Complete Proceedings 18th Nov 2015

Proceedings of the
First International
Symposium on Automotive and
Manufacturing Engineering (SAME)
Editors
Dr Mushtaq Khan
Dr Syed Husain Imran Jaffery
November 26-27
2015
ISBN: 978-969-8535-35-3
TABLE OF CONTENTS
Paper ID
Paper Title
Page
No
SAME-01
Cylinder Liner Temperature Mapping of Heavy Duty Diesel Engine
SAME-02
Performance Improvement of Compressor to Recover Flare Gas by

Factors Analysis Method
10
SAME-03
Bulge hydroforming of tube by rubber mandrel without axial feed:

Experiment and Numerical Simulation
19
SAME-04
Perspectives in Automotive Embedded Systems from Manual to Fully

Autonomous Vehicles
27
SAME-05
Optimization of TIG Welding Parameters using FULL FACTORIAL

and ANOVA Design of Experiments Approach
36
SAME-06
Virtual Engineering: A key Enabler for Reconfigurable Manufacturing

Systems
43
SAME-07
Simulation of Turbulent Flow over a Single and Tandem NACA 0012

Airfoils
49
SAME-08
Passively flapping energy harvesting eel in a viscous flow
56
SAME-09
CZM based Strength Prediction for Scarf Adhesive Joints with Varying
Bondline Angle
63
SAME-10
Design and Fabrication of Three-Dimensional Printing Machine
68
SAME-11
Development of Hybrid Mold casting for wheel Hubs
75
SAME-12
Comparative Study of Wax and Nitrate Salts as Thermal Storage for

Cooling Purpose of Automobiles
81
SAME-13
Energy Consumption in Orthogonal Machining of Al 6061-T6 Alloys
89
Proceedings of the First International Symposium on Automotive and Manufacturing Engineering (SAME)
SAME-01
Cylinder Liner Temperature Mapping of Heavy Duty Diesel Engine

F. Badshah1, M. U. Abdullah1, R. A. Mufti1, M. U. Aslam1
1
School of Mechanical & Manufacturing Engineering, National University of Sciences and Technology (NUST),
Sector H-12,
Islamabad, 44000, Pakistan.
Abstract In an internal combustion engine,

further improvements in the design and development
phase is difficult therefore different companies intend
to bring changes in other aspects like lubricant,
lubricating system, cooling system, fuel and fuel
system. The temperature of engine is monitored
indirectly by temperature sensor, but this
temperature is of cooling water in engine. To
improve the performance of engine a new technique
for measuring the real time temperature profile of a
liner under fired engine condition is proposed. To
study
the
temperature
profile
eighteen
thermocouples are mounted on three radial and
transverse locations. Signal conditioning unit is used
to transfer data to the computer for monitoring
temperature profile along the length of liner. The
profile depicted the trend of temperature variation
within the cylinder. The temperature of cylinder liner
corresponds to the different speed and load of the
engine.
Index TermsTemperature mapping,
engine, temperature profiles, Cylinder liner.
Diesel
I. INTRODUCTION
The power required by the car for motion comes from
engine which converts different forms of energy into
kinetic energy and supply it to wheels. In the early stages
of the development of cars, fossil fuels such as diesel and
gasoline were used as a source of energy. With the
passage of time, electric power started to be used in
special category of vehicles known as electric vehicles.
After the commercialization and overwhelming
acceptance, the number of cars on road keeps on
multiplying. Due to which prices of fuels also started to
rise.
At the end of 20th century, developed countries of the

world realized that deposits of different fuels which run
their cars are not inexhaustible. In addition to that, wastes
coming out of the cars in the form of emissions are
polluting the environment to an alarming extend.
Since then, automotive industry is taking
revolutionary steps to increase the efficiency of internal
combustion engines. Researchers from different countries
have investigated that thermodynamic inefficiencies and
mechanical losses are main contributors towards the
energy losses. Comparatively, thermodynamic losses are
more significant than mechanical losses but still
tremendous amount of improvement in terms of
efficiencies can be achieved by focusing on the
mechanical losses in the different parts of the vehicle.
Mechanical losses are mainly due to the friction which
comes into play when two bodies interact with each
other. High values of friction can result in decreased
efficiencies and wearing out of engine components.
As the current trend in car engine development is to
create smaller engines with higher specific power outputs
to meet the demands of both lower emissions and
customer satisfaction, the thermal loads in engines are
increasing [1].
Now a days the current trend in the advancement and
development in Automotive is to build a smaller and
lighter engine with extensive output in order to meet the
customer demands and to have a less emissions in the
engine. But due to this reason, the thermal loads &
thermal losses in an IC engines are increasing gradually.
To enhance the overall durability & the performance
of a diesel engine, Temperature of a cylinder liner is an
important factor which needs to be optimized [2]. In this
research, an attempt is made to measure and investigate
the temperature profile at the cylinder liner of a water
cool diesel engine. This project is aimed to study the
trend of temperature profiles on cylinder liner surface of
a heavy duty diesel engine. For this purpose eighteen

thermocouples were mounted on three circumferential
locations and six transverse locations. Signal
conditioning was transferred to the computer to log the
data regarding temperature profiles along the length of
liner. This temperature profile reflects the temperature
variation within the cylinder in real time. This
temperature of cylinder liner corresponds to temperature
of engine oil used. Oils of different viscosity could be
tested and are expected to behave differently at engine
operating temperature and thus will provide different
temperature profiles along the liner length. It is expected
that temperature will vary specifically along the liner
length & circumference and for the first time such data
will be available for relating the liner temperature with
engine conditioning.
The overall objective of this research work is to
provide a deep understanding on matter of temperature
mapping across the cylinder liner (or heat transfer) in a
heavy duty diesel engine by experimentally evaluating or
simulating the setup to gather the real time data which
could be used in the future not only to model Finite
element model of an IC engine, CFD (computational
fluid dynamics) of the said engine model but also to
compare the experimental results with the theoretical heat
transfer models (Not covered in this paper). The same
results could be verified by applicable empirical,
numerical & analytical approaches which will be carried
out in our future research work to apprehend the physics
& solution methodology of the liner thermal analysis.
The whole setup heavy duty diesel engine along with
the instrumented liner was setup on an engine
dynamometer DX-33 Taylor. The use of a dynamometer
(dyno) provides a means for testing an electric motor, or
engine, in various desired and controlled conditions.
Dynos produce measurements of a motors mechanical
force, speed, or power [3]. It is important to have a
method to accurately measure the performance of the
motors and ensure they operate as desired.
All dynos work to mimic load conditions by exerting
brake forces on a motor [4]. They can be AC or DC
motors. The type of dyno used will therefore depend on
the type of motor being tested and its intended purpose.
For testing car, engine dyno are used where the
dynamometer is mounted directly to the shaft of the
power source using a torque transducer, or flange [5]. In
our case, engine dyno was of water brake type.
Figure 1: Engine Dynamometer Water Brake Type

II. METHODOLOGY
A four stroke diesel engine wet Liner was taken out of
the engine. Machining of the liner was done at CNC
machine where 18 holes were drilled at specified position
(along the height and circumference) of the liner in order
to place & install the thermocouples on the said work
piece. These thermocouples were glued at the particular
position for the temperature mapping of the cylinder liner
of heavy duty diesel engine. The instrumented Liner was
fitted in the engine again which was coupled with the
help of coupling to a dynamometer in order to run and
simulate the road loading of either the engine so that the
activities of engine development i.e. calibration of
management controllers used in engine, investigations of
the combustion & tribology study could be carried out.
For the sake of development of data acquisition system,
hardware used for temperature measurements include
PCI 6221 Multifunction data acquisition device
connected to SCXI 1000 Chassis containing SCXI 1102
module. The thermocouples were attached with these
LabView data acquisition modules and the data was
plotted on the charts and logged in LVM file format. Test
were carried out at specific engine speed with some
engine load and data was saved for the 18 thermocouples
mounted across the length and circumference of the
Liner.
III. EXPERIMENTATION
Model EC100 engine is four stroke, direct fuel
injection system and water cooled .Valves are in cylinder
head and engine has two valves per each cylinder. Glow
plugs are also fitted for easy cold starting. The engine is
used in heavy duty trucks. Diesel is used as fuel in
engine. The wet sleeves/liners are used in engine. The
camshaft is fitted in cylinder block so it has gear to gear
timing. Some specifications of the engine are shown in

Table 1.
Table 1: Engine Specifications
Engine Model
HINO EC100
Type
Diesel, four stroke, Vertical, Six

Cylinder, Inline, Valve in head,
Water Cooled
coolant manifolds from head as per manual instructions.

Next task was to remove the cylinder liner from the
engine block. For this, piston along with the connecting
rod at position 6 was removed then cylinder liner was
pulled out of the bore with the help of SST liner puller.
To mount the thermocouples on the surface of
cylinder iner, marking was carried out first. The aim was
to cover the length of the liner as well as the sides. So the
cylinder liner sides were categorized for the sake of ease
as per the orientation of the engine ie thrust side, back
side, pump side and front side. Each side except the front
one was marked and drilled so that the thermocouples
could be inserted in. The detailed figures are shown in
Table II.
Bore & Stroke
97 mm x 113mm
Piston displacement
5.01 Liter
Compression ratio
20.3:1
Outer Diameter of Sleeve
110mm
Firing order
1-4-2-6-3-5
Inner Diameter of Sleeve
96.92mm
Injection timing
18 deg B.T.D.C
Thickness of Wall of Sleeve
6.54mm
Depth of Hole
5.84mm
Valve clearance
Intake 0.25mm
Distance from each T C to Sleeve inner Side
0.7mm
Each Hole Diameter
2.5mm
(Cold)
Injection Nozzle
Exhaust 0.25mm
120kg/cm2
Pressure
Engine oil Capacity
12.5 Liters
Cooling Capacity
27 Liters
(Engine &Radiator)
Electrical Voltage
24 Volt
Standard engine oil
1-5kg/cm2
pressure
Table 2: Machining Specifications
A. Arrangement of Thermocouples (TC) on Selected

Sleeve
Marking was carried out on the basis of the overall
engine orientation as discussed earlier with the help of
height gauge on the three selected sides with scriber on
the selected Liner. The three sides were from cooling fan
side, Fuel injection Pump Side which is denoted by (P),
side opposite of fuel injection Pump Side which is
denoted by (T) the third one is Flywheel Side which is
opposite to the front side which is denoted by (B) as
shown in the Fig.
P TC: Thermocouple on Fuel Injection Pump Side
B TC: Thermocouple on Fuel Flywheel Side
Engine was dissembled in order to have the cylinder

Liner separated. All the necessary wires were
disconnected ie ground wire, starter motor wire, pre
combustion chamber heater wire, and oil pressure switch
etc. and all the hose and pipes were removed in order to
lift up the engine away from the chassis. After that
cylinder head was dissembled via removing the tappet
cover, high pressure injectors pipes and intake, exhaust,
T TC: Thrust Side Thermocouple are on opposite of Fuel

Injection Pump Side
Marking was done at six different height levels as
shown in fig. similarly same marking was carried out at
two other sides as discussed earlier shown in fig. As 18
thermocouples were required to be mount on the surface
on the liner from top dead center TDC to bottom dead
center BDC so all points were marked and specified

accordingly. After marking of the cylinder liner at the
specified location where thermocouples need to be
installed, liner was drilled on CNC machine with the
nominal diameter dimension as specified in Table. The
thermocouples are allocated and categorized as per their
orientation and placement. For example a thermocouple
which is placed on number 5 level at flywheel side will
be designated as B.
P6: 6th TC from Top of Sleeve on Pump Side
Figure 2: Layout for Thermocouples Mounting
P3: 3rd TC from Top of Sleeve on Pump Side
Drilling process was done with CNC machine; 2.5mm

holes were required for each thermocouple. The hole
were drilled with a high speed dill bit at a depth of
5.84mm which is 0.7mm less to the thickness of Liner so
thermocouple were placed 0.5mm to 0.7mm away from
the inner side of the liner.
P2: 2nd TC from Top of Sleeve on Pump Side

P1: 1ST TC from Top of Sleeve on Pump Side
B6: 6th TC from Top of Sleeve on Flywheel Side
B3: 3rd TC from Top of Sleeve on Flywheel Side
Table 3: Thermocouple Layout
B2: 2nd TC from Top of Sleeve on Flywheel Side

B1: 1st TC from Top of Sleeve on Flywheel Side
T6: 6th TC from Top of Sleeve on Thrust Side
T3: 3rd TC from Top of Sleeve on Thrust Side
T2: 2nd TC from Top of Sleeve on Thrust Side
T1: 1st TC from Top of Sleeve on Thrust Side
Figure 4: Machining of Cylinder Liner
Figure 3: Layout block diagram
B. Instrumenting the Cylinder Liner with Thermocouples

Cylinder liner was instrumented with thermocouples
that produce mili-volts at the point of junction where the
two metals meet with the application of heat. This
voltage produced via thermoelectric seebeck effect are
somehow proportional, not always, for a certain range of
temperature. As thermocouples cover a high temperature
range with cyclic change or variation, it can be used in
IC engine to monitor the temperature variation under
fired conditions. Thermocouples that are used
commercially are cheap and can detect a wide range as
well as smallest change i.e. 0.1 degree Celsius. As
compared with other temperature measuring devices,
thermocouples are self-powered & no other excitation
voltages are required for their working, yet it has the
main limitation of accuracy as compared to RTDs. Any
two dissimilar metals could be used to produce voltages
after heating their junction but thermocouples are made
from typical alloys which repeatability and
reproducibility. They are usually calibrated &
standardized with a reference temperature of 0 degrees
Celsius so the practical instruments uses electronic CJC
methods (cold junction compensation) to calibrate the
instrument as used in our case. The Electronic
hardwares i.e. DAQ in our case, compensate the
variation in the readings, & hence improve the accuracy
of the instrument.
Figure 5: Cold junction compensation

K Type thermocouple (chromel with a 90 percentage of
Ni, nickel and 10 percentage of Cr, chromium} with the
combination of alumel (95% Ni, 2% Mn, 2% Al, 1% Si)
was used for the temperature mapping of the cylinder
liner with a resolution of 41 V/C approximately.
Thermocouples were inserted in the specified holes
which were drilled accordingly. This K type
thermocouples were inserted in such a way to expose to
the liner inner wall as much as possible in order to carry
the correct amount of heat. Wires from the
thermocouples were drawn out of the liner and directly
attached to the LabView data acquisition hardware
(DAQ).
C. Data Acquisition System

The Data Acquisition System (DAQ) used for the project
is based on National Instruments software and hardware.
The software used for the programming is LabView
(Laboratory Virtual Instrumentation Engineering
Workbench). DAQ hardware acts as a source of
interfacing the instruments with a computer. It has
analog to digital convertors that converts the analog
signal coming from the sensors to the digitized form so
that it can be read by computer. Many data acquisition
devices also have other functions ie digital to analog
conversions (DACs), digital input and output signals,
counter, timers etc.
Figure 6: Instrumenting the Cylinder Liner
displays a graph showing the real time data being

acquired on the chart.
The DAQmx physical channel constant lists all
physical Channels on devices and modules installed in
the system. For this program the physical channels to be
selected were the channels of SCXI 1303 which were
connected to the 18 Thermocouples. Order of selecting
solely depends upon the orientation of thermocouples.
These 18 numeric indicators display the current
temperature reading of the respective
thermocouple.
Indicators are placed as representation of actual
thermocouples position with respect to Top Dead Center
(TDC) and Bottom Dead Center (BDC).
E. Fitting of Engine on Engine Dynamometer
After instrumentation of the cylinder liner, cylinder
liner was fitted back to the engine block and it was
insured that the liner O-ring fits exactly in the bore seat
area. The challenge was to insert the liner along with
thermocouple wires due to less clearance between the
liner surface and the cylinder block bore diameters. After
fitting of liner, engine was assembled and mounted on
the engine dynamometer in order to test the engine. This
engine dynamometer DX-33 uses water brake method to
apply the load on the engine. The use of a dynamometer
(dyno) provides a means for testing an electric motor, or
engine, in various desired and controlled conditions.
Dynamometer produces measurements of a motors
mechanical force, speed, or power [3]. The engine crank
shaft was coupled to the dyno DX-33 drive shaft with the
help of coupling and universal joints that came with the
dyno. The whole setup is shown in the figure.
Engine was run at 1000rpm with 25% of Load
applied by the water brake system, and data was logged
for the 18 thermocouples mounted across the length and
circumference of the cylinder liner. Test was run to log
the data of 5min only after the initial startup of the engine
and applied load so that the temperature may not be
fluctuate and deviate due to variation in speed and load.
Figure 7: Software Development

D. Software Development for Data Acquisition System
The LabView software interface consists of front
panel and block diagram interface. The front panel
contains the output and input terminals while the block
diagram contains the code and algorithm of the program.
VI contains two tabs, the first one displays the Numeric
indicators, Controls, Error handler and instructions as
shown in the above photograph, while the second one
B. Temperature from Flywheel Side
a) Temperature at B6 and B5
We know that B6 and B5 are near to the top position the
sleeve/liner so the temperature is high at these positions
due to nearest position to the combustion chamber and
also Number sixth cylinder is at the end so here is small
path for cooling water so the temperature is high as
compare to the Pump and Thrust side of liner and
gradually temperature decreases to the down position
b) Temperature at B4 and B3
Thermocouple B4 and B3 are at center position of liner

and closed to water jackets so the temperature is low as
compare to TDC and BDC positions thermocouples but
more than the Pump Side TC P4 and P3 because cooling
water moves from Pump Side to back and then thrust
side.
c) Temperature at B2 and B1
Figure 8: Fitting of engine on Dyno DX-33
IV. RESULTS
Temperature data from instrumented liner was plotted
in Microsoft Excel when engine was rotated at 1000
RPM with 25% load. Three charts were generated against
three different sides of the liner i.e. Pump Side (P),
Flywheel Side (B) and Thrust side (T).
A. Temperature from Pump Side
a) Temperature at P6 and P5
We know that P6 and P5 are near to the top position the
b) Temperature at P4 and P3
Thermocouple P4 and P3 are at center position of liner

compare to TDC and BDC positions thermocouples.
c) Temperature at P2 and P1
Thermocouple P2 and P1 are at BDC position and are
away from cooling water so the temperature is more than
center position of the liner
Thermocouple B2 and B1 are at BDC position and are

center position of the liner but temperature is more as
compare to B4 and B3.
C. Temperature from Thrust side

a) Temperature at T6 and T5
We know that T6 and T5 are near to the top position the
temperature of liner is low because here is big cavity for
cooling water as compare to Flywheel Side of liner and
b) Temperature at T4 and T3
Thermocouple T4 and T3 are at center position of liner

compare to TDC position thermocouples but more than
the Pump Side TC P4 and P3 and also Flywheel Side
TC B4 and B3 because all sides hot water comes here
due to outlet pipes are connected here.
c) Temperature at T2 and T1
Thermocouple T2 and T1 are at BDC position and are

the Pump Side TC P2 , P1 and also from Flywheel Side
TC B2 ,B1 due to hot water passage.
Figure 9: Temperature profiles for 1000 engine speed at 25% load at Pump side
Figure 10: Temperature profiles for 1000 engine speed at 25% load at Thrust side
Figure 11: Temperature profiles for 1000 engine speed at 25% load at flywheel side
V. CONCLUSIONS
The liner temperature profile was measured by
mounting K-type Thermocouple temperature sensors
around the cylinder liner. In total eighteen
thermocouples were mounted, three in radial directions

and six in axial direction. Engine tests were carried out
under fired conditions at two different engine speeds
1000rpm and 1700 rpm and two load conditions
quarter and half load.
8
Advanced data acquisition system was configured

and programmed to acquire, analyses and log data
from different sensors. This temperature profile
reflects the temperature variation within the cylinder in
fired condition. The major conclusions are:
Highest Temperature: Temperature is higher at

TDC position due to near the combustion
chamber as compare to mid of liner and BDC
position.
Lowest Temperature: At mid position of liner
the temperature is lowest as compare to TDC
and BDC position due to nearest position to
water passages.
Mid Temperature: At BDC position the
temperature of the liner is between the Highest
and Lowest Temperature because of metal to
metal contact with the engine block and there is
no cooling water in this region.
The same results could be verified by applicable

empirical, numerical & analytical approaches
which will be carried out in our future research
work to apprehend the physics & solution
methodology of the liner thermal analysis
VI. REFERENCES
[1] GONERA, MARTIN, & O. SANDIN, "Thermal
Analysis of a Diesel Piston and Cylinder Liner
using the Inverse Heat Conduction Method.
[2] Sukumaran, Vipin, et al. Cylinder Block
Temperature Mapping and Development of
Cooling Cowl for Reducing the Maximum Liner
Temperature. No. 2012-01-0405. SAE Technical
Paper, 2012.
[3] Winther, J. B. Dynamometer Handbook of Basic
Theory Applications. Eaton Corporation, 1975.
[4] Miranda, Gabriel. "ECE 4007 L01/Dr. Smith Solar
JacketsMotor Control Dynamometer for Testing
Electric Motor."
[5] A.J. Martyr and M.A. Plint, Dynamometers and
the measurement of torque, in Engine Testing:
Theory and Practice, 3rd ed. Oxford: ButterworthHeinemann, 2007, pp.153-15
SAME-02
Performance Improvement of Compressor to Recover Flare Gas by Factors

Analysis Method
Naseem Ahmad1, Sahar Noor1, Rehman Akhtar1, Khizar Azam1
1
Industrial Engineering Department, University of Engineering and Technology

Peshawar, Pakistan
naseem_saddiqui@yahoo.com
Abstract- A large quantity of natural gas is flared in

combination with the exploration of crude oil at
Nashpa Oil and Gas Field, Pakistan. This flaring gas
expends huge amount of energy and causes ecological
deprivation and illness. The main centre for industrial
development and economics activities is oil and gas
sector.
The main purpose of this study is to develop the design
of experiment techniques, one factor at a time. two
factors at a time, three factors at a time etc to recover
large quantity of flared gas within the compressor
existing system. For analysis the data of gas and
compressor are collected from a number of sources
and meaningful information extraction. In addition, to
recover low pressure gas five factors are selected for
the compressor performance.
In this research, a new approach is applied on a gas
flaring to find the outlay effect method that carefully
reduce the negative environmental effects, modify the
existing system and to increase the economic benefits.
Key words- Oil and Gas Company, Natural Gas
Flaring, Factors Analysis, Environmental,
Economics
I.
INTRODUCTION
The oil and gas sector of Pakistan is a key sector in

nations economy. The two key components of energy
is oil and gas and both contributing around 80% to 83
% share to 70 million TOE (ton of oil equivalent) of
energy requirement in the country. The oil and gas
industry is growing quickly, and the oil and gas well
drilling activities in the country are going beyond the
average. The mixture of flow of crude oil, gas and
water is produced by oil wells. This mixture flow is
transfer to the plant area by its natural pressure for

separation to make the mixture useful. Firstly, the flow
is injected to three phase high pressure separator,
which separates the mixture of the three phase
separator. The gas pressure is 1100 PSI and this gas is
injected to dehydration plant to make gas for useful
purpose. The gas is shifted to SNGPL (Sui Northern
Gas Pipeline Limited) piping network after
dehydration plant. Water is sent to pit after separation
of crude oil mixture. The crude oil of high pressure
separator ,which has pressure of about 300-450 PSI is
injected in three phase medium pressure separator. The
crude of medium pressure separator is shifted to low
pressure separator and finally the crude is sent to crude
oil storage tanks. Gas of medium pressure separator
with the pressure of 300-450 PSI is injected in medium
pressure compressor. To increase the pressure of
medium pressure to 1100 PSI the gas is passed through
compressor. After compressor the gas is passed
through dehydration plant for useful purpose. The gas
which is transferred to low pressure separator from
medium pressure separator has low pressure (125-200
PSI) is flared due to its low pressure. Within the
existing system, it wants to recover the low pressure
gas that is being flared. The minimization of flaring
gas or waste gas is really environmental challenge by
which oil and gas industry is facing. Hence, this
process reduces negative ecological impacts, conserve
energy and natural assets, which are safe and sound for
employees, consumer and communities. Flares are
indispensible safety tools for the clean and cost
effective removal of combustible gases. Flares have
been operating in chemical plants, refineries, steel
mills and other industries for over sixty years. Purpose
of flare flow activity is crucial in establishing the
recovery system design capability. For projects
10
economics evaluating the average flare gas is

significant step.The objective of this research is to
develop a methodology to eliminate carbon emission
and recover flared gas to useable gas by analyzing
some factors of interest such as suction pressure,
temperature, specific gravity, RPM, and compression
ratio.
II. LITERATURE REVIEW
In oil and gas industries, the surplus natural gas is
burnt in open air. However, this leads to emission of
hazardous gases. The flaring gas creates hazards for
worker and nearby residents. Hence, it is important
measure in natural gas facilities and at drilling
operations. in petroleum industries the flaring systems
are deployed. Flaring of pure methane (CH4) produces
carbon monoxide (CO). In addition, Nitrogen present
in the air is oxidised during the combustion process
and form Nitrogen oxides, which is called NOx.
Emission of flare gas adversely affect the environment,
human health, lives stock plants and animals.
Consequences to that region depends upon on the
duration, magnitude and frequency. These gases
contribute to global warming. If the natural gas consist
of hydrogen sulphide (H2S) than emission includes
sulphur dioxide, carbonyl sulphide (COS) and carbon
disulphide (CS2), among others
oxidised during the combustion process and form
Nitrogen oxides, which is called NOx. Emission of
flare gas adversely affect the environment, human
health, lives stock plants and animals. Consequences to
that region depends upon on the duration, magnitude
and frequency. These gases contribute to global
warming. If the natural gas consist of hydrogen
sulphide (H2S) than emission includes sulphur dioxide,
carbonyl sulphide (COS) and carbon disulphide (CS2),
among others [1].Flaring of gas has been identifying as
an environmental hazard and economic waste.
Perinenet literature sources are available in the oil and
gas domain highlighting various issues. A case in
point, Erwin, Zhizhin, Pack, Bough,Tuttle, Elvidge,
Ziskin Ghosh (2009)
Conducted a fifteen years
record of global natural flaring of gas derived from
satellite [2]. They concluded that the flaring of gas
efficiency as the volume of flare gas per barrel of
crude oil produced and it is stable from last fifteen
years and its ranges from 140 to 170 billion cubic
meters (BCM). In addition, they found that the global
flaring of gas is declined by 19% since 2005, which is

