Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.

com, ISSN 1743-3509

Wave resistance prediction of a catamaran

by linearised theory
M. Insel," A.F. Molland& & J.F. Wellicome*
^Faculty of Naval Architecture and Ocean
Engineering, Istanbul Technical University, Turkey
^Department of Ship Science, University of
Southampton, Highfield, Southampton, SO 17 1BJ,
Hampshire, UK

ABSTRACT

The far field wave system and far field wave coefficients (Eggers coefficients)
of a Kelvin source are presented using linearised theory including the effects
of a canal depth and width. The wave resistance due to the distribution of such
sources is derived using a far field wave energy approach.
The trim and sinkage effects of the hull have been included by
supplying the dynamic hull surface in the calculation scheme. The effect of
transom stem has also been investigated by alternative methods. The developed
code has been tested on a mathematically defined hull form and a series of high
speed transom stern hull forms. Catamarans with separation ratios of 0.2 to 0.5
are used in the comparisons of the method with experimental results.

INTRODUCTION

Use of linearised wave resistance theory for the predictions of wave resistance
of catamarans can be justified due to the slender form of the demihulls. A
theoretical approach with extensions for transom stern, running trim and
sinkage corrections is described to calculate the wave pattern and the wave
resistance of a catamaran.
In the current work, the far field wave system of a Kelvin source in a
shallow water canal is presented. The wave resistance due to the distribution
of these sources is given using a far field wave energy approach. Far field wave
coefficients, i.e. Eggers coefficients^, are also presented. The hulls have been
represented by centerplane Kelvin source distribution conforming to thin ship
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

60 Marine, Offshore and Ice Technology

assumption. The trim and sinkage effects on the wave resistance have been
included by supplying the dynamic hull surface into the calculation scheme.
The effect of a transom stem has also been investigated by alternative methods.
The developed code has been tested on a mathematically defined hull
form. Various program parameters are investigated with this form. A series of
high speed transom stern hull forms varying in length to beam ratio are utilised
to investigate the results of the predictions on practical hull forms. Finally the
code has been applied to catamaran forms to calculate the wave interference
between the demihulls. A separation ratio range from 0.2 to 0.5 was utilised as
well as demihulls in isolation. The numerical results and the experimental
results are compared in detail.

BASIC ASSUMPTIONS AND BOUNDARY CONDITIONS

The initial assumptions for the mathematical model of the problem are: the
fluid is ideal; incompressible and homogenous; the flow is steady and
irrotational; the surface tension can be neglected; the wave height is small
compared with the wave length.
A cartesian coordinate system moving with the model and having its
origin on the free surface at the model centre is used. Ox is in the direction of
motion, Oy and Oz are to the starboard and vertical upwards respectively
(Figure 1). As the model advances in the positive direction in the x-axis at a
constant speed U in a canal having width of W and depth of H, the velocity
potential can be expressed as a summation of the free stream and flow
disturbance potentials.
(1)

Assuming the free surface is expressed as z=C(xj) and the underwater

geometry of the ship is represented by y=T/foz) (for -L/2<x<L/2 and -T<z<0),
the following linearised boundary conditions must be satisfied;
i) Laplace equation

dr dy* dy*
it) Free surface conditions
a) Dynamic free surface condition
2(+[%=0 atz-Q (3)

b) Kinematic free surface condition

tfC,-*j=0 at z=0 (4)

or by combining both conditions

(5)
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

Marine, Offshore and Ice Technology 61

iii) Bottom condition and Canal Side Condition

4>,=0 at z=-H $y=Q at y = ± W/2 (6)

iv) Radiation condition

litri <|> _ J o(l) for X<Q (j\
(**+y *)-*«>" 1 0 for x>Q

v) Hull surface condition

(#L+<k=0 at V==F/(JC.Z) (8)

The solution of this problem can be satisfied by representing the hull by

a distribution of sources satisfying the same boundary conditions.

FAR FIELD WAVE SYSTEM OF A SOURCE IN A SHALLOW CANAL

The velocity potential of a source with density of y^ and located

in shallow water with the depth of H and bounded by a canal having vertical
walls with width of W (Figure 1) can be obtained by considering images of the
source due to tank walls at yQ=y^+2nW and yQ=~yQ+(2n+l)W for
n =-«>... oo and the image due to tank bottom at Zi=-(2H+z<). The far field
velocity potential of a source can then be found and is given by INSEL^ as;

4W= z L" Cos(uj)+ "

(9)
Cosh(KJz+H)) ^ Cos(mnylW) for even m
Cosh(KJff) * Sin(mnylW) for odd m

where
*o) * ,, for even m ,+ ^
^\fa,,rfyo' ,, _ (10)

