1.0, (Fn/Fn, krit)f1
Fn, krit = d1+d2CB+d3CB2
KL = e1Le2
Detail description will be obtained from Hollenbachs
[1].
2.3.3 Van Oortmerssen Method

According to Oortmerssens [5, 6] formula
Speed
The speed can be decisive for the economic

efficiency of a ship and influences the main
dimensions inturn. The speed is determined largely in
accordance with the ideas and wishes of the ship
owner, and is thus outside the control of the designer.
The optimum speed, in economic terms, can be
related both to favourable and to unfavourable Froude
numbers. In General the trial speed will be considered
the normal basis for the optimization. However, the
service speed could be included in the optimization as
an additional condition.
RR
CR=C1*
+C2*
+C3*
+C4*
*
m= 0.14347*CP-2.1976
Where,
RR= Residual Resistance
= Displacement of Ship
CR = Coefficient of Residual Resistance
Detail description will be obtained
Oortmerssens [5, 6].
from
2.3 SQP Algorithm

2.3 Regression Based Resistance Method
The various methods used in the resistance calculation
are described below.
2.3.1 Holtrop Method

According to Holtrops [2,3,4] formula
RT
Where,
RTotal = Total Resistance
RF = Viscous Resistance
1+K1 = Form factor describing the viscous resistance.
RAPP = Appendage Resistance
RW = Wave Resistance
RB = Bow Resistance
RTR = Transom Resistance
RA = Model ship correlation Resistance
Viscous resistance RF =CF* *S*V2
Viscous
Coefficient
[7]
CF =
The sequential quadratic programming (SQP) method

is a general method for solving nonlinear optimization
problem with constrains. Let the optimization
problem be written as
Minimize
FX
Subject to
h X
0, j
1, M
1, M
g X
0, j
M
Where F is an objective function, and hi and gi are
equality and inequality constrains, respectively. When
the current design point is X(K), the next design point
X(K+1) is determined as follows. First, the following
quadratic programming problem is solved to obtain
the modification of d.
Minimize
dT H K .
F X K T.
Subject to
h
.
, j 1, M
h
g
.
,
j 1 M ,M
Here, the objective function F is approximated as the
quadratic function of X(K), and the constrains are
approximated as the linear functions of X(K). H is an
473
approximation
Lagrangian
to
the
Hessian
matrix
of
the
Table 1. Objective function, Design variable and

Design constraints of SQP.
Where u = ( , ,
. T and v =
T
.
are
the
Lagrangian
( , ,
multipliers for equality and inequality constrains,
respectively.
The next design point X(K+1) is obtained by
the line search along the vector d. The step-size d is
determined in such a way that the penalty function
with r, being a penalty parameter where
u=( , ,
. T and
v = ( , ,
. T are the Lagrangian
multipliers for equality and inequality constrains
respectively.
The next design point X(K+1) is obtained by
the line search along the vector d. The step-size d is
determined in such a way that the penalty function
with r, being a penalty parameter becomes smaller
than a set value.
Resistance
Coefficient
Objective
Function
Design
Variable
Length
(L)
Breadth
(B)
Draft (T)
Speed
Design
Constraints
0<L<75m
0<B<13.5m
0<T<3.5m
10<V<15 knots
2000<Displacement<2500
m3

Objective Function
Optimized
Parameter
Constrains
Flow Solver
Optimizer
0,
Finally, the next design point is computed by
The Hessian matrix is updated using the Broyden,
Fletcher, Goldfarb, and Shanon (BFGS), as follows:
Alternative Hull
Parameter
Figure-1: Optimization Process
Where
g
1
1.0
0.8
0.2
0.2
2.4. Design Variables and Constraints

For the inland waterways of Bangladesh, the
Length, Breadth, and Draught are very much limited
and sometimes speed. Generally optimization of ship
hull is carried out at a particular speed such as trial
speed, service speed etc. It is also possible to perform
optimization for multiple speeds by increasing the
objectives of multi-objective optimization. Hull
optimization is based on minimization of ship
resistance with following criterion. Table 1. Shows
the objective function, design variable and design
constraints in the optimization process.
A numerical method has been described for the hull

parameter optimization of inland vessel operating in
Bangladesh waterways with respect to ship resistance.
Combining the numerical method for ship resistance
with SQP program, improved design parameters can
be generated through a series of iterative
computations. The optimization process is described
with above Figure-1.
Figure-2 shows the resistance vs. length and
breadth of a typical ship for constant speed and draft.
From this Figure it is shown that the resistance
calculated by Holtrop method increases faster with the
increase of breadth rather than ship length.
Figure 3 shows the convergence history of ship
hull resistance in the optimization process of Holtrop,
Hollenbach and Van Oortmerssen method. Table 2
describes the optimized ship hull parameters of
Holtrop, Hollenbach and Van Oortmerssen method.
474
1.00
Van Oortmersseen
Holtrop Method
Rtotal/Original Rtotal
Draft = 3.5 m
Speed = 12 Knots
0.75
0.50
Holtrop
Hollenbach
0.25
0.00
0
100
200
300
400
500
600
700
800
900
1000
Iteration No.
Figure-2: Resistance of a ship hull versus length and

breadth by Holtrop method.
Figure-3: Convergence History of ship hull resistance

by Holtrop, Hollenbach and Van Oortmerssen method
Optimized Hull Parameter

Hollenbach
Van Oortmerssen
method
method
74.986
74.976
Length (m)
75.0
Holtrop
method
74.909
Breadth (m)
13.5
11.04
13.056
12.72
Draft (m)
3.50
3.037
2.986
3.342
Speed (Knots)
15.0
10.00
10.00
14.10
Block Coefficient
0.594
0.796
0.684
0.627
Midship
Coefficient
Water Plane area
Coefficient
Prismatic
Coefficient
Displacement (m3)
0.976
0.996
0.984
0.977
0.734
0.865
0.793
0.757
0.608
0.799
0.695
0.643
2105
2000
2000
2000
4. CONCLUSION
An integrated methodology for the basic preliminary
ship hull parameter evaluation and optimization of
these parameters for inland waterways of Bangladesh
has been presented. The waterways condition of
Bangladesh has been limited the option of thinking
about vessel of large dimensions. The main difficulty
of the numerical optimization lies in formulating the
objective function, design variables and all the
constraints. The optimization problem has been
Others Calculated
Parameters and
constraints
Initial Hull
Ship Parameter
Design Variables
and constrains for
Optimization
Table 2 Design parameters of optimized hull of different method
carefully formulated to give a stable inspection of