led by gas flaring reduction in Russia and Nigeria (the
only two countries with the highest level of gas
flaring). A combination of these work in literature
shows obvious absence of work, which has been aim
to assess the inequality between the flaring activities of
major oil and gas companies in Nigeria and the
variation between gas flared and produced. Odigure
and Abdul Kareem (2010) worked on the economic
benefit of natural gas consumption in Nigeria and this
case of study is based on processing of food industry
[3]. Their work concentrate on the measurement of
heat radiation from flaring of gas and the benefits of
replacement usual fuel and energy types, e.g.,
automotive gas oil (low pour fuel oil, automotive gas
oil diesel and electricity) with natural in economic
term by using mathematical principles in calculation
with the food processing industry case study in
Nigeria. Results revealed that about 30% on low pour
fuel oil, 69% on diesel and 69% on electricity saved by
company and if conventional energy and fuel is
substituted with natural gas it go round million of
dollars in five years. Onwukwe (2009) worked on gas
liquid technology in Nigeria [4] and explained the
view of gas to liquid (GTL) conversion technology is
the choice for sustainable natural gas utilization. In
their research, they found that this technology will
make possibility of the chemical conversion of natural
gas into (clean diesel, light oils, Naphtha, Kerosene) as
a marketable liquid product. Results revealed that
conversion will contribute to the exclusion of flared
gas and minimize the country overdependence
imported refined petroleum products. Hang, Blasing,
and Kimberly (2007) worked on monthly carbon
emission from natural flaring gas and cement
manufacturing in the United States [5]. They
established that the carbon emission and all fossil fuel
emission amounted to 10% it has no persistent and
clear annual model. Akpan (2009) performed a study
on flared gas to show that large quantities of gases are
used in chemical industries [6]. The method to control
the natural gas of the nation is to encourage the
establishment of gas based petrochemical complexes
that can consume huge quantity of volume needed to
eliminate gas flaring. It also suggested that Nigeria can
make sound foreign exchanges earnings on long term
by properly control the natural gas by establishing
petrochemical gas industries.
Nigeria has made
11
foreign exchange earnings by the LNG (liquified

natural gas ) technologies which involve liquefication ,
regasification, shipping and delieved into the pipeline
grid. Natural gas that includes methane is cooled and
by cryogenic process it is liquefied at a temperature of
-2600F and liquefied natural gas is formed. The
volume of natural gas is reduced to one six hundredth
and allow the transportation through a special liquefied
natural gas tanker ship over a long distances.Moore,
Boone, Hill and Randall (1985) developed a
methodology for gas flaring . Their research is based
on more than two years of engineering development on
several associated gas stream in large amount gas
liquids recovery, gas treatment, power generation and
residue gas compression are utilized effectively and
economically [7].
The total gas production of Pakistan is approximately
6.8 billion cubic feet (BCF) per day. Exploration and
production (E & P) companies in Pakistan support the
Petroleum Institute of Pakistan by taking initiatives in
the oil and gas sector. From rough estimate slow BTU
(British thermal unit) gas and gas reservoirs stand at 60
billion cubic feet per day. About 150-200 mmcfd
(million cubic feet gas per day) is flared in Pakistan
every day due to low pressure. As this flaring gas has
low BTU and the Oil and Gas Management, it is
essential to purify the gas. Hence, the management
installed mobile power houses to recover gas [8].
TABLE I: Selected Factors
For low pressure recovery of gas first one-factor-at-atime (OFAT) approach is deployed (i.e., suction
pressure was analyzed by varying and other factors
were held constant). Furthermore, factorial experiment
was conducted by varying various factors together.
One factor at a time: In this scenario, the suction
pressure is first analyzed and all the other factors are
unchanged. The suction pressure is changed from 125150PSIG while each factor is changed with respect to
its ranges keeping other factors constant. The range of
temperature is varied from 130-1400F. Range of
specific gravity is changed from 0.66 to 0.70 while the
range of compression ratio and RPM is changed from
2.70 to 2.90 and 950-1120 RPM, respectively.
TABLE II: One Factors is Changed
The objective of this research is to develop a

methodology to eliminate carbon emission and recover
flared gas to useable gas by analyzing some factors of
interest. These factors are suction pressure,
temperature, specific gravity, compression ratio and
RPM.
III. METHODOLOGY
In this research, five main factors were identified for
the performance of compressor improvement. Table 1
depicts the selected factors.
Table II depicts that only one factor (suction pressure)

is changed and the other factors are held constant..
Two factors at a time: In this section two factors were
varied witin the range while all the other factors kept
constant
12
TABLE III: Two Factor are Changed
Four factors at a time: In this secnario, four factors are

changed at a time while holding other factors constant.
All factors are changed within its range.
TABLE V: Four Factors are Changed
In Table III, two factors factors (suction pressure and

temperature) are changed and the other factors are kept
constant.
Three factors at a time: Three factors are changed at a
time with respect to their ranges and the other factors
held constant. A set of three factors is taken at a time
in this method
TABLE IV: Three Factors are Changed
Four factors suction pressure, temperature, specific

gravity and RPM are changed in Table V by four
factors at a time method and compression ratio is
unchanged in Table V. Each factor value is changed in
each case in the table V.
All factors at a time: All factors are changed at a time
with respect to their ranges in this section. The flow of
gas for each experiment is also given table VI..
TABLE VI: All Factors are Changed
In Table IV, three factors at a time method is given.

Suction pressure, temperature and specific gravity are
changed and compression ratio and RPM is
unchanged.The value of all the three factors are varied
in each case and the flow is also changed in each case
and is given in table IV.
13
In Table VI factors suction pressure, temperature,

specific gravity, RPM and compression ratio are
changed at all factor at a time method. In this method,
the value of all factors are changed in each experiment.
For each experiment the flow of gas calculation is
shown in table VI.
IV. DISCUSSION
Figure 1 depicts relation between suction pressure and
gas flow. The graph shows that suction pressure is
changed against gas flow while keeping other factors
constant. The gas flow in this graph starts from 2.6
mmcfd (million cubic feet per day) and is reached to
2.85 mmcfd. Initially, suction pressure is set on 125
PSI, temperature 130 0F, specific gravity 0.65, RPM
950 and compression ratio 2.90 for first stage. For
second stage, 2.85 compression ratio and 2.60 mmcfd
(million cubic feet per day) gas flow is achieved. In
next step, the suction pressure is set on 130 PSI and all
other factors are held constant, the gas flow achieved
through this set of experiment is 2.65mmcf. Further,
the value of pressure is changed to 135PSI 2.70
million gas flow is achived. Finally, the suction
pressure is set on 150 PSI the gas flow achived is 2.85
million. This figure 1 shows a straight graph which
reveals that the flow of gas continuously increasing.
when suction pressure and temperature is set on

135PSI and 1320 F, respectively while gas flow
decreased to 2.82 million. Finally, when suction
pressure and temperature is reached to 150 PSI and
1400 F than the flow of gas decreased than the gas
flow reached to 2.34 million.the flow of gas is
constantly decreasing when both factors are changed to
their respective ranges. In Table III the value of each
factor is given which shows that suction pressure and
temperature is changed and other factors are held
constant and figure 2 depicts relation between suction
pressure, temperature and gas flow. In figure 2 the
flow of gas is continuously decreasing and a straight
graph is shown from higher gas flow to lower gas
flow.
Fig 3 depicts relation between suction pressure,
temperature and specific gravity vs gas flow. In this
figure, three factors are changed at a time while
holding other factors constant. Initially, when the three
factors suction pressure, temperature and temperature
are set at 130 PSI, 132 F and 0.66 than 2.25 mmcfd of
gas flow is achived. In next step, the value of suction
pressure, temperature, and specific gravity
are
changed to 135 PSI, 1340 F and 0.67 respectively.
Gas Flow against Suction Pressure
Gas Flow against Suction Pressure and Temperature
2.9
3.5
2.85
3
2.5
2.7
3.52
Gas Flow
2.75
2.65
2.6
2.55
120
125
130
135
140
145
150
155
Suction Pressure
Fig. 1: Graph between suction pressure and gas flow

in one factor at a time method
Fig. 2 depicts a scenario in which suction pressure and
temperature are varied and the other factors are
constant. Initially, the gas flow reached to maximum
after that the gas flow goes on decreasing. When
suction pressure and temperature are changed initially
and set the value of suction pressure 130 PSI and
temperature 1320 F and other factors are held constant
than 2.98 million gas flow is achieved. In next step,
Gas Flow
Gas Flow
2.8
3
1.5
2.5
1
2
0.5
1.5
0
1 125
Gas flow
130
135
0.5
140
145
150
155
138
140
142
Suction Pressure
0
130
132
134
136
Temperature
Fig. 2: Suction pressure and temperature are changed

and gas flow is given. (two factors)
14
The gas flow reached to 2.34 million in this

experiment. Five experiments are performed in the
three factors at a time method. In the final step, the
value of suction preesure changed to 150 PSI,
temperature 1400 F and specific gravity to 0.70, the
flow of gas is reached to 2.61 million. Hence, three
factors are changed at time the flow of gas is start from
2.34 million and reached to 2.61million. The value of
each factor is shown in table IV and the gas flow
achived through each set of experiment is also shown .
In each case the gas flow is continuously increasing.
Hence, it concluded that by changing the factors to its
respective ranges and increasing the number of factors
will result in an increased gas flow. Figure 3 dipicts
that the gas flow is contionously increasing (i.e.,
revealing that when the three factors are changed
within the range the gas flow is increasing).
Gas Flow against Suction Pressure, Temperature and Specific gravity

2.65
2.6
Gas Flow
Gas Flow
Gas Flow
Gas Flow
2.55
Gas Flow
2.5
2.65
2.45
2.6
2.65
2.4
2.55
2.35
2.6
2.5
2.3
2.55
2.45
2.25
2.5
2.4
2.2
2.35
2.45
2.3 125
2.4
2.25
2.35
2.2
2.3 130
2.25
2.2
0.65
130
135
140
145
150
155
138
140
142
0.69
0.7
0.71
Suction Pressure
132
134
136
is increasing. Initially, when the suction pressure,

temperature, specific gravity and RPM are set at 130
PSI,1320F, 0.66 and 1000 than 2.4 million gas flow is
achived. Next, four factors changed again and and kept
the value of suction pressure 135 PSI, temperature
1340 F, specific gravity 0.67 and RPM 1030. In this
case, the flow of gas reached to 2.54 million. Further
change in factors is increasing the gas flow and finally
set the suction pressure, temperature, specific gravity,
and RPM at 150 PSI, 1400 F, 0.70 and 1120 so in this
scenario the flow of gas is reached to three million.
The flow is smoothly increasing in this set of
experiments. Hence, it is concluded that increasing the
number of factors is increasing the gas flow and
operation is running smoothly. Table V summarizies
results for this scenario.
Fig. 5 represents relation between all factors at a time
(suction pressure, temperature, specific gravity, RPM
and compression ratio) against gas flow. Flow of gas is
calculated by changing all the factor according to their
respective ranges. When suction pressure, temperature,
specific gravity, RPM and compression ratio set
initially at 130 PSI, 1320 F, 0.66, 1030 and 2.85/2.80,
than the gas flow reached to 2.5million. Further
increasing all factors values is increasing the gas flow.
Finally, when all factors value changed to 150 PSI,
1400 F, 0.70, 1120 and 2.70/2.75 , than three million
of gas flow is achived. The value of all factors are
shown in table VI and the gas flow in each case is
shown in table 5. From the graph (see Fig. 5) it is clear
that the gas flow is increasing smoothly. In Fig 5 the
gas flow in each case is increasing, a straight line
graph is made between all factors and gas flow. Hence,
this is the best case to recover huge amount of gas by
changing all factors at a time.
Temperature
0.66
0.67
0.68
Specific Gravity
Fig. 3: Suction pressure, temperature and specific

gravity are changed and gas flow is shown. (Three
Factors)
Fig. 4 illustrates relation between suction pressure,
temperature, specific gravity and RPM against gas
flow. In this scenario, four factors are changed within
the limit of range the gas flow is calculated. Fig. 4
reveals that when the factors are changed the gas flow
15
Gas Flow against Suction Pressure, Temperature, Specific gravity,
RPM and Compression ratio
Gas Flow against Suction Pressure, Temperature, Specific gravity and

RPM
3.5
3.1
3.5
3
0
2
1
130
135
140
145
150
155
138
140
142
Suction Pressure
132
134
136
1.5
0.5
Temperature
1
0
0.5
0.65
0.66
0.67
0.68
0.69
0.7
0.71
Specific Gravity
1000
1020
1040
1060
1080
3
3.1
2.9
3
2.8
2.9
3.1
2.7
2.6
2.83
3.1
2.5
2.7
2.9
2.43
2.6
2.8 125
130
135
2.9
1100
1120
1140
RPM
140
145
150
155
138
140
142
0.69
0.7
0.71
Suction Pressure
2.5
2.7
3.1
2.8
2.4
2.6
3 130
2.7
2.5
132
134
136
Temperature
2.9
0
980
Gas Flow Gas flow
130
Gas Flow Gas Flow
2
2.5
3.5
1.5
2
1
3
1.5
0.5
3.5
2.5
0
1
3
2 125
0.5
2.5
1.5
Gas Flow
Gas Flow Gas Flow
Gas FlowGas Flow
2.5
3
2.4
2.6
2.8 0.65
0.66
0.67
2.5
0.68
Specific Gravity
2.7
2.4
2.6
Fig. 4: Suction pressure, temperature, specific gravity

and RPM are changed and gas flow is shown. (Four
Factors)
ECONOMIC BENEFITS
Investment in oil and gas exploration and production
produce significant economic gain and other benefits
such as energy increased. The main factors which
determine future oil and gas production are prices,
geology and the available technology for exploration
and recovery. The total economic benefit of recovery
of gas in Nashpa Oil and Gas Field is $ 2.16 million
per year. This figure represents huge economy of a
single oil and gas field. The calculation is is under.
Three million gas is recovered through modified
system.
The price of 1 million cubic feet per day = $2,000
Per day income =
Three million gas recovered = 3 * 2000 = $6,000
Monthly income = $6,000 * 30 = $18,0000
Yearly income = $18,0000* 12 = $216,0000
2.5
980
1000
1020
1040
1060
1080
1100
1120
1140
RPM
2.4
2.65
2.7
2.75
2.8
2.85
2.9
2.95
Compression Ratio
Fig. 5: Suction pressure, temperature, specific gravity,

RPM, compression ratio, vs gas flow (All factors
changed)
ENVIRONMENTAL BENEFITS
Oil and gas industry, refinery and petrochemical
plants, based on type of manufactured products and
waste materials. Chemicals and industrial processes
also produce pollution. As a result, these contaminants
might polluting weather, soil, air and it causes serious
damage to humans and environment. Table VII and
table VIII showcontaminants quantity of flare gas of
2014 and 2015. By comparison of both tables it
concluded that by the recovery of flare gas the affect
of contaminant gases is reduced.
TABLE VII: Contaminant Quantity PER G/M3 of
Flare Gas of Nashpa Oil and Gas Field 2014
16
[1]
[2]
[3]
[4]
TABLE VIII: Contaminant Quantity PER G/M3 OF
Flare Gas of Nashpa Oil and Gas Field 2015
V.
SUMMARY AND
CONCLUSIONS
Flaring of gas is a huge problem for people,

environment and also economic problem. There is
emergent concentration to reduce gas flaring, due to
which the pollution emissions generated by flaring of
gas and possible large emission sources within a plant.
In this research , the method of recovery of flare gas
has been assessed by deploy design of experiment
techniques one factor at a time and furthermore,
factorial experiment was conducted by varying various
factors together for the recovery of flare gas. In this
research, the existing system is modified leading to
recovery of flare gas (three million). For modified
system suction pressure is set on 150 PSI, temperature
is 140 F, specific gravity is changed to 0.70,
compression ratio for both stages of the compressor is
set on 2.70 and 2.75 and RPM is changed to 1120.
recovery of large amount of flare gas minimizes the
negative environmental impacts. The CO is reduced
from 1.91 to 0.175, NO is reduced from 4.91 to 0.46,
SO2 from 1.52 to 0.30 and NOx is reduced 0.070.
REFERENCES
[5]
Opportunities in Pakistans upper stream oil and

gas sector Ministry of Petroleum & Natural
Resources Government
of Pakistan. www.mpnr.gov.pk
Definition of gas flaring, retrieved at
www.ehow.com on 3/6/2013, www.pcq.com.pk
Elvidge, D, Ziskin D, Baugh B, Tuttle B,
Ghosh
T., Pack D, Erwin E, Zhizhin M.A
Fifteen Year Record of Global Natural Gas
Flaring Derived from Satellite Data. Energies.
2009.2(3) 595 622
Abdulkareem, A and Odigure, J. Economic benefit
of natural gas utilisation in Nigeria: A case Study
of the food processing Industry. Energy
sources.2010. 5(1) 106 114.
Onwukwe S. Gas-to-Liquid technology:
Prospect for natural gas utilization in Nigeria
[6] Blasing T, Hand J. And Kimberly B. Monthly

carbon emissions for natural Gas flaring and
cement manufacture in
the United States. Tellus B. 2007.59(1) 15 28.
[7]
Akpan, S. The production and utilisation of
Natural gas resources in Nigeria: A review.
Nigerian annual international conference and
exhibition in August of the Society of
Petroleum Engineers in Abuja, Nigeria.2009.
[8]
Hill R, C- E Offshore production systems Inc,
Moore, J, Continental oil Co. OfIndonesia, Boone
D and Randall C. Associated gas utilisation with
liquid recovery.Offshore
technology conference. 1985.
[9] EPA. Environmental Protection Agency (1995)
Emission factors& AP42,Compilation of air
pollutant Emission factors. http://www.epa.gov
[10] American Petroleum Institute website,
Environmental regulations of the Explorationand
Production in the Oil and
Gas Industry.
Retrieved at www.ipo.org on 03/23/13.
[11] Douglas C. Montgomery Design and Analysis
of Experiments fifth Edition.
[12]
Daniel D. FREY and Hungjen WANG 2006

Adaptive One- Factor-at-a-Time Experimentation
and Expected Value of
Improvement.
[13]
ARIEL Heavy Duty Balanced Opposed
Compressors
Technical Manual For Models JGE, JGH, JGK
and JGT.
[14]
"Natural Gas Pipeline Development and
Expansion,"
U.S.
Energy
Information
Administration.
[15] Pell Research Report on Oil and Gas Pipeline
Construction cited with permission.
17
[16]
Imail OS, Umokoro GE (2012) Global

Impact Of Ga Flaring.Energy and Power
Engineering
4,
290302.
doi:
10.4236/epe.2012.44039.
[17]
API 521. 2007. Pressure-relieving and
depressuring systems. 5th Edition. American
Petroleum Institute.
[18]
Ismail O. Sahad and Umukoro G. Ezaina,
Global Impact of Gas flaring, Energy & Power
Engineer, 290-302, vol.4
issue 4, (July 2012)
18
SAME-03
Bulge hydroforming of tube by rubber mandrel without axial feed:

Experiment and Numerical Simulation
Faisal Qayyum1, Masood Shah1, Saad Ali1 and Umar Ali1
1
Mechanical Engineering Department, University of Engineering and Technology Taxila

Taxila, Pakistan
fysalqayyum@yahoo.com
Abstract Rubber forming has become an

important tool to manufacture high precision and
complex tube parts, but as the part complexity and
the choice of metal increases the calculations
required to accurately determine the final shape
become increasingly complex. The accuracy of the
final shape of the product depends upon the
different parameters, like material properties i.e.
Youngs modulus, elongation before rupture,
plastic properties of material, coefficient of friction
between the die and work piece.
In the current research tubes of AISI SS304 have
been hydroformed into bulge shape without axial
feeding mechanism. Urethane rubber instead of
hydraulic oil has been used to internally pressurize
the surface. Material properties for tubular SS304
were extracted from already published literature,
which was further used to carry out successful
simulation of the process in ABAQUS Standard
and to analytically calculate several geometric
parameters. Detailed simulation results have been
presented and tube wall thinning along the
periphery has been observed. Comparison between
analytical and simulation results show close
relationship.
Keywords-Tube Hydroforming; Bulge Shape;
Numerical Simulation; rubber mandrel
VII. INTRODUCTION
As compared to stamping process, the tube
hydroforming process is relatively complex,
development of a new hydroforming shape requires
expensive try-outs to establish the necessary process,

part and tooling designs. Establishment of process
goes back to 1939 when Grey et al. [1] investigated
manufacturing of seamless copper fittings with T
protrusions using a combination of internal pressure
and axial load. The hydroforming process got fame
after his work and researchers carried on further
research on this topic to make complex shapes. As the
part complexity increases, experience and design rules
may be limited in assisting engineers and designers.
Furthermore, there is not an extensively available
mathematical or numerical model for prediction of
final shapes and deformations yet for hydroforming
technology. Hence, in order to reduce development
time, effort and cost,
computer aided engineering (CAE) tools need to be
used effectively [2].
The finite element method (FEM) has been proven to
be a useful tool in the past for conventional metal
forming processes to obtain reliable and precise
information about the process, geometry and material
parameters [3]. Now a days use of FEM for tube
hydroforming process simulations has become a
standard
development
tool
after
numerical
investigations and experimental validations conducted
by many researchers since the early 1990s [4] and
application of current commercial FEM software, such
as LS-DYNA, PAM-STAMP, ABAQUS, MARC,
AUTOFORM and DEFORM into tube hydroforming
process was performed and presented successfully [5].
The feasibility of forming a specific part can be
predicted by analyzing thinning, thickening, stress and
19
strain distribution on a deformed tube. The effects of

different parameters can be investigated by varying
important dimensions or loading conditions on
common parts. Hence generic rules can be established
for future problems [3]. Kang showed that axial feed
has an important influence in order to obtain sound
products [6], he also showed that poor lubrication
often results in premature failure, due to excessive
local thinning. Bandar Alzahrani, Gracious Ngaile and
Chen Yang [7] established a model based on
equilibrium conditions, yielding criteria, geometrical
relationships, and a volume constancy condition. The
system of equations was derived and solved for
various process parameters. They validated their model
using finite element analysis and experiments. The
model has the ability to predict process parameters
such as stresses, strains, internal pressure, geometry
variables, and thinning rate distribution. M. Saboori,
et.al. [8] conducted free expansion testing of AISI SS
304 to calculate it more accurate 3D formability.
Tube hydroforming parts have various common
features like bulging, protrusion, and local expansions
[9]. Material properties were found essentially
affecting these tube hydroforming process output
features [10, 11]. A number of friction tests were
developed to determine the coefficient of friction [5], a
new analytical model to determine the friction
coefficient in the forming zone was proposed [12].
Limb [13] reported the effect of lubrication on the
bulge shape. Purit Thanakijkasem, et.al. [14] used
technique of Bright Annealing (BA) to enhance the
formability of SS 304 in tube hydroforming process.
Yeong-Maw Hwang and Yi-Kai Lin [15] proposed a
mathematical model considering an ellipsoidal surface
for tube hydroforming to analyze the plastic
deformation behavior of a tube during hydraulic bulge
forming. To verify the validity of their model, FEM
code DEFORM was used to simulate the forming
process. The effects of various forming conditions,
such as the strain-hardening exponent, the initial tube
length, etc., upon the forming pressures were
discussed. Abed Alaswad and Abdul Ghani Olabi [16]
performed finite element study using LS DYNA code
to investigate the effect of geometrical factors, bulge
height, stress distribution and thickness variation were
studied for the hydroformed part. B. Yang, et. al. [17]
used LS DYNA code to simulate bulge shape tube
hydroforming process and investigated the effect of
lading path on formation of bulge shape and wall
thickness distribution.
In this work, the bulge hydroforming of SS304 was

accomplished without axial feeding under different
loading and boundary conditions. Axial feeding is a
parameter which involved complex tooling and
control. Less axial feed results in failure of wall, while
greater axial feed results in wrinkling. Getting the
desired results without axial feed can save cost of
tooling and can simplify the production process. The
FEA analysis of tube hydroforming process for current
case was carried out using ABAQUS/StandardTM
software package because it provides detailed material
modeling and is relatively more accurate at high
strains. The aim of this research was to verify the
accuracy and use of ABAQUS/StandardTM for the
tube hydroforming under different conditions by
comparing the experimental data with the simulation
results, and see the effect of friction on various
parameters like material forming, corner filling, wall
thickness and final geometry of the finished product.
II. Experimentation
Two piece split die made from AISI 1060 steel was
used during experimentation. The die was
manufactured by initially cutting metal blocks on
DK7740 Wire EDM shown in Fig. 1, followed by
surface grinding. Cavity was machined into the split
die by using automatic control die sinking EDM Neuar
M30 shown in Fig. 2.
Figure 3.
DK7740 Wire EDM
20
Figure 6(a). Split view of hydroforming die with die

sinking electrode
Figure 4. Automatic control die sinking EDM Neuar
M30
To smooth down the rough surface obtained after die

sinking the cavity was polished using Dremel(R) tool.
The reduction in friction is especially important while
performing bulging without axial feed as M. Koc and
L.P. Lei through their research work [18, 19] have
showed that with low friction, more homogeneous
deformation can be achieved.
Both halves of split die with die sinking electrode and
tightening bolts is shown in Fig. 4 (a), Fully finished
die as shown in Fig. 4 (b) was mounted on 1 ton
hydraulic press which has been shown in Fig. 3 which
is available in Machine Tools and Manufacturing
Laboratory, Mechanical Engineering Department,
University of Engineering and Technology, Taxila.
Figure 4 (b). Polished cavity of die
AISI SS304 Tubular specimens of 47 mm length,

0.5mm thickness and outer diameter 18 mm as shown
in shown in Fig. 11 (a) were used, chemical
composition of material is shown in Table 1.
Table 1. Chemical composition of SS304
Table 3:
Chemical
Compositi
on of
SS304
Figure 5. Hydraulic press of 1 ton capacity
Comp.
Cr Fe
Mn
Ni
Si
Wt. %
0.0
8
20
10.
5
66.
3
AISI SS304 was specifically used during our

experimentation because it is highly ductile grade of
stainless steel, normally used in manufacturing
processes where high formability of material is
required. Tube and rubber mandrel were inserted into
the die cavity and the plunger was placed above. The
21
schematic diagram of the system is shown in Fig. 5.

