Вы находитесь на странице: 1из 63

SHIP RESISTANCE AND

POWERING
SD2710 INITIAL SHIP DESIGN
COURSE NOTES
v 1.0

KARL GARME
garme@kth.se
070-397 17 17
Stockholm January 2012

MARINA SYSTEM

CENTRE FOR
NAVAL ARCHITECTURE

TABLE OF CONTENTS
Table of contents........................................................................................................................................... 2
Ship resistance in calm deep waters ........................................................................................................... 3
Resistance basic equations and scaling laws ...................................................................................... 5
Viscous resistance................................................................................................................................... 10
Air resistance ........................................................................................................................................... 20
Appendage resistance ............................................................................................................................ 21
Non-ideal conditions .................................................................................................................................. 23
Shallow water .......................................................................................................................................... 23
Added resistance due to waves and wind ........................................................................................... 26
Fouling ..................................................................................................................................................... 28
Prediction of ship resistance and power requirements ......................................................................... 29
Systematic series and semi-empirical methods .................................................................................. 33
Standard ship propellers ........................................................................................................................ 38
Model tests ................................................................................................................................................... 52
Some systematic series of resistance data ........................................................................................... 55
References .................................................................................................................................................... 56
Appendix The exemple ship Protefs........................................................................................................ 57
Appendix Residual resistance coefficients .............................................................................................. 59
Appendix Propeller charts ......................................................................................................................... 62

SHIP RESISTANCE IN CALM DEEP WATERS


Resistance is all the energy transferred to the water from the hull and its appendages: rudders,
fins, bilge keels, propeller shafts and other installations. Resistance is clearly visible as waves and
eddies. The waves formed by the pressure variation generated by the moving vessel. Waves
would exist even if the ship was moving in a invicid fluid. Vortices in the boundary layer and in
the down stream wake, on the other hand, are viscous effects. Therefore, it is common and logic
to divide the resistance in the wave making resistance and viscous resistance. The relationship
between them is speed dependent and can, for a displacement vessel, appear as in Figure 1. At
low speed, the frictional resistance dominates while the wave making resistance prevails at higher
speed. For planing hulls, that are partly lifted out of the water as a result of the speed created
pressure, the trend of increasing wave making resistance can be broken and the friction drag
again becomes the dominant.

Figure 1. The relationship between the frictional drag and residual resistance (basically wave making
resistance). Figure from Byquist (1994).

Figure 2. Waves around a ship model. Student experiment in the GIH-swimming pool, spring-term 2005.

The vessels Figure 2-Figure 4 run in nearly calm water. Any disturbance of the water surface as
foam, eddies and waves is visible resistance.

Figure 3. Cruise ship in calm waters.

Figure 4. LNG tanker (Liquefied Natural Gas).

In ship design, it is important to appreciate the required installed power in order to meet the
contract trial speed (one of many specified requirements, sv. provtursfart) . The trials are carried
out off shore in calm weather. Talking about resistance in general such conditions are presumed.
The resistance, however, is of course strongly affected of the external circumstances: sea depth,
wind, wave and marine growth.
Model tests and semi-empirical methods are the means to estimate the ship resistance. Today is
also mathematical modeling (CFD) used frequently. The CFD is a great tool for comparing the
resistance of different hulls. Still, the correct resistance levels can not yet be calculated accurately,
besides calculations are made in model-scale and the scaling to full-scale remains.
In the procedure to scale model resistance data to full-scale, a number of resistance components
are expressed in the form of dimensionless coefficients as,
CT = (1 + k )CF + CR + CF + C AA

(1)

The coefficients are related to viscous friction and form drag, (1+k)CFS, wave making resistance,
CRS, roughness allowance, Cf, and air resistance, CAA. The resistance is calculated from,
R = CT

1
V 2 S
2

(2)

[ N]

where is the density of water, V the ship speed and S the ship wetted surface. Appendage
resistance is often treated as a separate term and can be added to the expressions (1) and (2).
Ship resistance is generally divided basically into these components, no exception here. Figure 5
shows examples of the relationship between the different resistance components for some ship
types. The ship types represent increasing speed from the slow-moving tanker where friction is
the dominating resistance to planing vessels where wave making resistance has been the largest
part.

Figure 5. Examples of the relative magnitude of resistance components for some types of ships. Figure
from Dyne (1991).

RESISTANCE BASIC EQUATIONS AND SCALING LAWS


Resistance, in everyday talk, on cars aerodynamic drag for instance, CD-values are often
mentioned. This is a dimensionless measure of the resistance as the force directed against the
direction of travel. The resistance coefficient is defined as,
C=

R
1
V 2 S
2

(3)

where R is the resistance [N], the density [kg/m3], V the body's speed [m/s] and S is a
characteristic area [m2] in case the resistance relates to a 3-d body. For a 2-dimensional body S is
a characteristic length [m] and R has the unit [N/m]. In equation (2) is thus the ship's wetted
surface the area of reference to the CD-value which is often indexed T indicating the total
resistance made up of several components.

Resistance is generally very difficult to calculate and CD-values are usually determined
experimentally or, for simple geometric shapes, taken from table works which, of course, once
were produced by systematic experimental series.
If the resistance coefficient is known the resistance can be determined by (3) and then the power
necessary for the body's forward motion, by,
P = R V

(4)

Figure 6 shows 2-d resistance coefficients as function of velocity in terms of the dimensionless
Reynolds number, Re. The CD-values relation to velocity is, as seen, far from linear. Especially
noteworthy is the rapidly decreasing resistance coefficient for the circular cylinder at Re
approximately 105. The flow changes character when the velocity is changing and we realize that
it is difficult to draw conclusions from conditions at one speed about the conditions when the
speed has changed a bit, thus it is not at all obvious how to scale a flow case. Note that the CDvalue decreases with increasing Reynolds number. This does not mean that the resistance
decreases. As you can see from Equation (3) the resistance is a function of the resistance
coefficient times the speed squared and the resistance generally increases with speed.

Figure 6. Two-dimensional resistance coefficients. From Massey (1990).

The explanation for the abrupt change in drag coefficient for the two-dimensional circular
cylinder is due to an abrupt change of character of the flow at approximately Re=105 in principle
as shown in Figure 7. In the separation zone downstream of the cylinder the flow is full of eddies
that require much energy and following the resistance is large. When the speed is approaching the
critical Reynolds number the separation point shift downstream and the wake becomes more
narrow and the resistance decrease. This change in flow pattern is due to fluid viscosity and the
shape of the body.

Figure 7. From Dyne (1991).

Look at the shapes in Figure 8. The resistance coefficient for the half-sphere, is nearly three times
greater when the curved side is facing downstream compared to if pointing upstream. Separation
initiates at sharp corners so the wake will look roughly the same in both cases. However, the
pressure on the upstream side will be significantly lower in the more streamlined case, reducing
the resistance considerably.

Figure 8. Three-dimensional resistance coefficients when 103 <Re <106. From Dyne (1991).

Example Parachute
Determine the diameter D, of a parachute used to ground a weight of 90 kg with a
maximum speed of 6 m/s. At Re>103, the parachute CD=1.3.
Solution:
Assume that the air density is =1.2 kg/m3 and kinematic viscosity =1.7 .10-5 m2/s and
the CD reference area is the projected area of the parachute to the ground, D 2 4
Set up the equilibrium equation and solve for the diameter:
R mg = 0

R
1

D
90 g 4

CD =
R = 1.3 1.2 62
= 6.3 m
D =
2
1
2
4
2
1
2 D

1.3
1.2
6
U
2

2
4

Check if the Reynolds number is ok: Re =

UD
6 6.3
=
= 2.2 106 >> 103 OK!
5
l
1.7 10

Resistance coefficients is just a dimensionless form of the resistance which must be calculated or
determined experimentally. The fluid mechanical problem describing the conditions can be
formulated with the continuity equation and an equation of motion. The continuity equation
represents mass conservation and the equation of motion is the momentum law (Newtons 2nd
Law). In ship hydro-mechanics these two are formulated with appropriate boundary conditions,
on a variety of ways depending on the simplifications that can be done. The momentum law is, in
its most general formulation, the Navier-Stokes equations. It consists of three equations and can
for incompressible flow, in vector form, and with all variables non-dimensional be expressed as,
U 2 Du
U2
U
2 2 u
gez
=
p +

L Dt
L
L
Trghetskrafter = Tryckkrafter + Visksa krafter Tyngdkraften

(5)

The equations are in principle the momentum law of the form,


mx&& = F

(6)

Dividing both sides of equation (5), with the factor of dimension [N/m3]1 ,

U2
L

(7)

we get a dimensionless equation, where the factor in front of each dimensionless term represent
the ratio between the term and the inertia force,
Du
Dt

Trghetskrafter
Trghetskrafter

Tryckkrafter
Trghetskrafter

O(1)

O(1)

2 u
UL
Visksa krafter
Trghetskrafter

O

UL

gL
ez
U2
Tyngdkraften
Trghetskrafter

(8)

gL
O 2
U

The equations are dimensionless with four unknown (velocities in three directions and pressure).
If two problems are geometrically uniform, and the dimensionless groups in the two last terms,
related to viscous forces and gravity, are equal, the equations become identical. The possibility of
scaling a model experiments thus lies in designing the model situation so that these dimensionless
groups are equal to their full scale values. The dimensionless groups representing the Reynolds
and Froude number which usually are written,
Re =

Fn =

UL UL

=
=
where

and

are
dynamic
and
kinematic
viscosity,

(9)

U
gL

1 The unit of the terms in equation (5) is [N/m3]. The equation is a partial differential equation demanding
boundary conditions. In case it had been expressed as an integral equation the terms unit would have been
[N], thus in principle the same as (5) but integrated over a volume.

If both Froude and Reynolds numbers are equal in two geometrically equal situations, we have
achieved dynamic similarity (sv dynamisk likformighet). When designing model experiments
dynamic similarity is the ideal. However, it is impossible in practices and one has to chose which
of the numbers that are most relevant for the situation to model. In a homogeneous medium
without free surface, waves do not occur, and dynamic similarity is achieved if the Reynolds
number is equal in model and full scale. However, for experiments in water with surface piercing
vessels, both the Froude and Reynolds numbers should equal their corresponding full-scale value
to really prevail dynamic similarity. Since this can not be achieved, the choice is to obey the
Froude model law, i.e. Fn is equal in model and full scale. The viscosity of water is very low and
the viscous forces are of secondary importance except in the absolute vicinity of the hull (or any
other solid boundary). Unfortunately, that is where the frictional forces arise which are a great
part of the total resistance. This means that the viscosity can not be neglected in resistance
calculations. Still, the error becomes quite small if viscous forces are neglected when calculating
the pressure force on the hull. The pressure control wave formation and vessel attitude. If the
Froude model law is obeyed the ship's attitude, wave generation and motion will be directly
scalable through that model law, but how the resistance, also a function of the Reynolds-number
dependent viscous forces, should developed to full scale is by no means unambiguous. If we, all
the same, assume that viscosity is irrelevant that term drops from the Navier-Stokes equations,
(5). The equations can in that case, if one integrates along a streamline, be expressed by the
Bernoulli equation for steady-state conditions,
1
u 2 + p gz = konstant lngs strmlinje
2

(10)

The Bernoulli equation is thus three equations of motion with four unknowns. The continuity
equation gives the additional equation for the unknowns to be determinable (just as when the
Navier-Stokes equations describe the motion).

