An Optimization Study For The Bow Form of High Speed Displacement Catamarans
An Optimization Study For The Bow Form of High Speed Displacement Catamarans
An Optimization Study For The Bow Form of High Speed Displacement Catamarans
Displacement Catamarans
D. Bülent Danışman1, Ömer Gören1 (M), Mustafa Insel1 (AM), Mehmet Atlar2 (M)
Abstract
Interest in high-speed marine transportation has acquired a wide variety of hydrodynamic research
activities on catamarans, including shape optimization for minimum total resistance. This study attempts to utilize
the mathematical programming in optimizing the bow form of twin hulls for minimum total resistance as well as in
analyzing the optimization by physical model tests. The total resistance is assumed to be composed of wave-making
and frictional components, which are formulated using Michell’s thin-ship theory and ITTC-1957 friction line,
respectively. The optimized hull form is analyzed by means of a computational flow solver before going through the
experimental analysis. The study demonstrates the capabilities of the optimization procedure, presented for
catamarans, in reducing the total resistance, as well as its limitations to be used as a design tool in a relatively high-
speed zone.
1
Istanbul Technical University, Dept. of Naval Architecture, Istanbul, Turkey.
2
University of Newcastle, Dept. of Marine Technology, Newcastle upon Tyne, U.K.
resistance is then taken as the objective function, which I ×J I ×J
is quadratic in ship’s half breadths. The linear inequality Cw = d mn y m y n = y T D y ( 2)
constraints used in the quadratic programming problem m =1 n =1
allow to implement the necessary design constraints into
the procedure. The standard quadratic programming where I×J denotes the total number of half-breadths
problem is then solved by Wolfe’s [12] algorithm. since I is the number of waterlines while J is the number
Although it is a well-known fact that in the high speed of stations. The coefficient matrix D is derived adopting
zone, the percentage of the wave pattern resistance in the same approach in Hsiung [14]. The quantity
total resistance does not generally exceed 15%, the calculated by equation (2) may not be considered as a
present study points out that by means of the good approximation of the wave resistance for high
optimization procedure it is possible to attain up to 5% speed vessels with transom stern. However this value,
gains in total resistance which may be regarded as which is of the same order as wave resistance, can
considerable for catamarans operating at high Froude relatively be minimized within the linear theory
numbers greater than 0.50. formulation.
The other component of the resistance that appears
in the objective function is the equivalent flat-plate
2. MATHEMATICAL PROGRAMMING frictional resistance, which depends on the wetted
PROBLEM FOR TWIN HULLS surface area approximated as
é 1 2 1 2 ù
The previous works of Gören and Calisal [13] and S =2 ê1 + 2 Fx ( x, z ) + 2 Fz ( x, z ) dxdz ( 3)
Gören et al. [11] is extended to cover twin hull ë
S0
geometries. Lunde’s [8] well-known formulation is used
to express twin hull wave resistance based on thin-ship Where, S0 is the projection area of the wetted surface of
theory; the ship on the center-plane. Fz is the derivative with
∞ respect to z. The use of tent functions is again made in
ρg (u 2 + 1) 2 (3) to give the frictional resistance coefficient, which is
RW = 4 2 1
(P 2 + Q 2 )
πc 2
(u + 2 ) 2
(1.a) based on the ITTC-1957 formula, in a quadratic form in
0
terms of the ship half breadths as in the following form:
ì é g 1 ùü
2 2
í 2 + cos ê2b 2 (u + 1)u (u + 2) 2 ú ýdu C F = c0 + c0 y + y T A y (4)
î ë c û F
In accordance with the assumption made for the total
Pü
T L
cos é g resistance, the objective function used in the
ù
ý= Fx ( x , z ) x(u 2 + 1) optimization procedure is taken as the sum of wave
Q sin êë c 2 resistance (2) and the frictional resistance (4):
0 0
ég ù CT = C F + CW = c0 + cb y + y T C y (5)
exp ê 2 ( z − T )(u 2 + 1) 2 dxdz (1.b)
ëc Note that C is a positive semi-definite, that is a
where c is the ship’s speed, ρ is the density of water and symmetric matrix.
g is the gravitational acceleration. L is the length of the The quadratic character of the objective
ship, T is the draft and Fx denotes the derivative of the function implies the employment of the quadratic
hull surface with respect to x. The coordinate axis used programming in which the design constraints can be
in the formulation is depicted in Figure 1. given by a set of inequalities. Thus the general form of
the quadratic programming is expressed as
to minimize p y + yT C y
( 6)
to satisfy A y ≤ B with y i ≥ 0
Lwl 25.25 m
Bmax 2.6 m
T 1.4 m
CB 0.410
10
6
7
6
7 5
11
9
8
12
6 5
10
12
7
10
8
7 5
9 9
11
8
9 6
8 10 5 6 10
12
8 7
7
9
11
8
9
6 6 5
10 11 9 5
6
7
11 10
11
13
4 7
12
8
13
5
6
4
12
14
1 2 13
13 4
9
76
1
2
1
10 8
10
9
11 11 10 6
12
11
4
8
-20 -10 0 10 20 30
X
Figure 3. Geometric modeling of optimized demihull and its free-surface vicinity.