every approach. The constraints are based on the
design parameter requirements. The objective
function is normalized with respect to its initial value.
For the optimization process, empirical formulae for
the calculation of design parameter used. For more
accurate or detailed optimization, require a more
separate work. The methodology may be used in the
preliminary design stages for selecting hull parameter
of inland vessel operating in Bangladesh waterways.
475
ACKNOWLEDGEMENTS
organizer for giving us the opportunity to present the
research work.
[6] Oortmerssen, G. van, A power prediction

method for Motor Boats, 3rd Symposium Yacht
Architecture, The Netherlands, also NSMB
publication 429.
[7] Proceedings 15th ITTC [1978], The Hague.
REFERENCES
[1] Hollenbach K.U., Estimating resistance and
propulsion for single screw and twin screw
ships, Ship technology research, 1998
[8] Roseman, D. P., Gertler, M., and Kohl, R. E.,

Characteristics of Bulk Products Carriers for
Restricted-Draft Service, Transactions SNAME,
Vol. 82, 1974.
[2] Holtrop, J and Mennen, G.G.J A statistical reanalyis of resistance and propulsion data
International Shipbuilding Progress, 1984.
[9] Schneekluth, H. and Bertram, V., Developments

in the Design for Efficiency and Economy,
Second
Edition,
Butterworth-Heinemann,
Oxford, UK, 1998.
[3] Holtrop, J and Mennen, G.G.J An approximate

power
prediction
method.
International
Shipbuilding Progress, Vol. 29. July, 1982.
[10]
[4] Holtrop, J and Mennen, G.G.J., A statistical

power
prediction
method.
International
Shipbuilding Progress, Vol. 25. October, 1978
[5] Oortmerssen, G. van, A power prediction
method and its applications to small ships,
International Shipbuilding Progress, Vol. 19,
1971.
Watson, D.G.M., and Gilfillan, A.W. , Some

Ship Design Methods, Transactions RINA,
Vol.119, 1977.
[11] Wolf Robert, Dickmann John, Boas Ryan, Ship

design using heuristic optimization methods.
Massachusetts Institute of Technology, USA,
2004.
476

A THEORETICAL METHOD FOR THE DESIGN OF

MARINE TANDEM PROPELLERS
Muhammad Refayet Ullah
Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh
E-mail : refayet@name.buet.ac.bd
ABSTRACT
The propulsive efficiency is reduced with increasing propeller power absorption and the consequent risk of
propeller cavitation and induced vibration. By fitting tandem propellers better performance can be achieved. In
this paper a procedure is described which was basically developed for calculating the self and mutually induced
velocities of the two-propeller system. The typical examples of the two-propeller system are tandem propellers
and contra-rotating propellers. The main difference between double and single propeller is the calculation of the
mutually induced velocities which vanishes for single propeller. The tandem propellers can be described as a
pair of conventional propellers fitted on the same shaft in series and rotating in the same direction. The lifting
line theory is applied for the calculation of the velocity field. For the calculation of the induced velocities the
circulation theory of propeller action based on Biot-Savart Law is used. Radial distribution of the self and
mutually induced velocities are calculated on the lifting line in such a way that each propeller is handled like a
single propeller. For optimization of the design the axial distance between the two propellers and angular
spacing of the blades of the two propellers are taken into consideration. The performance characteristics such
as thrust, torque and efficiency of individual and combined system as calculated are compared with the
available results.
Key words: Tandem propeller, propeller efficiency, lifting line theory, Biot-Savart Law.
1. INTRODUCTION
The tandem propellers can be described as a pair of
conventional propellers fitted on the same shaft in
series and rotating in the same direction. The well
known trend of reducing propulsive efficiency with
increasing propeller power absorption and the
consequent risk of propeller cavitation and induced
vibration means that the role of the conventional
propeller is virtually exhausted in some applications.
The tandem propeller can then be regarded as a
practical means of extending this difficult operating
range [1,2,7].
By fitting tandem propellers the variable
hydrodynamic loads transmitted to a propeller shaft
can be decreased [4].
The tandem propeller would utilize the same
turbines, reduction gears and shafting as any other
single screw installation and consequently would be
less expensive and complex than contra-rotating and
twin-screw installation.
2. THE IMPORTANT DESIGN

CHARACTERISTICS
The following additional characteristics are to be
considered compared with those in the case of a single
screw propeller system.

- thrust distribution between the forward propeller
and the aft propeller.
- distance between two propellers.
- diameter of two propellers.
- angular displacement.
- wake field of the forward propeller and the aft
propeller.
2.1 Thrust distribution

In order to investigate the effect of thrust
distribution between the forward and aft propellers on
the overall propulsion efficiency, it was planned to
design the tandem propeller system according to
various thrust combinations. The combinations may
be 40:60, 50:50, 60:40, 70:30 and so on. However,
only the thrust combination 50:50 for the forward and
aft propellers is considered for the present calculation.
2.2 DISTANCE BETWEEN TWO

PROPELLERS
The ratio of the axial spacing between the planes
of the forward propeller and aft propeller and the
diameter, L / D is the axial spacing ratio. It is thought
that the conventional propeller is more efficient than
the tandem propeller and that since in the limit when
477
the axial spacing ratio is zero the tandem propeller

becomes a conventional propeller i.e., the tandem
propeller would be more efficient as axial spacing is
smaller. According to various studies [5] the value is
between 0.2 to 0.25. In the present case it is taken as
L / D = 0.2
2.3 DIAMETER OF TWO PROPELLERS