Figure 9. (a) pressure distribution for invicid flow, (b) invicid flow, (c) viscous flow with boundary layers
without separation, (d) viscous flow with a boundary layer with separation. Figure from the PNA, (Lewis
1988).

In invicid flow, there is no separation and no vortices are present. The flow around a body could
in that case look like (b) in Figure 9. Consider the shape and think that the water flow approaches
9

from the right. The continuity equation states that the same amount of water coming in to a
constant control volume also has to leave it. For this to be true, the flow rate has to increase as
the control volume becomes narrower, ie. when the body becomes thicker. If the velocity
increases the pressure must decrease according to the Bernoulli equation (10). The pressure
distribution for the invicid case is shown in (a). If the pressure is integrated for this symmetric
case, one finds that the lift as well as the resistance becomes zero. If however, the body would be
asymmetric or the flow oblique, lift would generated. The resistance on the other hand would still
be zero and is always zero for bodies fully submersed in an invicid fluid. This is called
dAlamberts paradox and clearly shows that water can not be considered invicid when resistance
is concerned. The images (c) and (d) in Figure 9 illustrates viscous flow at slightly different aft
body shapes. In the boundary layer (sv grnsskikt) and in the wake the flow dominates by viscous
effects and it is here the resistance arise.
Friction is the shear stress between the body and the fluid. Shear stresses do not appear explicitly
in the Navier-Stokes equations, but are proportional to velocity with the viscosity as the
proportionality constant. Shear stress, , in the direction of the flow along a flat plate can be
expressed,
=

(11)

du
dy

where u is the velocity component in the x direction and is the dynamic viscosity. Figure 10
defines x and y. The friction force can be expressed, for a plate of width B and length L,

Ff = B

du

dy
L

(12)

dx
y =0

Figure 10. Velocity profile for an invicid (left) and viscous fluid for the flow along a flat plate.

VISCOUS RESISTANCE
The basis for determining the viscous resistance has traditionally been the friction force on a flat
plate, of the same length as the vessel, area equal to the vessel wetted surface, and with incident
flow parallel to the plate with velocity equal to the ships speed. The result is corrected with
respect to the vessel shape by a form factor. A roughness allowance is also calculated to get hold of the
resistance increase due to the ship steel or other construction materials are not completely
smooth, and certainly not as smooth as the models used in tank tests.

Plate skin friction & form resistance


In Figure 11, a flat plate is approached by a viscous fluid, such as water, with a homogeneous
velocity field upstream of the plate. On the plate the water particles have no velocity following
the no-slip condition (sv. vidhftningsvillkoret). Far out from the plate the particles with move
with the free stream velocity completely unconcerned by the plate's existence. Between these
10

extremes, in the boundary layer, particle velocity varies as illustrated to the right in Figure 10.
Boundary layer thickness, , defined as the distance along a normal vector from the surface to the
point where the velocity is 99% of the free stream velocity. The flow in the boundary layer is
controlled by viscosity, which we associate with the Reynolds number. As the boundary layer
changes along the plate length the choice of characteristic length to define the Reynolds number
is not obvious. Commonly, a local Reynolds number, Rex, is used where the distance from the
plate's leading edge is the characteristic length. At low local Reynolds number the boundary layer
flow is laminar. When the Reynolds number increases the boundary layer flow becomes
turbulent. At the transition point (sv omslagspunkten), where the character of the flow changes,
and down stream in the turbulent boundary layer the boundary layer thickness grows faster than
in the laminar part, see Figure 11.

Figure 11. Principle sketch of the boundary layer thickness as a function of the local Rex and the boundary
layer flow character. Figure from Dyne (1991).

The resistance will be larger if the boundary layer is turbulent, partly because to the frictional
force becomes larger (shear stresses on the surface, note the difference in velocity profile, Figure
12) but also because a thick boundary layer increases the pressure resistance (see form resistance).

Figure 12. The principal difference between the velocity profiles of laminar and turbulent boundary layers.
Figure from Massey (1990).

When the transition occurs depends on the local Reynolds number and also on the surface
roughness. The transition does not happen at exactly at the same Reynolds number across the
plate but initiates at the local unevenness. The Reynolds number at transition is called critical
Reynolds number and can typically be
Re x = 5 105

(13)

11

The difference between the friction coefficients for plates with laminar and turbulent boundary
layer is seen in Figure 13.

Figure 13. Friction coefficients for flat plates with laminar and turbulent boundary layers. Figure from
PNA Vol. II (Lewis 1988).

Ships are so long that the laminar part of the boundary layer is negligibly short. But when
frictional resistance is determined by model tests the boundary layer flow has to be forced to
turbulent in order to get as similar situations as possible. Models are so short and so smooth that
a disproportionately large part of the boundary layer would otherwise be laminar. Considering
sailboat rudders and fins, a large part of the boundary layer can be laminar and the resistance can
be minimized by maximizing the laminar part.
Plate skin friction can be approximated by the formula recommended by the ITTC-57,
CF =

0.075
(log10 Re 2)

(14)
2

where the Reynolds number is based on ship length.


Form resistance is a correction dealing with the fact that the viscous resistance is greater for a real
hull than for a flat plate of the same length and wetted surface as the vessel. The form resistance
is expressed by the form factor k as was introduced in the first term of equation (1),
(1 + k )CF

(15)

When water flows along a curved surface the boundary-layer growth is different from the plate in
Figure 11. Flow along a surface with negative pressure gradient, typically from the bow to the
maximum width of the hull, see Figure 14, move the transition point down stream, which is
mainly relevant to models, rudders or fins, and holds back the boundary layer growth (regardless
if laminar or turbulent). This effect is beneficial from a resistance point of view. A positive
pressure gradient has the opposite effect. A large positive pressure gradient due to small radius of
curvature can lead to separation (sv avlsning) and reverse flow. This happens in the stern
resulting in resistance increase.

12

Figure 14. Typical boundary layer along a water line. Figure from PNA Vol. II (Lewis 1988).

The overall effect of a curved surface is increased resistance. Due to the shape,the flow velocity is
larger than the free stream velocity from approximately the forward shoulder (sv frliga skuldran)
to the stern shoulder, which means larger friction compared to the flat plate. Boundary layer
thickness is also growing much faster in the stern, partly because the increasing pressure but also
because the width of the wet surface decreases. Boundary layer thickness affects the pressure
surrounding the ship, see Figure 15. The symmetry of the pressure with stagnation pressure at
the bow and stern that we saw in the friction-free flow (leading to dAlemberts paradox) is
becoming more and more asymmetrical with lower and lower pressure in the stern the larger the
boundary layer becomes. Hence, pressure resistance increases.

Figure 15. Illustration of the boundary layer thickness effect on pressure distribution. The top half of the
picture a) shows the pressure distribution and flow patterns in the ideal fluid. The lower when the viscosity
affects the flow. Figure from Byquist (1994).

Pressure change with depth means that the flow will not be in parallel with the water line plan.
This becomes especially clear in the bow and stern, resulting in vortex shedding which adds
resistance.
Form resistance is geometry dependent and can therefore to some extent be affected. Large
radius of curvature at the shoulders and a bulbous bow is advantageous. V-shaped rather than Ushaped stern sections (see Figure 16) can be beneficial. Pronounced V-shape at the stern
minimizes the resistance of the bare hull but single screw ships this leads to unfavourable
propeller conditions and stern design requires special considerations. Thus, optimal bare hull
resistance does not necessary result in the minimum power required because the flow depends on
propulsorn and hull interaction.

13

Figure 16. The figure shows the fundamental difference between U-and V-shaped frames. Figure from
Dyne (1991).

Calculating the form factor


The form factor k, is preferably determined by model experiments and the Prohaska method which
is based on low speed data when the wave making resistance is small. Following the theoretical
based assumption that the total resistance coefficient at low speed (Fn<0.1) can be written,
CT = (1 + k )CF + konst ( Fn )

(16)

the ratio between total and friction resistance coefficients is a linear function of the ratio of the
Froude number to the power of four and the friction coefficient,
CT
Fn 4
= (1 + k ) + konst
CF
CF

(17)

By plotting the ratios against each other, one can determine the form factor from the graph, see
Figure 17.

Figure 17. Prohaska-plot. Figure from Dyne (1991).

If you want to estimate the form factor lacking model test data, Watanabes (Dyne 1991)
approximate expressions for displacement vessels can be used,
k = 0.095 + 25.6

CB
L
B

(18)
B
T

The form factor is under debate. Because it embraces both pressure and viscous resistance it is
hardly independent of scale although it is normally used as if it was. It is also handled very

14

differently for different types of ship and at different test facilities, see for instance the discussion
in the ITTC (2005). For planing hull forms it is common to set k=0.

Roughness allowance
To the viscous resistance one can also count the third term in equation (1),
C F

(19)

known as roughness allowance (sv ytrhetstillgg). All formulas presented are more or less products
of experiment data. The roughness allowance term is there to compensate for the difference in
roughness between models and full-scale ship. In the re-scaling procedure of model test results to
full-scale ships, the following formula is used (by Bowden and recommended by the ITTC
(1978));
1/ 3

k
C f = (105 s
L

(20)

0.64) 10 3

where ks is a measure of the roughness with unit [m]. ITTC recommends ks=150m.

WAVE MAKING RESISTANCE


The energy continuously added to the wave system around the vessel a direct link to the wave
making resistance. Also near-surface underwater bodies, foils, fins or submarines generate surface
waves. The wave making resistance decreases for those cases with the distance to the surface.
The energy per unit area of the waves has a direct relation to the wave amplitude,
E=

1
g a 2
2

(21)

and large wash from a ship simply means large resistance.


The wave system is created by the pressure distribution on the submerged part of the ship.
Because of the dispersive nature of the surface wave, where the wave propagation velocity is
frequency dependent, waves of different wavelengths interfere to a particular wave pattern. The
wave system looks as in Figure 18, with transverse and diverging waves. The transverse waves
propagate with the speed of the ship i.e. the phase velocity is equal to the speed of the vessel,

c=


= =
k

2 g
2
,k =
=

(22)

g
=U
2

The divergent waves propagate with different angles with respect to the centreline of the vessel
but the over all pattern of both transverse and diverging waves confine within an area 19 28
(19.47) off the centreline.

15

Figure 18. Wave pattern in deep water. Figure from Dyne (1991).

The pressure distribution on the hull, which gives rise to the waves is a function of vessel shape
(length not the least) and speed. The shape affects pressure distributions appearance, and
initiates wave propagation. The speed determines the propagation speed and wavelength. Waves
developed along the entire ship but waves due to high pressure in the bow and stern dominate
the resulting pattern, see the principle in Figure 19.

Figure 19. Basic pressure distribution and wave formation along a surface piercing body. Figure from
Byquist (1994).

