27th Symposium on Naval Hydrodynamics

Seoul, Korea, 5-10 October 2008

Performance prediction and computation of slamming loads

on a planing craft using RANSE method
Manoj Kumar1, V. Anantha Subramanian2, Anant Lal1 and S. P. Singh1
(1Indian Register of Shipping, India, 2Indian Institute of Technology Madras,
India)
ABSTRACT
The paper presents free-surface RANSE computation
of a planing craft taking into account dynamic sinkage
and trim. The motion of the ship and wave dynamics
have been implemented in a commercial RANSE code,
Fluent 6.2 through user codes. The method has been
validated with towing tank experiments for a wigley
hull. The response of the boat to incident waves,
vertical acceleration and hydrodynamic impact loads
acting on it are also presented for regular and irregular
waves.
INTRODUCTION
Ship motions in waves can be predicted using
experiments or numerical methods. The numerical
methods [1] widely used are based on potential theory,
which assumes an irrotational ideal fluid. These
methods are considered as fast and robust tools in the
design stage as they allow large number of parameters
to be analyzed for the purpose of optimization. The
motion of simple bodies in waves can be computed
using these methods with reasonably good accuracy.
However, they are not suitable for flows, where
viscous effects or breaking waves play an important
role such as roll motion of a ship with bilge keel.
Strong nonlinear effects like slamming require more
sophisticated methods.
Earlier attempts to solve slamming problem were based
on the estimation of slamming pressure on a two
dimensional wedge section with a known velocity of
entrance. Wagner [2] presented a theoretical solution
for the water entry of 2D sections. Zhao et.al.[3]
presented a generalised Wagners solution. The above
methods are based the potential flow techniques.
In the recent past, the methods based on CFD
techniques are getting more popular. These methods
use the solution of Reynolds Average Navier Stokes

equations. The method is theoretically better equipped

to simulate the physics associated with the problem.
Capturing of two-phase interface is also better in these
methods.
Estimation of ship slamming pressure for the design
purpose is largely based on empirical formulae
provided by the classification societies. The empirical
formulae are mostly based on the operational
experiences. Another approach is to estimate the
slamming pressure using a formula of type p = kV ,
where V is the velocity of entrance, which is
sometimes calculated using the statistical methods e.g.
Ochi & Mottor 1973 [4]. The factor k can be
determined using the experiments or empirical
formulae given in the class rules. These empirical
formulae provides a quick assessment for the slamming
pressure, however many a times it is observed that the
pressure obtained from these formulations are
unrealistic. Application of such formulae for novel
designs is also questionable.
2

With the increase in computational power it is now

possible to carry out direct calculations based on
advance numerical methods. With the trend shifting
towards direct strength analysis and direct load
estimation, it is more logical to adopt to such direct
estimation tools.
In the recent past many researchers have carried out the
direct computations of slamming load estimation.
There are two approaches. In the first approach the ship
motions are estimated using ship motion code (usually
potential flow based) and then subsequently these
motions are used to estimate the slamming pressure at
the desired places. Hermundstad and Moan [5]
estimated the slamming pressures on a 2D ship
sections, here the ship motions were estimated by the
2D nonlinear strip theory, the motion thus obtained
were used to estimate the slamming forces/pressure

using the method proposed by Zhao et. al. [3]. A

similar approach was also used by Moctar et. al. [6],
however the main difference here was that in this case
a CFD technique was used to estimate the 3D
slamming pressure. The main assumption in this
approach is the assumption that slamming forces does
not affect the global ship motions.
The second approach is to solve coupled ship motion
and slamming problem using a CFD technique. One of
the drawbacks of this approach is high computational
resource, which has become relatively cheaper in
recent years. In the present paper this method has been
used to compute ship motion and slamming impact
loads on the ship hull in waves. The ship motion is
implemented in the RANSE solver through user code.
Rigid motion equations are solved ever time steps and
the mesh is updated. Calculations are performed for a
planing craft. The method to compute ship motions has
been validated with towing tank experiments for a
wigley hull. In order to validate the CFD method for
impact load calculation, computations were also
performed for a typical wedge section (Zhao et. al.
1996), which have been presented in Singh and Kumar
[7].
NUMERICAL METHODS