»fc (\-KJI * ° "

. (is)

The far field wave elevation can be found from free surface condition.
This corresponds to Eggers wave series and is given by INSEL ^ as;
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

62 Marine, Offshore and Ice Technology

r - v r *" r ^ < r Cos(m^y/W) for even m

"•
For practical calculations m can be truncated at a finite number M
(typically 100). Equation 14 gives the far field wave elevation of an arbitrary
source-sink distribution. £^, T]^, a^, P^ are the wave coefficients. (^ andi]^
represent the symmetric wave distribution relative to the tank centerline,
meanwhile a^ and P^ represent the asymmetric wave pattern. Hence for a
symmetric hull form relative to the canal centerline <x^, P^ are zero. By
substituting n=2m, (M=2N) the wave elevation (equation 14) becomes:
N
C=E [ «. Cos(<*s)+ TI. Si«(cv)] Cos&mylW) (15)
11=0

WAVE RESISTANCE OF A SHIP MODEL BSf A SHALLOW CANAL

The wave resistance of a ship model can be obtained from considerations of

energy changes as given in

Stnh&KJO)
(16)

In equation 16, ^ ^^ "Hm *™* ^m ^^ Pm represents wave resistance

due to symmetric wave pattern and asymmetric wave pattern respectively.

NUMERICAL EVALUATION OF THEORETICAL WAVE RESISTANCE

Based on the method described above, a computer program (WAVE2DF) for

the determination of theoretical wave resistance for a ship model in a shallow
water canal has been developed. The hulls are discretised into quadrilaterals.
Each quadrilateral is modelled by a surface source distribution over the
projection of the quadrilateral on the centerplane.

•^1234 f *
A test case was conducted by using a mathematically defined hull form,
WIGLEY HULL <*>. The hull form is given by:

Effects of Trim and Sinkage

Preliminary calculations and experimental results of the Wigley hull indicated
that trim and sinkage of the hull affects wave resistance. Hence trim and
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

Marine, Offshore and Ice Technology 63

sinkage must be included in the calculations. Program WAVE2DF has a

subroutine which reforms the input hull shape into running hull shape by taking
the trim and sinkage into account. The sections are represented by splines and
are offset by the trim and sinkage. Then the waterlines are interpolated from the
sections. Finally, panels are formed from these sections and waterlines.

Effect of Transom
Catamaran demihulls due to their high speed capabilities feature with transom
sterns. Flow at the transom stern displays different properties at the low speeds
and high speeds on the resistance. At the low end of the speed range, the
transom stem is wet and features a dead water zone-rotational flow. However
at the high end of the speed range, the transom stern is dry. Resistance features
of the wet and dry flow are also different. Various methods are introduced to
calculate transom stern drag. The methods reviewed here are :
a) Hydrostatic term approach-Dry Stern (by Chanel: It assumes that the
transom is running completely dry (fully ventilated). Transom drag is given by:
Rn=fjpg z dzdy (19)

b) Base drag approach-Wet Stern (by Hoerner*): From the tests on projectiles
and fuselages a drag component known as base drag was measured and
expressed as :
j (20)

c) Sink line distribution at transom stern-Dry Stern (by Yim *): Yim has
suggested that transom stem could be represented by a sink line at the bottom
of the transom. Sink strength is given by :
H(JWW=-^o (21)

As explained above some of the methods are applicable to only wet

transom stem while others are applicable to only dry transom stern flows.

FAR FIELD WAVE SYSTEM AND WAVE RESISTANCE OF A

CATAMARAN IN A SHALLOW CANAL

As a special case a catamaran made up two symmetrical demihulls is

investigated as it is the main concern for practical applications. For such a craft
two hulls will be located at y^-S/2 , y^=S/2. Equation 15 and 16 become <*>:

,, ,. (22)
n=0 W
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

64 Marine, Offshore and Ice Technology

, 2*0* ^

(23)
(1+
Sinh(2K^H)
where
S/2 SinQ}=Cos(iwSIW) (24)

It is noted that (^ and T|% can be obtained either theoretically using

equation 10 or by applying experimental values.