In some studies it is suggested that the diameter of
the aft propeller be smaller than the forward propeller,
which is true for contra-rotating propellers. In contrarotating case the diameter of the aft propeller is
reduced to absorb more rotational energy in the
slipstream and to avoid interference between the tip
vortex of the fore propeller and aft propeller. As
tandem propeller is different due to the fact that in
general tip vortices does not disturb each other and
also when the diameter of aft propeller is same as
forward propeller, the loading of the blades will be
reduced and the efficiency of the tandem propeller
improved. For the present case, the diameter of both
the propellers are considered the same.
2.4 ANGULAR DISPLACEMENT

The angular displacement, denotes the angle
between the aft propeller blade reference line and the
neighbouring forward propeller blade. The best
angular displacement of the aft propeller to the
forward propeller is that which allows the vortex
sheets of the forward propeller to pass midway
between the after propeller blades. This angular
displacement is determined by considering the angle
at which the flow left the forward propeller blades at
the 0.7 radius and the axial spacing between the
blades [5].
According to circulation theory the best angular
spacing is approximated by
= ( L / D ) / ( P1 / D ) 360 1800 / z
where, P1 / D at 0.7R = .x.Tan i for forward
propeller
z = number of blades
2.5 WAKE FIELD OF THE FORWARD AND

AFT PROPELLER
using a field point velocity program based upon

lifting line theory using Biot-Savart Law. The
computer program for the said purpose is
prepared by Ullah [8] and used for the present
calculation. The induced velocities on the
forward propeller are the interference velocities
of the after propeller on the forward propeller.
Conversely, the induced velocities on the after
propeller are the interference velocities of the
forward propeller on the aft propeller.
The physical parameters of each propeller in the
tandem set are now determined.
3.1 CALCULATION OF PROPELLER

INDUCED VELOCITY
The induced velocities by the vortex system
defining the propeller at any field point are calculated
by using the Biot-Savart law.
The velocity dv induced at a field point at a
distance d by a vortex filament of length dl and of
circulation is given by Biot-Savart law as,
dv = ( /4) . (dl d) / d 3
3.2 THE VELOCITY DIAGRAM

In calculating the performance of a tandem
propeller, the main flow features present are taking
into account, namely the interference between the
propellers.
In this method it is assumed that
-
the vortex sheets of the forward and aft

propellers are independent
the forward and aft propellers have the same

blade number, diameter and hub diameter
both the propellers are moderately loaded

and wake adapted. As moderately loaded, the
slipstream contraction effect is not
considered in the calculation.
The blades of a tandem propeller experiences the

self induced velocities ua / 2 (axial) and ut / 2
(tangential) and additionally the interference
velocities uai (axial) and uti (tangential) induced
from one set of blades on the other.
It is estimated [8] from the induced velocity field

by each propeller on the other i.e., interference
velocity.
The self induced velocities should be equivalent to

those induced by an isolated conventional propeller in
presence of velocity components VAf,a and 2nr.
3. DESIGN STEPS
The inflow velocity VAf,a can be expressed at any

radius r as follows.
VAf,a = VA + uai
In Figure 1, the velocity vector diagram is
presented for single propeller and tandem propeller
system.
The following design procedure is adapted,

-
A propeller producing half the required thrust is

calculated.
The induced velocities at selected control points
of the forward and aft propellers are calculated
478
averaged axial nominal wake distribution [3] are

shown in Table 1 and Figure 2 respectively.
4. A TYPICAL CASE STUDY

Design conditions have been selected according to
the discussion made in the previous section. The
selected design conditions and the circumferentially
Ua /2
Ut /2
VA = VS [1-(x)]
2nr
Single Propeller System
Ua /2
Ut /2
Uai
VA = VS [1-(x)]
VAf,a = VA + uai
2nr
Tandem Propeller System
Figure 1. Velocity diagram related to single and tandem propeller system.
Table 1. Design conditions for the propeller system

Effective horse power
Ship speed
RPM
Number of blades
Diameter of propeller
Thrust deduction fraction
Wake fraction
Thrust ratio
Distance between two propellers
Angular displacement
30,800 HP
19.25 Knots
105
5 forward and 5 aft
6860 mm forward and 6860 mm aft
0.143
0.335
50% by forward propeller and
50% by aft propeller
L/D = 0.2 or 1372 mm
61.1267 degree at 0.7 radius
479
0.9
0.8
0.7
1-w(x)
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.2
0.4
0.6
0.8
1.2
Figure 2. Circumferentially averaged nominal axial velocity distribution ( single propeller )
First a propeller producing half the required thrust

is calculated. The forward and aft propellers are
designed, taking into consideration the design of this
single propeller producing half the total thrust. The
radial distribution of hydrodynamic pitch angle and
the circulation of the single propeller is provided in
Table 2.
Using field point velocity program the axial
induced interference velocities are calculated and the
calculated velocities are provided in Figure 3. These
velocities in Figure 3 and the velocities of Figure 2
providing the inflow velocity to the forward and aft
propeller are presented as wake velocities for the
forward and aft propeller and presented in Figure 4.
Now the hydrodynamic design procedure is
followed for the forward and aft propellers separately
and the important design parameters are provided in
Table 3.
5. CONCLUSION
In this paper, some calculation results of the
authors recent study on the design of the tandem
propeller system have been briefly presented and
discussed together with specific example. Propulsive
efficiency is significantly improved by a tandem
system, and hence the same amount of fuel-saving
could be achieved. For the single propeller design
with half the total system thrust, the efficiency
achieved is 0.6106. Whereas, for the tandem propeller
system this efficiency achievement is 0.624 for
forward propeller and 0.6411 for aft propeller. Both
the propellers delivered half the total thrust. Thus the
performance of individual propeller can be observed.
Table 2. Radial distribution of hydrodynamic pitch angle and