Figure 5a. Computed wave pattern around the original demihull (Fn=0.65).
0.5
0.4
0.3
0.2
Wave Deformations (m)
0.1
20
NPL
0 18
-30 -20 -10 0 10 20 30 40 Optimal Level 1 2 3 4 5 6 7 8 9 10 11 12 13 14
F.P. A.P. 16 Deformation (m): -0.563 -0.492 -0.421 -0.350 -0.279 -0.208 -0.137 -0.066 0.076 0.147 0.218 0.289 0.360 0.431
-0.1 7 6
11
10
14
9
8 7 6
11
6
8
12
-0.2
8 12
10
7
Y
6
9
10 5
11
6
8
7
9
-0.3 8
6
8 9
12
7
5
10
9
5 6 8
11
9 7
6 8 5 9 13
10
12
10 9 7 5
10
-0.4 9 24
12 12 13
4 6
10
11
7
13
14
11
9
8 7
2 111
0 8 8 21 1 3 11 2 3
8
6
9 2 1 1
13
5 1
10
4
-0.5 9 10 4
14
9
3
6
7
5
Distance From Mid-ship (m) -20 -10 0 10 20 30
X
Figure 4. Computed wave elevations along the outer side of the demihulls Figure 5b. Computed wave pattern around the optimal demihull (Fn=0.65).
(Fn=0.65).
Monohull (Demihull in isolation) Twin Hulls, s/L=0.2
0.010
0.013
CT, optimized form (Exp.)
0.009 0.012 CT, optimized form (Exp.)
CT, original form (Exp.)
0.011 CT, original form (Exp.)
0.008 0.010
CT
0.009
CT
0.007
0.008
0.006 0.007
0.006
0.005 0.005
0.30 0.40 0.50 0.60 0.70 0.80 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
Fn Fn
Twin Hulls, s/L=0.3 Twin Hulls, s/L=0.4
0.013 0.012
0.012 CT, optimized form (Exp.)
0.011
CT, optimized form (Exp.)
0.011 CT, original form (Exp.) CT, original form (Exp.)
0.010
0.010
0.009
0.009
CT
CT
0.008 0.008
0.007 0.007
0.006 0.006
0.005 0.005
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
Fn Fn
Figure 7. Comparison of experimental total resistance coefficients for original and optimized models.
Mono hull (Demihull in isolation) Twin Hulls, s/L=0.2
0.09 0.12
Original form (Exp.)
0.08 Original form (Exp.)
Optimized form (Exp.)
0.07 0.10 Optimized form (Exp.)
CWP Distribution
CWP Distribution
0.08
0.08
0.06 0.06
0.04 0.04
0.02 0.02
0.00 0
0 10 20 30 40 50 60 70 80 0 20 40 60 80 100
Θ=(Wave angle) Θ=(Wave angle)
Figure 8. Measured wave spectrums for original and optimized forms, (Fn=0.65).
4. EXPERIMENTAL WORK ρgW é 2
( ) ( )æçç1 − cos 2(θ ) ö÷ù (8)
N 2
2 2 2 n
RWP = ê ξ 0 +η 0 + ξ n +η n ÷
2 ëê n =1 è ø
In parallel to the numerical investigation, the
physical model tests were carried out to investigate the In this case, resistance components are expressed
resistance characteristics of the original NPL form and as: RT = RV + RWP = (1 + k ) RF + RWP , (1+k) is being the
the optimized hull form in Ata Nutku Ship Model form factor. The results of the wave pattern resistance
Testing Laboratory (160m x 6.0m x 3.40m) of I.T.U., are shown in Figure 8 in terms of wave spectra only for
which is an ITTC Class Facility. The experiments were the case of s/L=0.3. The optimal form reduced the wave
performed for two sets of measurements, namely, total pattern for the wave angles up to 60°. It is also effective
resistances and wave pattern resistances. for all tested demihull separations. In Figure 9, it is
Associated models were built with a length of 2.5 possible to compare the numerical and experimental
meters and tripwires were used for the turbulence results for the gap ratio s/L=0.3. It should be remarked
stimulation, as shown in Figure 6. The test cases here that there is 5% excess of wetted surface area in
covered were the mono hull (i.e. one demihull in optimized form as compared to that of the original form.
isolation) and twin hulls having s/L ratios of 0.2, 0.3 The experimental and numerical analyses give the same
and 0.4, [19]. order of reductions in wave pattern resistance, where as
a shift is observed at crest and troughs of the resistance
curves. This may be attributed to the inherent phase
shifts in Dawson’s algorithm, which appear
considerably when there is interference. It is interesting
to note here that the present optimization is more
successful in reducing divergent waves rather than
reducing the interference effects in high-speed zone.