We used rubber mandrel to internally pressurize the
tube instead of hydraulic oil because less critical
sealing conditions are required in the case of rubber.
Localized pressurization technique was adopted to
successfully manufacture the part. Rubber mandrels of
smaller lengths which could force tube locally to file
the die cavity were used. Complete procedure was
done in 3 to 4 multiple pressurizing punches by
successively increasing mandrel length in each punch.
The punch had clearance of 0.1 mm from tube so
that no axial load acts upon the walls of the tube and
the tube wall might not get gripped against the die
which has been depicted in Fig. 5. The pressure was
applied in 3 steps with different size of rubber
mandrels for better coroner filling and to make sure
homogeneous material flow into the die cavity.
The final shape and geometry of fully developed part
was observed using as axial section as shown in Fig. 6
cut from a fully hydrofomed tube by using DK7740
Wire EDM. The section gave us the possibility of
measuring radii formation and thicknesses change
along the periphery of the part.
Figure 6. Sectioned view of hydroformed tube
VIII. SIMULATION
Numerical simulation of the tube hydroforming
process, specifically for our case was carried out on
ABAQUS/StandardTM. Detailed finite element analysis
was performed in 3-D axisymmetric environment. The
die was modeled as an analytical rigid surface which
does not require meshing and tube was modeled as a
deformable shell of 47 mm length and 1 mm thickness,
the elastic and plastic material properties of Stainless
Steel AISI304 were assigned to the tube section which
were taken from ASM Atlas of Stress Strain Curves
[20] as shown in Fig. 7.
Figure 7. Plastic region of stress-strain curve of AISI

SS304 at different temperatures process
Figure 5.
Schematic
hydroforming process
diagram
of
the
tube
Final assembly of die, tube and depiction of applied

internal pressure is shown in Fig. 8. Die and tube were
assembled together in the assembly mode using
displace instance command. The contact was defined as
frictional contact of frictional coefficient 0.1. A 4-node
bilinear
axisymmetric
quadrilateral,
reduced
integration, hourglass control element type CAX4R
was selected and the tube was freely meshed to 400
elements and 603 nodes. The die was constrained to be
rigid, and the tube was fixed from one end. Different
values of internal pressures were applied during
experimentation. It was observed that for better
22
forming of tube and acceptable corner filling internal

pressure of 50 MPa should be applied. Generalized
depiction of results obtained from simulation after
successful compilation of results has been depicted in
Fig. 9.
was also observed by doing several experiments that

there must not be sharp corners in the die. As shown in
Fig. 10(a) and Fig. 10(b) that both tubes have failed due
to excessive internal pressures, but instead of failing
from bulge due to excessive thinning of wall the failure
occurred at sharp corners.
As discussed earlier that AISI SS304 tubes were
successfully hydroformed into desired bulge shape. Fig.
11 (a) shows initial shape and size of tube before
hydroforming and Fig. 11 (b) shows the finals shape
and geometry of hydroformed tube. It can be observed
from this figure that reduction in length of tube occurs
when the material flows into the die cavity.
Figure 8.
conditions
Final assembly with applied boundary
Figure 10. Failed tubes of mateirlas having low

formability (a) copper (b) alluminum
Figure 9. Depiction of results from FEA analysis,

showing flow of material into the die cavity from free
side of the tube.
IX. RESULTS AND DISCUSSIONS

Experimental procedure for tube hydroforming has
been discussed previously. During our experimentation
we tried to hydroform tubes of Copper and Aluminum
as well. Results of those experiments have been
depicted in Fig. 10. It was observed that materials
having low formability and lower yield strength are
very difficult to hydroform beyond a certain range. It
Figure 11.
(a) AISI SS304 tube before
hydroforming (b) shape of tube after hydroforming
23
To get a deep insight of the process, a successfully

hydroformed tube was sectioned and was compared
with the results obtained from simulation. The
comparison has been showed in Fig. 12. Internal
bulge shape of 10mm has been formed while the
tube forms within 16mm.Outer radii of 3mm are
also fully formed in experimentation and
simulation.
R 3mm
R 3mm
10mm16mm
(a)
(b)
This study was utilized to calculate the analytical

values of hoop and longitudinal stress and strain and
the analytical results obtained have been compared
with the ones obtained from numerical simulation in
Table 2. The values obtained from these different
methods are close to each other within 10% error.
Table 2: Comparison of Results
Table 2:
Compariso
n of
Analytical
and
Numerical
results
Analytical
Parameter Calculatio
s
n
Hoop
Stress
8.75E+08
8.12E+08
Hoop
Strain
4.17E-03
3.12E-03
Longitudi
nal Stress
4.13E+08
4.91E+08
Longitudi
nal Strain
1.97E-03
2.08-03
Equivalent
Stress
7.58E+08
Figure 12. Comparison of Experimental results

with numerical solution a) numerical simulation
results b) experimental results.
Friedrich Klaas in 1987 [21] derived the conditions for
yield initiation of a cylindrical straight tube under axial
force and internal pressure and resulting stress state,
according to him the circumferential stress, Hoop stress
and effective stress values can be described as shown in
equation 1, 2 & 3.
Numerical
Simulation
6.67E-03
Variation in thickness of tube wall due to

hydroforming was also examined in simulation, which
has been shown in Fig. 13. It can clearly be seen from
the graph that tube wall gets extensively thin during
free forming at bulge, the wall gets thicker at outer
edges due to compressive stresses and remains almost
constant where the tube is in direct contact with the
die.
= 2 + 2
(3)
24
[3]
[4]
[5]
[6]
[7]
Figure 13. Wall thickness reduction obtained from

numerical solution a) Numerical Simulation b)
Thickness variation along path (all units are in mm).
X. CONCLUSION
After conducting experiments on die consisting of
sharp corners and smooth corners we concluded that
hydroforming of thin tubes is extremely difficult if the
die has sharp corners. By conducting experiments on
tubes of same geometry and thickness but of different
materials we concluded that material formability is
another very important factor in the hydroforming
process, if the material does not have enough
formability it will fail before it fully forms into the die
cavity. We used rubber mandrel to internally pressurize
the tube instead of using hydraulic oil, it was observed
that less critical sealing conditions are required if we
use elastomer instead of oil for hydroforming. In our
research we compared experimental results with results
obtained from simulation and analytical model of
Friedrich Klaas in 1987 [21] and a reasonable
agreement between these approaches was found.
Acknowledgment
The study was funded and supported by
Department of Mechanical and Aeronautical
Engineering, University of Engineering and
Technology Taxila, Punjab, Pakistan.
References
[1] Grey, J., A. Devereaux, and W. Parker, Apparatus
for making wrought metal Ts. US patent, 1939.
2(203): p. 868.J. Clerk Maxwell, A Treatise on
Electricity and Magnetism, 3rd ed., vol. 2. Oxford:
Clarendon, 1892, pp.6873.
[2] Ko, M. and T. Altan, Application of two
dimensional (2D) FEA for the tube hydroforming
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
process. International Journal of Machine Tools

and Manufacture, 2002. 42(11): p. 1285-1295..
Ko, M., et al., The use of FEA and design of
experiments to establish design guidelines for
simple hydroformed parts. International Journal of
Machine Tools and Manufacture, 2000. 40(15): p.
2249-2266.R. Nicole, Title of paper with only
first word capitalized, J. Name Stand. Abbrev., in
press.
Zhou, D. and R. Wagoner, Development and
application of sheet-forming simulation. Journal of
Materials Processing Technology, 1995. 50(1): p.
1-16.M. Young, The Technical Writer's
Handbook. Mill Valley, CA: University Science,
1989.
Alaswad, A., K. Benyounis, and A. Olabi, Tube
hydroforming process: A reference guide.
Materials & Design, 2012. 33: p. 328-339.
Kang, B.H., et al., Forming various shapes of
tubular bellows using a single-step hydroforming
process. Journal of Materials Processing
Technology, 2007. 194(1): p. 1-6.
Alzahrani, B., G. Ngaile, and C. Yang, Part 1:
Analytical modeling of symmetric multi-nose tube
hydroforming.
Journal
of
Manufacturing
Processes, 2013. 15(2): p. 273-286.
Saboori, M., et al., Evaluating the flow stress of
aerospace alloys for tube hydroforming process by
free expansion testing. The International Journal of
Advanced Manufacturing Technology, 2014. 72(912): p. 1275-1286.
Ko, M., Development of design guidelines for
part, tooling and process in the tube hydroforming
technology. 1999, Ohio State University.
Ko, M., E. Billur, and .N. Cora, An
experimental study on the comparative assessment
of hydraulic bulge test analysis methods. Materials
& Design, 2011. 32(1): p. 272-281.
Lianfa, Y. and G. Cheng, Determination of stress
strain relationship of tubular material with
hydraulic bulge test. Thin-walled structures, 2008.
46(2): p. 147-154.
Plancak, M., F. Vollertsen, and J. Woitschig,
Analysis, finite element simulation and
experimental investigation of friction in tube
hydroforming. Journal of materials processing
technology, 2005. 170(1): p. 220-228.
Limb, M., et al. The forming of axisymmetric and
asymmetric components from tube. in Proceedings
of the 14th International MTDR Conference. 1973.
sl]:[sn].
Thanakijkasem, P., et al., Effect of bright
annealing on stainless steel 304 formability in tube
hydroforming. The International Journal of
Advanced Manufacturing Technology, 2014: p. 19.
Hwang, Y.-M. and Y.-K. Lin, Analysis and finite
element simulation of the tube bulge hydroforming
process. Journal of Materials Processing
Technology, 2002. 125126(0): p. 821-825.
Alaswad, A. and A.G. Olabi. A numerical study of
the effect of geometrical factors on bi-layered tube
hydroforming. in Proceeding of 7th European LSDYNA conference. 2009.
Yang, B., W.G. Zhang, and S.H. Li, Analysis and
finite element simulation of the tube bulge
hydroforming process. The International Journal of
Advanced Manufacturing Technology, 2006. 29(56): p. 453-458.
Koc, M. and T. Altan, Prediction of forming limits
and parameters in the tube hydroforming process.
International Journal of Machine Tools and
Manufacture, 2002. 42(1): p. 123-138.
25
[19]
[20]
[21]
Lei, L., B. Kang, and S. Kang, Prediction of the

forming limit in hydroforming processes using the
finite element method and a ductile fracture
criterion. Journal of materials processing
technology, 2001. 113(1): p. 673-679.
Boyer, H.E., Atlas of stress-strain curves. 1997:
American Society for Metals.
Klaas,F., Aufweitstauchen von Rohren durch
Innenhochdruckumformen. 1987: VDI-Verl. D@:
usseldorf.
26
SAME-04
Perspectives in Automotive Embedded Systems from Manual to Fully

Autonomous Vehicles
Z. H. Khan1 and Arsalan Khan2

Dept. of Electrical Engineering, Riphah International University, Islamabad, Pakistan
zeeshan.hameed@riphah.edu.pk
Dept. of Control and Simulation, Center of Excellence in Science & Applied Technologies Islamabad,
Pakistan
arsalan1@mail.nwpu.edu.cn
AbstractThis paper describes the state of the

art in embedded system architecture as found in
automotive. As the autonomy increases, the manin-the-loop operation ends up with increased
system complexity and total emphasis on
electronics and software replacing human
driver. A modern automotive is composed of
multiple redundant ECUs connected via reliable
communication network which transfers multisensor data for real time computation of
command that is sent to actuators. Automotive
systems are safety critical which requires
comprehensive verification, validation and
testing (VVT) produces so as to ensure proper
functioning under all situations. Inspired from
the Google self-driving car, it is expected that
such high performance fully autonomous cars
will capture the future automotive market.
Index
TermsDrive-by-wire,
automotive
electronics, embedded systems, real time control,
controller area network.
XI. INTRODUCTION
In recent years, a technology breakthrough has
been achieved in vehicular automation to extend
previously manual, teleoperated and semiautonomous vehicles to fully robotic or autonomous
systems [1]. Thus, the trend in automotive industry
has transitioned from classical mechanical control to
drive-by-wire system (drive-by-wireless for
teleoperated vehicles) and finally to driver-less
vehicles [2-4] as shown in fig.1. In this paper, we
discuss some basic subsystems of such complex
systems with special consideration on real-time
communication, computation and control. While
ensuring successful design iteration, a multi-domain
integrated approach is required for these safety
critical hard real time embedded systems which have
to sense and control vehicles stability, tracking,
braking, steering and navigation. In-vehicle
embedded networks e.g. CAN, LIN, FlexRay etc.
are used as a communication bus to connect all
electronic control units (ECUs), sensors and
actuators in a dependable/fault tolerant architecture
[5]. Therefore, time dependency of critical
information must be ensured through intelligent
design of communication system [6].
27
The paradigm shift from semi-autonomous to

fully-autonomous vehicles has put a lot of burden on
In-vehicle embedded design perspective which
requires more number of coordinated ECUs, sensors
and actuators resulting in excessive complexity,
security and safety requirements as well as cost for
autonomous operation [7]. One of the greatest
inspirations for the commercial autonomous vehicles
has arrived from the unmanned ground vehicles
(UGVs) in defense and space applications. These
unmanned vehicles autonomously perform the
desired mission without the human intervention.
However, for unexpected scenarios, human
intervention via teleoperation mode is possible.
Some elegant examples of such systems include
Mars Rover, Bomb disposable robots, DARPA
Grand challenge automotive etc. Now-a-days, these
developed technologies are used in commercial
applications as witnessed by the self-driving cars.
The current research describes some experiences in
developing teleoperated vehicles and its comparison
with the autonomous systems from control and
automation perspective [8, 9].
Fig. 1.
Autonomy spectrum in Automotive Systems
In fact, DARPA Grand challenge laid the stone of

benchmarking in autonomous drive [10]. One such
autonomous vehicle is Stanley, who won the
competition of driver-less voyage in DARPAs
grand challenge held in 2005. The vehicle was an
actual Volkswagen Touareg modified with actuators
driven by onboard computers. It was also equipped
with multiple sensors i.e. 5 LIDAR lasers, Single
lens and Stereo cameras as well as 24 GHz RADAR
to replace a human driver [11]. Figure 2 shows the
vehicle which uses a path following algorithm using
machine learning approach running in real-time.
This algorithm uses real time images and obstacle
avoidance map using lasers. This simultaneous
localization and mapping (SLAM) algorithm shows
12.6% error due to false obstacles.
This paper is organized as follows: Section II

describes the details on autonomous vehicle
technology, Section III discusses the impact of
autonomous vehicle technology, Embedded
networks in automotive are presented in Section IV
with emphasis on CAN protocol and its variants
followed by the discussion on embedded
architectures used in automotive as found in Section
V. Some details on high performance computing in
automotive environment are forecast in Section VI,
while section VII concludes the paper.
TABLE I.
AUTONOMOUS CAR PIONEER PROJECTS
Project/Le
ader
E.
Dickmanns
PROMET
HEUS
CMU
Navlab
Carsense
Coun
try
Germ
any
Germ
any
USA
DARPA
demo III
ParkShuttle
DARPA
Challenge
II
DARPA
Challenge
III
USA
Vis-lab
Italy
Google
SDC
USA
EU
Germ
any
USA
USA
Platform/Theme
Year
1980
s
Autobahn
198795
11
different 1995
vehicles
Alfa 156 Sport 2000wagon 2.0
02
Collision
2001
avoidance demo
Automatic Shuttle 2002
Service
Autonomous drive 2005
in rough terrain
without traffic
Autonomous drive 2007
in
urban
environment with
traffic
Piaggio
Porter 2012
Electric vans
Lexus
RX450h 2013
SUVs
Mercedes-Benz
XII. AUTONOMOUS VEHICLE TECHNOLOGY

The history of automobiles dates back to 17th and
th
18 century when simple machines were invented
with tri-cycle structure. Steam engines were used at
those times as the only propulsion source, however,
28
soon some engineers worked to develop the first

ever electrical vehicle of ancient times.
Fig. 2.
Stanley- Winner of DARPA Grand Challenge

II
TABLE II.
Autono
my
Level 0
Level 1
Level 2
Level 3
LEVEL OF AUTONOMY AS DESCRIBED BY

NHTSA [12]
Description
Human driver has the complete
manual control
One function is automated, others
are manual
More than one function is automated
at any instant means higher
autonomy than Level 1
Driving is automated. Driver can do
other
activities
without
Fig. 3.
Level 4
compromising on safety
Autonomous car can drive it-self
without a human driver. Computer
has full control
Most of the development was witnessed in the

pre and post-world war-II era when more mature
vehicle designs and propulsion techniques were
implemented. In the modern age of automotive
market, designs are motivated by CAD simulations
to reduce aerodynamic drag, vehicle weight,
increase engine power, optimized fuel consumption
and more security and safety features to human
driver which is an integral part of a manually driven
automotive. Various benchmarks exist in the
autonomous driving category as shown in Table 1.
As suggested by National Highway Traffic Safety
Administration (NHTSA), different levels of
automation have been described for the automotive
as listed in Table 2.
Distributed embedded architecture of a Scania truck (simplified version)

29
Figure 1 depicts the autonomy spectrum in the

automotive development. As seen, the classical
approach was limited to manual control by human
driver. Later, some safety measures were
introduced to avoid accidents as per standards
enforced by SAE. These safety measures were
then combined in the form of a sub-system named
as driver assistance system to avoid
compromises on the vehicles stability. One such
example is that of an active yaw control or
dynamic stability control system which aids the
driver in case of over-steering and limits the
vehicle to skid out of the road. Other examples
include Anti-lock brake system (ABS), Collision
avoidance system (CAS) and the automatic cruise
control systems (CCS) to assist human drivers. A
comparison of their reliability requirements is
listed in Table 3.
In some cases, driving a ground vehicle is not
possible, so a remote driver uses the real-time
communication and control to send commands to
the remote vehicle and receive video from the
vehicle to carry on driving task [13]. Because of
man-in-the-loop operation of a teleoperated
ground vehicle, less number of sensors is required.
However, some autonomous features e.g. autoSTOP and collision-avoidance task are
implemented on-board in order to handle the
communication link failure scenario in which case,
the vehicle stops immediately [8]. In fully
autonomous vehicles, all sensing and actuation is
done on-board and it increases the system
complexity [14]. Real time handling of multisensor
data
and
implementation
of
computationally efficient fusion and decision
making algorithms require sufficient embedded
power and high performance computing platforms
[15]. On the other hand, creating intelligent
automotive systems can be utilized in automated
urban transport systems, automated highway
driving systems and finally the driverless car
technology as demonstrated by Vis-Lab, Google
and other companies.
This setup is developed and maintained at
GIPSA-lab, Grenoble, France [8, 16]. While
comparing remotely teleoperated vehicles with
fully-autonomous ones, one major difference is the
dependency of former on the communication
factor. Due to un-deterministic behavior as seen in
the presence of delay and packet losses, the
teleoperation system requires some autonomy to

minimize such effects, whereas, the later one is
independent of the communication factor after
paying the cost of increased autonomy.
TABLE III. COMPARISON OF RELIABILITY
REQUIREMENTS FOR AUTOMOTIVE SUBSYSTEMS
Subsystem
t r (s)
Antilock
Brake
System
Engine
Manageme
nt
Air Bag
< 2x104
Navigation
System
104
Reliabilit Comments
y
requirem
ents
high
Fail-safe
design
10-100
high
Robust
control
104
Very
high
moderate
DSP based
switching
Reliable
sensor in
case of GPS
outages
In teleoperation applications, a master station

drives a slave robot at a distance. The acceleration
and brake commands are sent via a wireless
network, while the streaming video is used to take
control decision in real-time. The test-bench setup
of a networked control ground vehicle NECSCar is shown in fig.4.
Fig. 4.
Bilateral teleoperation schematic
Comparing the hardware and software

complexity of an autonomous vehicle, lets take
the example of Stanley. It used 6 Pentium M
computers, a Gigabit Ethernet switch, Brake,
Throttle, Gear Shifter and steering controller,
IMU/GPS sensors, actuators and various
interfacing devices [17]. It consumes about 500 W
to power up this autonomous vehicle. On the
software side, the operating system is Linux with
six main functional groups: Sensor interface,
perception, control, vehicle interface, and user
interface. This architecture exchanges information
30
with in different modules to cooperate for

autonomous sensing and actuation. On the other
hand, if it was developed as a teleoperated vehicle
where driving commands were being sent based on
video received from the car mounted cameras, far
less computing was required.
XIII. IMPACT OF AUTONOMOUS VEHICLE
TECHNOLOGY
The automotive manufacturers used the recent
technologies for ensuring safety, comfort and
entertainment in compact aerodynamically profiled
cars. Following features are important for
discussion in order to understand the impact of
autonomous vehicle technology on men and
society.
1) Safety
The purpose of increasing autonomy is to
improve safety of the system. Each year, more
than 1.3 million people die in road accidents.
About 90% of these accidents are due to human
error. In driver assisted systems, some functions
are implemented to work autonomously e.g. ABS
in case of skidding, Airbag deployment after
collision detection and interrupted injection of fuel
if engine management system detects a broken
ignition plug which may lead to engine fire
otherwise [10].
Longitudinal Controllers: The longitudinal
control functions to control the vehicles
speed and assist in parking, cruise control
and pre-crash break. Full autonomous
system requires an integration of lateral and
longitudinal controllers.
Lateral Controllers: These control the
lateral dynamics of the vehicle e.g. Land
departure warning system (LDWS) and
Lane Keeping Assist System (LKA) are
installed in vehicles to keep the vehicle in
the chosen lane and avoid sideswipe
crashes. Parallel Parking Assistance System
(PPA) was first used in 2003 by Toyota to
assist drivers to park in by using a rear
camera display on dashboard screen.
2) Fuel Consumption
The embedded systems in automotive are
designed to work in an optimized manner.
Electronic fuel injection (EFI) systems are used for
controlled injection into combustion chamber
which ensures fuel savings as compared to manual

control.
3) Environment
The impact of using computer control for
critical engine sub-systems results in precise
timing
during
engines
intake-injectioncombustion-exhaust cycle. This result in complete
burning of fuel which minimizes the residual gases
and smoke expelled into the environment.
4) Weight reduction
Considerable weight saving is achieved with
embedded electronics as drive-by-wire system
removes mechanical links and related accessories
[18]. Due to embedded network, considerable
weight savings is achieved in terms of decreased
volume of cables and wires.
5) Reliability
Embedded architectures use fail safe strategy
for communication of critical information. This
ensures reliable operation [5]. Also, since
mechanical parts are lesser in number, less wear
and tear results in longer down time and mean
time between failures (MTBF).
6) Social Issues
Considering the socio-economic benefits of this
technology, it is evident that autonomous vehicles
will allow improved safety, better fuel economy,
efficient traffic flow and considerable savings in
time and maintenance. A considerable decrease in
number of accidents and lives lost due to them
have been reported after the implementation of
autonomous safety features in automotive [19].
However, social aspects are often neglected
when technologies are implemented. It seems that
if autonomous cars will be commonly used in
urban areas, no licensing requirement will exist as
humans will not be driving the vehicle. Another
important aspect of driver-less vehicles is who to
blame in case of an accident? Human drivers can
violate the signals and can do over-speeding but
autonomous cars could not. It may be very
difficult for traffic police to determine whose fault
is there in case of a crash [20]. In such cases,
standardization of safety measures in all
autonomous cars would also be mandatory. The
investigators will require not only the blackboxes of each vehicle but also the video from
nearby surveillance cameras to precisely find a
clue instead of relying on eye-witness only.
31
XIV. EMBEDDED NETWORKS IN AUTOMOTIVE

A modern automotive vehicle is regarded as
network on wheels due to electrification of
major subsystems. Controller area network (CAN)
is the most popular automotive network protocol.
It was proposed by Bosch GmbH in 1983. It was
then used in automotive communication systems to
interconnect various subsystems [21].
Later on, more variants of CAN came into the
market e.g. SAE J1939 (higher level protocol for
automotive), CANopen (open application CAN),
DeviceNet (distributed industrial automation),
CAN Kingdom (motion control applications),
SeaCAN (maritime applications) and CAN FD etc.
In CAN FD (flexible data-rate), the data rate is the
same as 1 Mbps but payload can vary longer than
8 bytes in each frame. TTCAN is the deterministic
version of the CAN protocol known as timetriggered controller area network. In this protocol,
nodes are synchronized by using a periodic signal
which is recognized by each node to align its
respective clock [22].
Fig. 5.
CAN message format
The worst case message frame size on CAN is

134 bits. It depends on how much data bits are
being transmitted. The maximum data rate on
CAN is up to 1 Mbps with in a distance of 40 m.
The CAN arbitration mechanism follows carrier
sense multiple access with arbitrary message
priority (CSMA/AMP). Each node is assigned a
priority according to which it gets access of the
bus. As such there is no destination address in the
frame sent by the transmitting node. Every node
gets all the traffic and it has to filter out the
relevant messages on the bus. Therefore, lowpriority nodes may experience excessive delay if
high priority nodes are more active. Error
detection in CAN is through bit stuffing with
frame check, ACK and cyclic redundancy check
(CRC).
Vehicle-to-Vehicle (V2V) communication

for cooperative awareness and driving
Fig. 6.
XV. IN-VEHICLE EMBEDDED ARCHITECTURE

In modern vehicles, computer assisted driving
system are common. Model based techniques for
safety critical systems are found to be applicable
on automotive system design too [23]. In usual
architectures a centralized controller is used with
parallel buses connecting various ECUs [24].
Figure 2 describes the distributed embedded
architecture of a SCANIA truck. It shows three
isolated buses connected to a coordinator system
(COO). The architecture is based on three CAN
buses and shown as Green, Yellow and Red buses
[22]. This color coding is used to exhibit the
relative importance of each bus. Green bus is the
least important one as it connects the audio
system, road information system, climate control
etc., while the red bus comprises of the most
critical ECUs including engine and brake
management, suspension control and exhaust
emission control. The middle (yellow) bus
connects the less critical data as compared to the
red bus e.g. locking and alarm control, instrument
cluster control, Bus chassis control etc. The central
ECU (i.e. COO) functions as a gateway to
interconnect three parallel CAN buses. A
diagnostic bus connected to green CAN is used for
fault detection, diagnosis and error reporting.
A higher level protocol above CAN used for
trucks is SAE J1939 protocol which enables plugand-play function. But due to fixed priorities,
optimization is difficult on this network. More
research is going on to obtain a robust protocol for
vehicle automation which should offer flexibility
as well as scalability.
It is important to note that the embedded
architecture is dependent on the application type.
CAN is being used in aerospace and even space
applications. Airbus A380 avionics system uses a
CAN bus for aircraft door position sensing. Also,
32
in SMART-1 spacecraft developed by Swedish

space agency, redundant CAN buses are used
which are designed such that they are resistant to
radiations effect as desired for space operation. In
addition to embedded networks, a CAN gateway is
used to inter-connect vehicles for vehicle to
vehicle (V2V) and vehicle to infra-structure (V2I)
communication as shown in fig. 6.
Yearly increase in automotive software

complexity shown by million lines of code of
automotive software
Fig. 7.
XVI. HIGH PERFORMANCE COMPUTING IN

AUTOMOTIVE ENVIRONMENT
As stated earlier, the increased autonomy comes
with a cost and that is complexity. Figure 7
describes how this complexity is increasing yearly
in an exponential manner. It shows that during
2004 to 2013, the number has increased from 5
million to about 150 million lines of software code
in automotive embedded systems due to increased
autonomy and added number of ECUs. Because it
is believed that in full-autonomous mode, software
drives the car, so the number of software lines will
increase even more [25]. Due to large number of
ECUs, cost is increasing. Also, due to insufficient
space, these units are difficult to package and place
inside the automotive. Non-standardization of
hardware and software results in multiple different
architectures in each brand.
In 2003, automotive giants agreed to combine
efforts in standardization under one umbrella for
future ECU software. It was named AUTOSTAR
(Automotive Open System Architecture). In
another consortium based in Japan named Japan
Automotive Software Platform Architecture
(JasPar), about 100 car makers and suppliers

joined hand to standardize automotive network
technology, middleware and software architecture
for automotive control system. This initiative in
2004 resulted in development of ISO 26262
guidelines which is a functional safety standard.
These standards emphasize on model based
design, component based and virtual platform for
software development.
It is required to work for reduction in the
number of ECUs in next generation automotive
and this can only be possible if low power, low
cost but reliable and high performance computing
platforms is used in automotive electronics [25].
One solution to this requirement is by using
graphics processing units (GPUs) to address the
real time data handling, processing and
communication. GPUs are high performance
systems where single GPU can be used to replace
all ECUs in the present day automotive
architecture. These GPUs are found suitable
processing units for multiple applications e.g.
general purpose processing, real fast, soft real-time
and hard real time. However, since GPUs are
additional processing units controlled by device
drivers which vary due to different vendors and
driver version, it is important to consider real time
application constraints while using them for
critical applications [26].
XVII. CONCLUSION
The paper describes in detail the development
history of automotive embedded systems with
emphasis on the autonomous driving vehicles and
their impact on society. It has been emphasized
that in order to lower the accident fatalities, the
role of self-driving vehicles is obligatory,
however, in addition to socio-economic benefits to
society, legal and social impact of these systems
must be considered before integration in our
society.
REFERENCES
[1]
C. D. Anderson and J. Anderson, Electric

and hybrid cars: A history: McFarland,
2004.
[2]
M. Bertoluzzo, P. Bolognesi, O. Bruno, G.