circulation of the single propeller
x
i ( deg. )
0.2
46.9666
0.3
35.5995
0.4
28.4378
0.5
23.8880
0.6
20.8934
0.7
18.6286
0.8
16.9563
0.9
15.4459
1.0
14.0907
1 = 0.6650
T = 1350.24 KN
CT = 1.6304
= 0.6106
480
G
0.0325
0.0455
0.0507
0.0477
0.0387
0.0296
0.0189
0.0114
0.0000
0.08
0.07
0.06
u / Vs
0.05
Forward propeller
0.04
Aft propeller
0.03
0.02
0.01
0
0
0.2
0.4
0.6
0.8
1.2
Figure 3. Axial induced velocity distribution at the forward and aft propeller plane.
1
0.9
0.8
1 -w(x)
0.7
0.6
Forward propeller
0.5
Aft propeller
0.4
0.3
0.2
0.1
0
0
0.2
0.4
0.6
0.8
1.2
Figure 4. Nominal axial wake velocity distribution at the forward and aft propeller plane.
Table 3. Important Design Parameters
x
Forward propeller
Aft propeller
i ( deg. )
C.CL / D
G
i ( deg. )
C.CL / D
G
0.2
47.9758
0.1059
0.0304
49.3636
0.1035
0.0283
0.3
36.5519
0.1088
0.0432
37.9105
0.1098
0.0409
0.4
29.2537
0.0969
0.0487
30.4756
0.0991
0.0464
0.5
24.5706
0.0803
0.0464
25.6516
0.0824
0.0443
0.6
21.4659
0.0623
0.0381
22.4165
0.0634
0.0362
0.7
19.1198
0.0464
0.0296
19.9820
0.0465
0.0277
0.8
17.3799
0.0299
0.0192
18.1556
0.0290
0.0175
0.9
15.8100
0.0177
0.0119
16.5132
0.0167
0.0106
1.0
14.4106
0.0000
0.0000
14.9642
0.0000
0.0000
1 = 0.7360
1 = 0.6870
T = 1349.64 KN
T = 1349.62 KN
CT = 1.3386
CT = 1.53637
= 0.6240
= 0.6411
Required thrust to be produced by each propeller, T/2 = 1352.46 KN
Required thrust to be produced by both propeller, T = 2704.92 KN
Actual thrust produced by both propellers = 1349.62 + 1349.64 = 2699.26 KN
481

REFERENCES
[1] Denny, S. B., A procedure for the design of
tandem propellers, AD A039027 (1974).
[2] Glover, E.J.. Contrarotating propellers and
tandem propellers, Proceedings of Symposium,
Advance
in
propeller
research
and
design,Gdansk, (1981).
[3] Hadler, J. B., Morgan, W.B. and Meyers, K. A.,
Advanced propeller propulsion for highpowered single-screw ships, Transaction of
SNAME, pp. 231-293 (1964).
[4] Meyers, K.J., Characteristics and performance
of tandem propellers, Thesis, Department of
Massachusetts Institute of Technology (1961).
[5] Qin, S. and Yunde, G., Tandem propellers for

high powered ships, Transaction of RINA, pp.
347-362 (1991).
[6] Rahman, N.A. and Islam, M.S., Parametric
design and comparison of non-cavitating and
fully-cavitating marine propeller in uniform
flow, B.Sc.Engg. Thesis, Bangladesh University
of Engineering and Technology (2009).
[7] Titof, I. A. and Biskup, B. A., Investigation into
the possibilities of tandem propeller application
with the aim of decreasing the variable
hydrodynamic loads transmitted to a propeller
shaft,The 11th ITTC Proceedings (1966).
[8] Ullah, M. R., A method for the calculation of
performance characteristics of the two-propeller
system: Part 1- Tandem propeller, A Report,
Department of Naval Architecture and Marine
Engineering, BUET, pp. 1-41 (2000).
482

DESIGN AND DYNAMIC BEHAVIORS OF MARINE STRUCTURES

Yoshiyuki Inoue
2
Professor Emeritus, Yokohama National University, Japan

E-mail: y-inoue@ynu.ac.jp
ABSTRACT
Understanding the dynamic behavior of ships in waves is essential not only for the design of hull structure but
also those of equipments and fittings referred in IMO which is the United Nations agency concerned with safety
of shipping and protection of the marine environment. Those out fittings such as Emergency Towing
Arrangements ETS , towing and mooring equipments, doorways and ventilators, hatch covers,
miscellaneous openings in freeboard and superstructure decks, windows and skylights are affected by dynamic
behaviors of ship in wave.
To design a securing system of container ship, induced loads due to ship motions in the sea should be the
most important matters, especially for huge container ships which are recently built for efficiently sea
transportation means.
This paper reviews a number of papers by the author on the dynamic tensions of towlines and motion
responses of various ships in short crested irregular waves to better understand and discuss the design criteria
of out fittings of ships.
Key words: dynamic behavior in waves, design of marine structure, ship motion,
1. INTRODUCTION
To protect the marine environment from pollution
as a result of shipping incidents, IMO[1] requires tank
ships operating in international to maintain emergency
towing equipment on board and to conduct emergency
towing drills. The International Convention for Safety
of Life at Sea (SOLAS) Chapter 5, Regulation 15.1
requires all tankers of 20,000 deadweight tons and
above and built after January 1996 to be fitted with
emergency towing arrangements at both ends of the
ship. The 75th session of the Maritime Safety
Committee (MSC-75) has also adopted Resolution
MSC-132(75) on Amendments to the guidelines on
emergency towing arrangements for tankers
(resolution MSC-35(63)) in 2002.
To provide significant improvements to the
structural safety of ships, in particular bulk carriers,
following on from recommendations of the United
Kingdom Report of the re-opened formal
investigation into the loss of the Derbyshire in 1980,
IMO MSC-77(2003)[1] has adopted a revised Annex
B to the 1988 Load Lines Protocol. The amendments
to Annex B include a number of important revisions,
in particular to regulations concerning strength and
others.
The MSC-80 (2005)[1] has adopted the
amendments to SOLAS regulation II-1/3-8
concerning towing and mooring equipment. The

