ARTICLE IN PRESS

Ocean Engineering 37 (2010) 947958

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Hydrodynamic characteristics of ship sections in shallow water with complex

bottom geometry
S. Sutulo, J.M. Rodrigues, C. Guedes Soares
Centre for Marine Technology and Engineering (CENTEC), Technical University of Lisbon, Instituto Superior Tecnico, Av. Rovisco Pais, 1049-001 Lisbon, Portugal

a r t i c l e in fo abstract

Article history: The method of boundary integral equation developed by the authors was applied for computing inertial
Received 11 September 2009 and damping characteristics of ship sections for the cases of multi-stepped and inclined bottoms.
Accepted 21 March 2010 Comparative calculations for three typical ship hull sections were performed and analyzed. The
Available online 25 March 2010
frequency-dependent data computed for these ship sections can be used to assess the bottom
Keywords: geometrys inuence onto the ship motions in waves by means of the strip theory. Limiting values of the
Shallow water hydrodynamics same characteristics corresponding to the close-to-zero frequency can also be used for estimation of
Oscillating ship sections hydrodynamic forces in manoeuvring over shallow and conned waterways.
Boundary integral equation & 2010 Elsevier Ltd. All rights reserved.
Stepped bottom
Inclined bottom

1. Introduction
Continuous growth of the average size and speed of vessels
results in repeated problems related to insufcient water
clearances affecting the navigation safety and increasing the
operation costs. Often these problems are related to the squat
sometimes resulting in undesirable contacts with the ground
which requires reduction of the ships speed deteriorating the
waterways capacity. Practical aspects related to the ship
manoeuvrability in restricted waters were discussed by Gray
et al. (2003) and by Landsburg et al. (2005). Another undesirable
effect is also related to touching the ground as result of ship
motions on waves. These motions depend relatively weekly on
the ships speed but mainly on the oncoming waves amplitude.
The remedy here is not to enter the dangerous zone until the sea
calms down but idle time means additional losses. Finally,
proximity of the seabed can alter the response of the ship to
control actions which increases the danger of steering errors
sometimes resulting in accidents and even disasters.
It is clear that it is extremely important to predict the ships
behaviour in shallow and conned waters. First, this can help to
make reasonable decisions about the channels dredging. Then, safe
levels of the sea state and ship speed can be estimated and
prescribed by the port authorities. Creation of adequate mathema-
tical models for ship manoeuvring with due account for the shallow-
water effects is also very desirable as if such models are implemented
in bridge simulators, navigators, and helmsmen can be better trained
and prepared to complicated manoeuvring phenomena.
 Corresponding author.
E-mail address: guedess@mar.ist.utl.pt (C. Guedes Soares).
Although most of the listed problems were dealt with more
than once, see Zhao (1986), Kobayashi (1995), and Beukelman and
Journee (2001) to name just a few, no one particular method can
be still recommended for preferable use. Especially weak is the
treatment of the shallow-water seakeeping problem. Beck and
Tuck (1972), Andersen (1979), Van Oortmerssen (1976), and
Svendsen and Madsen (1981) did some earlier work but the
recent literature has not been very prolic in this subject.
As the models computational efciency and speed are very
important for the envisaged applications, faster though simpler
hydrodynamic formulations must be currently preferred. The
largest gain can be obtained from using the strip method
whenever possible. This method exploits characteristics of the
ship hulls strips or sections for which the hydrodynamic problem
is two-dimensional and much simpler for solution then the full 3D
problem. Huge experience accumulated during decades indicates
on the sufcient accuracy of the strip method when applied to
slender ships. It is less suitable for catamarans or in other
situations when the ow bi-dimensionality is compromised.
Standard hydrodynamic characteristics of the ship sections are
those related to their small-amplitude oscillations within a
certain frequency interval and this predetermines the problem
considered in the present paper. It is also important to bear in
mind that low-frequency data corresponding to the horizontal
motions can also be useful for manoeuvring problems.
As the natural or even somewhat dredged seabed cannot be
always approximated with the horizontal at rigid surface, it is
important that the computational method be able to determine
sectional hydrodynamic characteristics in presence of arbitrary
rigid boundary. The only evident restriction important for keeping
consistency of the strip theory is gentle boundary surface slope
along the ships longitudinal axis.