This can be seen in Figure 10. Ultimately, the effective
power requirements of the original form and the
optimized form for s/L=0.3 are given in Figure 11 for
Figure 6. Bow form view of the two models: Original comparison. It is clear that there is at least 4% reduction
NPL form and its optimized counter part (at the back). in required effective power at Fn=0.65
n=0
CW
0.004
where W is the width of the tank, ωn = K n cos(θn ) , 0.003
2πn g 0.002
K n sin (θn ) = and K n − 2 sec2 (θn ) = 0. 0.001
W V 0
Here n, Kn and θn denote the nth harmonic, wave number 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
and wave angle, respectively. Thus, wave pattern Fn
1.60
Original form (Exp.) s/L=0.3 River Banks on The Resistance of a
Catamaran”, Proc. FAST’93, Yokohama, 1993.
RW cat/(2RWmono )
8000
FAST’97, Sydney, 1997.
6000
[11] GÖREN, Ö., HELVACIOĞLU, Ş. and INSEL,
4000
EHP Optimized (Exp.), s/L=0.3
M., “Bow Form Optimization of Displacement
2000 EHP Original form (Exp.), s/L=0.3
Ships by Mathematical Programming”, Ship
0 Technology Research, Vol. 27, No. 3, 1997.
[12] WOLFE, P., “The Simplex Method for
6.0 11.0 16.0 21.0 26.0
V(knot) Quadratic Programming”, Econometrica, Vol.
27, No.3, 1959.
Figure 11. Effective power for full-scale twin hulls
[13] GÖREN, Ö. and CALISAL, S. M., “Optimal
ACKNOWLEDGEMENT Hull Forms for Fishing Vessels”, 13th STAR
Symp., SNAME, Pittsburgh, 1988.
Funding for this project was partially provided by [14] HSIUNG, C.C., “Optimal Ship Form for
Istanbul Technical University (ITU) R&D Funds. A Minimum Wave Resistance”, J. Ship Research,
joint support of Scientific and Technological Research Vol.25, No.2, 1981.
Council of Turkey (TUBITAK) and the Royal Society, [15] SIDI, A., “The Numerical Evaluation of Very
given to the second author for his visit to the University Oscillatory Infinite Integrals by Extrapolation”,
of Newcastle, is gratefully acknowledged. Mathematics of Computation, Vol.38, No.158,
1982.
[16] BAILEY, D., “The NPL High Speed Round
REFERENCES Bilge Displacement Hull Series”, Marine Tech.
Monograph, Nr.4, RINA, 1976.
[1] INSEL, M. and MOLLAND, A. F., “An [17] GÖREN, Ö. and ATLAR, M., “A
Investigation into the Resistance Components Computational Study for the Wave Resistance
of High Speed Displacement Catamarans”, Analysis of Multi-Hull Forms”, University of
Trans. RINA, April 1991. Newcastle, Department of Marine Technology
[2] BRUZZONE, D., CASELLA, P., PENSA, C., Report: MT-1998-031, August 1998.
SCAMARDELLA, A. and ZOTTI, I., “ On the [18] DAWSON, C. W., “A Practical Computer
Hydrodynamic Characteristics of High Speed Method for Solving Ship-wave Problems”, 2nd
Catamarans with Round-Bilge Hull”, Proc. Int. Conf. Numerical Ship Hydrodynamics,
FAST’97, Sydney, 1997. Berkeley, 1977.
[3] SUZUKI, K., NAKATA, Y., IKEHATA, M. [19] DANISMAN, D. B.; “Bow form Optimization
and KAI, H., “Numerical Prediction on Wave of Twin Hulls for minimum Total Resistance”,
Making Resistance of High Speed Trimaran”, (in Turkish), MSc Thesis, I.T.U. School of
Proc. FAST’97, Sydney, 1997. Graduate Studies, Istanbul, 1997.
[4] LARSSON, L., JANSON, C.-E. and BRUN,
F., “A Numerical Investigation of Trimaran
Configurations”, Proc. FAST’97, Sydney,
1997.
[5] HSIUNG, C.C. and XU, H.,” Determining
Optimal Forms of a Catamaran for Minimum
Resistance by the Mathematical Programming
Method”, Schiffstechnik, Vol. 35, No.1, 1988.