regulation will require all ships to be provided with
arrangements, equipment and fittings of sufficient safe
working load to enable the safe conduct of all towing
and mooring operations associated with the normal
operation of the ship.
Following the SOLAS amendments, IACS[2] and
classification societies of shipping[3] have also
improved the safety of ship in waves.
Concerning these issues the author reviews his
papers on the dynamic tensions of towing lines and
motion response of various ships in short crested
irregular waves to understand and discuss the design
criteria of out fittings of ships in his paper[4].
Recently very large container ships are built and
become more economical ocean transportation means.
Very large induced loads on the containers due to ship
motions in the sea should be considered for stacking
the containers on the ships and their securing systems.
The author also reviews a recent study[5] of motion
responses of a mega container ship in short crested
irregular waves to give some information for the
safety of securing system of huge container ships in
the sea.
2. DYNAMIC TENSION OF TOWING

LINES
Towing marine structure to the required site in the
sea, towing a tanker in salvage operation and others
are indispensable in ocean engineering. In calm sea
conditions propulsive force is responsible for the
motion of the tow system against the tow rope pull
and external forces. However, the towing operations
become critical with the presence of wind, waves and
current.
A study of the dynamic analysis of tension in towline
by using the three dimensional 1umped mass method
is carried out[6]. In the paper a parametric study of
the dynamic analysis of tension of towline in the
various conditions of amplitude and frequency of
motions, pretensions and combination ratios of length
of wire and chain are presented.
From the numerical results it is recognized that
decrease of period of motion causes increase of
dynamic tension, increase of length of portion of
chain of total line decreases dynamic tension and
increase of initial tension or motion amplitude also
increases dynamic tension. And the most effective
way to reduce the dynamic tension is to use chain
together with wire rope for towline. On the other hand,
the length of the towline of wire rope is not so
effective to reduce the dynamic tension.
Figure 1 shows the influences of length of chain and
motion period on dynamic tensions of lines in case of
total length of 1000m and motion amplitude of the
towline end is 5m. From this we can understand that
the chafing chain for towing and mooring
arrangements is important means to reduce dynamic
tensions of lines in dynamic motions.
Figure 1 Influences of length of chain and motion

period on dynamic tensions of lines
A simulation study on the dynamics response of

tanker-tug tow system in the seas is also
investigated[7]. Three tankers each of 20,000dwt,
100,000dwt and 200,000dwt have been considered.
Two tugs of different pulling capacity of l00kN and
400kN are assumed for towing operation.
The principal particulars of the tankers are shown in
Table 1. The detai1s of the towline used onboard the
tankers are also given in the Table 1. These
parameters have been determined referring to the
equipment number recommended by rules and
guidance of NK[3].
The principal particulars of two tugs for the
computer simulations are shown in Table2. And the
calculating conditions of sea states are shown in Table
3.
Table 2 Principal particulars of tugs
tugs Length Breadth draft displace
(m)
(m)
(m)
ment
(tons)
A
16.0
6.1
3.0
180
B
27.0
10.0
4.0
900
Table 3 Calculation conditions of sea states

Beaufort significant wave
wind speed
scale
(knots)
height ( m) period( s)
3
0.6
5.2
8
4
1.0
6.0
12
5
2.0
7.1
18
6
3.0
7.9
24
7
4.0
8.4
30
8
5.5
8.8
36
Figure 2 towing configuration in horizontal plane with

external force
Table 1 Principal particulars of tankers

Tankers
A
B
C
pull
load
(kN)
100
400
Deadweight
(tons)
LxBxd
(m)
equipment
number.
20,000
100,000
200,000
135x21.3x8.6
224x35.9x13.7
293x47.0x17.3
1570
3210
5500
length
(m)
220
280
300
towline
diameter
breaking
(mm)
load(kN)
44
941
56
1471
56
1471
Figure 2 shows towing configuration in horizontal

plane with external forces. The dynamic response
motions in horizontal plane of tug and tanker tow
system are simulated by solving the coupling motion
equations of surge, sway and yaw of two ships
connected by spring in time domain. The
environmental forces and moments include wave
exciting force, current force and wind force. The
motion history of two ships influences the tension
variation on the towline. The dynamic tension history
is obtained from the non-1inear static characteristics
with the horizontal distance of the towline at each
time step.
Figure 3 shows (the maximum dynamic tension of
towline) divided by (the mean pull of tug) (Td/Ts)
which are obtained from the tension response for
various conditions in time domain. The dynamic
tensions on the towline of unit mass of 11kg/m with
length varying from 500m to 1000m are shown in the
figure for various Beaufort scales. Figure 4 shows the
results for tanker C. From this the dynamic towing
tensions for the bigger tanker C shows much higher
than those for smaller tanker A. It may be considered
that the motion response of tanker in the sea affects on
the dynamic tension of towline.
Figure 3 Dynamic tension for tanker A and tug A

mean pull 100kN
Figure 4 Dynamic tension for tanker C and tug A

mean pull 100kN
Figure 5 Dynamic tension for tanker A and tug B

mean pull 400kN
The mean value of pull load for the movement of the
surface vessels is very sensitive for the dynamic
tension of the towline. Figure 5 shows the dynamic
tension for tanker A towed by tug B. Due to a higher
mean pull load of 400kN by tug B the dynamic
tension is much increased even in the longer towline
by comparing with Figure 3.
Figure 6 Dynamic tension for tanker B and tug B