0029-8018/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.oceaneng.2010.03.008
 ARTICLE IN PRESS
948 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958

Such a method based on earlier publications by Yeung (1973,
1982) was developed by the authors, rst, for the deep water case
(Sutulo and Guedes Soares, 2004) and, second, for the uid of nite
depth (Sutulo et al., 2009). In the latest publication, the method was
applied to the at-bottom case where it was possible to verify it
comparing to some independent results and also the high-contrast
depth case was studied. It was found in the latter case that presence
of a variable depth seabed affects the contour hydrodynamic
characteristics in a non-trivial way and some other bottom
congurations approximating natural depth variations could be of
interest. In the present article, some additional results related to the
case of complicated though schematized bottom geometry: cases of
3, 4, and 8 levels (steps) and of a continuous inclined ramp. The main
part of the paper containing plots of added mass and damping
coefcients is preceded by a concise description of the theory
involved and computational method used which is, however,
reduced to the minimum still permitting independent reading of
this paper. Also, from the same reasons, shown are some selected
results for the at bottom and single-step case.

2. Formulation of problem and solution method

It is supposed that dened is a two-dimensional domain G

lled with a non-viscous incompressible uid and bounded with
Fig. 1. Contour and computational domain. the piecewise smooth boundary consisting of the smooth contour

Section 02 Section 10 Section 22

Fig. 2. Ship sections from container ship S-175 (not to the same scale).

1 1 1
H/T = 10.0 H/T = 10.0
H/T = 10.0 H/T = 15.0 H/T = 15.0
H/T = 15.0
0.8 H/T = 1.1 0.8 H/T = 1.1
H/T = 1.3 0.8 H/T = 1.1
H/T = 1.3
H/T = 1.3 H/T = 1.5 H/T = 1.5
H/T = 1.5 H/T = 1.7 H/T = 1.7
H/T = 1.7 H/T = 2.0 H/T = 2.0
H/T = 2.0
0.6 H/T = 3.0 0.6 H/T = 3.0
H/T = 4.0 0.6 H/T = 3.0
H/T = 4.0
H/T = 4.0 H/T = 5.0 H/T = 5.0
H/T = 5.0
22

22

22

0.4 0.4 Section 10 0.4

Section 02
0.2 0.2 0.2
Section 22
0 0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Dimensionless Frequency Dimensionless Frequency Dimensionless Frequency
1 2 0.6
H/T = 10.0 H/T = 10.0
H/T = 15.0 1.8 H/T = 15.0
H/T = 10.0
H/T = 15.0
H/T = 1.1 H/T = 1.1 0.5 H/T = 1.1
0.8 H/T = 1.3
H/T = 1.5
1.6 H/T = 1.3
H/T = 1.5
H/T = 1.3
H/T = 1.5
H/T = 1.7 H/T = 1.7
H/T = 2.0 1.4 H/T = 2.0 0.4
H/T = 1.7
H/T = 2.0
H/T = 3.0 H/T = 3.0 H/T = 3.0
0.6 H/T = 4.0
H/T = 5.0
1.2 H/T = 4.0
H/T = 5.0
H/T = 4.0
H/T = 5.0
1 0.3
22

22

22

0.4 0.8
Section 10 0.2
0.6
0.2 0.4 0.1
Section 02 Section 22
0.2
0 0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Dimensionless Frequency Dimensionless Frequency Dimensionless Frequency

Fig. 3. Inertial and damping dimensionless coefcients in sway for at seabed.

 ARTICLE IN PRESS
S. Sutulo et al. / Ocean Engineering 37 (2010) 947958 949

Fig. 4. Distribution of panels in the single-step case (section 10): leftgeneral view, rightsections neighborhood.

1.2 0.8

Section 02 Section 02
1 Section 10 0.6 Section 10
Section 22 Section 22

0.8 0.4

0.6 0.2
22

23

0.4 0

0.2 -0.2

0 -0.4
0 0.5 1 1.5 0 0.5 1 1.5
DimensionlessFrequency DimensionlessFrequency
0.2 1.4

1.2 Section 02
Section 10
0 Section 22
1

0.8
-0.2
24

33

0.6
Section 02
0.4
-0.4 Section 10
Section 22
0.2

-0.6 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.05 0.2

Section 02
0.15 Section 10
Section 22
0
34

Section 02
44

Section 10
0.1
Section 22

-0.05
0.05

-0.1 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 5. Dimensionless added-mass coefcients for the single-step bottom.