Buja, A. Landi, and A. Zuccollo, "Driveby-wire systems for ground vehicles," in
33
Industrial Electronics, 2004 IEEE

International Symposium on, 2004, pp.
711-716.
[3]
[4]
P. Ciaurriz, I. Diaz, and J. J. Gil,

"Bimanual drive-by-wire system with
haptic feedback," in Haptic Audio Visual
Environments and Games (HAVE), 2013
IEEE International Symposium on, 2013,
pp. 18-23.
J. Levinson, J. Askeland, J. Becker, J.
Dolson, D. Held, S. Kammel, et al.,
"Towards fully autonomous driving:
Systems and algorithms," in Intelligent
Vehicles Symposium (IV), 2011 IEEE,
2011, pp. 163-168.
Formula SAE Car," in World Congress,

2014, pp. 8457-8462.
[10]
U. Ozguner, C. Stiller, and K. Redmill,

"Systems for safety and autonomous
behavior in cars: The DARPA Grand
Challenge experience," Proceedings of the
IEEE, vol. 95, pp. 397-412, 2007.
[11]
S. Thrun, M. Montemerlo, H. Dahlkamp,

D. Stavens, A. Aron, J. Diebel, et al.,
"Stanley: The Robot that Won the DARPA
Grand Challenge," Journal of Field
Robotics vol. 23, pp. 661-692, 2006.
[12]
______. (2013, December 5, 2013).

Preliminary
Statement
of
Policy
Concerning
Automated
Vehicles.
Available:
http://www.nhtsa.gov/About+NHTSA/Pre
ss+Releases/U.S.+Department+of+Transp
ortation+Releases+Policy+on+Automated
+Vehicle+Development
[5]
R. Li, C. Liu, and F. Luo, "A design for

automotive CAN bus monitoring system,"
in Vehicle Power and Propulsion
Conference, 2008. VPPC'08. IEEE, 2008,
pp. 1-5.
[6]
S. Reiter, A. Burger, A. Viehl, O.

Bringmann, and W. Rosenstiel, "Virtual
prototyping evaluation framework for
automotive
embedded
systems
engineering," in Proceedings of the 7th
International ICST Conference on
Simulation Tools and Techniques, 2014,
pp. 1-10.
[13]
J. Agajo, V. Idigo, and I. Avazi, "Remote

Control via Can/Ethernet Gateway Using
GPRS,"
International
Journal
of
Computer Theory and Engineering, vol. 3,
p. 296, 2011.
[14]
H. Hodson, "Driverless cars in gridlock,"

New Scientist, vol. 225, pp. 20-21, 2015.
R. Mader, G. Griessnig, E. Armengaud, A.

Leitner, C. Kreiner, Q. Bourrouilh, et al.,
"A bridge from system to software
development for safety-critical automotive
embedded
systems,"
in
Software
Engineering and Advanced Applications
(SEAA), 2012 38th EUROMICRO
Conference on, 2012, pp. 75-79.
[15]
G. A. Elliott, "Realtime scheduling for

GPUs with applications in advanced
automotive systems," University of North
Carolina at Chapel Hill, 2015.
[16]
Z. H. Khan, J.-M. Thiriet, and D. GenonCatalot, "Drive-by-wireless teleoperation

with
network
qos
adaptation,"
International Journal of Engineering
Science and Technology, vol. 2, pp. 160169, 2011.
Z. H. Khan, J. M. Thiriet, and D. GenonCatalot, "Wireless network architecture for

diagnosis and monitoring applications," in
Consumer
Communications
and
Networking Conference, 2009. CCNC
2009. 6th IEEE, 2009, pp. 1-2.
[17]
M. D. Ventures, "Stanley: The robot that

won the DARPA Grand Challenge,"
Journal of field Robotics, vol. 23, pp. 661692, 2006.
J. Kalinowski, T. Drage, and T. Brunl,

"Drive-By-Wire for an Autonomous
[18]
A. Pruckner, R. Stroph, and P. Pfeffer,

"Drive-by-wire,"
in
Handbook
of
[7]
[8]
[9]
34
Intelligent Vehicles, ed: Springer, 2012,

pp. 235-282.
[19]
S. Batchu and S. P. Kumar, "Driver

Drowsiness Detection to Reduce the Major
Road Accidents in Automotive Vehicles,"
2015.
[20]
A. Forrest and M. Konca, "Autonomous

Cars and Society," Worcester Polytechnic
Institute, 2007.
[21]
______, "CAN Specification, version 2,"

ed. Stuttgard: Robert Bosch GmbH, 1991.
[22]
K. H. Johansson, M. Trngren, and L.

Nielsen,
"Vehicle
applications
of
controller area network," in Handbook of
networked and embedded control systems,
ed: Springer, 2005, pp. 741-765.
[23]
A. H. Khan, Z. H. Khan, and Z. Weiguo,

"Model-Based Verification and Validation
of Safety-Critical Embedded Real-Time
Systems: Formation and Tools," in
Embedded and Real Time System
Development: A Software Engineering
Perspective, ed: Springer, 2014, pp. 153183.
[24]
A. Angerd and A. Johansson, "Design and

implementation of a central control unit in
an automotive drive-by-wire system,"
Master thesis. Chalmers University of
Technology, Gothenburg, Sweden, 2013.
[25]
G. Macher, A. Holler, E. Armengaud, and

C. Kreiner, "Automotive embedded
software: migration challenges to multicore computing platforms," in Industrial
Informatics (INDIN), 2015 IEEE 13th
International Conference on, 2015, pp.
1386-1393.
[26]
G. A. Elliott and J. H. Anderson, "Realworld constraints of GPUs in real-time

systems," in 17th IEEE International
Conference on Embedded and Real-Time
Computing Systems and Applications,
2011, pp. 48-54.
35
SAME-05
Optimization of TIG Welding Parameters using full factorial and ANOVA

Design of Experiments Approach
Muhammad Ijaz Afzal1, Shahid Ikramullah Butt1, Mushtaq Khan1 and Syed Hussain Imran Jaffery1
1
Sector H-12, Islamabad, 44000, Pakistan
drshahid@smme.nust.edu.pk
Abstract The TIG welding is most commonly

used welding technique in the world. In this
research work, the optimization of the welding
process input parameters with Full Factorial
approach is applied to steel structures of 5mm
thickness with TIG welding Process. These input
variables/parameters were optimized by using
design of experiments technique in Minitab
software. ANOVA is applied for the optimization of
TIG welding input parameters. In this work, the
tensile strength was maximized using TIG welding
input process parameters current, voltage, speed and
shielding gas. It was found that current and voltage
have major contribution towards the tensile
strength.
Index TermsTIG Welding, Tensile Strength,
Full Factorial, ANOVA, Optimization
I. INTRODUCTION
The TIG welding is complex phenomenon
and a number of factors and variables effect the
output of the process [1]. The four basic variables
involved in any welding process are current,
voltage, speed of welding and shielding gas. In
this research work, the effect of these four input
parameters is studied on the output variable. The
output variable selected is tensile strength of the
welded test piece. The design of experiments
technique is used for the optimization of these

parameters for the maximization of output i.e.
tensile strength. The 5mm thickness High
Strength Low Alloy (HSLA) steel sheet is used in
this experiment. In this work, the Minitab
software package is used for Design of
Experiments.
K. M Eazhil, S.Mahendran, S.Ganesh
Kumar [5] optimized TIG welding on 6063
aluminum alloy using Taguchi method for the
maximization
of
mechanical
properties.
A.Sivasankaran [6] optimized TIG welding for
maximizing weld strength of aluminum 8011 using
Taguchi method. S.Akella, B. Ramesh Kumar [7]
did work for the optimization of weld process
parameters with Taguchi approach for the
transverse distortion control applied to MS
structures of 3 mm thickness with TIG weld
process. Vedprakash Singh, Vijay Patel [8]
discussed about the application of Full-factorial,
ANOVA and Regression analysis to analyze the
effect of process parameters (current, groove angle,
welding speed) on welding distortion for SS304
under GTAW welding process. C Prabaharan, P
Venkatachalam, K Suresh Kumar, K Lenin [9]
optimized the welding parameters in the gas
tungsten arc welding of an Inconel 825 alloy.
36
Enough data must be available before

planning of every experiment, so that science in
any phenomenon can be understood. First trial
and error approach is adopted to design
experiments.
In this approach, some experiments are
performed by trial and error method to get some
understanding of the phenomenon. The input
variables of every experiment are noted and
response to those variables is also noted.
Sometimes these experiments do not give positive
result and we have to conduct other experiments
by changing some input variables. Therefore a
number of experiments are performed to reach to
conclusion and to understand the problem
phenomenon in a better way.
For the optimal welding parameters, a full

factorial model is used in which 4 factors/input
variables, two levels of each factor are included and
in this way it becomes 16 experimentation
parameters. The high and low settings of these four
input variables are shown in table 1. The Minitab
software package is used for Design of Experiments
and all these values are put in the model. The model
designed 16 experiments with different input
variable combinations as shown in Table 2.
Design of experiments is a well-planned and

much better approach than the trial and error
approach to obtain good prediction about the input
parameters. First of all, the factors which are
important for explaining the process variation are
sorted out by previous experience. Then,
understanding is developed as how these factors
will interact in the process to find out the optimal
process performance.
II. FACTORIAL DESIGN OF EXPERIMENTS
The factorial design of experiments is used to
study simultaneous effect of several parameters on a
process.
Varying
different
factor
levels
simultaneously rather than one factor at a time,
saves considerable time and resources in terms of
cost, and at the same time interaction between
different factors can be studies. The screening is
done to reduce the input variables that affect the
quality of product and to focus only on few
important variables. Two level full and fractional
factorial design of experiment approach is used for
screening the important variables/factors that affect
the process response measures. The useful
information is gathered at relatively small runs per
input variable and indication is given about the
main tendencies to guide for further experiments
and optimum settings.
III. RESPONSE VARIABLE

The output variables or parameters from the
process are termed as response variable. Following
response variable is critical and need to be
evaluated with respect to integrity of the HSLA
steel.
37
Tensile Strength: Maximum weld strength measured

on the tensile sample (MPa).
alloying elements of base metal. Table 4 shows the

chemical composition of filler wire selected (H08).
To evaluate the tensile strength, the test pieces were

welded as per input combinations in Table 2. Then,
the tensile samples were prepared and tested to see
the response value in respect of input variables.
IV. FIXED PARAMETERS
Fixed parameters are those parameters which
were kept fixed in this research work. Fixed
parameters are listed below:
A. In this experiment, Flat Horizontal welding
Position was used
B. Geometry of joint was single V with included
angle of 70 degree with 1mm root gap and 1mm
root face.
C. The DCSP current was used in this experimental
work
D. Mixture of Argon and Helium gases was used as
shielding gas in this work [25l/min]
V. MATERIAL SELECTION
Chinese Grade D406A (HSLA steel) material was
used for the experiments in this work. High
Strength Low Alloy steels offer excellent
mechanical properties, and excellent resistance to
atmospheric corrosion. Major applications of HSLA
steels include aerospace, high pressure applications,
aircraft industry, and conventional industrial
applications. The thickness of sheet used in these
experiments was 5mm. The HSLA steel has
excellent formability due to low carbon value.
The composition of the HSLA material is shown in
the table 3
Tensile strength is the most important

parameter in the structural application of HSLA
steel. The tensile strength of the filler wire must be
comparable with that of base metal along with other
Table 5 displays mechanical properties of base

metal.
VI. EXPERIMENTAL PROCEDURE

The sixteen experimental conditions about the
input variables were designed by full factorial
design of experiments. The tensile samples were
prepared according to these input variables.
The fully automatic power supply source (SAF
NERTAMATIC) was used in this experiment for the
preparation of welded tensile samples. The
automatic movement system was used for the
precise movement of welding torch. The clamping
arrangement was also used for clamping and
positioning of the test-pieces for welding. The
welding parameters (current, voltage, welding speed
and shielding gas) were set for each experimental
condition as mentioned in Table 2.
The two tensile samples with each welding input
condition were machined and prepared according to
ASME Section IX. The tensile samples were tested
for tensile strength on universal testing machine.
The average value of the two results was taken for
each experimental condition. Figure 1 shows the
machined tensile samples. Figure 2 displays the
response (tensile strength) of the sixteen
experimental conditions obtained after tensile testing
of samples. The table 6 displays the maximum and
minimum values of tensile strength obtained from
these experiments.
38
Fig. 1 Tensile samples for strength testing as per

ASME IX
The relationship between input variables and

response/output is investigated and modeled using
the analysis of variance (ANOVA). ANOVA was
developed by Ronald Fisher [10] to interpret the
results of agricultural experimentation. The input
welding variables/parameters which can affect the
output/response of the system are checked and
investigated with ANOVA.
The importance of the welding input
variable/parameter is evaluated by percentage
contribution in the total sum of the squared
deviations [2].
The significance of the effect of input process
variable/parameter is decided by the F-test, named
after Fisher [2,3]. When F-test value is more than 4,
then the process input parameters are significant
and play important role in the output response [2].
The
relationship
between
input
parameters/variables and response variable/output is
modeled and investigated by full factorial analysis.
Fig. 2 Measured Tensile Strength of sixteen

Experiments
The ANOVA is used to analyze the tensile

strength and it predicted the observations which are
listed in Table 7. The analysis shows that input
parameters/variables have significant effect on the
output response.
VII. EXPERIMENTAL RESULTS, ANOVA

The tensile strength values obtained from
experiments (shown in Fig. 2) were put in the
Minitab software worksheet next to the
experimental conditions (shown in Table 2)
designed by the software. The effects of input
variables/factors upon the response/output are
discussed with respect to analysis of variance
(ANOVA), regression and optimization applied to
test results of tensile samples.
The F Value of Model (8.74) indicates that it is

significant model. The probability that the F Value
of Model is due to noise is only 0.20%. This noise
probability is calculated automatically by ANOVA.
F-value of model is compared with F-distribution
value obtained from tables based upon confidence
39
level (95% in this case), degree of freedom of

model and residual. The model is significant if FValue of model is greater than F-distribution value.
All the model input variables (terms) are
significant.
In Table 7, the current, voltage, speed and
shielding gas are all model terms. This model is
worthy to navigate the design space.
VIII. RESULTS AND DISCUSSIONS
Fig. 3 Effect of welding current upon strength
The graphs (Fig. 3 to 6) from ANOVA represent
the effects of altering the levels of each
setting/parameter on strength of the weld while three
other parameters are fixed. The strength of the weld
decreases with increase of current because the heat
input increases. The strength of the weld decreases
with increase of voltage because this also increases
the heat-input. However the strength of the weld
increases with increase of speed of welding because
this reduces the heat-input.
Now the numerical optimization for the
maximization of tensile strength is applied. The
optimization indicates that in TIG welding of high
strength low alloy steel, the tensile strength can be
maximized with the following conditions: the low
values of current, low values of voltage, high values
of speed of welding along with the shielding gas.
The Minitab allows for sensitivity analysis of
parameters by varying the input parameters and
evaluates the effect on the optimization plot. The
optimization plot shows the predicted weld strength
(761
MPa)
at
the
optimum
input
variables/parameters 230.00 Ampere current, 10.5
Volt Voltage and 18cm/min speed of welding as can
be seen in the figure 7.
Fig. 4 Effect of welding voltage upon strength
Fig. 5 Effect of welding speed upon strength
40
Fig. 6 Effect of shielding gas upon strength
Fig. 7 Optimization Plot of input parameters
IX. CONCLUSION
the weld tensile strength. Predicted properties at

optimum condition are verified with a
confirmation test as shown in Table 8.
The Full Factorial experimental design for

deciding the optimal welding parameters for the
maximization of weld tensile strength was
successfully applied.
The TIG welding
parameters for HSLA steel were optimized using
design of experiments for the output response
variable (tensile strength) of the welded test piece.
ANOVA was used to determine the level of
influence of the welding input parameters upon
41
REFERENCES
[1] Qureshi, M. Ejaz., 2008. Analysis of Residual
Stresses & Distortions in circumferentially
welded thin walled cylinders. Thesis (PhD).
NUST Islamabad.
[2] H.L.Lin, C.P. Chou. Optimization of GTA
welding process using the Taguchi method
and a neural network. Science and
technology of welding and joining. Volume
11, Issue 1 (01 February 2006), pp. 120-126
Engineering Volume 1, Issue 9, May-2014

ISSN (Online): 2347 4718
[9] C Prabaharan, P Venkatachalam, K Suresh
Kumar, K Lenin. Parametric optimization of
gas tungsten arc welding processes by using
factorial design approach. Journal of Scientific
& Industrial Research, Vol 73, June 2014, pp.
415-420
[10] Fisher RA. Statistical Methods for Research
Workers. Edinburgh, United Kingdom: Oliver
& Boyd; 1925.
[3] H.L.Lin, T. Chou and

C.P. Chou.
optimization of resistance spot welding
process using taguchi method and a neural
network. Experimental techniques. Volume
31, issue 5 (sept 2007), pp 30-36
[4] Dar Naeem ullah, Qureshi M.Ejaz, Hammouda
M.M.I; Analysis of Weld Induced Residual
Stresses and Distortions in Thin Walled
Cylinders; Journal of Mechanical Science and
Technology, KSME, Vol. 23, issue 4 (April,
2009)
[5] K. M Eazhil, S.Mahendran, S.Ganesh Kumar.
Optimization of Tungsten Inert Gas Welding
on 6063 Aluminum Alloy on Taguchi
Method International Journal of Research and
Scientific Innovation. Volume I, Issue III
(August 2014). ISSN 2321 2705
[6] A.Sivasankaran. Taguchi Optimization of
TIG Welding for Maximizing Weld Strength
of Aluminum 8011 International Journal of
Innovative Research in Science, Engineering
and Technology. Vol. 4, Special Issue 6, May
2015. ISSN(Online) : 2319 - 8753
[7] S.Akella, B. Ramesh Kumar. Distortion
Control in TIG Welding Process with Taguchi
Approach. International Journal of Advanced
Materials Manufacturing & Characterization.
Vol 3 Issue 1 (2013)
[8] Vedprakash Singh, Vijay Patel. Experimental
Investigation of GTAW for Austenitic
stainless steel using DOE. International
Journal for Technological Research in
42
SAME-06
Virtual Engineering: A key Enabler for Reconfigurable Manufacturing

Systems
Muhammad Usman Qadir1, Izhar Ul Haq1, Shahzad Anwar1, Kamran Shah1, Ashfaq Khan2
Institute of Mechatronics Engineering, University of Engineering and Technology, Peshawar, Pakistan
Department of Mechanical Engineering, University of Engineering and Technology, Peshawar, Pakistan

usmanqadir91@gmail.com, izhar@uetpeshawar.edu.pk
Abstract Due to global competition the product

lifecycle has reduced significantly, forcing
manufacturers
to
introduce
a
new
manufacturing paradigm for production systems.
Such production systems are easily upgradable
with new technologies in order to quickly launch
new products with high variety, variable capacity
and reduced lead time. To stay in todays
competitive
market,
the
concept
of
Reconfigurable Manufacturing Systems (RMS)
has evolved. These systems allow reacting
continuously to unpredictable changes in market
demands, processing technology, product variety
and allow quick adjustment of production
capacity and functionality by providing
configurable
or
reconfigurable
design
approaches. One of the key issues associated
with RMS is their design process in which there
is no provision for verification/validation of
manufacturing system during planning phase
resulting in increased lead time. Also, the
reconfiguration period for RMS (when switching
from one product to another) is significantly
high, often resulting in costly delays. This paper
address the above mentioned issue by proposing
virtual engineering as a key enabler to reduce
lead time and reconfiguration period for RMS.
By integrating virtual engineering with RMS the
time to market can be decreased and the
responsiveness of the manufacturing system can
be increased considerably.
Index
Terms
Reconfigurability,
Manufacturing Systems, Virtual Engineering,
Virtual Commissioning, Virtual Universe Pro
I. INTRODUCTION
Manufacturers for surviving in this global
world cannot afford to focus on design and
manufacturing of single product, instead
manufacturing technology today have to focus on
producing product families according to customer
demands in variable capacities [1]. The required
manufacturing systems can be described by
aggressive competitions on global scale driven by
rapid changes in process technology. High quality
variable capacity and responsiveness are the
objectives of todays market [2]. Therefore, to
sustain competitiveness in this dynamic market,
manufacturers are required to design production
systems with sufficient flexibility to produce
variety of products (within a product family)
without compromising on the quality [3]. In
addition these systems have to be easily
upgradable with the ability to absorb new
technologies, allowing manufacturers to quickly
launch new products with higher production rates
and rapid adjustment of production capacity [4].
Thus it is necessary to introduce new
manufacturing paradigms capable of responding to
fluctuating market demands. On the basis of
flexibility, production rate, process technology and
economic objectives, different types of production
43
systems have been developed which are dedicated

manufacturing
systems
(DMS),
flexible
manufacturing systems (FMS) and reconfigurable
manufacturing systems (RMS) [5].
DMS are generally equipped with fixed
automation and tooling. The economic objective of
such systems are to produce a specific part (or
product) in high capacity with no (or very limited)
variety and desired quality. These systems are
profitable and cost effective in stable market when
the demand of the product exceeds its supply. But
they are unable to respond to the dynamic market
scenarios. FMS on the other hand are designed to
produce variety of products in predictable market
conditions. These systems (FMS) are generally
equipped with fixed hardware and programmable
software. The economic objective of FMS is to
produce different kinds of parts within a specific
part family with reduced changeover time at
desired quality. However, they have achieved very
limited success because of high equipment cost
(since in most of the cases they include more
functionality than needed), low throughput and
complex design [6]
In order to address the limitations of DMS and
FMS by responding rapidly to variation in market
demands Koran et al. [7] proposed a new paradigm
of manufacturing system called as Reconfigurable
Manufacturing Systems (RMS). RMS allows not
only the production of variety of parts but offers
customized flexibility. Functionality in RMS can
be increased or decreased by adding or removing
hardware/software modules and new technologies
can be easily integrated into these systems,
therefore they are not subjected to obsolescence.
The comparison of DMS, FMS and RMS in
capacity-functionality coordinates is shown in fig.
1. It is evident that DMS have high capacity but
limited functionality. FMS have high but fixed
functionality with low capacity. While RMS lie
between DMS and FMS. The capacity as well as
functionality in RMS can adjusted (increased or
decreased) in response to fluctuating market
demands. Key features of RMS distinguishing it
from DMS and FMS includes modulatory,
integrability,
convertibility,
flexibility,
diagnosability and customization.
RMS have the ability to address the dynamic
market requirement of accommodating changes in
6 to 9 months (as product lifecycle has shortened).
These manufacturing systems can be upgraded to

introduce a variety of products, thus increasing the
lifecycle of manufacturing system.
Fig. 1. Mapping of manufacturing systems in

capacity-functionality coordinates
In literature most of the research for RMS is
related to justification for RMS [6], scheduling
architecture [1], design strategies [2], layout
selection & optimization [8] and capacity
management [9]. However, little efforts are
reported in the literature to make virtual
engineering as a key enabler for RMS. Also the
time constraints for RMS while switching from
one product to another is a key issue as well. The
rest of the paper is organized as following. Section
II describes the traditional engineering approach to
RMS and presents problem statement. In section
III a novel approach based on virtual engineering
is discussed while conclusion and future work is
presented in section IV.
II. TRADITIONAL ENGINEERING APPROACH TO
MANUFACTURING SYSTEMS
The design and build of RMS is one of key
areas in the product lifecycle. Design and build of
these systems involve geographically distributed
engineering teams of end-user (i.e. production
companies) and its supply chain partners (i.e.
machine builders and control vendors) [10].
Machine builders, also known as Original
Equipment Manufacturers (OEMs), are the Tier-1
suppliers to the end-users. They are responsible for
mechanical design, electrical design, control
design, building, testing, and installation of
production machines for the end-users. Machine
builder may also sub-contract certain parts of a
machine to various component builder or system
designers. The control vendors are the Tier-2
44
suppliers and fulfil control hardware/software

requirements for the end-user [11]. Collectively
all these partners are responsible for the
implementation and lifecycle support of the
automated systems as product and production
requirements changes [10].
Due to short product life cycle the requirements
of the dynamic market changes frequently. All
new business requirements (e.g. design and
manufacture of a new powertrains) are planned by
end users planning and business office in the
form of new business case for development of
new RMS or upgradation of existing one. In
response to new business case, the end users
planning and feasibility team contacts supply
chain partners to start simultaneous engineering
(SE) with end user domain experts. The product
design and process requirements are provided by
end users expert to machine builder and
component suppliers. The final design of RMS is
completed by machine builders subjected to the
approval of end users engineering teams. The
design activities of machine builders than take
place sequentially, beginning with mechanical
engineering followed by electrical and control
engineering. As a result the design activities for
machine hardware and control system remains
isolated from one another and overall verification
can only be carried out during final commissioning
after physical build. Figure 2 presents the
conventional design process for RMS.
The engineering process described above is typical
for most of the manufacturing industries and in
particular for automotive industries. However, this
mentioned approach has fundamental limitations.
These limitations are mostly caused by study and
SE phase. Currently, the impact of these
limitations is realized after installation,
commissioning and launch phase. During planning
and SE phase the designed processes are entirely
based on the expertise and knowledge of end
users engineering team. As a result inadequate
planning may be performed by domain experts
during this phase. This is due to fact that that there
is no common representation or visualization of
RMS throughout the lifecycle. In addition, within
the existing approach only 20% to 30% testing and
validation is achievable during planning phase,
thus typically leaving 70% to 80% testing and
validation to be carried out during commissioning

phase. As a result the planned production ramp-up
is seldom achieved in time and typically results in
increased lead time and costly delays. Another
weakness in the existing approach is that it does
not enable supply chain partners, in particular the
end user to initially design, build and reconfigure
the manufacturing system based on previous
system commonalities. This is despite the fact that
every new project contains about 70%
commonality with previous projects. Also, in the
existing approach, when the market demand
changes the system is reconfigured accordingly.
The existing reconfiguration process for a
manufacturing system, when switching from
product A to B is shown in fig. 3. During
reconfiguration the manufacturing system is
required to be shut down for significantly high
period due to lack of virtual integration between
mechanical and control system design. This
ultimately results in costly delays within
production ramp-up period. In order to address the
current mentioned limitations, a new engineering
approach based on integration of virtual
engineering with RMS is proposed focusing on
component based automation. The aim of the
proposed approach is to visualize the design and
verify RMS before actual build to reduce lead time
and to achieve the planned production ramp-up
time. In addition this research also addresses the
issue of reconfiguration when switching from one
product to another and enables end user domain
experts to design, build, reconfigure and verify the
manufacturing system based on previous
commonalities without physical build. This paper
also proposes virtual integration between
mechanical hardware and control system of RMS
during reconfiguration. This can enable
verification of reconfigured mechanical layout and
control system virtually thus reducing physical
shutdown period of RMS resulting in economic
benefits.
45
existing and the proposed approach. Figure 4

shows major transformations/migration paths from
existing practice to a more integrated and
concurrent engineering environment. This research
also proposes to develop a virtual library of
reusable, pre-defined and pre-validated standard
mechanism, which can be easily and quickly
reconfigured for any new process requirements.
Fig. 2. Traditional Design Process for RMS.
Fig. 3. Existing Approch for Reconfiguration in