COMPARISONS OF EXPERIMENTS AND NUMERICAL

CALCULATIONS

Extensive comparisons of the numerical method with experimental results,

given by Insel and Molland^, were conducted. Firstly comparisons are carried
out with mathematically defined form (Figure 2) in fixed to trim and sinkage
conditions. As this condition eliminates the problems of transom stern, running
trim and sinkage, numerical results can be assessed. Figure 3 compares the
experimental and theoretical results. The numerical results are rather oscillatory
for the fixed model. This effect is mainly due to exaggeration of the
interferences of bow waves and stem waves. In general, the theoretical
calculations are higher than the experimental results.
Secondly in order to investigate the running trim and sinkage effects the
mathematically defined form was tested free to trim and with sinkage. The
results are given in Figure 4. These results are very similar to the fixed results.
The oscillatory nature of the theoretical calculations is dominant.
A round bilge series (Figure 2) (derived from NPL round bilge series
by changing L/B ratio to 7,9 11 for C3, C4, C5 respectively) of catamaran
forms were used to investigate the catamaran interference for which the
experimental results are given in Reference 2 and 6. In Figure 5, 6 and 7 the
experimental results (Cw=Cr-(l+k)Cp) are compared with theoretical
calculations with and without transom corrections for C3, C4 and C5 demihulls
respectively. Wave resistance calculations are corrected by transom corrections
with a: indicating corrections by Chang's method, b: indicating corrections by
Hoemer's method, c: indicating corrections by Yim's method. For the low
speed end, the base drag approach by Hoerner is sufficiently good. For the high
speed end Yim's method gives good results. The hydrostatic term approach by
Chang always overpredicts the wave drag. Hence a combination of base drag
and Yim's method offers the best approach. However lack of information for
ventilation speed makes this approach very difficult, hence only Yim's transom
stern sink method has been utilised for the catamaran calculations.
Experimental results for catamaran combinations (demihull, separation
to length ratio of 0.2, 0.3, 0.4 and 0.5) are presented in Figure 8a, 9a and lOa.
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

Marine, Offshore and Ice Technology 65

The calculated wave resistance with transom correction is given in Figures 8b,
9b and lOb. As noted earlier, errors in low speeds are high (for Fn<0.4).
However for the higher speeds, the effect of interactions are found in the
calculations. Theoretical calculations for Froude number over 0.7 are correct in
magnitude as well. This would indicate Yim's method is really applicable to
high speeds with the current calculation scheme.

CONCLUSIONS

A numerical approach to multihull wave resistance has been developed

including the effects of the canal wall and shallow water. The current approach
is suitable for any multihull provided the hulls are symmetric relative to the
centerplane and fairly thin. The wave pattern and wave resistance of
catamarans is presented as a special case.
The effect of trim and sinkage is demonstrated as an important variable
for high speed multihulls. The transom effects for low speed and high speed is
investigated. The determination of ventilation speed is required for an effective
calculation scheme.
Comparisons between numerical and experimental results for a round
bilge series indicated that interference effects are predicted satisfactorily for
high speeds (over Fn=0.4) when a sink line distribution is used for the transom
stern.

REFERENCES

1. EGGERS, K.W.H., & SHARMA S.D., WARD L.W. An assesment of

some experimental methods for determining the wavemaking
characteristics of a ship form, Transactions ofSNAME, 1967
2. INSEL, M., A investigation into the resistance components of high
speed catamarans, Ph.D. Thesis, Department of Ship Science,
University of Southampton, 1990
3. CHANG, M.S., Wave resistance predictions using a singularity method,
Proceedings of the Workshop on Ship Wave Resistance Computations,
Maryland USA, Vol:H,1979
4. HOERNER, S.F., Fluid-Dynamic drag, Published by the author, 1965
5. YIM, B., Analyses of waves and the wave resistance due to transom
stern ships, Journal of ship Research, June 1969
6. INSEL, M., & MOLLAND A.F., An investigation into the resistance
components of high speed catamarans, Transactions ofRINA, 1992
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

66 Marine, Offshore and Ice Technology

Figure 1: Axis system and canal dimensions

Model C2 Model C3.C4-.C5

Figure 2: Mathematical and conventional model hull forms

Figure 3: Comparison of experimental and Figure 4: Comparison of experimental and

theoretical results (C2-Fixed) theoretical results (C2-free)

Figure 5: Comparison of experimental and Figure 6: Comparison of experimental and

theoretical results (C3-Demihull) theoretical results (C4-Demihull)
 Transactions on the Built Environment vol 5, © 1994 WIT Press, www.witpress.com, ISSN 1743-3509

Marine, Offshore and Ice Technology 67

Figure 7: Comparison of experimental and theoretical results (C5-Demihull)

Figure 8a: Experimental results for hull C3 Figure 8b: Theoretical results for hull C3

Figure 9a: Experimental results for hull C4 Figure 9b: Theoretical results for hull C4

Figure lOa: Experimental results for hull C5 Figure lOb: Theoretical reults for hull C5