Rigid body motion

The force on the hull is calculated by integrating the
RANSE predicted pressure and frictional shear on the
hull. Similarly the moment is calculated by integrating
the pressure and frictional shear on the hull along the
desired axis. The force and moment acting on the hull
are then fed into the solver to compute the translational
and angular motion of the center of gravity of the ship
hull. The governing equation [9] for the translational
motion of the center of gravity is solved in the inertial
coordinate system given by Equation 1.

vG =

(1)

1
fG
m

Where

vG

is the translational acceleration of the center

of gravity about earth fixed co-ordinate system, m is

the mass and f G is the force vector due to gravity,

pressure and frictional shear. The angular acceleration
of the body

wG

is more easily computed using body

coordinates.

wG = L1 ( M B wB L wB )

(2)

L is the inertial tensor, M B is the moment vector of the

body about the body co-ordinates due to the pressure

The Finite Volume commercial CFD code, Fluent 6.2

has been used to carry out the computations. The
numerical approach involves discretizing the spatial
domain into finite control volumes using a mesh. The
governing equations of mass and momentum are
integrated over each control volume, such that the
relevant quantity (mass, momentum, energy, etc.) are
conserved in a discrete sense for each control volume.
The volume of fluid (VOF) method has been used to
capture the free surface. It was originally proposed by
Hirt and Nichol [8], who used a finite-difference
donor-acceptor scheme to capture the moving free
surface interface of a fluid. In the present evolved
form, it is widely used to compute the free surface. The
location of the two fluids is specified using a volume
fraction function, q , with q =1 inside one fluid
and q = 0 in the other. The cell for which q lies
between 1 and 0 contains the interface between qth
fluid and one or more other fluid.
For two phases, = 1 1 + (1 1 ) 2
For n phases, = q q and q = 1
n

A single momentum equation is solved throughout the

domain, replacing density in the original conservation
equations with the modified density based on the
volume fraction of each phases. The turbulence was
modeled using Reynolds Stress model.

and frictional shear on the hull, and wB is the rigid

body angular velocity vector. Once the angular and
translational accelerations are computed from Eq (1)
and (2), the angular and translational velocities are
obtained by numerical integration, which are used in
the dynamic mesh calculations to update the rigid body
orientation.

Domain and Grid

Figure 1: Body plan of the hull

The details of the ship hull chosen for the numerical

simulations have been given in Table1. The body plan
of the hull, Model 4667-1 [10], is shown in Figure 1.
The domain used for the numerical computation
consists of a rectangular box, which includes both
water and air. The ship hull motion has been realized
by imposing motion, obtained by solving rigid body

motion equations, to a box shaped sub-domain as

shown in Figure 2. The volume mesh in the rest of the
domain is free to deform. This choice of sub-domain
ensured error free grid deformation from moderate to
extreme hull movements. The domain is discetized into
hexahedral cells using Ansys ICEM-11.0. A meshed
domain, with 125398 hexahedral cells, was well able to
capture the flow and the motion. The same grid was
used in all simulations. A sufficiently refined grid near
the hull was ensured. Figure 3 shows the blocking
strategy used to generate the hexahedral grid. The
surface mesh on the planing hull is shown in Figure 4.
Figure 5 shows the mesh on the center plane of the
hull.

Figure 5: Mesh on the centre plane of the hull

PLANING HULL IN CALM WATER

The domain extended to 1.0 L in the front of the hull,
3.0 L behind the hull and 1.5 L each on the both sides
of the hull. The distance from the ship to the bottom
and top is 1.0 L and 0.7 L respectively. The entire outer
boundary, except the outlet and the top, has been
modeled as velocity inlet (Figure 2).

Figure 2: Domain and boundary conditions

Simulations for various Froude number, covering

around 0.3 to 2.6, were carried out in calm water. The
numerical method allowed the hull to have two degrees
of freedom, namely trim and sinkage. The numerically
computed resistance, trim, sinkage and wetted surface
area have been compared with experimental results in
Figures 6, 7, 8 and 9 respectively. Figure 10 shows the
free surface profile on the centre plane of the hull for
various Fn. It can be observed from Figure 10 that
initially the hull behaves like a displacement hull. At
higher Fn a typical dry transom is observed and the
hull plans on the water surface. Figure 11 shows the
free surface contour around the hull for various Fn.