The result is that the waves sometimes heighten and sometimes weaken each other depending on
the speed of the vessel. The wave amplitude in the resulting pattern becomes more or less high,
and following varies the also the resistance. Consider Figure 20. The 4.88-meter long ship-shaped
body has a speed corresponding to Fn=0.238 which means a speed of 1.65 m/s and according to
Equation (22) waves with wavelength 1.74 m. If speed is increased the wavelength increases and
vice versa. You realize that the wave systems in the figure, if the speed is varied, sometimes will
be reinforcing with a resistance peak as a result (hump) and sometimes partially extinguish each
other, resulting in a local resistance minimum (hollow). Figure 21 shows an example of a typical
wave resistance curve. It also shows that the interaction effect, in particular, is connected to the
transverse wave system.

16

Figure 20. Interaction of waves formed by a ship-like shaped model. Figure from PNA Vol. II (Lewis
1988).

Figure 21. Wave making resistance from the transverse and diverging waves. Figure from PNA Vol. II
(Lewis 1988).

As we have seen the boundary layer affects the pressure distribution around the vessel, especially
in the stern, and there is thus dependence between the viscosity and the wave generation. Still, it
is important to remember that the waves would occur even in an absolute invicid fluid. However,
in experiments, effort is made to model the boundary layer in order to receiving as accurate wave
formation as possible. The resistance measurement is often called residual resistance, which
explains the indexing of the second term in (1), CR.
The wave making resistance is determined by model tests or calculated using empirically based
methods. Since the wave formation essentially follows potential theory, it is likely that CFD
calculations works well even quantitatively for this resistance component. One, on the other
hand, purely viscous resistance contribution, that has to do with wave generation is often called
wave breaking resistance in the literature. If the waves around the bow become so large that they
break, energy disappears from the resulting wave system. Breaking waves does not follow linear

17

wave theory and escapes the formed wave pattern. Nevertheless, the vessel continuously adds
energy to the breaking bow wave, which means resistance. With appropriate bow design a major
unstable bow wave can be avoided. Forebody shape can affect the resistance in several ways. By
forming the bow with a bulbous bow is the resistance in many cases reduced. The bulbous bow
creates, at best, bow wave of its own, which weakens the real bow wave, i.e. decrease pressure
and waves. The bulbous bow can also by changing the bow wave reduce wave breaking resistance
and reduce the form resistance (see viscous drag) because the lines in the bow can be made
softer. The bulbous bow shall be located just below the water surface, it may otherwise, especially
at low speed, increase resistance.
The wave making resistance can in principle be reduced by minimizing the Froude number,
Fn =

(23)

gL

which requires a lower speed or greater length. If the vessel are made slimmer (if stability allows
and the cargo can be placed aboard), the pressure distribution would become smoother with less
pronounced wave formation.

18

Example Humps & Hollows and the maximum speed of displacement ships
Determine the speeds, expressed in Froudestal, Fn, when humps and Hollows occur.
Also, determine an expression for the maximum speed of a displacement vessel.
Assume that the water is deep and that the transverse wave system generated by the
moving ship is the sum of a bow and a stern wave. Let the distance between the waves
origins be L, see figure.

Figur from Byquist (1994).


Solution
Wave propagation velocity: c f =

(if depth> ).
2
2

For the transverse waves: cf= ship speed U and =


The stern system is pronounced when =

L
n

2 U 2
.
g

n = 1,2,3,... and weakened =

Maximum wave resistance:

L 2 U 2
1
Fn =
=
Fn 2 =
n
g
2 n

Minimum wave resistance:

2L
2 U 2
Fn 0.32, 0.25, ...
=
(2n + 1)
g

2L
.
(2n + 1)

1
0.40, 0.28, ...
2 n

The wave resistance increase dramatically when L . Seeing this condition as a


measure of the maximum speed of displacement ships giving:
=L=

2 U max 2
U max =
g

Umax in knots: U max 1.25

gL
1.25 L [m/s]
2

3600
L 2.43 L [kn]
1852

The table shows the approximate maximum speed for a number of different waterline
lengths. For L 3-7 m, Umax end up pretty close to what the rule of thumb that says,
namely that the maximum speed in knots is equal to the waterline length in metre.
Waterline length [m]
3
5
7
10

Umax [kn]
4.2
5.4
6.4
7.7

19

AIR RESISTANCE
Air resistance means the resistance of air to the ship's superstructure due to the speed of the
vessel. Depending on the superstructures form, the resistance coefficient can vary significantly,
roughly,
0.5 < CD < 1.2

(24)

If the CD is known (e.g. from wind tunnel tests), resistance can be calculated as,
RAA =

1
AV 2 CD AT
2

(25)

where A is the air density and AT is the projected frontal area of the superstructure (Figure 22).

Figure 22. Projected frontal area AT. Figure from Dyne (1991).

By making (25), dimensionless with the water density and the ship's wetted area, the air resistance
coefficient is expressed as,
C AA =

RAA

A
= A CD T
1

S
V 2 S
2

(26)

The ITTC recommends the following expressions which comes from assuming CD=0.85 and a
ratio of the densities to 1/850,
C AA = 0.001

AT
S

(27)

Air resistance is a minor part of the total resistance, about 2-5 % in calm water, Molland et. al
(2011). However, including the effect of wind the share can be about 10% for oceangoing car
carriers and container vessels. Ongoing studies at KTH Centre for Naval Architecture (in cooperation
with Wallenius, SSPA and Starcs experimental aerodynamics), Figur 23, shows on a potential for 4050% air resistance reduction for those types of vessels meaning a fuel save of several tones a day.

Figur 23. A PCTC (Pure Car and Truck Carrier) at sea and investigated in wind tunnel.

20

APPENDAGE RESISTANCE
Appendages are all that stands out from the hull, such as rudders, bilge keels, stabilizing fins,
propeller shafts and logs. It can be difficult to determine the appendage resistance by model tests
due to scale effects. One can imagine that the model (physically or mathematically) appendixes
and hull separately but a difficulty is to know the flow conditions at the appendixs position on
the vessel. Equations (1) and (2) expresses the resistance components of the bare hull: viscous
friction and form resistance, residual resistance, surface allowance and air resistance. Appendage
resistance is estimated for the various appendages and added to the bare hull resistance.
For some appendages, such as bilge keels, there exist simple formulas for estimating their
resistance contribution. In Peck's formula (PNA, Lewis 1988), (28), is resistance supposed to
consisting of the effect of increased wetted surface and on adverse effects on the flow by the
presence of the bilge keel. In case the bilge keel is a flat plate the two resistance contributions are
equal.
RBK =

1
2z
V 2CF (2
) S BK
x+ y
2

(28)

where SBK is bilge keel (sv. slingerkl) wetted surface. The variables x, y, z and L are defined in
Figure 24. L is the bilge keel medium length and should be used in the calculation of plate
friction coefficient CF. Bilge keels should be placed so that they follow the stream lines. This can
be difficult to achieve, not least in view of load condition variations.

Figure 24. Bilge keel dimensions to Peck's formula. Figure from PNA vol. III (Lewis 1988).

On fast ships propeller shaft often stands out at an angle through the hull. This is a necessity to
get the propellers down to sufficient depth since the hull depth of the highest speeds is very
small. The disadvantage is resistance increase due to the oblique flow at the shaft and shaft
supports. Shaft resistance, according to Dyne (1991) is expressed as,
1
V 2 (1.1 sin 3 + C f ) Laxel d axel
2
1.327
Cf =
if Red < 105
Re d
Raxel =

(29)

1
V 2 (0.6 sin 3 (2.25 ) + C f ) Laxel d axel
2
1
1700
Cf =

if Red > 5 105


2
( 3.46 log (Re ) 5.6 ) Red
Raxel =

10

(30)

The PNA shows slightly different expressions and also versions for the intermediate Reynoldsnumber range. The PNA also proposes expressions to calculate the appendage resistance of the
rudder and fins for instance the Peck and Hoerners formulas. In Holtrop & Mennens semi
empirical method appendage resistance is determined with the formula and coefficients in Figure
25. To calculate the appendage resistance in this way is at least as reliable as using the formulas of
the PNA. If higher accuracy is required model tests or a combination of CFD calculations and
model tests is needed.

21

Figure 25. Appendage resistance as calculated in Holtrop & Mennens semi-empirical method. From
International Shipbuilding Progress, Vol. 29, 1982.

It is worth noting the large difference between the factors (1+k2) for rudder, rudder behind stern
and stabilizer fins in the table in Figure 25. The latter is located in relatively undisturbed flow, while
the rudder is in the wake. The rudder can be seen as the aftmost end of the hull, or the flat plate,
where the local friction coefficient is relatively low. The relationship between these factors is
shown clearly when comparing Cf values calculated from the Reynolds number based on vessel
length and the fin/rudder chord length.
When using formulas as the above or others that can be found in the literature, one must be
extremely cautious on the conditions of applicability.

22

NON-IDEAL CONDITIONS
SHALLOW WATER
Water depth affects resistance, in particular the wave making resistance. Limited depth also
makes the displaced water pass the hull at higher speed than otherwise. The effect is lower
pressure and following increased draught, changes in trim and increased resistance as a result.
The change in running attitude is called squat. Proximity to canal sides, docks or other vessels
causes the same effect. In narrow passages a ship can be sucked towards a side wall or hitting the
sea floor, see Figure 26.

Figure 26. Schematic change of pressure distribution as a ship comes close to a boundary. From Dyne
(1991).

The wave formation change around the hull at speeds characterised by Fnh1, Froude number
based on water depth. In particular the trim growth can be seen as an effect of this, see Figure
27.

Figure 27. Illustration of squat change of trim and draught around the critical speed Fnh=1. From Dyne
(1991).

When is the water deep? In linear wave theory the wave potential can be expressed as,
=

g a cosh(k ( z + h))
sin(kx t )

sinh(kh)

(31)

where h is the depth, a the wave amplitude, k the wave number, frequency, x and z coordinate
directions in or normal to the horizontal. Inserted in the linearized free-surface condition (the
combination of the kinematic and dynamic conditions) at z=0,
23

2
t

+g

=0
z

(32)

the wave propagation velocity, the phase velocity (cf= /k), can be identified,
cf 2 =

g
tanh(kh)
k

(33)

The energy propagation velocity, the group velocity (cg=d/dk), is expressed as,
1

kh
cg = +
cf
2 sinh(2kh)

(34)

and the limits for the deep and shallow water will follow (e.g. using lHpitals rule),

g
cf =

k , d h

1
c = c
g 2 f

c f = gh
, d h 0

cg = gh

(35)

When the depth is large a single wave propagates twice the speed of a group of waves, if the
depth is small, on the other hand, are the wave energy and wave shape propagating at the same
rate regardless of wavelength, but as a function of depth. Depth dependence means that the wave
propagation speed decrease with depth, which is the main cause for waves to always approach a
beach in parallel and with increasing amplitude eventually breaking.
In order to find more practical limits than infinite and zero depth the expression for phase
velocity, (33), is used and the wavenumber k expressed in terms of wavelength , by =2/k,
cf 2 =

g
2
tanh
h
2

(36)

Now the limits of deep and shallow water can be expressed as relationships between depth and
wavelength. The hyperbolic function approaches 1 when the argument approaches infinity and
the argument itself as the argument approaches zero. If we plot the argument of the hyperbolic
function, one can see that,

2 2
tanh
h
h
25

tanh( ) 1
h
2

(37)

24

Figure 28. The function tanh plotted to its argument.