 ARTICLE IN PRESS
950 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958

SC, the free surface SF, and the rigid bottom SB (Fig. 1). The contour
can perform sinusoidal vibration with the frequency o which is
the wave encounter frequency in seakeeping problems. These
vibrations will generate free-surface waves traveling away from
the oscillating contour to innity if the domain is not restricted in
the horizontal direction from both left and right (SL and SR in
Fig. 1). If the domain is bounded from at least one side in the
horizontal plane, part of the generated waves will reect from the
boundary and standing waves can also appear. However, in order
to apply the boundary integral equation method, the domain
must always be bounded in the horizontal direction at least with
articial boundaries which are also designated with SL and SR but
differing from real rigid boundaries by special boundary
conditions.
The problem is considered in linear formulation which means
that the displacements of the vibrating contour are neglected
while its velocities are taken onto account and the free surface
remains non-deformed and is recognized as such only by means
of the imposed boundary condition. The co-ordinate system Oxy
as shown in Fig. 1 is also xed in space. Due to linearity of the
problem, the time-dependent velocity potential f(x,y,t) can be
represented as f Ref ^ x,y eiot  where f
^ is the complex
amplitude potential independent of time t.
The boundary-value problem is formulated for the complex
potential f^ which satises the Laplace equation and a set of
boundary conditions which are (1) non-penetration Neumann
condition on any rigid boundary: homogenous on the bottom and
other xed boundaries and dened by the velocity distribution on

1.4 0.6

1.2 0.4
Section 02
Section 10
Section 22 0.2
1
23 0
22

0.8
-0.2
0.6 Section 02
-0.4 Section 10
Section 22
0.4 -0.6

0.2 -0.8
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.2
Section 02
Section 10 1.6
0.1 Section 22

Section 02
0 1.2 Section 10
Section 22
33

-0.1
24

0.8
-0.2
0.4
-0.3

-0.4 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.04 0.2

Section 02
0.02 Section 10
0.15 Section 22

0
44
34

0.1
-0.02

Section 02 0.05
-0.04 Section 10
Section 22

-0.06 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 6. Dimensionless damping coefcients for the single-step bottom.

 ARTICLE IN PRESS
S. Sutulo et al. / Ocean Engineering 37 (2010) 947958 951

the oscillating contour; (2) mixed free-surface linear boundary
condition containing the deep-uid wave number k0 o2/g as
parameter; and (3) another mixed boundary condition on each of
ctitious vertical boundaries depending on the wave number k
corresponding to the local depth which can be different for the
left and right boundary.
The formulation outlined above is presented with all details in
Sutulo et al. (2009) as well as the solution method. In brief, the
latter is based on solving the integral boundary equation with
respect to the surface distribution of the complex potential f ^ P,
where P is the point lying on the domains boundary. The equation
is solved through representation of the whole boundary as a set of
small rectilinear segments with constant value of the potential on
each of them. The boundary conditions is then imposed in the
least-square sense on each of these segments which results in the
set of linear algebraic equations whose solution gives the required
distribution of the potential. However, of practical interest is only
distribution on the contours boundary as this denes hydro-
dynamic reactions on the oscillating contour. Namely, it can be
shown with the help of the linearized Bernoulli equation that the
complex amplitude of the hydrodynamic pressure is p^ riof ^,
where r is the uid density. Integration (summation in the
discretized form) of this pressure over the contour with an
appropriate weight will result in the hydrodynamic force in
moment.
However, in strip methods the pressure is rarely used
explicitly but the total hydrodynamic forces and moments are
expressed through the complex added masses
Z where T is the sections draft, B is the sections beam, and the
@fj i
m^ ij o r fi dS mij o nij o
SC @n o coefcients are non-dimensionalized similarly but they are also
where the indices i and j correspond to various generalized
motions and forces, mij is the (real) added mass coefcient per se,
and nij is the damping coefcient. The value i, j 2 corresponds to
sway i.e. to the motion along the axis Ox and force projection on
the same axis; i, j 3 describes the heave (along the z-axis) motion
and force, and i, j 4 relates to roll i.e. rotation around the origin O
and the corresponding roll moment. The remaining values of the
indices i.e. 1, 5, and 6 do not appear in the two-dimensional
problem.
The indices i and j can be different and thus hydrodynamic
properties of the contour are completely described by two 3  3
matrices. However, it can be proven that any of these matrices is
symmetric, so in total 12 scalar values are sufcient. This number
reduces further when symmetry about the y-axis is present. If
both the contour and the bottom conguration are symmetric, all
the coefcients with indices 23, 32, 34, 43 become zero and there
only remain 4 added masses and equal number of damping
coefcients.
The added mass coefcients are presented in this study as
functions of the non-dimensional frequency o0 (o2B/2g) in the
following non-dimensional form:
mij
m0ij at ij 22,23,32
rpT 2
m
m0ij ij 3 at ij 24,42,34,43 2
rpLr
m33 m
m33
0
, m044 444
rpB=22 rpLr
1 generalized reference length Lr max(T,B/2). The damping