mean pull 400kN
Figure 7 Dynamic tension for tanker C and tug B

mean pull 400kN
In cases of the same tug Bpull load 400kNfor

tanker B or tanker C no much differences are seen in
the dynamic tension as shown in Figures 6 and 7.
From these results it can be seen that during the
towing in the rough seas the dynamic tension of
towline is much higher than mean pull load by the tug
in case of without chain or some shock absorbers for
towing wire rope even the length of towline is more
than 1000m.
To reduce such high dynamic tensions, the additional
chain may affect significantly as shown in Figure 3.
From these simulations the dynamic tension
increases when, order of sea state is increased, the
mean pull is increased and the length of the towline is
decreased.
Rules and guidance of NK[3] show the following
criteria as shown in Table 4 which are based on IMO
guidelines for emergency towing arrangements on
tankers. As heavy and long chain have more effective
to reduce the dynamic tension of towline in the rough
seas, grade 2 chain which is heavier than grade 3 has
smaller safety factor for the design load as shown in
the Table4. But this may not be definitely considered
the dynamic tensions referred in this paper and may
be considered corrosion margin and others.
Table 4 design criteria for emergency towing
arrangements on tankers
chafing chain
deadweight Emergency
nominal
diameter,
Towing
Arrangements breaking test load
grade 2
grade 3
20,000tons 1,000kN type
62mm,
52mm,
and over
2,060kN
2,110kN
50,000tons 2,000kN type
90mm,
76mm,
and over
4,090kN
4,300kN
3. RELATIVE WAVE HEIGHTS OF SHIPS

IN SEVER SEAS
A ship will face various sea states during her voyage
in ocean route. So the necessity of accurate
predictions of ship motion, wave load, deck wetness
and others in these sea state are not only necessary
from the view point of seakeeping performance but
also from the serious damage of excessive wave load
in severe sea state which may even cause sinking of a
vessel.
For determination of deck load and bottom
slamming force, relative wave height may be
considered as an essential parameter. A parametric
study on relative wave heights of ships in short
crested irregular waves are carried out[8,9]. In the
paper linear potential theory has been used to describe
the fluid motion and 3-D Green function method with
forward speed has been used to determine
hydrodynamic forces for surface ship advancing in
waves. The time domain simulations of relative wave
heights of typical container ships, bulk carriers, pure

car carriers (PCC) and general cargo ships in short
crested irregular waves have been carried out for
sever sea states as shown in Table 5. Empirical roll
damping has been taken into account in time domain
analyses of motion responses of ships. The short
crested irregular waves are generated by using ISSC
wave spectrum and directional distribution function of
cosine square. And for ensuring longer time
simulation of the random sea waves, unequal
frequency spacing method is used.
Table 5 Sea State for Numerical Simulation
sea state
mean
(s)
moderate gale
strong gale
hurricane
6
10
10
period
significant
wave
height
(m)
5
10
15
As the numerical examples, typical bulk carriers of

different sizes are present in this paper. The principal
particulars of these ships are shown in Tables 6.
Table 6 Principal particulars of bulk carriers
Items
Bulk-14K
Bulk-140K
LBP (m)
120.0
260.0
B (m)
18.90
40.20
D (m)
10.56
22.18
D (m)
7.69
15.58
13783
135661
(m3)
8.0
9.8
U (knots)
0.790
0.8746
Cb
Cw
0.85
0.85
Cm
0.997
0.997
KG (m)
L.C.G. (m)
GM (m)
K xx
5.563
1.611
1.89
34.0 % B
12.20
3.629
4.02
34.0 % B
K yy
26.0 % LBP
26.0 % LBP
K zz
26.0 % LBP
26.0 % LBP
Figures 8~11 show an example of time domain

analysis for bulk carrier 140K in a short crested
irregular wave for strong gale condition at cursing
speed of 9.8 knots. Figure 8-10 show the motion
responses of heave, roll and pitch in this sea state.
Roll motion is induced even in head seas because of
short crested irregular waves. Figure 11 shows
corresponding relative wave height at forward
perpendicular FP position of the ship.
After calculating relative wave height in time
domain at side of the main hull, the maximum and the
1/3 highest mean values have been determined along
the ship length.
Heave motion (m)
-1
-2
-3
1200
1400
Tim e (s)
1600
1800
Figure 8 Heave motion of Bulk140K in strong gale

6
Roll motion (deg.)
-2
-4
-6
1200
1400
1600
1800
Tim e (s)
Figure 9 Roll motion of Bulk140K in strong gale

4
3
Pitch motion (deg.)
2
1
we can determine the corresponding height above

where the weather deck load becomes zero. This
corresponding height for the deck load may be
considered as above the maximum relative wave
height. The rule also requires closing appliances for
the openings of outfitting on deck and where exceeds
the some height of openings above the decks, such
closing appliances may be omitted. This height for the
closing appliances omitted may be considered as
above the 1/3 highest mean values of relative wave
height. These two kinds of heights are referred to the
relative wave heights in various sea states.
Figures 12~14 show the results the maximum and
the 1/3 highest mean of relative wave heights along
the ship length of Bulk-14K for 3 different sea states.
The small bulk carrier Bulk14K shows the relative
wave heights are below the corresponding height for
the deck load and also the height for the closing
appliances omitted in moderate gale. But In hurricane
the relative wave heights are over the reference
heights and bottom slamming may be occurred. In
strong gale the maximum values exceed the
corresponding height for the deck load at fore and aft
and bow slumming may be appeared some times.
Figures 15-17 show the results for the large bulk
carrier Bulk140K. The 1/3 highest mean as well as
the maximum value are below the corresponding
height for the deck load and also the height for the
closing appliances omitted and may not be appeared
bottom slamming even in hurricane condition.
[m]
0
-1
-2
-3
-4
1200
1400
1600
1800
Time (s)
Figure 10 Pitch motion of Bulk140K in strong gale

10
Relative wave height (m)
F.P. position
Figure 12 Relative wave height of Bulk14K in

moderate gale
[m ]
-5
-10
1200
1400
1600
1800
Time (s)
Figure 11 Relative wave height of Bulk140K at F.P.