RMS.
III. PROPOSED APPROACH TO MANUFACTURING
SYSTEMS
The advent of globalization is rapidly changing
business plans within all manufacturing sectors.
To deliver agility through modularity and
reconfigurability for future RMS, this research
work has proposed and developed a new
engineering approach. Primarily this approach is to
provide 1) ease of reconfiguration, 2) visualization
and 3) lifecycle support. The application domain
of the proposed approach offers building RMS and
its automation systems with the use of new
engineering tools from generic library of modules.
Such engineering tools can potentially enable 100
% virtual design and build of RMS and its
validation and verification prior to actual build.
This is one of the key differences between the
Fundamentally, all new business requirements

can be easily and quickly managed by virtually
defining, designing and building new production
automation systems during early lifecycle phases.
In response, the domain experts of end user can
virtually design and build manufacturing system,
thus enabling them to visualize the system to be
designed. As shown in fig. 4, during virtual design
and build of manufacturing system the mechanical,
electrical and control system phases can be
performed in parallel. The shifting of these design
activities from physical build phase to early design
phase allow end user engineering teams to validate
their system without actual build, thus enabling
them to decide whether proposed system will meet
production targets. In addition the virtual model
enables domain experts to optimize manufacturing
system requirements in terms of their cycle time,
kinematics and control behavior. The virtual
modelling is also capable of virtually validating
and verifying design of manufacturing system.
Additionally, during virtual modelling the end user
domain expert team can model the complete set of
operations by allowing workpiece to ultimately
produce virtual product. Key difference between
existing and proposed approach is at study and SE
phase. During this stage all the mechanical,
electrical and control design will be completed and
nearly all the elements, components and subsystem requirements will be verified. Whereas in
the existing approach all the major design
activities are carried out during physical build
stage. The visualization and verification of RMS
virtually improve the robustness of manufacturing
system due to early identification and fixation of
issues related to mechanical, electrical and control
system design which are currently realized and
addressed during installation and commissioning
phase. As a result significant time saving can be
achieved during physical build, installation and
commissioning of the RMS. In result, planned
production ramp-up can be achieved on time.
46
This paper also proposes virtual integration

between machine hardware and control system of
manufacturing system during reconfiguration. This
can enable verification of reconfigured mechanical
layout and control system virtually thus reducing
physical shutdown period of RMS. In the existing
approach, during mechanical reconfiguration the
manufacturing system is physically shutdown. The
mechanical layout is than changed, optimized and
validated. According to the most optimum layout
control system is designed. The control system is
than integrated with mechanical hardware and
further its testing is performed. This
reconfiguration period is significantly high and
most of the time results in costly delays. Through
integration
of
virtual
engineering,
the
reconfiguration period can be reduced.
Fig. 4. Proposed design process for RMS.

The proposed approach for reconfiguration of
RMS is presented in fig. 5. Initially, the
manufacturing system is designed and build for the
production of product A. At later stages market
demand changes gradually from product A to
product B. The manufacturer can then redesign
the manufacturing system through reconfiguration
to produce product B. This reconfiguration is
performed virtually to design and optimize
machine layout along with development and
verification of control logic. During this virtual
reconfiguration the manufacturing system can
continue to produce product A resulting in
economic benefits to manufacturer. Once the
mechanical and control system design is verified
virtually, the manufacturing system can be shut

down for a shorter period and reconfigured
quickly.
Fig. 5. Proposed reconfiguration process for

RMS.
To support the proposed approach, software
Virtual Universe Pro [12] is used in this research
due to its various functionalities provided for
virtual engineering of the RMS. The software
supports virtual engineering to a great extent and
consists of number of modules. Mechanical
models can be designed based on CAD drawings.
The control logic for these mechanical models can
be designed and tested. A pre-defined library of
validated manufacturing mechanisms can be
developed. New mechanisms can also be designed
and added to the library. The control logic can be
designed in ladder logic, sequential flow charts
(SFC) and functional block diagrams (FBD).
Virtual universe pro also offers integration with
number of PLCs including Siemens S7-1200 and
S7-200 etc. in order to integrate machine hardware
with virtually built models. In addition, using this
application control behaviors can also be
generated.
IV. CONCLUSION AND FUTURE WORK
Rapidly changing global business plan are
norms within all manufacturing sector. To
facilitate and accommodate unforeseen business
changes within manufacturing industry, the
concept of RMS has evolved. In this paper, virtual
engineering is identified as one of the key enabling
technologies for future manufacturing systems
(RMS). In response, a new proactive engineering
approach has been introduced to quickly design,
build and reconfigure RMS. Most prominent
aspect of the proposed approach is to better
provide lifecycle support and to reduce lead time
by strengthening the virtual engineering approach.
47
The paper also addresses the issue of high

reconfiguration period of RMS. Further, the
proposed approach will be tested on industrial
testbed in order to demonstrate key benefits of the
virtual engineering to industry and academia.
1.
2.
3.
V. REFERENCES
D. H. Liles and B.L. Huff, "A computer
based production scheduling architecture
suitable for driving a reconfigurable
manufacturing system," Computers &
Industrial Engineering, 1990. 19(1): p. 1-5.
M. R. Abdi and A.W. Labib, "A design
strategy for reconfigurable manufacturing
systems
(RMSs)
using
analytical
hierarchical process (AHP): a case study",
International Journal of Production
Research, 2003. 41(10): p. 2273-2299.
S. E. Chick, T. L. Olsen, K. Sethuraman,
K. E. Stecke and C. C. White, "A
descriptive multi-attribute model for
reconfigurable machining system selection
examining buyer-supplier relationships,"
International
Journal
of
Agile
Management Systems, 2000. 2(1): p. 3348.
4.
Y. Koren, "Reconfigurable manufacturing

systems: Key to future manufacturing. in
Journal of Intelligent Manufacturing,"
2010. Citeseer.
5.
Y. Koren, et. al., "Reconfigurable

manufacturing systems," CIRP AnnalsManufacturing Technology, 1999. 48(2):
p. 527-540.
6.
R. Singh, N. Khilwani, and M. Tiwari,

"Justification for the selection of a
reconfigurable manufacturing system: a
fuzzy
analytical
hierarchy
based
approach," International Journal of
Production Research, 2007. 45(14): p.
3165-3190.
7.
Y. Koren, Z. J. Pasek, A. G. Ulsoy and U.

Benchetrit, "Real-time open control
architectures and system performance,"
CIRP Annals-Manufacturing Technology,
1996. 45(1): p. 377-380.
8.
M. R. Abdi, "Selection of a layout

configuration
for
reconfigurable
manufacturing systems using the AHP,"
ISAHP, Honolulu, Hawaii, 2005.
9.
A.
M.
Deifand
W.
ElMaraghy,
"Investigating optimal capacity scalability
scheduling
in
a
reconfigurable
manufacturing system," The International
Journal of Advanced Manufacturing
Technology, 2007. 32(5-6): p. 557-562.
10.
A. Colombo, et al. "Service-oriented

architectures for collaborative automation.
in Industrial Electronics Society, 2005,"
IECON 2005. 31st Annual Conference of
IEEE. 2005. IEEE.
11.
M. H. Ong, "Evaluating the impact of

adopting a component-based approach
within the Automotive Domain," 2004,
Loughborough University.
12.
Virtual Universe Pro,

http://www.iraifrance.com/#!virtualuniverse-pro/c8wn
(accessed
Sept.
30,2015).
48
SAME-07
Simulation of Turbulent Flow over a Single and Tandem NACA 0012 Airfoils
Mudassar Hussain1 and Emad Uddin1
1
Sector H-12, Islamabad, 44000, Pakistan
emaduddin@smme.nust.edu.pk
AbstractNACA 0012 airfoil has been used as

standard airfoil for a long time. It is used in a lot of
applications
ranging
from
aircrafts
to
turbomachinery. The incentive of this report is to
investigate the flow past a single and two in-tandem
NACA 0012 airfoils at high Reynold number. The
drag and lift coefficient have been obtained using
ANSYS Fluent. The results have been validated by
comparing them with experimental work. The
results obtained by simulation and experiment are
in conformity with each other.
Index TermsSimulation,
NACA 0012, Airfoils, Tandem
Turbulent
Flow,
I. INTRODUCTION
ing design is a critical process. When a body moves
Win air, it experiences a resultant force that tends to
resist its motion. However, there is a certain class of
body for which lift is significantly greater than the
drag. This type of body is quite suitable for use as a
wing structure for the flight of an aircraft[1]. When an
aircraft moves through air, the air moves above and
below the wing of the aircraft. The upper surface of the
wing is designed in such a manner that the air rushing
over the top speeds up and stretches out. From
Bernoullis principle we know that, where speed is
high pressure is low and vice versa. Since the air above
the wing is moving with a high velocity so it creates a
low pressure region above the wing. In a similar
manner a high pressure region is created below the
wing. Since the air moves from a region of high
pressure to a region of low pressure, therefore the air
beneath the surface tends to lift the wing. The more the
velocity, the more the lift.
II. LITERATURE REVIEW

Flow over an airfoil is a typical case of Fluid
Structure Interaction (FSI). Peskin used the term
Immersed boundary methods for simulating heart
mechanics. Peskin [2] extended the previous work by
solving Navier-Stokes equation for dynamic
meshing/moving grid problems.
Many schemes have been presented for simulating
flows in complex geometries. J. Mohd-Yusof [3] came
up with a discrete time immersed boundary method for
simulation of flow in complex geometries. The scheme
proposed by him was tested on the laminar ribbed
channel. E.A. Fadlun et al. [4] developed a efficient
technique for simulating unsteady three-dimensional
flows in complex sceenarios.
Peskin [5] formulated the mathematical structure of
the immersed boundary method in order to simulate
FSI especially in biological fluid dynamics. With the
enhancement and improvements in Immersed
boundary methods, Gianluca Iaccarino and Roberto
Verzicco [6] came up with the application of
Immersed Boundary (IB) method to simulate
incompressible, turbulent flows around complex
configurations. They used standard RANS or the more
sophisticated LES approach for turbulence modeling.
III. PROBLEM SPECIFICATION
The incentive of this research was to investigate the
flow past a single and Two-in-Tandem NACA 0012
Airfoils. First, the flow around a single NACA 0012
airfoil was considered and then the flow around twoin-tandem airfoils was investigated. The freestream
velocity was calculated to be 43.821 corresponding to
a Reynold number of 3 106. The operating
parameters are summarized in TABLE IV.
49
In order to have a better control of the mesh point

distribution, the top and bottom edge of the airfoil
were split into two halves. This was done in order to
have a non-uniform grid spacing for x < 0.3 and
uniform spacing for x > 0.3. There are four boundaries
in this problem. Farfield 1 and 2 are the velocity-inlet.
Farfield 3 is the pressure outlet, while airfoil is
assumed to have the wall type boundary. The mesh is
shown in
Fig. 1.
TABLE IV
OPERATING PARAMETERS
Input
Value
Velocity of flow
43.821 m/s
Operating temperature
300 K
Operating pressure
101325 Pa
Model
Spalart Allmaras (1
equation)
Density of fluid
1.225 kg/m3
Kinematic viscosity
1.4607 10-5
Reynolds number
3 106
Length
1m
Angle of Attack
-8,-4,0,6,12,17
Fluid
Air
V. METHODS
In order to discretize the governing equations for fluid
flow various methods have been developed. Since
ANSYS Fluent was used to carry out the simulation of
the airfoils, so finite volume method is used for the
solution of the problem. A simple scheme has been
used for the pressure-velocity coupling. In finite
volume method, the property values are stored at nodal
points, so an interpolation is required to establish
values at faces. In case of our problem, second order
upwind scheme for momentum and modified turbulent
viscosity have been used. Green-Gauss cell based
gradient scheme has been used. Pressure staggering
option has been used as a pressure interpolation
scheme.
VI. RESULTS
After the simulation results were compiled and
analyzed. A summary of these results is presented
below:
A. Contours of Pressure Coefficient
The contours of pressure coefficient for various angles
of attack are presented below:
IV. GRID GENERATION
Fig. 2. Contours of pressure coefficient for -8 angle of

attack
Fig. 1. Mesh showing the four boundaries. Farfield 1

and 2 are velocity-inlet type boundary and farfield 3 is
pressure-outlet. Airfoil is having wall type boundary.
From the above contour, we can see that there are two
regions. The region at the leading edge is a high
pressure region while a low pressure region on the
upper surface of airfoil. From Bernoulli equation, we
know that where speed is high pressure is low and vice
versa. A similar behavior is observed in case of
contours for other angles of attack. These contours are
shown below.
50
Fig. 3. Contours of pressure coefficient for -4 angle of

attack
Fig. 4. Contours of pressure coefficient for 0 angle of

attack
Fig. 6. Contours of pressure coefficient for 12 angle

of attack
Fig. 7. Contours of pressure coefficient for 17 angle

of attack
From the above contours, we can see that the pressure
at the leading edge increases with an increase in angle
of attack. At higher angle of attack, eddy formation
and vortex shedding phenomena becomes significant
which will be evident from the contours of velocity.
B. Contours of Velocity
The contours of velocity for various angles of attack
are presented below:
Fig. 5. Contours of pressure coefficient for 6 angle of

attack
51
Fig. 11. Contours of velocity for 6 angle of attack

Fig. 8. Contours of velocity for -8 angle of attack
If we compare the above contour to that of pressure
contour in
Fig. 2, we can see that we have a higher velocity on
the lower surface of airfoil. This is in accordance with
Bernoullis principle which states that, where speed is
high pressure is low. A similar behavior is observed at
various other angles of attack.
Fig. 9. Contours of velocity for -4 angle of attack
By comparing the above contours, we can clearly see

that the effects of turbulence at higher angles of attack
become significant. The vortex shedding phenomenon
is significant at higher angles of attack. The
comparison between contours of velocity for 12 and
17 angle of attack clearly indicates that with increase
in angle of attack, flow separation starts occurring at a
very earlier point. In case of 12 angle of attack flow
separation starts to occur from the centre of airfoil
while in case of 17 angle of attack it starts to occur at
the leading edge of the airfoil. The formation of
vortices at the trailing edge of airfoil in case of 17
angle of attack is also clear from the contours of
velocity.

52
It can see that at around 400000 cells, the problem is

grid independent.
D. Variation of Lift and Drag Coefficient with Angle
of Attack
The contours of pressure coefficient and velocity have

been discussed before. With the help of simulation
variation of lift and drag for various angles of attack
was seen. The results are shown below. Fig. 16 shows
the variation of lift coefficient with angle of attack
while shows the variation of drag coefficient with
angle of attack.
C. Grid Independence
In order to identify the number of cells at which the
problem is independent of grid, case was simulated for
various number of cells. The variation of lift and drag
with number of cells is shown in Fig. 14 and Fig. 15
respectively.
Fig. 16.Graph showing the variation of lift coefficient

with angle of attack. We can see that the lift coefficient
increases with increase in angle of attack.
Fig. 14. Graph showing the variation of Lift coefficient

with number of cells. We can see that at around
400000 cells, the problem is grid independent.
Fig. 17. Graph showing the variation of drag

coefficient with angle of attack.
It can see that the drag coefficient shows a parabolic
variation. For negative angle of attack the drag
decreases while for positive angle of attack it
increases.
VII. VALIDATION
Fig. 15. Graph showing the variation of Drag

coefficient with number of cells.
The results were validated with the experimental data

of Abbott et al [7] and Charles L. Ladson [8]. The
results obtained using simulations were in conformity
with the results of experimental data. Fig. 18 and Fig.
19 show the comparison between the results of
simulation and experimental data.
53
Fig. 20 show the variation of lift coefficient on two

airfoils by varying the distance between them. From
the figure we can see that at small distance the lift is
higher and then it decreases. This is because the
vortices or eddies that are generated lose their energy
as they shed from one airfoil to the second airfoil. Fig.
21 show the variation of lift coefficient on two airfoils
by varying the distance between them.
Fig. 18. Comparison between turbulent models

simulation results and experimental data from Abbott
et al and Charles L. Ladson of the Lift coefficient
curve for NACA 0012 airfoil
Fig. 21. Graph showing the variation of drag

coefficient by changing the distance between two
airfoils.
Fig. 19. Comparison between turbulent models
simulation results and experimental data from Abbott
et al and Charles L. Ladson of the Drag coefficient
curve for NACA 0012 airfoil
The contours of velocity for various distances are

shown in Fig. 22-25.
VIII. TWO-IN-TANDEM NACA 0012 AIRFOILS

After the simulation of flow for single airfoil, the
variation of lift and drag coefficient for two-in-tandem
NACA 0012 airfoils was investigated. The variation
was seen by varying the distance between two airfoils.
Fig. 22. Contours of velocity for distance of 0.01 m.
Fig. 20. Graph showing the variation of lift coefficient

by changing the distance between two airfoils.
54
through simulations have been validated with the

experimental works. The simulation results and
experimental results are in conformity with each other.
REFERENCES
Fig. 23. Contours of velocity for distance of 0.5 m.
Fig. 24. Contours of velocity for distance of 1 m.
[1]
H. Glauert, The Elements of Aerofoil and

Airscrew Theory: Cambridge University Press,
1983.
[2]
C. S. Peskin, "Numerical analysis of blood

flow in the heart," Journal of Computational
Physics, vol. 25, pp. 220-252, 11// 1977.
[3]
J. Mohd-Yusof, "Combined immersedboundary/B-spline methods for simulations of

flow in complex geometries," 1997.
[4]
E. Fadlun, R. Verzicco, P. Orlandi, and J.

Mohd-Yusof, "Combined immersed-boundary
finite-difference
methods
for
threedimensional complex flow simulations,"
Journal of Computational Physics, vol. 161,
pp. 35-60, 2000.
[5]
C. S. Peskin, "The immersed boundary

method," Acta Numerica, vol. 11, pp. 479-517,
2002.
[6]
G. Iaccarino and R. Verzicco, "Immersed

boundary technique for turbulent flow
simulations," Applied Mechanics Reviews, vol.
56, pp. 331-347, 2003.
[7]
I. H. Abbott and A. E. V. Doenhoff, Theory of

Wing Sections: Including a Summary of Airfoil
Data New York: Dover Publications, Inc.,
1959.
[8]
C. L. Ladson, "Effects of independent

variation of Mach and Reynolds numbers on
the low-speed aerodynamic characteristics of
the NACA 0012 airfoil section," 1988.
Fig. 25. Contours of velocity for distance of 2 m.

IX. CONCLUSION
This paper investigates the flow past a single and twoin-tandem airfoils. We have seen that the lift
coefficient increases with the increase in angle of
attack while the drag coefficient varies in a parabolic
manner. The effects of vortex shedding are significant
at high angle of attack. The stall angle has been found
to be 17 in case of single airfoil. We have also seen
that the lift and drag decrease with the increase in
distance between two airfoils. The results obtained
55
SAME-08
Passively flapping energy harvesting eel in a viscous flow

1
Emad Uddin , Muhammad Saif Ullah Khalid , Aamir Mubashar , Hyung Jin Sung
1 School of Mechanical and Manufacturing Engineering (SMME), National University of Sciences and Technology
(NUST), Islamabad, Pakistan.
2
Department of Mechanical Engineering, KAIST 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Korea
emaduddin@smme.nust.edu.pk
AbstractInspired by the energy harvesting eel, a

flexible flag behind a D-shape half cylinder in a
uniform viscous flow was simulated by using the
immersed boundary method. The downstream flag
was strongly influenced by the vortices shed from
the upstream cylinder during vortex-vortex and
vortex-body interactions. The flexible flag was
subjected to passive flapping to optimize geometric
and flow parameters, which were subsequently
used in an actively flapping flag. In this study,
dependence of bending energy of the passively
flapping flexible flag on various parameters, i.e.,
length of the flag, diameter of the D-shape half
cylinder, streamwise gap between the cylinder and
the flag, and bending coefficient of the flexible flag
was examined in detail. Constructive and
destructive vortex interaction modes and,
unidirectional and bidirectional bending were
found that explained the variations in the bending
energy on the downstream flag.
KeywordsImmersed boundary method, fluid
structure interaction, energy harvesting eel, immersed
boundary method
1. INTRODUCTION
Limited availability and the environmental impact
of fossil fuels motivate the development of renewable
energy sources. Considerable efforts are focused on
use of renewable energy from natural resources such
as flowing water, rain, tides, wind, sunlight,
geothermal heat and biomass. From a hydrodynamic
point of view, experimental evidence has supported the

assertion that an appropriate synchronization of
positioning a flag in the wake of a bluff body can be
advantageous by increasing the hydrodynamic
resistance due to fluid-structure interaction, thereby,
introducing significant energy harvesting [1-5].
Nevertheless, the phenomenon of the hydrodynamic
interactions of a flag in the wake of a bluff body has
not yet well understood, and a more detailed
understanding of the mechanisms would be desirable.
Interaction of vortices and a flexible flag in wake
of a bluff body is a common occurrence in various
natural phenomena, e.g. stationing of a fish behind a
sea-plants, and sea-stones etc. The application of
flapping wings to extract energy from uniform flows
was first proposed by McKinney and De Laurier [6].
Both experiments and theoretical investigations
discovered that a foil submerged in the free surface
could be propelled by using the energy from the oncoming waves [7-10]. Allen et al. [1] examined the
response of a piezoelectric membrane to vortex
shedding behind a bluff body for harvesting flowenergy. Tang et al. [11] analytically studied the
response of a flutter-mill for electricity generation.
These
bio-inspired
energy
converters
are
environmentally friendly in terms of noise generation
thus reducing impact on the navigation of aquatic
animals. These oscillatory devices are not only
structurally robust, but also they sweep a rectangular
cross section of flow. The swept area for a single
device can thus be wide and shallow, allowing large
systems to be installed in shallow water. Subsequently,
56
multi-mega watt devices can be envisaging for a wider

range of the resource areas.
The current study provides details on fluid dynamics
around an eel, modeled as a flexible flag in the wake
of a D-shaped half cylinder subjected to twodimensional uniform viscous flow. There are a few
earlier analytical and numerical studies on the energy
harvesting eel but the detailed investigation and
optimization of various flow and geometric parameters
is the missing link [12]. This study provides necessary
insight for better understanding of the dynamics of a
flapping flexible flag in the wake of the D-shaped half
cylinder subjected to viscous flow and the optimization
of the parameters for energy harvesting eel for the
passive mode of flapping. Dependence of bending
energy on the streamwise gap between a D-shaped half
cylinder and the downstream flag, the flag-length and
the bending rigidity of the downstream flag was
examined in detail.
2. PROBLEMFORMULATION
A flexible flag in the wake of a D-shaped half cylinder
was subjected to two-dimensional viscous flow. A
schematic of the geometric configuration and the
coordinate system is shown in Fig. 1. Distance
between the D-shaped half cylinder and the flexible
flag was varied by adjusting the streamwise gap (S)
between these two bodies. The head of the downstream
flag was fixed under a simply supported boundary
condition. The boundary condition of a free end was
considered at the tail of the flag. Center of the rigid Dshaped half cylinder is placed at the origin of the
coordinate system. Here, its diameter is denoted as D.
An Eulerian coordinate system was used for the fluid
domain, while a separate Lagrangian coordinate
system was employed to monitor the motion of the
flexible flag. The flexible flag in a viscous flow was
modeled through an improved immersed boundary
method [13], in which the governing equations of the
fluid flow and the flexible flag were solved in each
coordinate system, and the interactions among
components was calculated using a feedback law. The
fluid motion was governed by the incompressible
NavierStokes and continuity equations,
where u is the velocity vector, p is the pressure, f is the

momentum force applied to enforce the no-slip
conditions along the immersed boundary, and the
Reynolds number (Re) is defined as Re = VL/,
where , V, L, and represent the fluid density, freestream velocity, the flag length, and the dynamic
viscosity of the fluid, respectively. The flag motion
was governed by
where u is the velocity vector, p is the pressure, f is the

momentum force applied to enforce the no-slip
conditions along the immersed boundary, and the
Reynolds number (Re) is defined as Re = VL/,
where , V, L, and represent the fluid density, freestream velocity, the flag length, and the dynamic
viscosity of the fluid, respectively. The flag motion
was governed by
where s denotes the arc length, X = X (s,t) is the

position, T is the tension force, shows the bending
rigidity, and F denotes the Lagrangian force exerted on
the flexible flag by the fluid. Equation (3) was nondimensionalized by the flag density, the flag length L,
and the free stream velocity. It should be pointed out
that the tension force can be determined from the
inextensibility condition of flag [13]. The bending
rigidity is defined as EI, with E being the Youngs
modulus and I the second moment of area, which gives
the non-dimensional value . The boundary conditions
applied at the fixed end and the free end were
57
3. RESULTS AND DISCUSSION

A schematic of the problem setup and the coordinate
system for the downstream flag in the wake of the Dshaped half cylinder subjected to a uniform flow is
provided in Fig. 1.
where and are large negative free constants [13].