250

Figure 3: A part of the Blocked domain

Resistance (N)

200
150
100

Exp

CFD

50
0
0

0.5

1.5
2
Froude No.

Figure 6: Total Resistance curve

Figure 4: Surface mesh on the hull

2.5

Fn =0 .312

6
5
Trim (Degree)

Exp

CFD

4
3
2
1

Fn = 0.532

0
-1

0.5

1.5

2.5

Froude No.

Figure 7: Trim
0.1

Fn = 0.532

Sinkage (m)

0.08
0.06
0.04

Exp

CFD

0.02
0
-0.02

0.5

1.5

2.5

Fn = 0.532

Froude No.

Figure 8: Sinkage

Fn = 0.532

Wetted Surface Area (m 2)

1.6
1.4
Exp

1.2

CFD

1
0.8
0.6
0.4

Fn = 0.532

0.2
0
0

0.5

1.5

2.5

Froude No.

Figure 9: Wetted surface area

Table 1: Particulars of the hull (model) used for calm water

analysis
LP (m)

BPA (m)
BPX (m)
BPT (m)
Mass (kg)

2.438
0.487
0.596
0.381
100.3

Fn = 0.532

Figure 10: Hull orientation and free surface profile on the

centre plane of the hull for various Fn

Fn =0 .312

Fn = 0.532

Fn = 0.532

Fn = 0.532

WATER ENTRY OF WEDGE

In order to validate the RANSE based slamming load
computation, simulation of the water entry of a wedge
section was carried out. The details of the wedge
geometry, entrance velocity etc, are given in Zhao et al.
[3]. In their experiment, the vertical velocity during the
water entry was measured. In the numerical simulation,
the vertical velocity was taken as per Figure 12, which
is obtained by fitting a polynomial to the curve
presented in Zhao et al. [3]. Numerical simulations
were carried out for both 2D and 3D cases. The details
of the numerical procedure are given in Singh and
Kumar [7].
The comparison of 2D and 3D simulations is shown in
Figure 13. The results are compared with experimental
results of Zhao et al. [3]. It can be observed that 3D
result is closer to the experimental result. It is also
observed that the 2D simulation over-predicts the
force. Hence a 3D simulation is essential.
Figure 14 shows the pressure distribution along the
depth of the wedge at time t = 0.0158 sec. The pressure
is expressed in term of pressure coefficient Cp given as
p pa
Cp =
0.5V 2 , where p is the absolute pressure, p is
a

the atmospheric pressure, is the fluid density and V

is the entrance velocity.

Fn = 0.532

Fn = 0.532

Vertical Velocity (m/sec)

6.5

6.0

5.5

5.0

4.5
0

0.005

0.01

0.015

0.02

0.025

Time (Sec)

Fn = 0.532

Figure 11: Free surface profile around the hull for various Fn

Figure 12: Vertical velocity during the water entry of Wedge

predicted wave profile agrees closely with the

experimental observation though the wave elevation in
the trough of the wave is little over predicted by CFD
(Figure 16).

10000
9000
Exp

8000

3D

2D

Force (N)

7000
6000
5000
4000
3000
2000
1000
0
0

0.005

0.01
Time (Sec)

0.015

Figure 13: Comparison of 2D and 3D simulation with

experiment
10
9

2D

Press Coeff.

3D

Exp

Figure 15: Velocity vectors on a plane in the tank

7
6
5

0.08

CFD

0.06

Exp

0.04

1
0
0

0.2

0.4

0.6

0.8

Non-Dimentional Depth

Free Surface (m)

0.02
0
1

-0.02
-0.04

Figure 14: Pressure coefficient along the depth at t = 0.0158

sec

-0.06
-0.08

Time (Sec)

MOTION COMPUTATION AND VALIDATION

Modeling of Waves
The accurate generation of waves in the tank is very
essential for the accurate prediction of the ship motion.
Two of the commonly used ways of wave generation in
the numerical tank has been discussed here.
1.
2.

Simulation of an actual wave tank with a wave

maker.
Wave particle velocity definition at the inlet
boundaries.