position in strong gale
The minimum deck load for weather deck is required
by rules and guidance of NK[3] as a function of
vertical distance from the designed maximum load
line to the weather deck at side. From this requirement,
Figure 13 Relative wave height of Bulk14K in strong

gale
[m ]

hurricane
[m ]
highest mean value is almost below the height at

where closing appliances are not necessary as well as
corresponding height to the minimum requirement of
deck load.
IACS S27 Strength requirement for deck fittings
and equipment gives the velocity of water over the
fore deck is 13.5 m/s as the applied loading for air
pipes, ventilator pipes and their closing devices[2].
This dynamic pressure corresponds to the sea water
pressure head of 9.3m. And IACS S21 Evaluation of
Scantlings of Hatch Covers and Hatch Coamings (Rev.
4) also gives hatch cover load model[2]. By IACS
S21, the pressure at the forward perpendicular may be
water pressure head 5.0m for Bulk-14K and 6.0m for
Bulk-140K respectively. And the pressure at midship
part is 3.4m for both of Bulk-14K and 140K. These
water pressure heads are close to the heads of
corresponding to the 1/3 highest mean values of
relative wave height in strong gale for Bulk-14K and
in hurricane for Bulk-140K respectively. This may
suggest the regulations require more severe sea
conditions for the larger ships than the smaller ships.
4. MOTION RESPONSES OF A MEGA

CONTAINER SHIP IN SEAWAYS
moderate gale
[m ]

strong gale
[m ]

hurricane
From these numerical simulations, except in
hurricane condition, the numerical result of the 1/3
A ship will face various sea states during her voyage

in ocean route. Large motion acceleration of ship in
severe sea state may cause serious damages of cargoes
and lashing equipments especially for the huge
container ships. Since ship motion accelerations in sea
ways are functions of ship geometry, mass distribution,
ship speed and wave conditions, the assessment of
seaworthiness characteristics of a ship requires large
number of variations covering all possible wave
directions and ship speeds.
In the paper[5], by using the same computer code
based on 3D Green function method with forward
speed[8,9], some numerical simulations of
accelerations due to ship motions in waves are carried
out for a typical mega size container ship. The
numerical results of the maximum and the 1/3 highest
mean of transverse and vertical accelerations along
the ship length are determined from the time domain
analyses.
The principal particulars of the ship are
shown in
Table 7. The accelerations of ship
motion are calculated on the weather deck and the top
of containers 15m above the deck. The simulations are
performed in the hurricane with the significant wave
height of 10m and mean wave period of 15s. All
calculations are carried out for the main wave
direction of attack angle of 180 deg (head sea), 150
deg (bow sea) and 120 deg (oblique sea).
The numerical simulations of the ship have been also
performed in the significant wave height of 10m and
mean wave period of 20s. The wave condition of the
significant wave with mean wave period of 20s seems
to be too large swell in the ocean. DNV and NK may
consider the design wave height of 10.75m for the
wave period of 15s and 8.28m for 20s respectively.
Table 7 Principal particulars of a mega container ship

LBP
B
D
D
320 m
42.80 m
24.40 m
14.00 m
115,584 m3
24.5 knots
0.602
Cb
KG
L.C.G.
GMT
GML
K xx
17.40m
-3.86 m (aft)
2.60m
530 m
16.1 m
K yy
79.3 m
K zz
79.3 m
Figures 18~23 show the results the 1/3 highest mean

and the maximum of transverse and vertical
accelerations on deck height and the top of containers
along the ship length at side. In these figures design
criteria values by the guidelines for container lashing
forces are also shown to compare with the results of
numerical simulations in the strong gale. From these
numerical results, we can see the maximum values are
greater than 2 times of the 1/3 highest mean values at
both ends of the ship in the wave condition of strong
gale. This may be appeared from beating of coupling
motions of roll, sway and yaw for the transverse, and
heave and pitch for the vertical accelerations
respectively.
Figure 18 The 1/3 highest mean transverse

acceleration at 15m above the deck in the hurricane
Figure 19 The maximum transverse acceleration at

15m above the deck in the hurricane
Figure 20 The
1/3 highest mean transverse
acceleration at deck height in the hurricane

deck height in the hurricane
Figure 22 The 1/3 highest mean vertical acceleration

in the hurricane
consider yaw motion in their guidance. Classes may

consider the design transverse acceleration at the
maximum roll motion at the resonance wave period
and the design vertical acceleration at the maximum
pitch motion. LR[10] considers the force from wave
impact and shipping green seas where the form and
proportions of the ship are such that these may occur.
In general the design load for the securing
arrangements in the forward 0.25LBP are to be
suitable for forces increased by 20% except where it
can be shown that the containers are adequately
protected by break-water or similar structure[10].
Corresponding to this requirement the transverse
acceleration by LR is increased by 20% in the forward
0.25 LBP in Figures 18 to 21, 25, and 26.
In huge swell condition, the transverse accelerations
by guidance of classes show partially smaller than the
maximum values of the simulations as shown in
Figure 26, 27. In these huge swell conditions the
maximum roll angle reaches 35degrees[5]. We may
not need to consider such severe huge swell sea state
for the container securing systems.
The vertical accelerations of GL guidelines show
lower than the simulation results. GL guidelines show
the influence of the container fore to aft for the
transverse acceleration and show the same tendency
to the simulation results.
Figure 23 The maximum vertical acceleration in the

hurricane
Figure 25 The maximum longitudinal acceleration in

the hurricane
Figure 24 The 1/3 highest mean longitudinal
acceleration in the hurricane
We can recognize the transverse accelerations at the
bow are larger than those at stern as shown in Figures
18 to 21. The transverse acceleration due to yaw
motion is added to those of sway and roll as plus and
minus to fore and aft. So that at stern the transverse
acceleration on deck is larger than those at the top of
containers 15m above deck.
The transverse acceleration at forward location of
the ship at the mean wave period of 15s is higher than
those at the mean wave period of 20s which is more
close to the natural period of roll as shown in Figures
26, 27. This is caused by the effect of yaw motion to
roll and sway motions at higher frequency than the
natural frequency of roll. LR, ABS and DNV may not
Figure 26 The 1/3 highest mean transverse