Between these limits the phase and group velocities are described of equations (33) and (34). A
wave with a 4s period and a 25m wavelength, which is common in the Baltic Sea, will be affected
by the sea floor when the depth is less than about 12.5m and at depths as small as 1m, its
propagation speed will be stated by the depth.
Ship-generated waves depend largely on the speed of the vessel. Apparently, the ship's speed was
also important for when the water is considered to be shallow. The wash wave system is always
stationary relative to the vessel thus; the whole wave pattern propagates with the ship speed U. In
deep water the appearance of the wave pattern is independent of the ship's speed which is not
the case in shallow water. For a displacement vessel at such high forward speed that the
wavelength of the transverse waves approaches the vessel length will shallow water effects begin
to realise when the depth is less than half the vessel length.
Vessel speed is related to the wave phase velocity in shallow water by,
Fnh =

(38)

gh

which is the Froude number with depth as the characteristic length. Figure 29, illustrate how the
wave pattern change with Fnh in shallow water. Up to Fnh of approximately 0.4 the wave pattern
looks like in deep water. When the ratio increases, the angle limiting the wave system increases up
to 90 at Fnh=1. At this so-called critical speed the wave system degenerate to waves
perpendicular to the vessel, propagating with the speed of the vessels, see Figure 29. The
resistance increases radically. Fnh=1 is often associated with the M=1 in aerodynamics. If the Fnh
is increased further, the wave system will be made up by diverging waves only, confined by a limit
angle that decreases with increasing Fnh, and the wave resistance decreases. At both sub-and
super-critical speeds, the resulting wave pattern is looked into a limited area. At the critical speed,
there is no interaction between the wave components and the energy of the individual waves
constitutes the resistance.

25

Figure 29. The figure shows the wave pattern in shallow waters at different Fnh. The graph on the right is
the resistance curves plotted for different water depths. Note that the horizontal axis is expressed in Fn,
converted to Fnh the resistance peaks are located approximately at Fnh=1. Figure from PNA Vol. II (Lewis,
1988).

If we assume that the depth is 10 m, we see from (38) that the critical speed is about 20 knots.
Intuitively, we realize that the consequences of this depth-speed ratio, differs for the small
pleasure boat, and for instance a tanker. The capability to reach super-critical speed depends on
the ship's wash and resistance increase at the critical speed which is dependent on ship type and
size. Increased waves and resistance increase in shallow water and around the critical speed is a
real problem, partly through increased power requirements and fuel consumption and partly due
to damages to beaches, piers and property because of the waves. Suggestions on how to
appreciate the additional resistance due to shallow water are discussed in PNA Vol. II (Lewis,
1988).

ADDED RESISTANCE DUE TO WAVES AND WIND


Wind and waves add to resistance to similar extent (Lewis 1988). As an estimate one can use the
expression for aerodynamic drag, with the relative wind speed instead of the speed of the vessel.
C AA = 0.001

RAA =

AT
S

(39)

1
2
(V + Vvind ) S C AA
2

(40)

26

It should be noted that a wind is in principle always blowing. At a wind speed of 5m/s, "gentle
breeze" and at 20 m/s "fresh gale", is the wind speed between 10 and 40 knots which obviously for
most ships means a significantly higher air resistance.
The resistance increase in waves can be considerable, and can affect both the schedule and fuel
consumption. The PNA (Lewis 1988) summarize it as an increase in power demand due to:
 Added resistance
- direct wind and wave action
- indirect effect associated with ship motions
- rudder action
 Reduced propulsive efficiency (D)
- increased propeller loading
- ventilated propeller
- unsteady propeller conditions
- reduced hull efficiency
An expression for calculating the added resistance of a ship in regular heading waves is,
RAW (e ) =

k
2e

(b

33 ( x ) U

da33 ( x)
)VR ( x)2 dx
dx

(41)

where a33 and b33 are coefficients for added mass and damping. U is the speed of the vessel and L
its length, k and e is the wave number and frequency of encounter and VR is the vertical relative
velocity between the water particles and the hull. The relative speed is of great importance as
illustrated in Figure 30. In the range where the ratio between wavelength and ship length is close
to 1, the ship motion most accentuated and dominates the relative velocity. For short waves the
added resistance is dominated by wave motion.

Figure 30. The figure shows the relationship between added resistance and incident wavelength for a ship
in regular waves. Figure from Faltinsen (1990).

To estimate how much extra installed power is needed for a route one can use equation (41) (or
something like that) and handle the added resistance on essentially the same way as is convention
for evaluating motion responses. The expression, (42), corresponding to 0-th spectrum moment
and can be considered as the average resistance increase in the encounter spectrum S (e ) .
Rw medel = 2

Rw (e )

S (e ) d e

(42)

27

The term

Rw (e )
2

is not a response amplitude operator (RAO) in the way it is defined for the motion

response, but has that role in equation (42).


Schneekluth & Bertram (1989) refers to a more simplified approach that could be applied in the
design. The method is applicable to tanker and container ships and provides an estimate of speed
reduction due to weather and the relative wave direction.

FOULING
When estimating the necessary power to fulfil contract trial speed surface roughness is taken into
account. When a ship has been waterborne for some time, "roughness" has increased due to
fouling. Fouling affects fuel consumption and consequently involves both economic and
environmental consequences. Figur 31, provides an example of the increase in power required
compared to when the ship was new. Note that the surface becomes irreversibly rougher with
time despite cleaning and repainting. Modern paints have made the figures on the axes (Figur 31)
slightly inaccurate, the time between docking is longer and the increase in shaft horsepower less.
Nevertheless, the resistance increases due to fouling is evident. ITTC (2005) argues that the need
for additional available power (sea margin or service allowance, sv sjtillgg) is greatly affected by
fouling and that the choice of paint system makes a big difference.

Figur 31. Increase in required shaft power due to fouling. Figure from Dyne (1991).

28

PREDICTION OF SHIP RESISTANCE AND POWER REQUIREMENTS


In the initial design phase the power needed for the contracted trial speed is estimated, i.e., the
power requirements during ideal conditions, PS, shaft power. The required installed power is then
appreciated with a sea margin or service allowance (sv sjtillgg) to compensate for wind, waves,
marine growth, etc. A standard value for sea margin is 15%. This power level is known as NCR,
Normal Continuous Rating.
The power requirements at ideal conditions can be expressed, in relation to the bare hull power
PE, effective power (sv slpeffekt), if ignoring a few percent loss of power in the shaft bearings,
as
PS PD =

PE
D

(43)

where D is denoted propulsion efficiency (in the order of 0.7-0.8 for the merchant ships at
Fn<0.3, Milchert (2001) and generally in the range of 0.6-0.85, Dyne (1990)). Propulsion
efficiency contains several elements, including the effect of the propeller operating in the wake,
which increases the efficiency compared to a propeller working in undisturbed water. Therefore
can hull and particularly stern design affect the propeller working condition and thus the required
power. Propulsion efficiency is not an efficiency coefficient in the ordinary sense since it can be
larger than 1 due to the difference in power requirements between a towed hull and a hull
propelled with a propeller. Very briefly, one can say that the propeller, which generates the
propulsive force primarily by suction on the face of the blades, lowers the pressure around the
stern and consequently increases the resistance. This effect is summarized in the thrust deduction
factor (sv sugfaktorn),
t=

R
T

(44)

where R is the resistance increase and T the propeller thrust force. The relationship between
bare hull resistance and propeller pressure force can be expressed as,
T = R + R = R + tT T =

R
1 t

(45)

When the propeller is located as close to the hull that the resistance is affected, it works in the
viscous boundary layer, which means lowered incident velocity to the propeller and increased
propeller efficiency. The difference between the ship speed and (medium)-incident velocity to the
propeller is called wake field (sv medstrm), and is expressed by the wake fraction (sv
medstrmsfaktor),
w = 1

VA
VS

(46)

where VA is the incident velocity to the propeller and VS is the speed of the vessel. Generally,
the benefits of the wake field are greater than the disadvantages of the propeller induced lowered
pressure. The relationship between these is expressed with the hull efficiency,

29

H =

1 t
1 w

(47)

which therefore usually is greater than 1. The hull efficiency is a factor of the propulsion
efficiency, which can be expressed,
D = o R H

(48)

where R is called the relative rotary efficiency. It is approximately one and has to do with
differences in propeller torque in case the propeller operates in the wake field or not. From (48)
we see that the hull efficiency can do, and normally does, that the propulsion efficiency is larger
than the propeller open water efficiency o. The thrust deduction factor and the wake fraction can
be determined by model tests and are used to calculate the propulsion efficiency and the
appropriate propeller dimensions.
To determine the bare hull power, PE, the tools are published and self-acquired data from model
test series on similar hull shapes and full-scale measurements from vessels similar to what is being
designed. At the ship testing facilities experiments have been conducted in which individual
parameters systematically been varied. The results are compiled in so-called systematic series and
can give the designer a picture of the effect of changing a given parameter. The systematic series
is in principal tables where one can evaluate one parameter at a time. A more sophisticated form
of compiling experimental results is to create a formula language that takes into account a range
of parameters at once. These methods are often called semi-empirical and ask for a range of hull
parameters and velocity as input.
The designer can get a good idea of the bare hull resistance, and power requirements, through a
rational use of information on full-scale and model ships, semi-empirical methods and systematic
series. As part of the study, CFD calculations could be of great value. To carry out model tests to
estimate the resistance is still the most reliable method but is rarely made for the initial estimate
of power requirements.
All compilations of experimental data use parameters to describe the ship's hull. With these
design parameters (sv konstruktionskoefficienter) the rather complex hull geometry is
summarized by a number of measures that can be connected to the ship's resistance, stability and
seaworthiness. Here follows a few that the Naval Architect should know,
Block coefficient
CB =

is the displaced volume. CB is sometimes denoted .


BTL pp , where

(49)

Figure 32. The geometric interpretation of the block coefficient CB. Figure from Rawson & Tupper (2001).

30

Midship section coefficient


AM
BT

(50)

Figure 33.

Dimensions that define the midship section area and water line area. Figure from Rawson & Tupper

CM =

(2001).

Waterplane area coefficient


AW
LWL B

CW =

(51)

Figure 34. Definition of the waterline length LWLand the water line area AW. Figure from Rawson &
Tupper (2001).

Prismatic coefficient

AM L pp , the prismatic coefficient is occasionally denoted .

CP =

(52)

The block coefficient expresses the fullness of the hull while the prismatic coefficient also
provides a measure of the distribution of the displacement. A large CP says that the displacement
is fairly evenly distributed over the vessel length, while a low value implies that the displacement
is concentrated to the area around mid ships.
Table 1. Typical range of design parameters for some ship types.
Passenger ship
Container ship Roro ship
Bulk carriers
CB
0.55-0.65
0.58-0.68
0.57-0.70
0.85-0.89
CM
0.97-0.985
0.98-0.99
0.98-0.99
1.00
CW
0.70-0.75
0.75-0.82
0.77-0.83
0.88-0.91
CP
0.60-0.70
0.65-0.75
0.67-0.75
0.84-0.89
0.24-0.36
0.22-0.26
0.23-0.30
0.15-0.18
Fn

Oil tanker.
0.85-0.95
1.00
0.88-0.91
0.85-0.95
0.12-0.18

Frigates
0.45-0.50
0.70-0.80
0.68-0.76
0.56-0.64
0.44-0.58

Table 1 shows that the most robust ships, bulk carriers and tankers have almost identical
prismatic and block coefficients. This is due to rectangular mid frames with full width and
maximum draft. A low Cw means that the hull is slender with full width far aft. The waterline area
coefficient gives, similar to the CP, a picture of the ship slenderness. Ship slenderness is otherwise
often expressed with the slenderness parameter (sv slankhetstal),
=

LWL

(53)

1/ 3

31

The slenderness parameter is for conventional ships between 5 and 7, but can for fast ships be up
to 10. Low wave making resistance is promoted by slender forms, for fast ships should the bow
be lengthy and all design coefficients small. Such a form means that LCB is located a bit aft of
amidships. The coefficients are of course hardly the free choice of the designer but controlled,
above all, by the ship type and its purpose. Still it may be influence somewhat and then so that
the combination of resistance, stability, manoeuvring and seakeeping become as good as possible.