Fig. 7. Schemes of the multi-step and continuous-ramp bottom.

 ARTICLE IN PRESS
952 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958

divided by the dimensional frequency o. As only non-dimen-
sional parameters will be shown below, all primes will be
dropped.
3. Sample contours and previous numerical results
Considered were three characteristic sections of the container ship
model S-175 whose particulars can be found in ITTC (1983). Section
02 is a bulbous bow section, Section 10 is the midship section, and
Section 22 is a typical V-shaped afterbody section (Fig. 2).
It was possible to validate the method in the case of the
horizontal at bottom with various water depth H. While all
results of this validation are given in Sutulo et al. (2009), their
subset for the sway case, which is the most important in
manoeuvring problems, is also presented here (Fig. 3).
In this gure symbols show the results obtained with the
present method while the corresponding lines refer to the patch
method developed by Soding (2005), see also Bertram (2000).
The agreement must be recognized as quite satisfactory: the
observed discrepancies have no practical meaning and correspond
to those typically observed when codes of this type are being
compared.
In this particular case, the two methods used have been
developed absolutely independently and are different in every
detail: the patch method is not based on a boundary integral
equation for the potential but on the staggered distribution of
point sources with the body boundary condition formulated for
the induced velocity. The free-surface boundary condition and
the radiation conditions are also applied in the different way.
Finally, the patch method is desingularized while the present one
is not: the desingularization implies staggering hydrodynamic

0.8 0.02

3 Steps 0.01
0.6 4 Steps
8 Steps
Ramp
0
23
22

0.4
-0.01

0.2 3 Steps
-0.02 4 Steps
8 Steps
Ramp

0 -0.03
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0 1

3 Steps
-0.1 0.8 4 Steps
8 Steps
Ramp
33
24

-0.2 0.6

3 Steps
4 Steps
-0.3 8 Steps 0.4
Ramp

-0.4 0.2
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.01 0.2

3 Steps
4 Steps 3 Steps
0.005 8 Steps 0.15 4 Steps
Ramp 8 Steps
Ramp
34

44

0 0.1

-0.005 0.05

-0.01 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 8. Added-mass coefcients for multi-step bottom, section 02.

 ARTICLE IN PRESS
S. Sutulo et al. / Ocean Engineering 37 (2010) 947958 953

singularities beneath each surface/contour present in the pro-
blem. The stagger distance is an additional more or less arbitrary
parameter inuencing the results of computation.
In general, the authors experience in dealing the boundary
integral equation method indicates that while it is rather difcult
to achieve full inherent convergence (the result depends on the
number of panels on the free surface, distribution of their sizes,
and on the position of articial far-away boundaries), the method
is sufciently robust and, once validated for some situation, keeps
bringing enough accurate and dependable results in other
circumstances. Thus, it made sense to use the method for
exploring some cases which could not be treated with the existing
patch-method code or any other available program.
First, the case of the large-contrast single-stepped bottom was
studied which cab be considered as idealization of a deep channel
dredged in a very shallow area (Fig. 4). It is assumed that in the
shallower part H/T1.1 while in the deeper part H/T15.0 which
practically represents deep water.
Plots of added mass coefcients for all three sections are given
in Fig. 5 and similar plots of damping coefcients are presented in
Fig. 6. The main difference in behaviour with the at bottom case
is determined by the presence of a considerable number of local
extrema never observed over a horizontal bottom. Apparently this
is due to the presence of two different uid depths, see Sutulo
et al. (2009) for more details.
4. Numerical results for multiple bottom steps
Further results also correspond to a two-depth bottom but,
rst, with a smaller contrast: H/T 1.1 for the shallower part and
H/T 2 for the deeper part. Also, these two parts are connected in