acceleration at 15m above the deck in huge swell
REFERENCES

15m above the deck in huge swell
Though the numerical simulations have been
performed only for a typical huge container ship in
limited wave conditions, they may suggest that
necessity of considering the effect of yaw motion to
the transverse acceleration. So the following
simplified correction to the guide lines of classes for
the container securing systems may be suggested as
C =1+ (x 0.75LBP) / LBP
for
x 0 . 75 L BP .
Where, x is longitudinal distance from aft

perpendicular to the considered center of gravity of
the container.
Due to this correction the design load for the
securing arrangements in the forward are to be
lineally increased up to 25%. This correction has been
applied for NK new guideline[11].
5. CONCLUSION
Though the design criteria of equipments and fittings
of ships are not only based on the dynamic behaviors
of ships in the sea but also other environmental and
functional loads, and so many factors, the author
however reviews his papers on studies of dynamic
tension of towline and motion responses of ships in
short crested irregular waves to understand and
discuss these design criteria. The time domain
simulations of motion accelerations of a typical mega
size container ship in short crested irregular waves
have been carried out to compare guidelines of classes.
By the numerical simulation studies, we can see
some design criteria are reasonable and may be
improved in the consequences of dynamic behaviors
of ships in the sea.
[1] IMO home page: http://www.imo.org/index.htm

[2] International Association of Classification
Societies (IACS): Requirements concerning
STRENGTH OF SHIPS, 1988/Rev. 2004
[3] Class NK: Rules and Guidance 2008
[4] Y. Inoue : On the Design of Fittings and Dynamic
Behaviors of Ships in Waves, Class NK
Technical Bulletin, Vol. 26, 2008.12
[5] Y. Inoue & K. Ayada: A study on motion
responses of a mega container ship in seaways,
10th International Marine Design Conference
(IMDC), Trondheim, Norway, 2009.5
[6] Y. Inoue, T. Nakamura, H. Miyabe: On the
Dynamic Tension of Towline in Ocean Towing,
10th OMAE, pp.547 -553, Stavanger, Norway,
1991.6
[7] Y. Inoue, S. Surendran, T. Shimizu: Computer
Simulation on the Dynamic Tension of
Emergency Towing Lines of Tankers in the Seas,
The 5th STAB, Florida, USA,1994.5
[8] Y. Inoue & N.M.Golam Zakaria: A Design Aspect
of Relative Wave Heights of Ships in Sever Seas
by 3-D Source Technique with Forward Speed,
RINA International Conference Design &
Operation for Abnormal Conditions, London,
UK, 2005.1
[9] Y. Inoue, N.M. Golam Zakaria & R. Nakai: A
Parametric Study on Relative Wave Height of
Ships in Short crested Irregular Sea with
Forward Speed, 24th OMAE, OMAE2005-67191,
Halkidiki, Greece, 2005.6
[10] LR: Rules and Regulations for the Classification
of Ships, July 2007, Cargo Securing
Arrangements
[11] Class NK: Guide to container stowage, lashing
and securing arrangement, 2009

Martec2010 160329072110

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Martec2010 160329072110

Загружено:

Авторское право:

Доступные форматы

Paper List

IMO ballast water convention and enforceability of G2

http://www.buet.ac.bd/name/martec2010/2010/contents.htm[3/29/2016 1:19:10 PM]

Proceedings of MARTEC 2010

POTENTIALS, PRIORITIES AND PROSPECTS OF SHIPBUILDING

moving to countries with lower wages of required

Table-1: Comparison of labour costs

Proceedinggs of MARTEC 20010

A compparison of prodductivity (Table-2) shows thaat

1.3 Lower Effective Cost

Proceedings of MARTEC 2010

Weighted labour rate

Weighted average output cost

on a reasonably attained even keel of 4 m may be

1.5 Higher Level Skills

2.2 Meghna basin

2.3 Mongla basin

2.4 Karnofuli basin

Excluding sand bars and shoals here and there,

The Karnofuli River has been unfortunately

Proceedings of MARTEC 2010

2.5 Long coast line & river falls

3.5 Backward linkage industries

The long coast line and mouths of rivers falling to

Bangladesh is now exporting a few items meeting

2.6 Weather conditions

3.2 Training institutions

3.6 Access to the shipping world

3.3 Linkage industries

3.7 Courageous entrepreneurs

Forward and backward linkage industries both of

Courageous entrepreneurs have created a

3.4 Forward linkage industries

Proceedinggs of MARTEC 20010

3.8 Statuttory quality control

4.1 Worlld shipbuildiing capacity

Proceedinggs of MARTEC 20010

These vessels are of great importance

Fig..- 5 : Change inn economic baackground

Proceedings of MARTEC 2010

development. Bangladesh aims to attain middle

7.1 National shipping cost

7.2 Training centers

7.3 Other benefits

Shipbuilding may be defined to encompass the

Building ships for export is a priority to the

Proceedinggs of MARTEC 20010

8.3 Contrributions of lin

Fig.-6: Groowth pattern off shipyards in Bangladesh

8.4 Beneffits of developed national sh

Proceedings of MARTEC 2010

national shipping lines. Assuming an addition of 10

in 10 years and the number one in the field in 30

Proceedings of MARTEC 2010

VAN GOGH, GAUGUIN, FLUIDITY AND THE UNDERGRADUATE

engine, the camera, and the computer have brought

2. THE INTERNAL COMBUSTION

Proceedings of MARTEC 2010

farmer in terms of the skills and education he requires.

effective use of computers in education, of which the

To have a better understanding of the changes brought

Still cameras and movie cameras have changed the

We can try to determine what those additional