Figure 35. Ships 710 with V-frames and 725 with U-frames. Figures from Harvald (1983).

The shapes of the frames affect the vessel's characteristics. V-shape provides low pitch
acceleration and low slamming loads in a seaway. V-shape generally gives better stability
properties. A body plane of a V-V shaped hull is shown in Figure 35 (hull 710). U-U shape
(Figure 35, hull 725) may provide favourable resistance characteristics low-speed vessels.
The ship aft shall provide as a smooth ending as possible of the vessel. This is more difficult to
achieve than it sounds. Optimum bare hull resistance is obtained with a pronounced V-shaped
stern. The form is unfortunately unsuitable for single screw ships leasing to low propulsion
efficiency. The aft must therefore be modified to improve the propulsion efficiency, either by a
fuller, more U-shaped, stern, or with an appropriate stern bulb that create a favourable wake field
from the propeller point of view. It is also important that the bilge curvature at the aft shoulder,
the transition between the ship's middle section with constant cross-section and the stern,
increases carefully down stream to avoid separation and bilge vortices. This is especially difficult
for bulky vessels (CB0.8) and implies that a plump stern is only appropriate in low speed.
When it comes to bow shapes the bulbous bow or cylindrical bow is probably the most striking,
see Figure 36. The cylindrical bow is appropriate for the really big tankers (VLCC and ULCC).
Bulbous bows are fitted to most ships. They affect the flow around the ship in a favourable way,
for instance because the pressure disturbance from the bulbous bow interact with the bow wave
so that the wave making resistance decreases. There are a wide range of bulbous bow shapes. The
form is sometimes chosen because the design team has concluded that it minimises the
resistance, but just as often based on manufacturing aspects. Generally, the bulb should be
located just below the water surface. When the ship is in ballast condition or unfavourably loaded
the bulbous bow may not give the intended resistance reduction, but rather adds resistance.

32

Figure 36. Two examples of ships fitted with a bulbous bow and a tanker with cylindrical bow.

SYSTEMATIC SERIES AND SEMI-EMPIRICAL METHODS


Many systematic experiment series was carried out during the 50 -, 60 - and 70's and the results
are still in use, either in its original form or as a basis for semi-empirical methods that take into
account several hull parameters simultaneously, and often also the effect of bulbous bow and
appendages. Semi-empirical methods, for example the widely used method by Holtrop &
Mennen for displacement vessels or the Savitsky method for planning hulls, is easily
implemented and gives a quick estimate of power requirements as soon as the main dimensions
are decided.

Figure 37. Residual resistance (in principle the wave making resistance) for slenderness parameter 6.0 as a
function of forward speed. in the diagram is the prismatic coefficient, usually denoted Cp. From
Guldhammer & Harvald (1974).

33

An assemblage of systematic series originating from several model test facilities was published in
1974 by Guldhammer & Harvald. Figure 37, exemplifies the work. The plot is valid for one
particular slenderness parameter and shows graphs of residual resistance for a range of prismatic
coefficients, as function of forward speed. Humps and hollows are clearly seen and one realizes
that the risk the quality of the resistance estimate is at stake in speed ranges where the CR-curve
has a pronounced slope. Graphs for slenderness parameter 5, 5.5 and 6.0 are reproduced in the
Annex. (All in all Guldhammer & Harvald (1974) consists of graphs for slenderness parameters
4.5-8 in steps of 0.5.)
For displacement and semi-planing ships, Fn<approx 0.8, the method published by Holtrop
(1984) is well established. The method was first presented in Holtrop & Mennen (1978) and later
modified by the two in 1982 .The first edition of the method was based on systematic
experiments on 191 models. The latter version, which mainly improves the prediction of the
wave making resistance in the higher speed region, is based on data from experiments on a total
of 334 models. The large amount of model test data has provided a statistical basis that, through
regression analysis, is transferred to a formula language. Holtrop & Mennens method is one of
the most frequently used to estimate the power requirement during the design of ships. The
method provides the required shaft power, but also resistance components, the components of
the propulsion efficiency, thrust and wake factor. Input to the method is main dimensions,
design parameters and some geometric information about the bulbous bow, stern and
appendages. The structure is schematically presented in Figure 38.

Figure 38. The structure of the Holtrop & Mennen method. Figure from Nreskog (1995).

The resistance is divided in similar terms as for the model test,


RT = (1 + k ) RF + RAPP + RW + RB + RTR + RA

(1+k)
RF
RAPP

(54)

hull form factor


frictional resistance according to the ITTC
appendage resistance (see Figure 25)

34

RW
RB
RTR
RA

wave making resistance


increased pressure resistance connected to the bulbous bow vicinity to the water surface
increased pressure resistance due to transom stern
ship-model correlated resistance (in principle surface allowance)

The hull form factor is expressed by means of the ship's main dimensions, displacement, LCG,
and a coefficient dependent of the stern shape, see Figure 39.

Figure 39. Stern shape correction of the Holtrop & Mennen method. From Holtrop & Mennen (1984).

The table in Figure 39 shows that the method assumes some sort of most common or normal aft
shape and then introduces corrections based on the variation from this. The resistance coefficient
RTR is non zero for hulls with a transom stern. The coefficient is expressed by the transom
under-water area, forward speed and the water line area coefficient. The ship-model correlated
resistance coefficient RA includes surface allowance and air resistance which in the scaling
procedure of model tests is expressed with dimensionless coefficients CF and CAA.
By means of systematic series, you can get an idea of how sensitive the resistance is to a single
parameter. The slope of the resistance curve at the point you reach with the presumed geometry
shows how sensitive resistance estimate is to an error. A selection of charts from the
Guldhammer & Harvald (1974) is reproduced in appendix and there is a list of systematic series
at the end of the section on model tests.

Example-Estimation of ship resistance on the basis of main data


Estimate the vessel resistance, bare hull power (effective power) and power requirement.

The vessel Protefs has the following dimensions (based on RINA (2004), see appendix):
Lpp=217 m
S=10800 m2
B=32.2 m
Sroder=100 m2
T=12.5 m
AM=402 m2
V=14.4 kn
=74200 m3
The wetted surface area was calculated with the following approximate formula,
S=1.025Lpp(CBB+1.7T) from Guldhammer & Harvald (1974).

35

Assume the water kinematic viscosity and density to be,


=1.2.10-6 m2/s
=1025 kg/m3
Solution
Determine the resistance coefficient, CT = (1 + k )CF + CR + CF + C AA , and add-on
appendage resistance, in this case rudder resistance.
Determine the dimensionless numbers necessary for the analysis
1852
1852
14.4
217
14.4
VL
V
3600
3600 = 0.16
Re L =
=
= 1.3 109
Fn =
=
6

1.2 10
gL
g 217
CB =

= 0.85
BLT

LW

1/ 3

217
742001/ 3

CP =

= 0.85
AM LPP

= 5.2

Viscous resistance
CF =

1) Skin friction of a flat plate

2) form factor(Watanabe)

k = 0.095 + 25.5

0.075

( log10 Re L 2 )

CB
L
B

B
T

= 1.48 103

= 0.20

1/ 3

k
CF = 105 s 0.64 103 = 0.29 103 ,

ks=150m as recommended by the ITTC-78.

3) surface allowence

Wave making resistance (Residual resistance)


Use the chart from the Guldhammer and Harvald (see appendix) where is the
prismatic coefficient.
L
With = CP = 0.85 , Fn = 0.16 and 1/ 3 = 5.2 , the graphs for slenderness 5 and 5.5 give

after interpolation, CR = 1.00 103 .


Unfortunately, the CR-charts were drawn assuming that CR=CTM-CfM. This means that
the residual resistance coefficient is dependent on scale and must be recalculated by first
calculating the CTM and then determining "our" CR as CR=CTM(1+k) CfM. This new CR is
equal in model and full scale (see the re-scaling of model tests acc. ITTC-78). This
manoeuvre requires that we can calculate CfM. This in turn requires that we know the
size of the models in the experiment. Assume that the model length at the systematic
testing was 4 m and correct for the fact that the graph represents a residual resistance
coefficient CRM=CTM CfM. We are following ITTC-78 and want to have residual
resistance as CRM=CTM(1+k) CfM where CRM=CRS. The model length gives the scale
factor =54.25 and
U M = U S / Re M = 3.35 106 C fM = 3.67 103 CTM = (1.00 + 3.67) 103

and finally, CRS=CRM=(4.67-(1+0.20)3.67)10-3=0.27.10-3 (wave generation is extremely


small at such low Fn, friction dominates).

36

Air resistance
1
luftV 2CD AT . The term can be made
2
dimensionless by a factor related to the forward speed of the vessel, wetted surface area
luft AT
1
CD can then be
V 2 S . The coefficient, C AA =
and the density of water,
2
S
A
included in the total resistance coefficient. ITTC recommends: C AA = 0.001 T
S

Air resistance can be expressed as, RAA =

Suppose, for example AT B 1.5T = 604 m2 giving. C AA = 0.06 103 (it is generally a
small part of the resistance).
Hull total resistance coefficient is then;
CT = (1 + 0.2) 1.48 103 + 0.27 103 + 0.29 103 + 0.06 103 = 2.40 103

and resistance,
RT =

1
V 2 S CT = 729 kN
2

Estimation of the appendage resistance, rudder


Form factor of symmetric lifting
profiles,
2t
k=
= 0.3
c

Estimate rudder dimensions


based on the known rudder area
and the ship's draft. Suppose
(e.g.) t=0.75 m and c=5 m
1
Rroder = V 2 Sroder CD
2
where CD = (1 + kroder )CF , roder
For a rudder that acts in the
wake (unlike for example a
stabilizing fin) the Reynolds
number is based on the vessel
length when the skin-friction
coefficient is determined,
CF =

0.075

( log10 Re L 2 )

= 1.48 103

The rudder resistance becomes, Rroder =

1
V 2 Sroder CD = 5.4 kN
2

Overall resistance, RT = 729 + 5.4 = 734 MN


Effective power, PE = RT V 5.4 MW eller 7400 hk!