Equation (8) provides an interpolation of the Eulerian
fluid velocity at the Lagrangian points. In physics, (6)
(8) represent a stiff spring system. This system
connected the Lagrangian points on the immersed
boundary to the surrounding fluid particles. After
obtaining the Lagrangian force F, the expression was
transformed to the Eulerian form using the smoothed
Dirac delta function
where =1 is based on the nondimensionalization steps, and denotes the structure

region. After discretization, the force was applied over
a width of several grids, which supported the smoothed
delta function. The computational domain for fluid
flow ranges from -2 to 6 in the streamwise (x)
direction and from -4 to 4 in the spanwise (y) direction.
Positions along both of these directions were
normalized by the flag-length. The Eulerian grid size
for the fluid was 512350 in the streamwise and
spanwise directions, respectively, and the Lagrangian
grid size for each flag was 64. The Eulerian grid was
uniformly distributed along the x- direction, and it was
uniform in the y direction for 22, but stretched
otherwise. A far-field boundary condition was applied
at the top and the bottom boundaries as well as at the
inlet of the fluid domain, whereas the convective
boundary condition was used at the outlet. The
computational time step was set to 0.0002, which
resulted in a CFL number of 0.1. Validation of the
flow solver employing the immersed boundary
method, as well as the structure solver for the flag
motion, has been provided in Ref. [13].
The computational configurations were set to be the

same as described in the previous section. A
quantitative characterization of the resultant
parameters was supplemented by qualitative
evaluations of the flag deformations and the vorticity
contours within the flow that provides an intuitive
understanding of the vortexflexible body interactions.
Three parameters were chosen for optimization within
the following ranges: the streamwise gap between the
D-shaped half cylinder and the downstream flag
(12), the bending coefficient of the flag
(0.020.08),
and
length
of
the
flag
(1.22.0).
Bending energy () of the downstream flag as a

function of the length and the bending coefficient of
the flag is plotted in Fig. 2a. For a fixed streamwise
gap of = 1.6, contours of the bending energy
clearly marks two regions; one representing the
maximum bending energy in the range of
58
1.51.9 for 0.040.075, and the second

region showing that low bending energy levels are
constituted by rest of the ranges of , and . The
bending energy in the high value region is four times
larger than the low value region.
Details of the variation in bending energy with the

change in the flag-length and the bending rigidity can
be unfolded by analyzing the evolution of the tail
position and the flag deformation shapes (Fig. 3), the
power spectrum (Fig. 4) and the vorticity contours
(Fig. 5). Four points corresponding to the low and high
bending energies are shown for different and
.
Fig. 3a and 3b show the tail position evolution and flag
deformation shapes with increase in the bending
rigidity from 0.020 to 0.060 for the same L/D ratio.
The tail position evolution shows the increase in the
flapping frequency (seven peaks in 30 seconds as
compared to six peaks) with the increase in the
bending rigidity as well as it is accompanied by
changes in the flag deformation shape.
The flag deformation shape remains constrained at

increased bending coefficient (Fig. 3b) as compared to
the low bending coefficient (Fig. 3a) by virtue of the
more stiffened flag. Power spectra of the
corresponding two cases are shown in Fig. 4. With an
increase in stiffness of the flag, the dominant
frequency of the flag is increased from 0.19 to 0.22.
For the less flexible flag in Fig. 3a and 4a, the flag gets
significant deformation due to the influence of the
upstream vortices. It changes the flag shape
significantly, resulting in decreasing the dominant
frequency.
59
The increase in the stiffness causes more resistance to

the upstream vortices, which decreases the effect of the
upstream vortices resulting in the low deformation of
the flag as well as increase in the dominant frequency
that is closer to the natural frequency of the flag in a
uniform flow.
The power spectra of the two low bending region

points (Fig 3 & 4 (c,d)) show that the dominant
frequency is lesser than that of the higher bending
energy case. The dominant frequency is 0.1 for
=2.0 and the bending stiffness =0.05 (Fig.
4d). With a decrease in the flag-length, increase in
bending rigidity for =1.25 and =0.06, the
deformation switches from bi-curvature bending to the
unidirectional curvature. This bending is also
accompanied by the higher dominant frequency of
0.175 (Fig. 4c).
The vorticity contours are shown in Fig. 5. The

flapping flag is in the shear layer shed by the D-shaped
half cylinder for the length ratio above =1.5, for
the length below 1.5, the flag affects the upstream
vortices by increasing the shear resistance to the
upstream vortices which delays/blocks the upstream
vortices and causes the two vortices to combine and
create a shear layer for the downstream flag. For the
flag with reduced length and at the same streamwise
gap, the vortices are completely shed from the Dcylinder in the upstream shed vortices interacts with
the flag in the constructive manner where the positive
vortices on the downside of the D-shared half cylinder
interacts with the downside (positive vortices) of the
flag and similarly negative vortices shed from the
upper side of the D-cylinder interacts with the upper
side (negative vortices) of the flag. This effect is same
for all the cases as the constructive mode. This mode is
dependent of the gap distance [14]. Production of shear
60
layer causes suction zone in front of the downstream

flag. Beal et. al. [15] reported the suction zone for
<2.0 that reveals its dependence on ratio
by the action of increased shear resistance. The suction
region generates an extra force on the flag which is
higher for =1.75, and =0.06 shown by the
absence of the vorticies in-front of the flag. While for
the other cases (Fig. 5), presence of the vortices in
front of the flag causes lesser effect of the suction zone
resulting in the lower bending energy.
The second map (Fig. 2b) showed the relation between
the length and the gap. Details are shown in the tail
position evolution, the power spectra and the vorticity
contours (Fig. 6, 7 & 8). The tail position evolution
shows the lowest energy region at =1.25 and
=1.75 (Fig. 7a) that corresponds to the lowest
dominant frequency (=0.15). As the energy is
increased at =1.5 and =1.25 (Fig. 7c), it is
accompanied by a slight increase in the dominant
frequency to =0.18. The bending energy is increased
at =2.0 and =2.0 (Fig. 7d). It is
accompanied by a small increase in the dominant
frequency to =0.20. For the highest energy point at
=1.75 and =1.75 (Fig. 7b), it has the
highest dominant frequency of =0.22. The tail
position evolution shows two interesting findings. One
is the frequency of the tail on the left side of Fig. 8 that
is complementing the power spectrum (Fig. 7). At the
optimal location (Fig. 8b), there appear seven peaks in
30 seconds, while for the other lower energy cases,
there are six peaks in 30 seconds. Other observation is
the flag deformation shape, for the length ratio of the
flag <1.75, unidirectional curvature deformation
of the flag is observed (Fig. 8a & c). For the length
ratio >1.75, there exists bidirectional curvature
deformation
(Fig.
8b
&
d).
The
unidirectional/constrained deformation is attributed to
the lower bending energy. The bending rigidity and the
gap is same but length is different causes jump from
the unidirectional to the bidirectional curvature
deformation (figure 8a, b) which shows that the
curvature change is function of the length of the flag.
At the same bending rigidity, the shorter flag length
causes constrained motion, as the shorter flag length
results in the less effective area for exertion of the
viscous forces by the fluid, while as the length is
increased to L/D=1.75 provides the sufficient surface
area of the flag to exert more viscous forces on the flag
to cause the bidirectional curvature flapping. This map
(figure 2b) gives the optimal length and the gap
distance for maximum energy harvesting (L/D=1.75,

S/D=1.75) while optimal bending rigidity is =0.06.
4. CONCLUSIONS
We examined the dynamics of energy harvesting
flexible flag behind a half cylinder undergoing passive
flapping modes subjected to a uniform viscous flow by
using the immersed boundary method. Analysis was
carried out under different conditions by varying
, , and . For passive flapping, an
increase in bending stiffness for 0.04 0.08 and
the length ratio of the flag (1.5 / 1.9) cause an
increase in the bending energy. Increase in bending
stiffness is accompanied by an increase in the flag
deformation, and the dominant flapping frequency also
is increased. Constructive mode is observed for the
optimal bending energy region when the streamwise
distance is increased. Also, unidirectional curvature is
attributed to the lower bending energy while the
bidirectional curvature is attributed to the higher
bending energy. By analyzing the passive flapping, we
obtained the optimal length of the flag, gap distance
between the flag, D-cylinder, and the bending rigidity.
The flag-in-fluid system suggested a mechanism by
which advantages may be gained via the vortexflexible body interactions in the energy harvesting eel.
REFERENCES
[1] J.J. Allen and A. J. Smits, 2001, Energy
harvesting eel, Journal of Fluids and Structures, vol.
15(34), pp. pp. 629640, 2001.
[2] A. Barrero-Gil, G. Alonso and A. Sanz-Andres,
Energy harvesting from transverse galloping,
Journal of Sound and Vibration, vol. 329(14), pp.
28732883, 2010.
[3] M. M. Bernitsas, K. Raghavan, Y. Ben-Simon and
E. M. H. Garcia, VIVACE (Vortex Induced Vibration
Aquatic Clean Energy): a new concept in generation of
clean and renewable energy from fluid flow, J.
Offshore Mech. Arct. Eng., vol. 130, pp. 041101,
2008.
[4] Z. Peng, and Q. Zhu, Energy harvesting through
flow-induced oscillations of a foil, Phys. Fluids, vol.
21, pp. 123602, 2009.
[5] K. Singh, S. Michelin and E. De Langre, The
effect of non-uniform damping on flutter in axial flow
and energy harvesting strategies, Proc. R. Soc. A, vol.
468, pp. 36203635, 2012.
61
[6] W. Mckinney and J. Delaurier, The wing mill: an

oscillating-wing wind-mill, Journal of Energy, vol. 5,
pp. 109115, 1981.
[7] T. Y. Wu, Extraction of flow energy by a wing
oscillating in waves, Journal of Ship Research, vol.
16, pp. 6678, 1972.
[8] T. Y. Wu and A. T. Chwang, Extraction of flow
energy by fish and birds in a wavy stream,
Proceedings of the Symposium on Swimming and
Flying in Nature, Plenum Press, New York, 1975.
[9] H. Isshiki and M. Murakami, A theory of wave
devouring propulsion, Journal of Society of Naval
Architecture Japan, vol. 156, pp. 102114, 1984.
[10] J. Grue, A. Mo and E. Palm, Propulsion of a foil
moving in water waves, Journal of Fluid Mechanics,
vol. 186, pp. 393417, 1988.
[11] L. Tang, M. P. Paidoussis and J. Jiang,
Cantilevered flexible plates in axial flow: Energy
transfer and the concept of flutter-mill, Journal of
Sound and Vibration, vol. 326, pp. 263-276, 2009.
[12] J. Eldredge, Dynamically coupled fluidbody
interactions in vorticity-based numerical simulations,
J. Comput. Phys., vol. 227, pp. 91709194, 2008.
[13] W. X. Huang, S. J. Shin and H. J. Sung,
Simulation of flexible filaments in a uniform flow by
the immersed boundary method J. Comput. Phys.,
vol. 226, pp. 2206-2228, 2007.
[14] S. Kim,W. X. Huang and H. J. Sung,
Constructive and destructive interaction modes
between two tandem flexible flags in viscous flow
Journal of Fluid Mechanics, vol. 661, pp. 511-521,
2010.
[15] D. Beal, F. Hover, M. Triantafyllou, J. Liao and
G. Lauder, Passive propulsion in vortex wakes,
Journal of Fluid Mechanics, vol. 549
62
SAME-09
CZM based Strength Prediction for Scarf Adhesive Joints with Varying
Bondline Angle
Aamir Mubashar1, I. A. Ashcroft2, Erum Aamir3
Department of Mechanical Engineering, School of Mechanical and Manufacturing Engineering, National University
of Sciences and Technology, H-12, Islamabad, Pakistan,
2
Faculty of Engineering, University of Nottingham, Nottingham, NG7 2RD, United Kingdom

3
Muhammad Ali Jinnah University, Islamabad, Pakistan

aamir@smme.nust.edu.pk
Abstract In this research, an investigation of

the effects of triaxial stress state in the adhesive
layer of scarf joints on the failure prediction was
carried out. Three types of scarf joints with zero,
9.5 and 26.5 angle bondline were manufactured
using steel adherends and a structural epoxy
adhesive FM73-M. The joints were tested to failure
under tensile loading. A finite element based model
of the joints was used to predict the failure
strength. The model consisted of continuum as well
as cohesive zone elements. The cohesive zone
elements were used to represent the adhesive layer
in the scarf joint and are capable of
phenomenological modeling of damage and failure.
Using the same set of cohesive zone model
parameters, the strength of the scarf joints was
predicted. The finite element modeling showed that
the middle of the adhesive layer had higher damage
than the edges of the bondline. The predicted
strength of the joints was in good agreement with
the experimental results for zero and 26.5 bondline
joints while lower failure load was observed for 9.5
bondline joints.
Keywords Scarf Joint; Cohesive Zone Model;
Stress Triaxiality; Adhesive Joints
I.
INTRODUCTION
Prediction of failure and joint strength is a topic of

great interest because of the increasing use of structural
adhesive joints in automotive and aerospace sectors.
Several type of joints configurations are used to test the

experimental joint strengths including double cantilever
beam (DCB), single lap joint and double lap joint. The
stress state in the bondline changes with the changing
loads and joint type. Triaxial stress becomes important,
especially in case of adhesives with hydrostatic stress
dependence. Using the finite element analysis, Wahab
et al [1] showed that the triaxial stress state changes
along the bondline of single lap and butt adhesive
joints. Scarf adhesive joints were selected for study of
failure under different values of stress triaxiality as the
state of stress triaxiality remains same along the
bondline in a scarf adhesive joint [2]. Based on the low
cycle fatigue data, a damage law was defined and its
parameters were calibrated as functions of stress levels
and triaxiality.
The damage and failure in adhesive joints has been
predicted by using cohesive zone modeling (CZM) [35]. Mubashar et al [6] used a cohesive zone model
based approach to predict the failure strength and
damage propagation in single lap adhesive joints that
were conditioned at 50C in water. A bilinear cohesive
zone model was used with moisture concentration and
moisture history dependence. The predicted results,
based on the generic methodology used for simulations,
were in good agreement with experimental failure
results of the joints. The strength of single and double
lap adhesive joints subjected to moist environments
was also predicted by Liljedahl et al [7]. The joints
used aluminum and carbon fiber reinforced plastic
(CFRP) composite adherends. The cohesive zone
parameters were determined by using mixed mode
flexure test specimens. Good agreement with
63
experimental results was observed and it was shown

that the cohesive zone parameters determined from one
type of specimens may be used to predict the strength
of other types of adhesive joints when the loading
modes are similar.
In this research, scarf joints with three different
bondline angles were selected and tested
experimentally to determine their failure strengths. The
three bondline angles produce three different stress
triaxiality conditions in the adhesive bondline. The
failure strength of the scarf joints was predicted using
finite element method. A cohesive zone modeling
approach was used where the damage in the bondline
was predicted in the form of a damage parameter. Same
set of cohesive zone parameters was used to determine
the strength of all three types of scarf joints to explore
the effect of stress triaxiality on the cohesive zone
parameters.
II.
EXPERIMENTAL METHODS
Three types of scarf joints with bondline angles of

0, 9.5 and 26.5 degree were manufactured and tested to
determine their failure response. The geometry and
bondline angles of the scarf joints were selected so that
the triaxial stress state in the bondline are different
from each other [2, 8]. Triaxial state of stress is
important in the failure of the used adhesive joints due
to the hydrostatic stress dependence of the adhesive.
The adherends were made from steel and an epoxy
based structural adhesive FM73-M was used to join the
adherends. FM73-M is a rubber toughened epoxy and
is supplied in film form. It has a polyester knit carrier
for support and handling. The configuration and
geometry of the scarf joints is shown in Fig. 1 where
represents the bondline angle. The adherends were 5
mm thick. The surface of the adherends was prepared
for bonding by degreasing followed by sand blasting.
The sand blasted adherend surfaces were coated with
BR172 corrosion inhibiting primer, which was air dried
for 30 minutes. The air dried joints were then cured for
1 hour at 120C. Adhesive was applied to the prepared
bondline surfaces and the joints were cured at 120C
for one hour.
The adhesive overflown to the sides of the
adherends was removed after curing. The prepared
joints were stored in desiccators. Tensile testing of the
scarf joints was carried out on Instron 3366 dual
column testing machine where the scarf joints were
loaded to failure and force-displacement data was
recorded.
Figure 1.
Configuration and geometry of scarf

joint. All dimensions are in mm.
III.
FINITE ELEMENT MODELLING
The scarf joints were modeled using commercially

available finite element code Abaqus. Three
dimensional models of the scarf joint were prepared as
the triaxial stress state in bondline of the scarf joint
changes with the change in bondline angle. The
adherends were modeled using 8 node linear brick
elements with reduced integration. A biased mesh with
a dense mesh in the middle and a coarse mesh at the
ends was used, as shown in Fig. 2. A linear elastic
material model was used for steel adherends, as the
adherends are of much high stiffness than the bondline
and did not deform during the experimentation.
The geometric dimensions of the bondline are very
small as compared to the overall dimensions of the
scarf joints and it can be modeled as a zero thickness
layer of cohesive elements, as shown on left in Fig. 2.
These were 8 node three dimensional cohesive
elements with their thickness direction perpendicular to
the bondline thickness. Bilinear cohesive zone law was
used for the cohesive elements, which is represented in
traction () and separation () form in Fig. 3. The
parameters used in the cohesive zone law are cohesive
energy Gcoh, stiffness K, tripping traction c, critical
displacement c and final displacement f.
Damage initiation in the adhesive layer was based
on a stress criterion, which is based on the ratio of
stress at any time increment to the maximum stress
reached. Damage initiates when this ratio reaches one
as given below:

(1)
max n , s , t 1
nm sm tm
where nm is the maximum normal stress and sm,
tm are maximum stresses in transverse shear
directions. n, s and t are corresponding stresses at
any time increment.
64
IV.
Figure 2.
Scarf joint meshing scheme. The zero
thickness adhesive layer is shown on the left.
Since no damage occurs under compression, a
Macaulay Bracket is used in Eq 1. Values of different
parameters used for the cohesive zone law are given in
Table 1. The values are based on the cohesive zone law
studies conducted for single lap joints [6], however, a
lower value of tripping traction was used for scarf
joints, as compared to the single lap joints, owing to the
difference in stress state of the two joints. The scarf
joint was held from one end and displacement loading
was applied to the other end to obtain the failure
response of the scarf joints.
THE COHESIVE ZONE MODEL PARAMETERS
FOR SINGLE LAP JOINT.
TABLE I.
Tripping
traction c
(MPa)
32
Figure 3.
Stiffness
K (N/mm3)
1 x 105
Cohesive
energy Gcoh
(kJ/mm2)
2.9
Bilinear cohesive zone law.
RESULTS AND DISCUSSION
The experimental load-displacement response of the

scarf joints is shown in Fig. 4. It can be observed that
the failure load of the scarf joints increased as the
bondline angle was changed from zero to 26.5. The
force-displacement response of the scarf joints becomes
non-linear after an initial linear force increase with
increasing displacement. The non-linear increase in
force is observed when adhesive starts to deform
plastically. The initial linear portion of the forcedisplacement curve is increased in joints with 9.5 and
26.5 when compared to the zero bondline scarf joints.
It was also observed that with increasing bondline
angle of the scarf joints, the strain to failure increases
and the scarf joint with 26.5 bondline angle exhibits
more plastic deformation than the zero bondline angle
scarf joints. The higher joint strengths with changing
bondline angles may be attributed to the increased
adhesive area and changed stress state in the bondline.
The stress triaxiality in the bondline, which is the ratio
of hydrostatic stress to the von-Mises stress, changes
from 3.45 to 1.36 as the bondline angle changes from
zero to 26.5 [2]. This shows that the effect of shear
stresses is more in scarf joints with an angular bondline
than a scarf joint with zero bondline angle. It may also
be observed from Fig. 4 that for a particular applied
force, the displacement 26.5 bondline is more than the
zero degree bondline scarf joints. This is owing to the
increased plastic deformation of the scarf joints with
greater bondline angles. The scarf joints fail suddenly
as the load approaches the failure load.
The equivalent von Mises stress in the middle of
the bondline, at maximum failure load, is plotted in Fig.
5. It may be seen that the stresses are uniform along the
bondline for zero and 9.5 bondline angle scarf joints.
However, a small variation in stress magnitude is
observed along the bondline for 26.5 angle scarf
joints, which is due to the change in loading direction.
Stress levels at joint failure are also higher for joints
with higher bondline angles. This occurs as the loading
conditions change from pure tensile to combined
tensile and shear loading. With increasing bondline
angle, the area of bondline interface also increases.
The cohesive zone based model predicted the
damage propagation and failure load of the scarf joints.
The cohesive zone model used a scalar damage variable
to represent the damage initiation and propagation. The
value of damage parameter is between zero and one,
where one means that the material has failed. The
damage variable used in Abaqus is SDEG. The
predicted damage evolution in the bondline of the three
selected scarf joints with varying bondline angles is
shown in Fig. 6. The upper adherends of the scarf joints
65
are removed in Fig. 6 to make the zero thickness

cohesive element layer visible. The damage variable,
SDEG, is plotted as contours of varying levels of
damage. Since the elastic material model used for
adherends is not capable of predicting damage, the
lower adherend is depicted in a single colour, which
does not relate to the legend shown in Fig. 6. It may be
seen that the damage initiates in the middle of the
bondline and propagates towards the edges of the joint.
The phenomenon is symmetric in the zero bondline
angle scarf joints but becomes asymmetric in the scarf
joints with 26.5 bondline angle. This happens as the
loading conditions on the bondline change with change
in scarf angle of the joint and with increase in scarf
angle from zero to 26.5, the loading does not remain
perpendicular to the bondline.
The predicted failure loads of the three types of
scarf joints are plotted in Fig. 7. It is observed from
published experimental results that adhesive joints
exhibit variations in failure loads for joints of similar
type and manufactured under similar conditions [9].
These variations may be attributed to various factors
including material variability, porosity in bondline,
curing of the adhesive and environmental factors. To
include the effects of variability in the adhesive joint
failure strength, vertical variation bars of 5% of the
mean failure strength were added in Fig. 7. The
cohesive zone model based predicted failure strength of
the scarf joints is in good agreement with the
experimental results for zero and 26.5 bondline angle
scarf joints., Lower than experimentally observed
failure strength is predicted by the cohesive zone model
for 9.5 bondline angle joints. However, if the
variability in the joint failure strength is considered, the
predicted strengths for all bondline angles are close to
the experimental results.
Figure 5.
(a)
Experimental force-displacement
response of scarf joints.
(b)
(c)
Figure 6.
Damage predicted by CZM in the
scarf joints with bondline angle of (a) 0 (b) 9.5 and
(c) 26.5.
Figure 4.
Experimental force-displacement
response of scarf joints.
66
joints," Int. J. Adhesion and Adhesives, vol. 26, pp.

325-336, 2006.
[5] S. Li, M. D. Thouless, A. M. Waas, J. A.
Schroeder, and P. D. Zavattieri, "Competing
failure mechanisms in mixed-mode fracture of an
adhesively bonded polymer-matrix composite,"
Int. J. Adhesion and Adhesives, vol. 26, pp. 609616, 2006.
Figure 7.
Comparison of experimental and finite
element method predicted failure loads for scarf joints.
V.
CONCLUSIONS
The failure strength of scarf adhesive joints with

varying bondline angle was determined experimentally
and predicted using cohesive zone modeling. The
experimental results showed that as the bondline angle
increased, the scarf joint strength also increased. This
may be attributed to increase in bonded area and
change of loading conditions. The load-displacement
response of the scarf joints was nonlinear for most of
the loading duration. The cohesive zone modeling
provided damage and failure strength predictions of the
scarf adhesive joints. A single set of cohesive zone
parameters gave good predictions of the failure strength
of scarf joints with bondline at three different angles.
REFERENCES
[6] A. Mubashar, I. A. Ashcroft, G. W. Critchlow, and

A. D. Crocombe, "Strength prediction of adhesive
joints after cyclic moisture conditioning using a
cohesive zone model," Engineering Fracture
Mechanics, vol. 78, pp. 2746-2760, 2011.
[7] C. D. M. Liljedahl, A. D. Crocombe, M. A.
Wahab, and I. A. Ashcroft, "Modelling the
environmental degradation of adhesively bonded
aluminium and composite joints using a CZM
approach," Int. J. Adhesion and Adhesives, vol. 27,
pp. 505-518, 2007.
[8] I. Hilmy, M. M. Abdel Wahab, I. A. Ashcroft, and
A. D. Crocombe, "A finite element analysis of
scarf joint for controlling the triaxiality function in
adhesive bonding," Key Engineering Materials,
vol. 385-387, pp. 17-20, 2008.
[9] A. Mubashar, I. A. Ashcroft, G. W. Critchlow, and
A. D. Crocombe, "Moisture absorption-desorption
effects in adhesive joints," International Journal of
Adhesion and Adhesives, vol. 29, pp. 751-760,
2009.
[1] M. A. Wahab, I. Hilmy, I. A. Ashcroft, and A. D.

Crocombe, "Damage Parameters of Adhesive
Joints with General Triaxiality Part I: Finite
Element Analysis," Journal of Adhesion Science
and Technology, vol. 25, pp. 903-923, 2011.
[2] M. A. Wahab, I. Hilmy, I. A. Ashcroft, and A. D.
Crocombe, "Damage Paramters of Adhesive Joints
with General Triaxiality Part 2: Scarf Joint
Analysis," Journal of Adhesion Science and
Technology, vol. 25, pp. 925-947, 2011.
[3] M. F. S. F. de Moura and J. A. G. Chousal,
"Cohesive and continuum damage models applied
to fracture characterization of bonded joints," Int.
J. Mech. Sci., vol. 48, pp. 493-503, 2006.
[4] A. D. Crocombe, Y. Hua, W. K. Loh, M. A.
Wahab, and I. A. Ashcroft, "Predicting the residual
strength for environmentally degraded adhesive lap
67
SAME-10
Design and Fabrication of Three-Dimensional Printing Machine

Jasim Arif Ali1, Syed Zohaib Ali Naqvi1, Rehan Afzal1, Tufail Ahmed Memon1, Muhammad Fahad1, Maqssod
Ahmed Khan1
1
Department of Industrial and Manufacturing, NED University of Engineering and Technology, Karachi,
Pakistan.
mfahad@neduet.edu.pk
Abstract
This paper outlines the fabrication of threedimensional (3D) printing machine using fused
deposition modeling technique. The purpose of this
research was to fabricate a low cost 3D printer to be
used in schools and households. In 3D printing a
computer aided design (CAD) model is converted into
G-codes which are used by the controller to print the
model. The three commonly used techniques for 3D
printing are fused deposition modeling (FDM),
selective laser sintering (SLS) and stereolithography
(STL). The fabricated 3D printer is based on FDM
technique. In fused deposition modeling a plastic
filament is used which is melted and deposited by a
nozzle layer by layer on a bed forming a replicate of
CAD model. Detailed discussion on design of our 3D
printer is made involving the steps of manufacturing of
some major components. Furthermore costing of the
individual component of printer is mentioned. In the
end comparisions were made between a commercial 3D
printer and the fabricated one.
Keywords: Fabrication, 3D printer, FDM, design.
VI. INTRODUCTION
This paper is the outcome of the project done in final
year engineering project-making. The aim is to
manufacture a low cost three-dimensional printing (3DP)
machine to promote the use of 3DP in Pakistan. 3DP is
an additive manufacturing technique in which layers of
plastic are deposited one over another to form an object
which is replicate of a computer aided design (CAD)
model. First a CAD model is designed which is imported

in slicing software; this software slices the image into
small layers which are then printed by the threedimensional (3D) printer. The CAD model needs to be
converted into an STL file format. This file format is
usually used for 3DP. Three commonly used techniques
for 3D printing are Selective laser sintering (SLS),
StereoLithography (SL) and Fused Deposition Modeling
(FDM).
In 1980 Hideo Kodama of Nagoya Municipal
Industrial Research Institute first invented the technique
now called as stereo lithography [1]. In 1984 Chuck Hull
of 3-D Systems Corp patented the stereo lithography
method of printing objects [2] and designed the STL file
format. The STL file format is the basis of almost all the
open source slicing software available today and
different types of infill pattern also called as air gaps are
defined based on STL file format. The 3-D printing
technology has evolved since then and researchers have
concluded that 3-D printing could help in sustainable
development [3].
A. Selective Laser Sintering (SLS)
In this process a laser of high power is used to sinter
particles of materials like plastic, ceramic or metal to
form a three-dimensional model. The CAD model is
traced by the laser on surface of powdered bed. The
main components of SLS printers are as follows;
68
High power CO2 Laser

Build Platform
Leveling Roller
Control System
Powder Cartridge
The high power laser sinters the particles by tracing a

cross sectional area of the part. The platform is moved
down by a single layer thickness and the material is
pushed by the leveling roller across the build platform.
In this way complete part is built. 3D systems are
leading developers of SLS printers. [4]
B. STEREOLITHOGRAPHY (SL)
The basic principle of SL is using a laser and a
photopolymer. Four main components of the SL printers
are as follows;
An Ultraviolet (UV) laser.