In the first method one of the outer walls of the domain

corresponding to the wave-maker, is specified with the
known time varying velocity of the wave-maker.
Velocity vector on a plane in the tank, obtained using
this method, has been shown in Figure 15 for shallow
water. The second method, which involves specifying
velocity at the inlet boundaries, is computationally
cheaper. This method has been used in the present
work. Comparison of numerically and experimentally
obtained free surface height at a wave probe in the
domain has been shown in Figure 16. The CFD

Figure 16: Comparison of numerically and experimentally

obtained free surface height at a wave probe in the domain

Wigley hull in head Waves

The inlet velocities are derived from Airy wave theory.
The velocity components (u, w), for regular plane
progressive wave, are given in Eq (3) and (4).
u = U + Ae kz cos(kx t + )
w = Ae sin(kx t + )
kz

(3)
(4)

The wave travels in positive x-direction, z points

positive upwards with z = 0 at the undisturbed free
surface. A, k, and are the wave amplitude, the
wave number, the circular frequency and an arbitrary
phase angle respectively. The elevation of water is
given at the inlet according to Eq (5).

= A cos(kx t + )

(5)

Table 2: Test cases

Case

Fn

/L

H /

0.3

2.75

1/55

0.3

0.99

1/30

0.20
0.15
Swan

0.10

Exp

CFD

Figure 17: Boundary Conditions for Head wave condition

Heave (m)

0.05
0.00
2

10

-0.05
-0.10
-0.15
-0.20
Time (sec)

Figure 19: Heave time history for case 1

6
Swan

Figure 18: Volume fraction of air on xy plane at z=0; Red

represents air and blue water

Pitch (Degree)

Exp

CFD

2
0
3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

-2
-4

Simulations corresponding to test cases given in Table

2, for which experimental data [11] were available,
were carried out. A grid, consisting of 44268
hexahedral cells, has been used for the simulation. A
sufficiently refined grid near the hull was ensured to
obtain grid independent solution. Figure 18 shows a
typical wave profile around the hull.

-6

Time (sec)

Figure 20: Pitch time history for case 1

0.05
CFD

0.04

Exp

Swan

0.03
0.02

Heave (m)

The domain extended to 0.75 L in the front of the hull,

1.0 L behind the hull and 0.6 L each on the both sides
of the hull. The distance from the ship to the bottom
and top is 0.5 L and 0.3 L respectively. The size of the
domain used is very small as compared to the
conventional domain size generally used for these
kinds of problems. The smaller domain reduces the
computational cost considerably. The entire outer
boundary, except the outlet and the top, has been
modeled as velocity inlet (Figure 17).

0.01
0.00
-0.01

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

-0.02
-0.03
-0.04
-0.05

Time (Sec)

Figure 21: Heave time series for case 2

The motions after a few seconds, when the ship hull

oscillates about the equilibrium, have been shown in
the Figure 19 and Figure 20 for case 1, and Figure 21

and Figure 22 for case 2. It can be observed from these

figures that CFD results compare well with
experiments and those obtained through potential flow
solver, SWAN.

The planing hull geometry and the domain used for the
calm water analysis were scaled down to for use in
the current analysis. This model was scaled down to
reduce the simulation time of the transient analysis.
The details of the scaled down hull, and the head wave
it encounters, have been given in the Table 3 and Table
4 respectively. Figure 23 shows the wave elevation on
a waveprobe located at the midship, 0.8L away form
the hull centerline. The acceleration at a point on the
hull in the forward region is plotted in Figure 24.
Figure 25 shows the impact pressure at the same
location. A typical wave profile around the planing hull
has been shown in Figure 26.

0.06
0.04

Wave Height (m)

PLANING HULL IN REGULAR HEAD WAVE

0.1
0.08

0.02
0
-0.02 2

2.5

3.5

4.5

-0.04
-0.06
-0.08
-0.1
Time (Sec)

Figure 23: Wave elevation time history on a Waveprobe for

regular head wave

60
50
Acceleration (m/sec 2)

40
10.0
8.0

CFD

Exp

Swan

6.0

Pitch (Degree)

4.0

20
10
0
-10 2

2.5

3.5

4.5

-20

2.0

-30

0.0
-2.0

30

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

-40
Time (Sec)

-4.0

Figure 24: Acceleration of a point in the forward portion of

the hull for regular head waves

-6.0
-8.0

16000

-10.0

14000

Time (Sec)

Table 3: Particulars of the Hull used for head sea analysis

Pressure (Pa)

12000

Figure 22: Pitch time series for case 2

10000
8000
6000
4000
2000
0

LP (m)