37

STANDARD SHIP PROPELLERS


To choose a propeller for the vessel, of efficient power PE, so-called open-water propeller charts
(sv frigendekaraktristikor) is a useful tool. The charts describes the propeller thrust T and
torque Q, as it rotates in a homogeneous flow with speed VA. The open-water graphs are
determined experimentally and for initial-design purposes published experimental series can be
used (a selection from the SSPA standard series Lindgren & Bjarne (1967) is shown in the
appendix). The incident flow velocity, thrust and torque are expressed dimensionless with the
following expressions,
J=

VA
Dn

-Advance number (sv framdriftstal)

(55)

KT =

T
D 4 n2

-Thrust coefficient

(56)

KQ =

Q
D5 n2

-Torque coefficient

(57)

where D is the propeller diameter and n is propeller speed in revolutions/s.

Figure 40. KT, KQ and the propeller open-water efficiency o, for a propeller of the SSPA standard propeller
series. The propeller 4.60 has four blades and blade-area ratio 0.60. The family of curves are associated
with different pitch ratio P/D. The chart comes from Lindgren & Bjarne (1967). Note that the scale of KQ
is decreasing upwards in the figure!

The curves in Figure 40 describe thrust, shaft torque and the relationship between them in this
case for a four-bladed propeller. Thrust represents (multiplied by speed) beneficial effect from
the propeller that can be used to overcome the resistance and torque (multiplied by angular
frequency) the input power. The relationship,

38

KT J
= o
K Q 2

(58)

represents the propeller efficiency. Figure 40 shows that the efficiency curves have a clear
maximum. The propeller should of course be chosen so that the normal conditions are near that
maximum. An efficiency curve is very steep at higher advance numbers than the operating
optimum, and due to the uncertainties in the resistance prediction it could be wise to choose a
propeller slightly to the left of the efficiency optimum.
In ship design, we are seeking one or more propellers to drive the ship, and a suitable engine that
can deliver sufficient power at an appropriate speed. The propeller should have reasonable
dimensions, number of blades and blade geometry. The blade design is complex and chosen
according to fluid mechanical properties, profile lift and drag, structural properties and the risk
for cavitation. In the propeller charts the geometry is described by the pitch ratio P/D and blade
area ratio expressed as AD/A0 (for SSPA:s standard propellers, Lindgren & Bjarne (1967)) and
AE/A0 (Wageningen B-Screw Series, Lammeren et al. (1969)), where A0 is the propeller disc area
(D2/4) and AD and AE are Developed and Expanded area respectively. These two dimensions are
different, but represent approximately the surface spanned by local chord length of the blade, at
every cut from the hub to the blade tip, twisted to the propeller disc plane. Blade area ratio
describes the total surface of the blades in relation to propeller disc size. With too low blade area
ratio risk is increased for noise and propeller damage caused by cavitation. If the blade area is
unnecessarily large, on the other hand, the profile drag becomes large and the propeller efficiency
is worse than needed to be.
Propeller pitch, P, has to do with the blade angle relative to the propeller shaft, but is a length
defined by the distance made by a section of the propeller if it was bolted through a solid
medium, see Figure 41. The pitch depends on the distance to the hub and the local angle of
attack and can be constant or vary along the propeller blade. In the latter case the propeller pitch
is defined as the pitch at 70% of the radius.

P = 2 r tan( )

Figure 41. Pitch or geometric pitch. The distance a blade profile would have moved in the shaft direction if
bolted through a solid medium.

The propeller does not bolt through the water but move a shorter distance than the pitch
indicates. The difference is called slip (sv slirning) and is defined in Figure 42. Slip depends on

39

propeller load. From the propeller charts of Figure 40 one can see that if the propeller load
increases, (perhaps the resistance increased as the ship entered shallow waters, with higher waves
and strong head wind), the speed will decrease and result in increased slippage. If the vessel is
tied up (Bollard pull) the propeller produce maximum thrust and torque but with poor efficiency
and with slip=1.

slirning =

P n VA
Pn

Figure 42. Slip (sv slirning) is the difference between the geometric pitch and the distance that the propeller
actually moves. In the figure is SLIP indicated as a velocity but is usually expressed as dimensionless by
division with the speed Pn, where P is the pitch and n the propeller speed. In the figure is SLIP=P.n-VA.

The propeller charts in Figure 40 shows that the efficiency increases with the pitch ratio. There is
of course a limit. According to Tornblad (1990) maximum efficiency is reached at =45
(P/D=2.2) but in practice is pitch ratio larger than P/D1.5 not used. Large pitch means that the
profile is highly loaded and the risk for cavitation increases.
Cavitation is a matter of low temperature boiling of water due to local low pressure. From fluid
mechanics, we know that increased flow velocity and reduced pressure is linked which is clearly
demonstrated by the invicid case of Bernoulli's equation (10). Thermally the boiling point and
pressure are connected and for the flow around for instance a propeller, the low pressure can
turn the boiling point to or below the water temperature, typically 0-15C, and the water changes,
locally, physical state. The local boiling is called cavitation. This phenomenon implies that the
suction on the face of the propeller blade is limited. When the pressure rises and vapour is
released into liquid again, the process can be so dramatic that it erodes the propeller and induces
noise and vibration in the vessel. As a propeller blade rotates in a heterogeneous velocity field,
the local conditions of cavitation are constantly changing, which means that vapour bubbles are
formed and implode continuously with resulting noise and propeller damage. Would cavitation
be stationary, the draw back would be limited to a major loss in propeller efficiency, which in
itself would be reason enough to avoid cavitation.
As we have seen, the pressure on ship hulls decreases (e.g. Figure 15) from the bow to the
maximum breadth and increases again as the vessel is narrowing in the stern. On the suction side
of the propeller blade (the convex side) is of course the pressure distribution similar. This means
that the pressure minimum occurs where the profile is at its thickest, and that a low thicknesscord ratio (t/c) reduces the low pressure and the risk of cavitation. A slim profile with a more
evenly distributed suction, however, has to have a certain thickness for sufficient strength, which
puts demands on the chord length. Ship propellers have therefore rather wide blades and the
wider the greater becomes the blade area ratio. Large blade area means less risk for cavitation but
large blade area also implies large profile drag. Blade area should therefore be large enough but
not exaggeratedly large so that the efficiency will suffer.

40

Choosing a propeller
When selecting a propeller from a standard series you initially estimate the number of blades and
the blade area ratio, BAR, to find an appropriate propeller chart. Then you choose the largest
possible propeller diameter based on the ship draught. The thrust force is determined from bare
hull resistance, thrust deduction factor and wake fraction. Based on this input the propeller with
the pitch ratio giving the best propeller efficiency is chosen from the chart. The risk for
cavitation is eventually controlled by semi-empirical methods. If the propeller is at risk the choice
may be corrected by reducing the pitch ratio or increasing the blade area ratio. If, on the other
hand, the propeller is well below the cavitation limiting curve the blade area ratio can be reduced
and a propeller with a more optimal pitch ratio chosen.
In initial design the goal is to select a propeller and a main engine that could work. Ship
propellers are produced with fixed or adjustable blades, i.e. with or without the ability to change
the pitch. Usually a single propeller is used but limited space or the power requirement can lead
to solutions with two or more propellers. Here, the suggested starting point is to run the ship
with a single propeller from a standard series of various combinations of number of blades, Z,
and blade area ratio BAR (AE/A0AD/A0 where AE denotes expanded area and AD developed area.
A0 is the propeller disk area). These variables must initially be selected even if the propeller
choice then may require iteration between different propeller versions. In both the Wageningen BScrew series and the SSPA standard series the charts are selected based on number of blades and
blade area ratio. The series has propellers with 2-7 blades but with reference to Tornblad (1990),
we can limit our interest to 3-4 blades for smaller vessels and to 4-5 for the large ships. The
efficiency is generally better with fewer blades but pressure fluctuations become larger. The latter
implies that the risk for vibration and noise are increasing. For vibration problems are also the
excitation frequency essential which depends on both the propeller speed and the number of
blades.
Blade area ratio should be large to reduce the risk for cavitation, but small to keep the profile
drag low and thus the efficiency up. Usually a propeller series does not include so many different
BARs so a strategy is to chose the smallest and change in case cavitation turns out to be a risk.
The propeller should be chosen as large as possible. Large diameter and low rpm give high
efficiency (Milchert 2000). The propeller shall operate in the wake field which limits the size and
consequently the distance to the keel line. Neither should the propeller be too close to the hull or
rudder. Consider Figure 43 where the measurements are plotted together with the distances to
the hull, rudder and keel -clearance (sv klarning).

Figure 43. Propeller size and distance to the hull, rudder, keel line (clearance) and to the sea surface.

The requirement for the distance c is mainly to avoid excessive inducement of noise and vibration
to the hull. The following recommendations from the classification society DNV reproduced of
Schneekluth & Bertram (1998), give a clear and useful picture of the estimates a, b, c, e, see Table
2.

41

Table 2. Recommendation of clearence, Schneekluth & Bertram (1998).

a>0.1D
a>2(AE/A0)D/Z
b>(0.35-0.02Z)D
c>(0.24-0.01Z)D
c>(0.3-0.01Z)D
0.1D<e<0.2D

single screw
twin screw
single screw
twin screw

Schneekluth & Bertram (1998) writes that the distance between the blade tip and the water
surface should roughly be a propeller radius and it can be concluded that hD. The distance may
seem generous, but should ensure that the propeller is under water also in ballast condition. They
further write that the distance between the lowest position of the blade tip and the keel line
should be 0.1-0.2D. Roughly, the maximum propeller diameter then becomes about 60% of the
design draft. Harvald (1983) gives the rule of thumb that the maximum diameter should be less
than two-thirds of the draft at stern, TA. This is guidelines and one should be aware that the
clearance requirements may lead you to a smaller diameter.
The thrust force T is required to express the thrust coefficient KT. T is the bare hull resistance
added with the resistance increase described by the thrust deduction factor t (see equation (45))
which is a way to describe the effect of the propeller reducing the pressure around the stern. The
thrust factor can be empirically determined through combined systematic experiments where the
hull resistance is measured without propeller (towing tests, sv slpfrsk), the propeller
characteristics are studied for the propeller in uniform flow (open water tests sv frigendefrsk)
and by experiments in which the ship model is propelled by a model propeller (self propulsion
test, sv sjlvdriftsfrsk). Figure 44, Figure 45 and Figure 46 show thrust deduction factor, wake
fraction and these two combined to hull efficiency h (by equation (47)) which is results from the
towing and self propulsion tests of cargo ships and data from the propeller open-water tests.
Data are displayed for a variety of ship fullness (block coefficient CB or , see (49)) and
slenderness (slenderness parameter, see (53)).
Figure 44 shows that the thrust deduction factor, for this series, only depends on the slenderness
parameter. It is reasonable to believe that slenderness, in general, has a more significant impact
on the thrust deduction factor than the body fullness and we conclude that t decreases with
increasing slenderness. The smaller t the lesser resistance increase due to the propeller and a long
slender hull is thus beneficial for both the bare hull resistance (low Fn) and the propulsion.

42

Figure 44. Thrust deduction factor t as a function of the slenderness parameter for the SSPA Cargo Liner
Series. The models were run at self propulsion test (Williams 1970) with propellers from the SSPA standard
series (Lindgren and Bjarne 1967). pp in the upper graph is the block coefficient usually denoted CB. Note
that the upper graph shows results for a ratio between the propeller diameter D and the ship length
between perpendiculars Lpp of 0.04. The lower graph gives the corrections on the thrust deduction factor in
case D/Lpp deviates from 0.04. Chart from Williams (1970).