Perforated platform.
Photopolymer liquid plastic that is filled in a
tank.
A Platform that is run by computer.
An Ultra Violet (UV) laser beam is directed by a
computer guided mirror on the surface of the UV
photopolymer resin. The model is built one layer at a
time from supplied 3D CAD data [5].The process starts
by immersion of the perforated platform into the liquid.
The UV laser strikes the photopolymer and hardens it.
The cross sectional area of part is traced by the laser.
After each layer the platform descends for the formation
of next layer and at the end of the process the part is
lifted out of the vat. 3D systems are leading developers
of SL printers.
C. FUSED DEPOSITION MODELING (FDM)

The basic principle of FDM printers is melting of a
plastic filament and extruding it out through the nozzle.
The major components of FDM printers are as follows;
Extruder
Nozzle
Plastic filament
Bed
The nozzle is guided by axes which are controlled by
computer. The plastic filament used could be Polylactic
acid (PLA) or Acrylonitrile Butadiene Styrene (ABS).
Pure PLA has a glass transition temperature (Tg) of
about 55 C and melting point (Tm) of about 180 C [6].
The glass transition temperature of ABS is
approximately 104 C (219 F), due to amorphous nature
this material does not display a melting point [7]
.However 220 C is usually used as a melting point for
3DP .The melted filament coming out from the nozzle
forms the cross sectional area of the part in the form of
layers. After each layer the bed on Z axis descends to a
distance equal to the layer thickness and the process is
repeated resulting in a completed part. Stratasys is the
leading developer of FDM printers.
FDM technology was less expensive and
complicated, therefore it was adopted to develop desktop
printers.
69
D. DESKTOP PRINTERS
Desktop printers like Prusai3, Ultimaker, and
Makerbot etc. all are Cartesian co-ordinate printers. The
success of desktop printers is due to their high quality as
compared to their costs. Some of these printers like
Repraps are self replicating as many of their parts are
made of plastic and can easily be 3D printed.
Reprap forum provides all the basic information needed
for making a desktop 3-D printer.
FDM technology has become so powerful that
researchers at MIT have achieved glass printing through
a nozzle [8] plus the advancements made in the
development of different filaments have opened up gates
for 3D printed electronics board as well [9].
In Pakistan 3DP is not as common as in other
countries. Pakistan does not have any 3D printer
manufacturing firm. The common misconception
prevailing in Pakistan is that 3D printing is expensive
and time consuming which is untrue. In the following
study an attempt has been made to modify certain
components of a desktop 3D printer and make a low cost
3D printer in Pakistan with all the possible resources
available in the country.
distribute the load on all the shafts. This is shown in Fig.

2
Following are the most important components of the
X-Y axes assembly;
Steel rods
Timing pulleys and Belts
Sliders
1) Timing Pulleys
The timing pulleys are also manufactured in NED
university of Engineering and technology(UET). The
diameter of the pulleys is fifteen millimeters with a pitch
of two and half millimeter as shown in Fig. 3. Our first
material choice was Aluminum as it is light weight and
has good strength to weight ratio, but the aluminum
available could not withstand cutting forces and was
fracturing, therefore we chose Brass for manufacturing
the pulleys that can easily bear the cutting forces and is
also corrosion resistant.
The pulleys are manufactured in three Steps
Steps of Manufacturing
VII. METHODOLOGY
The methodology we adopted for fabricating the
machine is discussed in this section. Fabrication of
major components is also discussed in detail.
A. Frame
The construction of the frame is inspired from
Ultimaker as showm in Figure.1, however few changes
are made to minimize the cost. The frame is erected out
of steel conduits which are welded to form a cubic
structure. Brackets are used to hold motors and rods.
Ball bearings are tight fit into brackets. These brackets
are fixed with nuts and bolts on the conduits.
B. X- Y Axes Assembly
The X-Y axes are controlled using NEMA17 Stepper
motors coupled with Steel rods that are mounted with
timing pulleys and belts in order to make the nozzle
move in the desired direction. Also a slider facilitates the
smooth motion of the nozzle and helps to equally
Grooves were cut into a Brass rod of diameter

15mm and length of 150mm
Another brass rod was used to make the top and
bottom flanges of the pulley so that the belt may
not slip.
70
The grooved rod was cut into 8 pulleys and the

flanges are press fit on them.
2) Sliders
The slider facilitates the smooth movement of the
nozzle. Its function is to grip the timing belt and move
with it. The material of slider is Aluminum. They are
fitted with LM8UU bearing for smooth linear movement
on the surface of the steel rod. While on the other face of
the slider an eight millimeter diameter hole is drilled to
hold the extruder assembly as per the design (see Fig. 6).
On the top surface of the slider a Carbon Steel plate is
screwed, this plate has grooves with the same pitch as
the belt for gripping the belt.
71
C. Extruder
The cold end of extruder is made up of Aluminum. It
has an adjustable bearing assembly as shown in Fig. 7.
The bearing could be adjusted such that the filament
could be driven properly into the nozzle. A stepper
motor with a knurled bush provides the necessary grip to
push the filament forward. The cold end of extruder is
manufactured completely in Pakistan, at NED University
of engineering and technology, manufacturing lab while
the hot end (consisting of heat sink, 0.4mm nozzle, and
40 watt heater) is imported from China.
The cold end of extruder assembly is divided into two
separate sub assemblies
1. The motor assembly
2. The adjustable bearing assembly
Steps of Manufacturing
Aluminum block was faced according to the

dimensions of the design.
A 15 mm hole was drilled into the front face of
the block for LM8UU bearing.
Another hole of 8mm diameter (visible in the
side view of the block in Fig. 6) was drilled.
Grooves were cut into a Steel plate of 2.5 mm
pitch for gripping the belt.
1) The motor assembly

Steps of manufacturing of motor assembly are given
below :
A rectangular block of Aluminum was faced to
required dimensions by using an end mill cutter.
Then the center hole was drilled.
After that four M3 holes were drilled as per the
given dimensions.
Then a 3mm hole was drilled in the two small
rectangular blocks using a drill press.
Motor was then mounted on the plate.
2) The adjustable bearing assembly
Steps of manufacturing of adjustable bearing assembly
are given below:
The Aluminum block is milled to form a U-shaped

plate, with holes on the top right corner to fit into
the motor assembly.
Another separate Aluminum plate was used to
mount the bearing with a threaded bush in the
center so that the bearing rests on the plate and
inner ring of the bearing does not move.
A slot was then cut on the top side of the bearing
plate
This bearing plate slides into the U shaped housing
and the slot was screwed into a hole and can be
adjusted by loosening the screw slightly.
72
The two assemblies are then screwed together using

M3 socket head screws.
D. Electronics
We have used RepRap Arduino Mega Pololu Shield
(RAMPS) for our 3D printer. It is commonly used
electronics board for desktop 3D printers. It uses an
Arduino board which serves as a brain of printer, Pololu
drivers DRV8825 are used for driving motors and ATX
power supply is used.
VIII. COSTING
The manufacturing of complete printer costs 25,182
Pakistani Rupees (PKR). Most of the things were
fabricated in Pakistan. The price breakdown of major
components is discussed in this section. Costs of
individual components are mentioned in the separate
tables, fabrication as well as raw materials costs are
taken into account while documenting.
GT2.5 Pulleys
4,000
Total price
15,480
TABLE III. COST OF EXTRUDERS COMPONENTS

Serial
no.
Component
Quantity
Total cost
in PKR
Hot end
assembly
800
Cold end
assembly
800
Total price
1,600
TABLE I. COST OF EACH FRAME COMPONENT

Serial
no.
Component
Quantity
Total cost
in PKR
Steel conduits
12
1000
Steel brackets
500
Aluminum
bed
500
Total price
TABLE IV. COST OF ELECTRONICS

Serial
no.
Component
Quantity
Total cost
in PKR
1
2
3
RAMPS
Stepper motors
ATX power
supply
Total price
1
4
1
4,600
1,200
600
2,000
6,400
TABLE II. COST OF EACH X-Y AXES COMPONENT

Serial
no.
Component
Quantity
Total cost
in PKR
Ball bearings
80
Linear
bearings
10
1,500
Steel rods
12
6,000
Sliders
2,000
GT2.5 belts
700
Lead screw
1,200
TABLE V. OVERALL COST

Components
Serial
no.
1
2
3
4
Frame
X-Y assembly
Extruder
Electronics
Total price
Total cost
in PKR
2,000
15,480
1,600
6,400
25,480
IX. COMAPARISONS
In this section we will compare our 3D printer with
commercial desktop 3D printer in terms of cost,
73
structure and build envelop. The comparisions drawn are

based on our experimentation with Prusa i3.
REFERENCES
The average cost of a commercial desktop printer

such as Prusa i3 is 400 US Dollars(USD), [10] while our
printer costs around 25,500 Pakistani Rupees (255
USD).
[1] Hideo Kodama A Scheme for Three Dimensional

Display by Automatic Fabrication of Three
Dimensional-Model IEICE TRANSACTIONS on
Electronics (Japanese edition), Vol.J-64C, No.4, pp,
237-234 April 1981
[2] Charles W. Hull Apparatus of Production of ThreeDimensional
objects
by
Stereolithography
US4575330 11 March 1986
[3] Joshua M. Pierce, Christine Morris Blair, Kristen J.
Laciak, Rob Andrews, Amir Nosrat, Ivana Zelinka Zovko 3-D Printing of Open Source Appropriate
Technologies for Self Directed Sustainable
Development. Journal of Sustainable Development
3 (4): 17-29. Do: 10.5539/jsd. v3n4p17. Retrieved
2012-01-31.
[4] Anatomy of SLS Process (2015, October 10)
Retrieved
from
http://www.3dsystems.com/quickparts/learningcenter/sls-anatomy
[5]Anatomy of SL Process (2015, October10) Retrived
From
http://www.3dsystems.com/quickparts/learningcenter/sla-anatomy
[6] Ewa Rudnik (2007)Compostable Polymer
Materials Edition-2, Elsevier.
[7]Melting and Boiling points of ABS and PLA
Retrieved
From
http://www.nrri.umn.edu/NLTC/ABS07.pdf
[8] Rankin, T. M., Giovinco, N. A., Cucher, D. J., Watts,
G., Hurwitz, B., & Armstrong, D. G. (2014). Threedimensional printing surgical instruments: are we
there yet?. Journal of Surgical Research, 189(2),
193-197.
[9] Church, K. H., Tsang, H., Rodriguez, R.,
Defembaugh, P., & Rumpf, R. (2013, July). Printed
circuit structures, the evolution of printed circuit
boards. In IPC APEX EXPO Conference
Proceedings.
[10] Prusa i3 cost (2015, October 5) Retrived from
https://www.3dprintersonlinestore.com/full-acrylicreprap-prusa-i3-kit/
The structure of Prusa i3 is quite fragile as it is made

up of Acrylic for this reason we have used Iron conduits
to provide rigidity to the frame.
The build envelope of Prusa i3 is 200mm x 200mm x
180 mm whereas the build envelope of our 3D printer is
200 mm x 200 mm x 300 mm.
TABLE VI COMPARISONS BETWEEN PRUSAI3
AND OUR 3D PRINTER
Parameters
Cost
Prusa i3
400 USD
3D Printer
250 USD
Material of structure
Acrylic
Iron
conduits
Build Envelope
200 mm x
200 mm x
200 mm x180 200 mm x
mm
300 mm
X. CONCLUSION
The effort was made to fabricate a low cost 3D printer to
promote the use of these printers in schools for teaching
purposes, also for promoting research and development
environment in our country. The cost was reduced by
designing a lean structure, few components were
inexpensive in China therefore they were imported. As
the project is in its partial fulfillment stage therefore
final phase which is the calibration phase is still left to
be done. We hope that by our efforts the awareness and
use of 3D printers will increase in our country.
ACKNOWLEDGMENT
We would like to express our deep gratitude to the lab
technicians of NEDUET, Industrial and manufacturing
department, for their support and help in the fabrication
of 3D printer.
74
SAME-11
Development of Hybrid Mold casting for wheel Hubs

Waseem Arshad1, Muhammad Adnan Hanif1, Muhammad Usman Bhutta1 and Usman Abdullah1
Department of Mechanical Engineering, School of Mechanical and Manufacturing Engineering (SMME), National
University of Sciences and Technology (NUST), Islamabad, Pakistan
Abstract Some products for automotive industries

are difficult to manufacture using the conventional
metal casting processes. In such situation different
methods are combined to form a hybrid casting
process. The hybrid casting process is a technique in
which different casting process are combined to get a
cast product. Therefore, the hybrid casting technique
gives advantage over conventional casting process. In
this work we discuss the hybrid casting process
where sand mold is combined with permanent mold
before pouring. This hybrid casting process is used to
cast wheel hubs to get better properties. Sand Molds
are made from green sand while cores are made
using NaSi sand. As we combine permanent mold
with sand mold to get the hybrid mold casting
process it enables a broader range of products to
have a superior surface finish and lesser porosity
when facing with a thick section products. The
hybrid casting process gave close tolerance than the
traditional casting process, better surface finish on
the permanent mold side and lower porosity as
compared to the conventional casting process.
Index TermsHybrid casting, automotive, wheel
hubs.
XI. INTRODUCTION
In a casting manufacturing process molten metal is
transferred into a mold which has the desired/anticipated
shape and after solidification the metal gets the shape of
the mold cavity [1]. Most of the castings are made from
sand casting process. Approximately all casting alloys
can be sand cast. Sand casting is widely used for metals
that have high melting temperatures such as nickel, steel
and titanium. Its adaptability allows the casting ranging
in dimensions from small to very large and in fabrication,
quantities from one to thousands [2]. Green sand molding

is the least expensive among all the molding techniques
as sand is replenished after each batch [3] and chances of
hot tearing of the casting are low [4]. In sand casting
there is a limit to complexity of design [5] and it has poor
dimensional accuracy as the size of the casting increases.
Also, surface finish is poor as compared to other
techniques [6].
Permanent mold casting process is the process in
which pouring is done under gravity in a permanent mold
[7] which is made after machining grey or nodular iron
[8]. Permanent mold casting is suitable for a broad
variety of shapes [9]. Permanent mold temperature is
raised using a flame and mold is filled with molten metal
[10]. Good dimensional accuracy and surface finish are
the main advantage of permanent mold casting [11].
Today most of the wheel hubs are made from ductile
iron. With elongation of more than 18% and high
strength up to 1000 MPa provides the opportunity to
choose ductile iron. Ductile iron has cost savings benefit
over malleable iron and steel [12] [13]. The metallic
charge used for making ductile iron generally consists of
steel scrap, pig iron and ductile iron scrap [14].
Sometime carbides are formed in ductile iron due to the
following reasons [15]:
Elements in the charge that promote carbide
formation
1. Unnecessary Magnesium
2. The cooling rate is high
3. Melt having Low C.E
4. Bad or no inoculation
Carbon, Carbon equivalent, Manganese, Magnesium,
Silicon, Aluminum and Phosphorous are the important
elements that should be controlled wisely [16].
75
XII. METHODOLOGY AND EXPERIMENTATION

The hybrid mold casting process is developed in
which one half of the casting is made from sand mold
and other half is made from permanent mold. Sand mold
was prepared from green sand. Green sand consists of Si
sand, Bentonite, coal dust and water. Table 1 shows the
various parameters determined using different sand test.
No
TABLE 1. MOLDING SAND PARAMETERS

Sand Test
Value
Compactability
35%
Specimen weight
142g
Compression strength
21.7N/cm2
Splitting strength.
4 N/cm2
Permeability
116
Moisture
4.1%
Cope Hardness
90 (B scale)
A permanent mold is made from cast iron. Hybrid

casting process is developed for wheel hubs. In this
experiment cope is made from sand mold while drag is
made from permanent mold as shown in Figure 1. Sand
mold was prepared from a mixture of silica sand,
bentonite and coal dust. The Permanent mold is started
from polyurethane pattern in a sequence in which
negative sand mold is prepared from which positive sand
mold was prepared. Cast iron melt was poured into the
positive sand mold to get the negative permanent mold.
Permanent mold heating was done up to 120oC.
Figure 1. Permanent Mold Drag

Graphite as a stripping agent is used which helped us
in the smooth knocked out finish of the casting. The
Coating was ignited to remove the moisture and to get
the thermoplastic layer. Core was made from the Si sand
and sodium silicate. The core was coated with mixture of
graphite, spirit and phenolic resin as shown in figure 2.
The coating was ignited to obtain the thermoplastic layer.
Permanent mold was coated with the mixture of
graphite, spirit and phenolic resin and then the coating
was ignited to obtain a thermoplastic layer which helped
in the stripping of the casting.
Ductile iron was obtained from steel scrap in which
carbon and silicon is added to get the desired
composition as shown in Table 2. Molten metal having
the carbon between 3.7-4% and Silicon between 2.7% to
2.9% is treated with FeSiMg after which FeSi is added
for inoculation.
TABLE 2. DUCTILE IRON COMPOSITON

Element
Percentage
Composition
3.7-4.1%
Si
3.1-3.3%
Mn
0.1-0.3%
Mg
0.03-0.06%
<0.03%
76
Ductile iron having the desired composition is poured

into a hybrid mold as shown in Figure 3. Molten metal
pouring temperature was at about 1480OC.
Figure 7. Ductile Iron Pouring

Casting was knocked out after 3.2 minutes and then
shot blasting was done to remove the sand which was
stuck to casting mold. After shot blasting casting was
sent to paint shop for painting. Figure 4 shows the final
shape of wheel hubs.
Figure 4. Wheel Hubs

XIII. RESULTS AND ANALYSIS
Figure 2. Core preparation
There were no defects in the casting of the hub.

Wheel hub is cut into eight equal parts to check its cross
section. Each section showed that there is no porosity and
all the sections are perfectly sound.
77
Cross section was cut to get hardness and

microstructure samples at different points. The last
poured hub was sectioned through their casting gate
locations to produce longitudinal cross section, as shown
in Figure 5. Cross section was subjected to Brinell
hardness (HB) testing. The results of the hardness testing,
given in Table 3, revealed that hub generally exhibited
variation in hardness values at the locations tested. The
results show that cross section cast in permanent mold
has high hardness as compared to sand mold.
samples taken from the permanent mold side shows that

microstructure is predominantly carbidic as shown in
figure 7. The nodule count is higher in the permanent
mold side as compared to sand mold.
TABLE 3. WHEEL HUB HARDNESS
Section
position
Hardness (BHN)
Sand mold
1
171
181
171
190
Permanent Mold
5
441
294
341
441
Figure 8. Wheel Hub Cross Section

The removed longitudinal cross section is polished
and etched in natal 3% solution. Cut section was further
divided into eight sections to observe the overall
microstructure at different points. The microstructure of
the wheel hub cast in sand mold is shown in Figure 6.
The four samples that were taken from the sand mold
sides show that microstructure is predominantly ferriticpearlitic and some carbides are found while the four
78
Figure6. Sand Mold Microstructures at 100X

(Predominantly Ferritic-Pearlitic)
Figure 8.
Permanent Mold microstructure at 100X

(Predominantly Carbidic)
XIV. CONCLUSION
79
In this work the hybrid mold casting process is

developed in which sand mold and permanent molds are
combined to improve the quality of the wheel hubs. The
hardness of the cross section, taken from the wheel hubs,
shows that permanent mold side has a much higher
hardness as compared to sand mold side. This is
concluded because the microstructure of the permanent
mold is predominantly carbidic, which give rise to
hardness. Moreover, the surface finish on the permanent
mold side is good as compared to sand mold side. In sand
mold side microstructure is predominantly ferriticpearlitic due to which sand mold side has more ductility.
Dimensional control is easy on the permanent mold side.
[13]
Jones, J. and B.G. Corbett Jr, Restraining
mechanical joint gasket for ductile iron pipe, 2008,
Google Patents
[14]
Karsay, S.I., Ductile iron production1967:
Quebec Iron and Titanium Corporation.
[15]
A. Javaid, J.T., M. Sahoo and K.G. Davis, AFS
Transactions. Vol. First Edition 1999. 107.
[16]
Hughes, I.C., BCIRA, International Centre for
Cast Metals Technology, Great Britain1990.
REFERENCES
[1] Callister, Materials Science and Engineering. 2007
[2] Groover, M.P., Fundamentals of Modern
Manufacturing Materials, Processes and Systems.
Vol. Fourth Edition.
[3] M. M. Avedesian, H.B., ASM Specialty Handbook:
Magnesium and Magnesium Alloys. p. 61.
[4] Davis, J.R., ASM Specialty Handbook Cast Irons.
1996: p. 494
[5] Wanlong Wang, H.W.S., James G. Conley, Rapid
Tooling Guidelines For Sand Casting. 2010.
[6] Aluminum and
Davis1993. 784.
Aluminum
Alloys,
ed.
J.R.
[7] Yu, K.-O., Modeling for Casting and Solidification

Processing. 2002: p. 373-374.
[8] John Gilbert Kaufman, E.L.R., Aluminum Alloy
Castings: Properties, Processes, and Applications.
2004.
[9] Jack V. Michaels, W.P.W., Design to Cost. 1989.
[10]
Hybrid
Casting:
http://www.industrialmetalcastings.com/casting_hyb
rid_mold_casting.html.
[11]
Donald Askeland, P.F.,
Engineering of Materials2005.
The
Science
&
[12]
SORELMETAL Ductile Iron Advantage,
http://www.sorelmetal.com/en/ductile/main_advanta
ge.htm
80
SAME-12
Comparative Study of Wax and Nitrate Salts as Thermal Storage for

Cooling Purpose of Automobiles
Umair Saleem, Muhammad Adnan Hanif and Adeel Waqas
Department of Mechanical Engineering, School of Mechanical and Manufacturing Engineering, National University
of Sciences and Technology (NUST), Islamabad, Pakistan
Department of Energy System Engineering, Centre for Advance Studies in Energy, National University of Sciences
and Technology, Islamabad, Pakistan
Umair_saleem41@hotmail.com
AbstractEnergy Demand of the world is increasing

exponentially. Transportation sector holds the major
share in this energy demand. During summer
conditions in countries like Pakistan, cars utilize
more fuel due to use of car air conditioning unit.
Thermal storage can prove to be helpful in reducing
this cooling demand by storing extra heat inside the
car and reducing car temperature and thus reducing
cooling load of the car. This paper presents a
comparative study between two phase change
materials; nitrate salts and wax for energy storage
for cooling purpose of automobiles. A comparison
between their thermal storage capacities, their
charge and discharge time is made using CFD tool
like COMSOL.
Index Terms CFD, COMSOL, AUTOMOBILE

AC, PCM
VI.
INTRODUCTION
Energy demand is increasing exponentially with

transportation sector as major stake holder. In countries
like Pakistan especially during summer season, cars
utilizes more fuel due to excessive use of air
conditioning unitError! Bookmark not defined.[1]. In
a stationary car parked under the sun inside temperature
can reach up to 70C[2]. Thermal comfort is one of the
basic human needs and Thermal storage can prove

helpful in this regard as it will store the excess heat
inside the car and thus reduce cooling load[3]. Among
all the thermal heat storage Latent Heat Thermal Energy
Storage (LHTES) using Phase Change Materials (PCM)
is useful as it provides heat at constant temperature
during discharge time and store a lot of heat during
charge time at constant temperature.[4]
This paper focuses on a comparative study of PCM Wax
and Nitrate Salt as a thermal storage to enhance its
operational capability. Comparison is performed
between their charge and discharge time using
COMSOL Multiphysics as a CFD tool for this
comparative study and results of this comparison are
discussed below.
VII.
RESEARCH APPROACH
The methodology followed in this research incorporates

study of two phase change materials, i.e., wax and
sodium salt primarily sodium nitrate. Study comprises of
detailed three dimensional analyses, review of the
physics involved, definition of computational domain
and numerical solution of the physics over the domain.
Heat transfer with phase change analysis has been
performed in two phases. Firstly, phase change material
experience change in its phase due to compartment high
temperature of 70oC followed by retaining previous state
by heat loss in PCM at temperature of 25oC. In the end
81
results of comparison of wax and sodium salt, i.e.,

sodium nitrate are presented.
TABLE II. COMPUTATIONAL DOMAIN DIMENSION
Tubes [mm]
A. Phase change materials
Two phase change materials, i.e., wax and sodium

nitrate were selected for comparison. The required
properties of both materials are presented in Table 1
below.
TABLE I.
Diameter
100
Length
1500
Thickness
10
Cylinder [mm]
Diameter
500
Length
1500
PHASE CHANGE MATERIAL PROPERTIES
Properties
Wax
Sodium
Nitrate
Unit
Thermal
conductivity
2.132
0.9
W/(m*K)
Heat capacity at
constant pressure
3.43
1.75
J/(kg*K)
Density
775
2260
kg/m^3
Ratio of specific
heats
Fig.1. Computational domain