BPA (m)
BPX (m)
BPT (m)
Mass (kg)

1.219
0.243
0.298
0.19
12.5

-2000

2.5

3.5

4.5

-4000

Time (Sec)

Figure 25: Impact pressure at a point in the forward portion

of the hull for regular head waves

Table 4: Wave Details

Fn
/L
H /
Wave Amplitude (A)
Encounter Frequency ( )

0.75
1.5
0.05
0.0457 m
14.696

Figure 29: Impact pressure at a point in the forward portion

of the hull for irregular head waves

Figure 26: Hull in a regular head wave

PLANING HULL IN IRREGULAR HEAD WAVE

CONCLUSIONS

ISSC wave spectrum, corresponding to Hs = 0.15 m

and peak frequency 1.2 Hz, was used to simulate
irregular wave in the numerical tank with the planing
hull model. Figure 27 shows wave elevation in the
tank at the wave probe. The acceleration and impact
pressure at a point on the hull in the forward region are
plotted in Figures 28 and 29 for irregular head waves.

A computational method, to predict motion and

slamming load on a planing hull due to its motion in
regular head waves has been presented. Some of the
conclusions drawn are as follows:
- It is evident from the experimental comparison of the
RANSE predicted resistance, trim and sinkage results
that, the body motion coupled RANSE method can
predict the calm water resistance and dynamics of a
planing craft quite accurately. The tool can be used to
check the performance of planing crafts or similar
vessels in the design stage.
- The method appears to be suitable for impact load
computation. More validation studies especially impact
load validation in waves, for which experimental data
are not widely available, should be carried out.

0.1
0.08
0.06
Wave Height (m)

0.04
0.02
0
-0.02 1.0

2.0

3.0

4.0

5.0

6.0

REFERENCES

-0.04
-0.06
-0.08
-0.1
Time (Sec)

1.

Bertram, V. (2000), Practical Ship Hydrodynamics,

Butterworth-Heinemann

2.

Wagner. H. (1932), Uber stoss- und

Gleitvergange
an
der
Oberflache
von
Flussigkeiten, Z. A. M. M. , Vol 12, No 4, pp
149-163.

3.

Zhao, R., Faltinsen, O.M. & Aarsnes, J.V.(1996),

Water Entry of Arbitrary Two-Dimensional Sections
with and without Flow Seperation, Proc. 21st Symp.
Naval Hydrodynamics, Trondheim, Norway, pp 408-423

4.

Ochi M. K. & Mottor L. E. (1973), Prediction of

Slamming Characteristics and Hull Responses for
Ship Design, Trans. SNAME, Vol 81, pp 144
176
Hermunstad O.A. and Moan T. (2004),
Numerical and Experimental Analysis of Bow
Flare Slamming on a Ro-Ro Vessel in Oblique
Waves, 25th Symp. On Naval Hyd., Canada.
EL Moctar O.(2004), Prediction of Slamming
Loads for Ship Structural Design Using Potential
Flow and RANSE Codes, 25th Symp. On Naval
Hyd., Canada.

Figure 27: Wave elevation time history on a Waveprobe for

irregular head waves
100

Acceleration (m/sec)

80

60

40

20

5.
0
1.0

2.0

3.0

4.0

5.0

6.0

-20
Time (Sec)

Figure 28: Acceleration of a point in the forward portion of

the hull for irregular head waves

6.

18000

7.

16000
14000

Pressure (Pa)

12000
10000
8000
6000
4000
2000
0
1.0

2.0

3.0

4.0
Time (Sec)

5.0

6.0

Singh S. P. and Kumar M., Hybrid method to compute

ship slamming, 10th International Symposium on
Practical Design of Ships and Other Floating
Structures, Houston, Texas, USA, pp 412-418, 2007
8. Hirt, C.W. and Nichols, B.D., "Volume of Fluid (VOF)
Method for the Dynamics of Free Boundaries," Journal
of Computational Physics 39, 201, 1981.
9. Fluent 6.2 User Manual, 2005, Ansys Fluent Inc.
10. Eugene P. Clement, Donald L. Blount, Resistance Tests
of a Systematic Series of Planing Forms, Trans.
SNAME, 1963
11. L.J.M. Adegeest, 1994, Model Tests, Nonlinear Hull
Girder Loads in Ships, pp 15 2