43

Figure 45. Wake fraction w as a function of the slenderness parameter for the SSPA Cargo Liner Series.
The models were run at self propulsion test (Williams 1970) with propellers from the SSPA standard series
(Lindgren and Bjarne 1967). pp in the upper graph is the block coefficient usually denoted CB. Note that
the upper graph shows results for a ratio between the propeller diameter D and the ship length between
perpendiculars Lpp of 0.04. The lower graph gives the corrections on the wake fraction in case D/Lpp
deviates from 0.04. Chart from Williams (1970).

The wake fraction w, defined in (46), describes how much the boundary layer surrounding the
ship has slowed down the mean velocity approaching the propeller in relation to the ship speed.
A lower speed is beneficial to the propeller and w should be as large as possible. In Figure 45, we
note that w depends primarily on fullness of the vessel and that w increases with block coefficient,
but the wake fraction also depends on the slenderness parameter and w decreases with increased
slenderness. The wake fraction is therefore most advantageous for short plump vessels.

44

Figure 46. Hull Efficiency h as a function of the slenderness parameter for the SSPA Cargo Liner Series.
The models were run at self propulsion test (Williams 1970) with propellers from the SSPA standard series
(Lindgren and Bjarne 1967). pp in the upper graph is the block coefficient usually denoted CB. Note that
the upper graph shows results for a ratio between the propeller diameter D and the ship length between
perpendiculars Lpp of 0.04. The lower graph gives the corrections on the hull efficiency in case D/Lpp
deviates from 0.04. Chart from Williams (1970).

From this discussion on the thrust deduction factor and wake fraction follows that there is no
obvious best hull form. The overall effect of the propeller; -increased pressure resistance and
lower incident velocity is expressed by the hull efficiency (defined in (47)) given for the
experimental series in Figure 46. The graph in Figure 46 shows that the hull efficiency is greater
than 1 and that the hull efficiency increases with both slenderness parameter and block
coefficient. A hull efficiency larger than 1 says that the resistance increase due to the propeller is
less than the advantage of a lower propeller incident velocity than the vessel speed.

45

With the approach to choose a propeller in the initial design phase from a standard propeller
series, it is appropriate to recapitulate that we need to do assumptions on numbers of blades and
blade area ratio. Thrust is also needed which can be expressed by (45) if the thrust deduction
factor t is known. Figure 44 shows the thrust deduction factor for a systematic experiment. If the
projected vessel has a similar hull shape and is within the range of the slenderness parameter and
block coefficient, data from the graph could be used. In Holtrop & Mennens semi-empirical
method, data from a series of model tests are compiled and expressions for thrust deduction
factor and wake fraction, based on regression analysis, are given. More accessible estimates can
be made with the following expression of Tornblad (1990),
w = 0.5 CB 0.05 1-prop. ftg

(59)

w = 0.55 CB 0.20 2-prop. ftg


t = 0.6 w 1-prop. ftg.

(60)

t = 1.25 w 2-prop. ftg.

Pitch ratio & propeller speed from propeller charts


Assume that we have chosen the design forward speed and ship main particulars, that we choose
the largest possible propeller diameter, the wake fraction and thrust deduction factor from model
series and have calculated the bare hull resistance. It is now time to use the propeller charts. With
the propeller thrust T from (45) and the incident velocity to the propeller VA from (46) follow
the advance number J and the thrust coefficient except for the propeller speed n. If we express
the ratio,
KT

J2

T n2 D4

(V ( nD ) )
A

)=

T
T
KT =
J2
2
2
D VA
D 2VA 2

(61)

one can express KT as a function of the square of the advance number independent from the
propeller speed n. The expression of KT, the KT-parabola, is drawn into the propeller chart
(exemplified in Figure 47).

Figure 47. KT -parabola for Protefs plotted in the SSPA 5.60 propeller chart (5 blades, blade area ratio
AE/A0=0.60).

46

The parabola represents the required thrust and the task is to determine the advance number and
pitch ratio leading to the best propeller efficiency. In the propeller chart of Figure 47 a KTparabola for the bulk carrier Protefs, previously used to exemplify the resistance calculation, is
drawn. From the figure one can find the propeller efficiency for different pitch ratios by reading
the efficiency curve at the advance number where the KT-parabola intersects the KT-curve for the
corresponding pitch ratio (the reading of P/D=0.6 is plotted in Figure 47). Propeller efficiency
and propeller speed as a function of pitch ratio is plotted in Figure 48 (the speed is determined
from the J since the diameter is selected). Maximum efficiency is obtained for an SSPA 5.60 with
pitch ratio 0.8-1.0 at propeller speed with 1.20 to 1.02 rev/s = 72-61 rpm.

Figure 48. Efficiency and propeller speed of the SSPA 5.60 plotted as a function of pitch ratio. A diameter
of 7.5 m was used to calculate the propeller speed.

Check on risk for cavitation


At this stage we should do a check to make sure that the propeller is, at least not obviously bad
from a cavitation point of view. As the propeller choice has been made a certain blade area ratio
(0.60 in this example) is the check an assessment if the blade area ratio is reasonable and whether
the optimal pitch ratio can be used. A simple approximate method for estimating the risk of
cavitation is to use the Burrill diagram shown in Figure 49 where the cavitation number of SSPA
5.60 in the example has been plotted.
The coefficient c and the local cavitation number at 70% of the radius, (0.7R), are calculated
according to Lewis (1988);
c =

T
T
=
2
2
2
1/ 2 APVR (0.7)
1/ 2 AP VA + ( 0.7 D n )

(62)

where AP is the projected propeller area which has the following approximate relationship to the
expanded area AE (or developed area AD) as (Lewis 1988);
AP AD = 1.067 0.229 P D

(63)

This can be rewritten in terms of the known BAR,


Ap = BAR A0 (1.067 0.229 P / D )

(64)

The local cavitation number can be expressed,

47

(0.7 R ) =

pa + gh pv

1/ 2 VA 2 + ( 0.7 nD )

(65)

where the numerator is the difference in pressure between the static pressure at the propeller
shaft and water vapour pressure Pv (use 2 kPa at 15C). The propeller shaft depth, h, is defined in
Figure 43. Pa is the atmospheric pressure. The denominator represents the dynamic pressure with
respect to the blade relative velocity at 70% of the radius.

Figure 49. Burrils cavtaion diagram. From PNA Lewis (1988). The results from the Protefs-example is
drawn for pitch-ratio P/D=0.7, 0.8, 0.9 and 1.0.

Exemple propeller to the Protefs


Select a propeller to Protefs (the ship that was used as an example for resistance
calculation) and calculate the shaft power.
Choose Z = 5 (because it's a big ship) and a propeller from the SSPA standard series.
Blade area ratio is 0.6 for all five-bladed propellers of the series.
The vessel Protefs has the following main particulars:
Lpp=217 m
B=32.2 m
TKVL=12.5 m
=74200 m3
CB=0.85
slenderness parameter=5.2
Choose:
propeller diam. D=0.6TKVL=7.5 m
shaft depth h=D=7.5 m

S=10800 m2
Srudder=100 m2
AM=402 m2
V=14.4 kn
R(14.4 kn)=734 kN

Determine wake fraction and thrust deduction factor. The block coefficient is too large
for the graphs in Figure 44 and Figure 45 why the expressions (59) and (60) is used
instead giving w=0.38 and t=0.22. With wake fraction and thrust deduction factor
known, the propeller thrust and the incident velocity at the propeller can be determined
from (45) and (46) to T=941 kN and VA=4.6 m/s. Now KT is determined as a function of
the advance number squared according to (61) and drawn into the propeller chart. The
calculated values are tabulated in the first column of the table on the next page and the

48

parabola is plotted in Figure 47. From the propeller chart the efficiency and
corresponding propeller speed is determined for different pitch ratios, see the middle
section of the table below and Figure 48. The maximum propeller efficiency is thus 0:59
for pitch ratios in the range of 0.8-1.0.
KT-parabola
(se Figure 47)
J
KT
0.4 0.12
0.5 0.19
0.6 0.28
0.7 0.38
0.8 0.49

o & n as fkn. of P/D data from Figure 47


(se Figure 48)
J
P/D
o
n=VA/JD
0.42
0.6
0.53
1.46
0.47
0.7
0.58
1.30
0.51
0.8
0.59
1.20
0.55
0.9
0.59
1.12
0.60
1.0
0.59
1.02
0.63
1.1
0.57
0.97

Kavitationskontroll
(se Figure 49
c
(0.7R)
0.16
0.19
0.22
0.27

0.70
0.82
0.94
1.11

The cavitation numbers for the pitch ratios P/D=0.7, 0.8, 0.9 and 1.0 have been
calculated (using equations(62), (63) and (65)) and drawn into the Burril-diagram in
Figure 49 where one sees that the selected propeller SSPA 5.60 ends up below the
recommended limit for merchant ships. Thus, targeting maximum efficiency, you can
choose the SSPA 5.60 with P/D=0.9 at speed 1.12 rev/s = 67 rpm and a propeller
efficiency of 0.59.
Propeller power can be determined by the expressions (43) and (48),
1 t

PD = PE / (o R H ) = R V / o R
=
1 w

1 0.22

= 735 103 7.4 / 0.59 1


= 7328 kW
1 0.38

(66)

Correct the diameter to an appropriate engine speed


In the example, the required power for the propulsion is about 7350 kW. With sea margin (15%)
and margin for temporary high load (10%) an engine capable of delivering approximately 9300
kW is required, which is low for two-stroke machines with a speed as low as 67 rpm. From
manufacturers, for instance MAN B & W, it is easy to obtain data, via the internet, on engine
power, rpm and weight which is important to know early in the project, see for example MAN B
& W (2005a). There is also guidance on how to handle engine layout diagrams and to choose
engine, see MAN B & W (2005b). Studying the relationship between power and engine speed it is
seen that there is a larger assortment of engines in a suitable power range at speeds around 100
rpm. During the design process, it is common that the propeller diameter is decided from an
appropriate engine speed. Looking at the dimensionless numbers J and KQ and note that the
relationship between shaft torque and engine power is described by,
PD = 2 n Q

(67)

it is seen that KQ can be expressed in relation to J independent of the diameter D as,


KQ J 5 =

Q n3
VA5

(68)

Prerequisiting that we know the engine speed, bare hull resistance and an estimate of the
propulsion efficiency D. Propeller series data is often presented in relation to KQ/J5 with a line
indicating the maximum open water efficiency, see Figure 50. Assuming that the propulsion
efficiency and thus power requirements are as in (66) and not significantly influenced by the fact

49

that the propeller speed is modified and that we want to use an engine that can deliver those
7328 kW at for example 100 rpm, the propeller diameter can be determined from Figure 50.

Figure 50. Examples of propeller data presented in relation to the diameter independent quantity KQ/J5.
From Lindgren & Bjrne (1967).