B. Computational Domain
COMSOL Multiphysics has been selected for the

simulation of heat transfer in fluid. It uses a finite
element discretization over an unstructured grid.
Geometry selected for simulation consists of three tubes

which are placed in a large cylinder. Tubes contain
vegetable oil at 300oC while phase change material is
present in a large cylinder at 25oC. Dimensions of
computational domain selected for simulation are
presented in Table 2 and Figure 1.
C. Mesh Structure
Free tetrahedral meshing is used to create an

unstructured mesh with tetrahedral elements for the three
dimensional models of both phase change materials.
High mesh density was in regions where gradients of
temperature were expected to be higher based on the
physics of the problem. Statistics of the generated
meshes for both cases are shown in table 3 while Figure
2 depicts the mesh structure.
82
TABLE III. MESH STATISTICS
300o C. With time, heat is being stored in a phase change

material. Table 4 shows the conditions that are defined
in this case.
Property
Value
Minimum element quality
0.08798
Average element quality
0.6563
Equation
58988
Boundary
Condition
Patch
Tetrahedral elements
Triangular elements
11404
Initial values
All patches
T=To=25o C
Edge elements
1610
Tubes
T=To=300o C
Vertex elements
56
Temperature
1
Number of elements
58988
Temperature
2
Inlet and outlet of

large cylinder
T=To=25o C
Element volume ratio
0.00076
Symmetry
-n.(-
Mesh volume
0.2895 m^3
External surface of
large cylinder
Maximum growth rate
4.855
Average growth rate
1.916
TABLE IV. BOUNDARY CONDITIONS FOR CASE I
Case II: In this case vegetable oil is at room

temperature, 25o C while phase change materials are at
300oC transfer their restored heat to vegetable oil
retaining previous state. Table 5 shows the conditions
that are defined in this case.
TABLE V. BOUNDARY CONDITIONS FOR CASE II
Boundary
Condition
Patch
Equation
Initial values
All patches
T=To=25o C
Symmetry
Tubes
-n.(-
Temperature
1
Inlet, outlet and

external surface of
large cylinder
T=To=300o C
Fig.2. Free tetrahedral meshing

E. Governing Equations
D. Boundary Conditions
Heat transfer with phase change analysis has been

performed in two directions. Phase change material
experiences change in state while in second direction it
retains its previous state.
Governing equations which define the physics of this

problem are given below;
cp
T
c p u.T .(k T ) Q Qvh W p
t
k k phase1 (1 ) k phase 2
Case I: Phase change materials, i.e., wax and sodium
salt is maintained at room temperature of 25o C while
vegetable oil which is flowing through the tubes is at
83
c p c p. phase1 (1 )c p. phase 2 L
d
dT
phase1c p. phase1 (1 ) phase 2c p. phase2

c p. phase1 (1 )c p. phase 2
Wp
In Eq. 1 shows material density,
shows work
Qvh
shows stored latent energy, Q is the stored

heat, T shows temperature difference. In Eq.2 k
input,
shows the thermal conductivity,
cp
b.
Case I: error in non linear solver of wax phase

change material
is the specific heat at
constant pressure and shows the direction for phase

change.
F. Solution and Results
Relative tolerance for both cases is 0.01 while results are

computed using direct solver PARDISO to obtain results
of heat transfer with phase change in COMSOL
Multiphysics utility.
c.
Case II: error in non linear solver of sodium nitrate

phase change material
d.
Case II: error in non linear solver of wax phase

change material
Error plots for both the cases are shown in figure 3.
Fig.3. Relative Error

a.
Case I: error in non linear solver of sodium nitrate

phase change material
Case I: Fig. 4 & 5 shows charging of PCM, i.e., heat is

transferred from the hot medium and stored in PCM.
Figure 4 shows the temperature rise of sodium nitrate
from room temperature to high temperature after a time
interval of every 15 min and the whole outer cylinder
that contains sodium nitrate is charged and is at 550 K in
84
105 minutes and stored 8.5 KJ/kg of energy. Similarly,

figure 5 shows charging of wax at the same time interval
as for sodium nitrate. Wax is charged faster, i.e., 45
minutes as compared to sodium nitrate and stored energy
8.8 KJ/kg. Similar amount of energy was released by
both materials during discharging
Fig.5. case I temperature surface plot of wax
Case II: A Fig. 6 & 7 show discharging of the PCM, i.e.,

stored heat in the PCM is transferred to the cold region.
Figure 6 shows discharging of Sodium nitrate after a
time interval of 15 min and shows that oil traveling in
the tubes is heated after 90 minutes. Similarly, figure 7
shows discharging of wax at same time intervals and in
45 minutes complete oil is heated. Figure 8 & 9 shows
how sodium nitrate and wax is converted back to solid as
heat is released respectively. As time passes PCM is
converted back to solid and liquid form shown by red
cylinder is reduced with time and after complete melting
blank surface is present.
Fig.4. case I temperature surface plot of sodium nitrate
85
Fig.6. case II temperature surface plot of sodium nitrate

Fig.8. case II isothermal contours of sodium nitrate
Fig.7. case II temperature surface plot of wax
86
Constant temperature in last 15 min shows that heat is

being stored in the form latent heat thus temperature
remains the same.
Fig.11 shows the charging profile of WAX compared to

the graph of sodium nitrate shown in figure 10 wax is
charged faster in 45 min and a constant temperature
shown in the graph in the next time intervals. This shows
more heat is stored in wax.
Fig.9. case II isothermal contours of wax

Fig.11. Charging Profile of WAX
VIII. PRACTICAL EFFECT ON THE VEHICLE IN USE
This system comprises of an insulated case having PCM

in it and pipes for the flow of air and will be placed on
the dashboard of the cabin of automobiles. Flow
mechanism attached to it will suck the hot air from the
cabin and extracts heat from it and release cold air back
in the cabin. After reaching the desired temperature flow
mechanism will stop and no further air heat will be
extracted or transferred into the cabin. This stored heat
in PCM can then be dissipated into the atmosphere.
IX.
Fig.10. Charging Profile of Sodium Nitrate
CONCLUSIONS
Simulation performed and the results obtained shows

that wax is charged and discharged faster as compared to
sodium nitrate, i.e., in almost half of the time taken by
sodium nitrate. Charging graphs shows that after
reaching their melting point both store further energy as
latent heat. In case of discharging time taken by wax is
87
same as for charging but in case of sodium nitrate it

reduces to 90 min from 105 min. This shows that it can
be extremely effective to be used in Automobiles for
cooling purposes.
ACKNOWLEDGMENT
We would like to thank the School of Mechanical and

Manufacturing Engineering & Centre for Energy System
for their continues support and for providing students a
platform to apply their knowledge gained during their
professional studies
References
[1]
K. Mahmoud, E. Loibner, B. Wiesler, C.

Samhaber, And C. Kumann, "SimulationBased Vehicle Thermal Management SystemConcept And Methodology," SAE Technical
Paper 0148-7191, 2003.
[2]
A. Mezrhab And M. Bouzidi, "Computation Of

Thermal Comfort Inside A Passenger Car
Compartment," Applied Thermal Engineering,
Vol. 26, Pp. 1697-1704, 10// 2006.
[3]
H. O. Nilsson, "Thermal Comfort Evaluation

With Virtual Manikin Methods," Building And
Environment, Vol. 42, Pp. 4000-4005, 2007.
[4]
J. Jeon, J.-H. Lee, J. Seo, S.-G. Jeong, And S.

Kim, "Application Of PCM Thermal Energy
Storage System To Reduce Building Energy
Consumption," Journal Of Thermal Analysis
And Calorimetry, Vol. 111, Pp. 279-288, 2013.
88
SAME-13
Energy Consumption in Orthogonal Machining of Al 6061-T6 Alloy

Salman Sagheer Warsi1, Syed Husain Imran Jaffery1, Mushtaq Khan1, Riaz Ahmad1, Sohail Akram1
1
School of Mechanical and Manufacturing Engineering (SMME), National University of Sciences and Technology
(NUST), Sector H-12, 44000, Islamabad, Pakistan
salman.warsi@smme.nust.edu.pk
Abstract Recent researches in machining have

revealed that electricity consumption of machine tools
accounts for 90% of their environmental impact.
Therefore, minimization of energy consumption will
yield both environmental and economic gains. The
research work presented here focuses on the analysis
of energy consumption in the machining of aluminium
alloy Al-6061 T6, which is one of the widely used
materials in manufacturing industry. Full factorial
experimental plan was developed and performed for
orthogonal machining of Al-6061 T6 to investigate the
effect of machining parameters (cutting speed and
feed) on the energy consumption and energy efficiency
of the machining process. All the experiments were
performed by unused inserts, therefore, tool wear
effect was not considered for power calculations. The
analysis of the results revealed a strong correlation
between the energy consumption and the machining
parameters (feed and speed). The Material Removal
Rate (MRR) also showed good correlation with
different cutting parameters and power consumption of
the machine tool.
Key words Cutting parameters, Energy
consumption, Energy efficiency, Full factorial
experiments, Orthogonal Machining
XV. INTRODUCTION
Despite
many
recent
developments
in
manufacturing technology, machining holds a
prominent place as the most widely used production
process [1] and reduction in environmental impact of
machining processes is expected to yield considerable
gains for the betterment of environment. Therefore,
various efforts can be found in literature to explore the
numerous perspective of machining such as;
environmentally conscious process planning [2],
evaluation of energy consumption data of machine tool
[3], energy efficient controls of machine tool [4],
standardized energy consumption measurement [5],
evaluation of combined efficiency of a machine shop
[6], energy savings through enhanced utilization of

cutting tools [7]. Two important aspects in the
environmental analysis of the machining operations are
the modelling of the energy consumption during
machining and optimisation of machining operation.
Modelling of energy consumption is currently being
investigated at machine tool level, component level and
system level [8].
Mechanistic models are developed by applying well
understood physical principles [9]. The mechanistic
modeling of the energy consumption was initially done
by using exergy framework [10] and [11]. The
proposed framework aimed at explaining the power
consumption ( ) in machining, as a function idle power
( ) of machine and a product of specific process
energy ( ) and process rate ( ). The exergy framework
model was modified by dividing the total power
consumed in machining in to three parts namely; the
constant power, the power to drive spindle and
worktable during cutting and the power to
accelerate/decelerate the spindle [12]. The above
mentioned two approaches were blended and altered to
present a more detailed energy consumption model
[13]. Total energy in a single pass turning operation
was modelled using the approach similar to the
traditional machining cost modelling [14]. Energy
consumption was related with the G and M codes used
in the CNC programming[15]. This methodology was
furthered by incorporating the machinability of
workpiece in the energy model[16].
A machine tool is a complex system that is
susceptible to losses which cant be explained through
well-defined mechanical theories. In order to overcome
these complications, empirical modelling of energy and
efficiency of machine tools was also explored by the
researchers. Model for efficiency of a vertical milling
machine tool was statistically developed through
response surface methodology [17]. The relationship
between material removal rate (MRR) and specific
energy consumption (SEC) was statistically modelled
using response surface methodology [18]. A mixed
89
energy modeling approach was proposed in [19] that

comprised of mechanistic and empirical modeling. The
impact of depth of cut on specific energy consumption
was statistically modelled in [20]. Specific energy
consumption as a function of material removal rate
during turning operations was empirically modeled
[21].
The trend towards sustainability and environment
has also influenced the optimisation studies and
sustainable objectives like energy/power minimisation
and minimisation/elimination of cutting fluid can be
found to be incorporated in these studies. Therefore,
another emerging approach towards minimization of
energy consumption in the machining processes is
through optimization of cutting conditions. Design of
experiment (DOE) techniques like Taguchi methods
and Response Surface methods (RSM) along with other
multi-objective optimization methods like grey
relational analysis (GRA), composite desirability
function analysis have often been used in optimization
of machining parameters (speed, feed and depth of cut)
to minimize energy and power consumption in a
machining process.
The present research analyses the energy
consumption during the orthogonal machining of
aluminum alloy Al-6061 T6 at high cutting speed of
1000 m/min and feed rates up to 0.4 mm/rev. The effect
of machine parameter settings on efficiency of the
machining process has also been analyzed and
optimum machining conditions have also been
proposed.
XVI. EXPERIMENTAL DETAILS
Orthogonal machining was performed in this study
and for this purpose Al-6061 T6 pipes of 200 mm
diameter having wall thickness of 4mm were used as
the workpiece. Sandvik cutting tool SCACL 1616 K
09-S with uncoated inserts of CMW 09 T3 04-H13A
was used. In order to simulate orthogonal cutting,
cutting tool was fed in perpendicular direction to the
workpiece and the length of the cutting edge was
greater than the wall thickness of the workpiece. Full
factorial experimental plan was used to explore the
energy consumption pattern during the machining at
varied cutting speed and feed. Table 1 shows the range
of cutting parameters. Initially 16 experiments were
done based on these conditions. The results obtained
were analyzed and were further refined by performing
additional experiments. In total 70 different cutting
conditions were tested and the results were analyzed.
TABLE 1. Full factorial design of experiments

Parameters
Level
1
Level
2
Level
3
Level
4
Feed (mm/rev)
0.1
0.2
0.3
0.4
Cutting Speed
(m/min)
250
500
750
1000
Electric power was measured and logged through

Yokogawa Power Analyzer CW 240, which was
installed at the main bus of the machine. First power
measurements were made during the air cut which
didnt involve material removal (denoted as Pair) and
then the power was measured during the actual
machining operation (denoted as Pactual). The difference
of two powers provided the power specifically required
to remove material (denoted as Pcut). This cutting
power was divided by material removal rate (MRR) to
obtain specific cutting energy (SCE). Similarly, total
specific energy (TSE) was evaluated by dividing the
actual power by material removal rate. Finally,
efficiency of the machining operation was defined as
the ratio of SCE and TSE.
XVII. RESULTS AND DISCUSSION
Specific cutting energy and total specific energy
were computed and their relationship with the cutting
parameters were established.
A. Energy Consumption Trends in Machining
Total specific energy (TSE) and specific cutting
energy (SCE) were analyzed for different cutting
conditions and specific efficiency of the machining
process was evaluated. Fig. shows the trends followed
by TSE, SCE and s, when cutting speed was gradually
increased while maintaining the feed at a constant
value. The specific efficiency ( s) of the machining
process greatly improved while working at higher
cutting speeds as the contribution of the productive
SCE towards TSE gradually increased. It has been
observed that milling at high speeds reduced energy
consumption if subjected to the appropriate cutting load
[22]. In orthogonal machining, depth of cut was fixed
so the cutting load was gradually increased by
increasing the feed rate. For this purpose, cutting speed
was fixed at a high value of 1000 m/min and feed rate
was gradually increased from 0.1 mm/rev to 0.4
mm/rev as shown in Fig..
90
TABLE V. Average values of Responses (v= 1000m/min)

Speed
Feed
MRR
SCE
TSE
(m/min) (mm/rev (mm3/s) (J/mm3) (J/mm3)
)
Fig. 1. Energy consumption trends and specific

efficiency of machining process (feed fixed)
TSE and SCE consumption reduced as the cutting
load was increased consequently increasing the specific
efficiency (s) of the machining process. Balogun and
Mativenga (2014) pointed out that specific energy was
related to feed and the roughing operations should be
done at higher feed rates greater than the tool radius.
Fig. depicts that s generally improved when feed was
increased but the difference between the s at f=0.3
mm/rev and that at f=0.4 mm/rev was marginal. Error!
Reference source not found. presents the comparison
of MRR, SCE and TSE at these cutting conditions.
Fig. 2. Energy consumption trends and efficiency of

machining process (speed fixed)
Although material removal rate increased by 33%
when the feed was increased to 0.4 mm/rev from 0.3
mm/rev but SCE and TSE reduced just by 7% and 10%
respectively. This might be due to the fact that at such
high level of feed, MRR substantially increased but at
the same time machine drew more power to
compensate for this high feed which averaged out TSE
and SCE.
1000
0.1
7.11
0.63
1.26
1000
0.2
13.94
0.57
0.86
1000
0.3
20.58
0.55
0.77
1000
0.4
27.32
0.52
0.69
B. Main Effect Plots

The influence of cutting speed and feed on cutting
power (Pcut), specific cutting energy (SCE) and specific
efficiency (s) was further investigated through main
effects plots as shown in Fig. 3-5. The values of SCE
and Pcut were judged against the criterion of smaller the
better, whereas, s was judged against the larger the
better criterion. The two main effect plots in Fig. and
Fig. depicted an almost opposite trend. The smallest
value of Pcut was achieved with minimum cutting
velocity of 250 m/min and minimum value of SCE was
achieved at higher cutting speed of 850 m/min.
Minimum cutting feed of 0.05 mm/rev resulted in the
smallest value of Pcut and minimum value of SCE was
achieved at highest cutting feed of 0.4 mm/rev. These
results indicated that Pcut and SCE were contradictory
in nature and different settings of cutting parameters
were required for their minimization. For specific
efficiency (s), setting of cutting parameters at highest
levels has resulted in highest specific efficiency. The
results of mean effect plots highlighted the importance
of choosing the appropriate response in studies
pertaining to the environmental impact of machining.
Considering only cutting power as a response variable
would be misleading as it doesnt take into account the
amount of material removed and would result in setting
of cutting parameters at lower specific efficiency
levels.
91
Levels of cutting parameters identified through main

effect plots were used for confirmatory tests. The
setting of cutting parameters at lowest level would
result in low power consumption but the process would
not be efficient from environmental point of view,
therefore, these setting were not pursued further. For
minimum SCE, cutting speed should be 850 m/min and
feed should be 0.4 mm/rev. As discussed above, that
improvement in specific efficiency of machining
process was marginal when feed was increased from
0.3 mm/rev to 0.4 mm/rev, therefore, maximum feed of
0.4 mm/rev was disregarded and the next best level of
feed at 0.325 mm/rev was used for confirmatory trial
and it resulted in SCE of 0.5 J/mm3 that was indeed the
lowest value of SCE.
Fig. 3. Effects of Cutting Speed on Cutting Power and

SCE
Fig. 5. Effects of Cutting Speed and Feed on

Specific Efficiency (s)
XVIII. CONCLUSIONS
Following conclusions can be made from analysis of
energy consumption trend:
Fig. 4. Effects of Cutting Feed on Cutting Power
and SCE
1. Energy minimization was not achieved by just

rotating the workpiece at higher speeds,
92
however, when complimented with higher

cutting feeds significant energy reduction was
observed.
2. Maintaining high speeds while keeping the
feed at low levels had less pronounced effect
on efficiency as compared to high cutting
speed complimented with high cutting feed.
3. Specific efficiency (s) of the machining
process increased when the feed was increased
from 0.1mm/rev to 0.3 mm/rev. The rise in
specific efficiency then plateaued between the
feeds of 0.3 mm/rev and 0.4 mm/rev. This
little or no gain in specific efficiency on higher
feed rates pointed out that optimal result
would be achieved if the feed rate was kept
around 0.3 mm/rev. Moreover, cutting inserts
carry substantial cost and it might not be cost
beneficial to subject them to extreme cutting
environments for such a marginal gain in
efficiency.
4. SCE should be chosen as a response variable
in energy consumption analysis studies as it
takes in to account the material removed
XIX. REFERENCES
[1]
[2]
N. a Abukhshim, P. T. Mativenga, and M. a

Sheikh, An investigation of the tool-chip
contact length and wear in high-speed turning
of EN19 steel, Proceedings of the Institution
of Mechanical Engineers, Part B: Journal of
Engineering Manufacture, vol. 218, no. 8, pp.
889903, 2004.
P. Sheng, M. Srinivasan, and S. Kobayashi,
Multi-Objective
Process
Planning
in
Environmentally Conscious Manufacturing: A
Feature-Based Approach, CIRP Annals Manufacturing Technology, vol. 44, no. 1, pp.
433437, 1995.
[3]
a. Vijayaraghavan and D. Dornfeld,

Automated energy monitoring of machine
tools, CIRP Annals - Manufacturing
Technology, vol. 59, no. 1, pp. 2124, 2010.
[4]
P. Eberspcher and A. Verl, Realizing energy

reduction of machine tools through a controlintegrated
consumption
graph-based
optimization method, Procedia CIRP, vol. 7,
pp. 640645, 2013.

[5]
T. Behrendt, A. Zein, and S. Min,

Development of an energy consumption
monitoring procedure for machine tools, CIRP
Annals - Manufacturing Technology, vol. 61,
no. 1, pp. 4346, 2012.
[6]
Q. Wang, F. Liu, and C. Li, An integrated

method for assessing the energy efficiency of
machining workshop, Journal of Cleaner
Production, vol. 52, no. September 2009, pp.
122133, 2013.
[7]
F. Schultheiss, J. Zhou, E. Grntoft, and J. E.

Sthl,
Sustainable
machining through
increasing the cutting tool utilization, Journal
of Cleaner Production, vol. 59, pp. 298307,
2013.
[8]
Z. Yingjie, Energy efficiency techniques in

machining
process:
a
review,
The
International
Journal
of
Advanced
Manufacturing Technology, vol. 71, no. 58,
pp. 11231132, Dec. 2013.
[9]
D. C. Montgomery, Design and Analysis of

Experiments. John Wiley & Sons, 2006.
[10]
T. Gutowski, J. Dahmus, and A. Thiriez,

Electrical
Energy
Requirements
for
Manufacturing Processes, in 13th Cirp
International Conference on Life Cycle
Engineering, 2006, pp. 623628.
[11]
T. Gutowski, J. Dahmus, A. Thiriez, M.

Branham, and A. Jones, A Thermodynamic
Characterization of Manufacturing Processes,
in The 2007 IEEE International Symposium on
Electronics and the Environment, 2007, pp.
137142.
[12]
M. Mori, M. Fujishima, Y. Inamasu, and Y.

Oda, A study on energy efficiency
improvement for machine tools, CIRP Annals
- Manufacturing Technology, vol. 60, no. 1, pp.
145148, 2011.
[13]
V. Balogun and P. Mativenga, Modelling of

direct energy requirements in mechanical
machining processes, Journal of Cleaner
Production, vol. 41, pp. 179186, 2013.
93
[14]
M. F. Rajemi, P. T. Mativenga, and A.

Aramcharoen,
Sustainable
machining:
selection of optimum turning conditions based
on minimum energy considerations, Journal of
Cleaner Production, vol. 18, no. 1011, pp.
10591065, 2010.
[15]
Y. He, F. Liu, and T. Wu, Analysis and

estimation of energy consumption for
numerical control machining, Engineers,
Part B: Journal , 2011.
[16]
A. Aramcharoen and P. T. P. Mativenga,

Critical factors in energy demand modelling
for CNC milling and impact of toolpath
strategy, Journal of Cleaner Production, vol.
78, pp. 6374, 2014.
[17]
F. Draganescu, M. Gheorghe, and C. Doicin,

Models of machine tool efficiency and
specific consumed energy, Journal of
Materials Processing , vol. 141, no.
September 2002, pp. 915, 2003.
[18]
W. Li and S. Kara, An empirical model for

predicting
energy
consumption
of
manufacturing processes: a case of turning
process, Proceedings of the Institution of
Mechanical Engineers, Part B: Journal of
Engineering Manufacture, vol. 225, no. 9, pp.
16361646, 2011.
[19]
L. Li, J. Yan, and Z. Xing, Energy

requirements evaluation of milling machines
based on thermal equilibrium and empirical
modelling, Journal of Cleaner Production,
vol. 52, pp. 113121, Aug. 2013.
[20]
V. a. Balogun and P. T. Mativenga, Impact of

un-deformed chip thickness on specific energy
in mechanical machining processes, Journal of
Cleaner Production, vol. 69, pp. 260268,
2014.
[21]
S. Velchev, I. Kolev, K. Ivanov, and S.

Gechevski, Empirical models for specific
energy consumption and optimization of
cutting parameters for minimizing energy
consumption during turning, Journal of
Cleaner Production, vol. 80, pp. 139149,
2014.
[22]
O. I. Avram and P. Xirouchakis, Evaluating

the use phase energy requirements of a machine
tool system, Journal of Cleaner Production,

vol. 19, no. 67, pp. 699711, 2011.
94
ISBN: 978-969-8535-35-3

Complete Proceedings 18th Nov 2015

Загружено:

Сведения о документе

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

Complete Proceedings 18th Nov 2015

Загружено:

Авторское право:

Доступные форматы

Proceedings of the

Cylinder Liner Temperature Mapping of Heavy Duty Diesel Engine

Performance Improvement of Compressor to Recover Flare Gas by

Bulge hydroforming of tube by rubber mandrel without axial feed:

Perspectives in Automotive Embedded Systems from Manual to Fully

Optimization of TIG Welding Parameters using FULL FACTORIAL

Virtual Engineering: A key Enabler for Reconfigurable Manufacturing

Simulation of Turbulent Flow over a Single and Tandem NACA 0012

Passively flapping energy harvesting eel in a viscous flow

Design and Fabrication of Three-Dimensional Printing Machine

Development of Hybrid Mold casting for wheel Hubs

Comparative Study of Wax and Nitrate Salts as Thermal Storage for

Energy Consumption in Orthogonal Machining of Al 6061-T6 Alloys

Cylinder Liner Temperature Mapping of Heavy Duty Diesel Engine

Abstract In an internal combustion engine,

At the end of 20th century, developed countries of the

a heavy duty diesel engine. For this purpose eighteen

Figure 1: Engine Dynamometer Water Brake Type

timing. Some specifications of the engine are shown in

Diesel, four stroke, Vertical, Six

coolant manifolds from head as per manual instructions.

Bore & Stroke

Outer Diameter of Sleeve

Inner Diameter of Sleeve

Thickness of Wall of Sleeve

Each Hole Diameter

Standard engine oil

Table 2: Machining Specifications

A. Arrangement of Thermocouples (TC) on Selected

Engine was dissembled in order to have the cylinder

T TC: Thrust Side Thermocouple are on opposite of Fuel

center BDC so all points were marked and specified

Figure 2: Layout for Thermocouples Mounting

P3: 3rd TC from Top of Sleeve on Pump Side

Drilling process was done with CNC machine; 2.5mm

P2: 2nd TC from Top of Sleeve on Pump Side

Table 3: Thermocouple Layout

B2: 2nd TC from Top of Sleeve on Flywheel Side

Figure 4: Machining of Cylinder Liner

Figure 3: Layout block diagram

B. Instrumenting the Cylinder Liner with Thermocouples

Figure 5: Cold junction compensation

C. Data Acquisition System

Figure 6: Instrumenting the Cylinder Liner

displays a graph showing the real time data being

Figure 7: Software Development

Thermocouple B4 and B3 are at center position of liner

Figure 8: Fitting of engine on Dyno DX-33

Thermocouple P4 and P3 are at center position of liner

Thermocouple B2 and B1 are at BDC position and are

C. Temperature from Thrust side

Thermocouple T4 and T3 are at center position of liner

Thermocouple T2 and T1 are at BDC position and are

thermocouples were mounted, three in radial directions

Advanced data acquisition system was configured

Highest Temperature: Temperature is higher at

The same results could be verified by applicable

Performance Improvement of Compressor to Recover Flare Gas by Factors

Industrial Engineering Department, University of Engineering and Technology

Abstract- A large quantity of natural gas is flared in

The oil and gas sector of Pakistan is a key sector in

transfer to the plant area by its natural pressure for

economics evaluating the average flare gas is

flaring of gas is declined by 19% since 2005, which is