With data from the example, the propeller efficiency is o=0.55 and the diameter D=6.4 m. Thus,
a smaller, slightly less efficient propeller than was the result at 67 rpm. The authentic ship Protefs
has a fixed pitch propeller with a diameter of 6.76 m. The engine is direct-acting with speed 105
rpm at maximum power output. The propeller is smaller than we first thought which might be
due to demands fore clearance to the hull (see Table 2 and Figure 43), which we omitted, or that
there were arguments for the particular engine: MAN B & W 5S60MC-MK6 (with 10 200 kW at
105 rpm). The selected engine is about 10 m high, over 7 meters long and weights nearly 320
tonnes, dimensions that affect the whole ship arrangement and therefore is important to know
early in a project. For smaller vessels and less power outtake the engine speed is higher. The
propeller and engine speeds do not match and a reduction gear has to be installed between
engine and propeller shaft.

Engine layout diagram


Figure 51 shows an engine layout diagram with the propeller curves for light and heavy loaded
propeller fitted into the engine operating area. Ship resistance is normally calculated for calm sea
and deep waters which are represented by the dotted line 6 in the figure. The design point PD is
the required power PD, calculated from the design cruise speed. It is well-known that the hull and
the propeller will deteriorate with age and it is therefore expected that the propeller speed will
decrease slightly, this can be further accentuated by the resistance increase due to wind and
waves. If the propeller is loaded more heavily and the advance number goes down, the KT and KQ
increase. If the propeller speed then decreases in proportion to the increase of the KQ the engine
can deliver as much power as for the lighter-loaded propeller, but the ship's forward speed drops
slightly. For a suitable engine, the propeller speed at curve 2 shall typically be 5% lower than for
the line 6. In order to keep cruising speed at larger resistance increase due to waves and wind, the
propeller speed must be increased and more power must be delivered from the engine. The point
SP in the layout diagram represents NCR (normal Continuous rating) which is the regular power
outtake i.e. the calculated shaft power, Ps, increased by typically 15% (Service allowance or sea
margin). An additional margin (often 10%) is added and the point MP is reach MCR (Maximum

50

Continuous rating) which is a maximum level that should not be exceeded. The engine operating
area is defined by the diamond with the corners L1-L4. Propeller power and speed requirements
should thus match the engine's operating area, where the L1 point can be seen as the MCR of the
engine.

Figure 51. Engine layout diagram. With power and engine speed expressed as log (P) and log (n) can both
the engine operating area, limited by L1-L4, and propeller power curves (for limited variation in speed) be
described by straight lines. Point L1 is the largest acceptable MCR.

51

MODEL TESTS
Model tests with surface piercing or near-surface models which generates surface waves, are
scaled by the Froude model law, i.e. the Froude number should be the same for the model as it is
for the full-scale situation modelling. Typically, the model is equipped with turbulence triggers in
order to give the boundary layer the same character as in full scale. By choosing the model speed
following the Froude model law and by modelling the shape and character of the boundary layer,
the wave formation and vessel attitude will be geometrically well modelled. Still, dynamic
similarity is not achieve, (the Reynolds number will be different in model and full scale), and the
resistance measurement that is partly a viscous effect, must be scaled to full scale in a way so that
the proportion of viscous resistance and wave making resistance becomes reasonable adequate.
Quantities that are directly related to the pressure distribution like ship motions and waves are in
contrast scaled straight according to the Froude model law.

Scale factors
Froudes model law

Geometrical similarity
LM =

AM =

VM =

LS

AS

VS

length

UM =

area

tM =

US

tS

velocity

time

volume

Requirements on the model experimental set-up


The model should be chosen as large as possible for best results. If the model is too large relative
to the towing tank (sv slprnna), the velocity field and thus the pressure will be affected when
the water that the hull displaces is referred to the limited spaces between the hull and tank
bottom or edges. The degree of influence is associated with the model speed, hull length and
cross section. In order to avoid influenced by the proximity to solid boundaries the main
particulars of the model hull should be chosen so that the following conditions are met,
L pp < h

(69)

B
2

(70)

L pp <

Tm Bm <

Bh
200

(71)

where Lpp, Tm and Bm is the model length between perpendiculars, draught and maximum breadth
respectively. The width and depth of the towing tank is denoted B and h.

52

Towing test
Usually the model is towed without appendages (bare hull) but sometimes a rudder is fitted to
improve course keeping stability. The rudder is in such case considered as a part of the hull, the
air resistance is negligible and the model's surface is smooth. In terms of resistance coefficients
the measurement is described by,
CTM = (1 + k )CFM + CRM

where the index M refers to the model scale. The model is equipped with some devise to initiate
turbulence in order to create a boundary layer as similar to the full-scale situation as possible. The
towing force shall act in the model's centre line, if possible at the centre of gravity with a
minimum of imposed torque. The tow speeds are determined by scaling the speed range of the
full-scale ship. Model test speed is stepped through the speed range. Measurements at the lowest
speeds are used to the Prohaska method for determining the form factor.
At the tow tests is, besides the resistance also ship sinkage & trim are measured, i.e. the ship's
change in running attitude, vertical displacement and inclination relative to the horizontal plane.
Water temperature is also measured to find out the present water density and viscosity.

Scaling measurement data to full-scale ITTC-78 procedure


This procedure describes the rescaling of the resistance data from a model test with a
displacement hull2. The method is recommended by the ITTC (1978) (International Towing
Tank Conference). In principle: -Wave making resistance is modelled correctly because the
experiment is designed with respect to the Froude model law, the viscous resistance is estimated
with plate skin friction coefficient and shape factor which preferably is determined from results
of runs at low speed.
1. Measure the total model resistance CTM at the same Fn as in full-scale
2. Calculate the frictional resistance CFM of the model in accordance with the ITTC-57
CFM =

0.075
(log10 Re LM 2)

(72)
2

3. Determine the form factor k


4. Determine the residual resistance CRM
CRM = CTM (1 + k )CFM

(73)

5. State that (a consequence of the Froude model law)


CRS = CRM

(74)

6. Calculate the frictional resistance CFS of full-scale ship,

For planing hulls the rescaling procedure is basically the same with the complication that the wetted
surface varies with speed and that the form factor becomes something else (usually k=0 for planing hulls).

53

CFS =

0.075
(log10 Re L 2)

(75)
2

7. Calculate roughness allowance CF as,


1

(76)

k 3
C f = (105 s 0.64) 103
L

where ks is the ship roughness. ITTC recommends ks=150 m.


8. The ship's air resistance coefficient is determined by,
C AA = 0.001

AT
S

(77)

where AT is the projected frontal area of the ship superstructure.


9. Calculate the full-scale total resistance coefficient CTS,
CTS = (1 + k )CFS + CRS + CF + C AA

(78)

10. The ship total resistance and effective power is calculated as


RS = CTS

1
S VS 2 S S
2

PE = RS VS

(79)

[ N]

[W]

(80)

Figure 52. Illustration of the ITTC-78 rescaling procedure.

54

Self propulsion test


Self propulsion tests are conducted with the geometrically similar model hull propelled by a
model propulsion installation. An important aim is to determine the wake fraction and thrust
deduction factor necessary to, by means of propeller charts, select a propeller and to determine
the propulsion efficiency.
Scale effects are naturally present for the propellers too. The lifting force and the resistance
developed on the propeller blades scale differently since pressure and friction resulting from
different scale laws (assuming that we have not been able to achieve true dynamic similarity). This
means that the open water propeller charts which is determined by model experiments have to be
recalculated to apply in full scale.
ITTC has recommendations for the implementation and re-scaling of the self propulsion test and
the open-water test.

SOME SYSTEMATIC SERIES OF RESISTANCE DATA


The collection is taken from Dyne (1991).

55

REFERENCES
Byquist T. Exempel p analysmetoder och modeller inom farkostteknik och tillmpad mekanik, (M. Huss
editor), TRITA-FKT kompendium 94/04, KTH 1994.
Dyne G. Fartygs motstnd Kompendium i kursen modellfrsks- och propellerteknik, CTH 1991.
Dyne G. Fartygs framdrivning Kompendium i kursen modellfrsks- och propellerteknik, CTH 1990.
Faltinsen O. M. Sea Loads on Ships and Other Floating Structures, Cambridge University Press, 1990.
Guldhammer H. E. & Harvald Sv. Aa. Ship Resistance, Akademisk frlag Kpenhamn, 1974.
Harvald Sv. Aa. Resistance and Propulsion of Ships, John Wiley & Sons, Inc. ISSN 0275-8741, 1983.
Holtrop & Mennen, A statistical power prediction method, International Shipbuilding Progress, Vol.
25, July 1978.
Holtrop & Mennen, An approximate power prediction method, International Shipbuilding Progress,
Vol. 29, July 1982.
Holtrop, A statistical re-analysis of resistance and propulsion data, International Shipbuilding Progress,
Vol. 31, November 1984.
ITTC, Report of Performance Committee, Proc. 14th ITTC, 1978.
ITTC, The Specialist Committee on Powering Performance Prediction, Proc. 24th ITTC, 2005.
van Lammeren, W. P. A. et al., The Wageningen B-Screw Series, SNAME Transactions Vol. 77 p269317, 1969.
Lewis (editor) Principles of Naval Architecture, The Society of Naval Architects and Marine
Engineers, 1988.
Lindgren & Bjrne, The SSPA Standard Propeller Family Open Water Characteristics, SSPA
meddelanden nr 60, 1967.
Massey B. S. Mechanics of Fluids, Chapman and Hall Ltd, 1990.
Man B&W, http://www.manbw.com/category_000077.html, 2005a.
Man B&W, Basic Principles of Ship Propulsion, http://www.manbw.com/article_003859.html,
2005b.
Milchert T. Handledning i fartygsprojektering, KTH, 2001.
Molland A., Turnock S. & Hudson D., Ship Resistance and Propulsion, ISBN 978-0-521-76052-2,
Cambridge University Press, 2011.
Nreskog J. On the Use of Concurrent Engineering in the Design of Ships, TRITA-FKT Report 98/10,
KTH 1998.
RINA, Significant Ships of 2004, The Royal Institution of Naval Architects, 2004.
Rawson K. J. & Tupper E. C. Basic Ship Theory, Butterworth and Heinemann, 2001.
Savitsky D. Chapter IV Planing Craft, Naval Engineers Journal, February 1985.
Savitsky D. Hydrodynamic Design of Planing Hulls, Marine Technology, 1964.
Schneekluth & Bertram, Ship Design for Efficiency & Economy, second edition, ButterworthHeinemann 1998.
Tornblad, Fartygspropellrar och fartygs framdrift, Marinlaboratoriet KaMeWa AB Kristinehamn, 1990.
Williams, The SSPA Cargo Liner Series Resistance, SSPA meddelanden nr 66, 1969.
Williams, The SSPA Cargo Liner Series Propulsion, SSPA meddelanden nr 67, 1970.

56

APPENDIX THE EXEMPLE SHIP PROTEFS


From: RINA, Significant Ships of 2004, The Royal Institution of Naval Architects, 2004.

57

58

APPENDIX RESIDUAL RESISTANCE COEFFICIENTS


From Ship Resistance Effect of Form and Principal Dimensions, Guldhammer & Harvald, Akademisk
forlag Copenhagen, 1974. The work comprise data for the slenderness parameter range 4-8.

59

60

61

APPENDIX PROPELLER CHARTS


From: Lindgren & Bjrne (1967), The SSPA Standard Propeller Family Open Water Characteristics,
SSPA Meddelanden Nr. 60, 1967.

62

63

Вам также может понравиться