0.8 0.02

0.01
0.6

0
0.4 23
22

-0.01
3 Steps 3 Steps
0.2 4 Steps 4 Steps
8 Steps -0.02 8 Steps
Ramp Ramp

0 -0.03
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0 3.6
3.2
-0.05
3 Steps 2.8
4 Steps 3 Steps
8 Steps 2.4 4 Steps
-0.1 Ramp 8 Steps
Ramp
2
33

-0.15 1.6
24

1.2
-0.2
0.8
0.4
-0.25
0
-0.3 -0.4
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.01 0.1

0.08
0.005

0.06
44

0
34

0.04
3 Steps 3 Steps
-0.005 4 Steps 4 Steps
8 Steps 0.02 8 Steps
Ramp Ramp

-0.01 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 9. Damping coefcients for multi-step bottom, section 02.

 ARTICLE IN PRESS
954 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958

a more natural way i.e. with a moderately abrupt decline
approximated in 4 different ways: 3 levels (or steps), 4 levels, 8
levels, and as a continuous ramp (Fig. 7). The decision to perform
a comparative study of these approximations is motivated by the
following. Although the developed method can deal with any
shape of the bottom, generation of approximating polygon can be
tedious and not always easily automated. Also, the number of
segments needed for a close t can be too large which will slow
the computations down. So, the question of how rough can be the
seabeds approximation is of certain practical value but no
information of this kind could be found in the literature.
The results are presented in Figs. 813. As is clearly seen from
the plots, inuence of the bottom shape is, in general, not
negligible although its strength differs considerably depending on
the sections shape, motionforce index and on the oscillation
frequency.
First of all, it must be noted that the oscillatory dependency on
the frequency discovered earlier for the high-contrast single step
is also observed here and qualitatively it does not depend on the
local bottom shape, being determined by the depth difference far
left and far right from the contour.
There is some difference in the bottom shapes inuence on the
hydrodynamic characteristics of sections 02 and 22 on one side,
and on those of the section 10 on the other side. In the rst case,
the inuence of the number of steps simulating the depth change
is small and often even negligible. For section 02 this inuence is
nevertheless somewhat more pronounced for the added masses
with indices 22, 24, and 34 at lower frequencies and for the added

1.2 0.4

1
3 Steps 0.2
4 Steps
8 Steps
0.8 Ramp
23 0
22

0.6
3 Steps
-0.2 4 Steps
0.4 8 Steps
Ramp

-0.4
0.2

0 -0.6
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.14 1.2

0.12 3 Steps
1 4 Steps
8 Steps
0.1 3 Steps Ramp
4 Steps
0.08
8 Steps 0.8
Ramp
24

33

0.06 0.6
0.04
0.4
0.02

0 0.2
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.04 0.04

0.03 3 Steps
0.035 4 Steps
8 Steps
Ramp
0.02
34

0.03
44

0.01
3 Steps
4 Steps
8 Steps
Ramp
0.025
0

-0.01 0.02
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 10. Added-mass coefcients for multi-step bottom, section 10.

 ARTICLE IN PRESS
S. Sutulo et al. / Ocean Engineering 37 (2010) 947958 955

1.6 0.2

1.4
3 Steps 0
4 Steps
1.2 8 Steps
Ramp
-0.2
1

22

23
0.8
-0.4
3 Steps
4 Steps
0.6 8 Steps
Ramp
-0.6
0.4

0.2 -0.8
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.1

4
0.08 3 Steps
4 Steps
8 Steps 3 Steps
Ramp 4 Steps
0.06 3 8 Steps
Ramp
24

33
0.04 2

0.02 1

0 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

0.01 0.01

0 0.008 3 Steps
4 Steps
8 Steps
Ramp
-0.01 0.006
34

44

3 Steps
-0.02 4 Steps 0.004
8 Steps
Ramp
0.002
-0.03

0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 11. Damping coefcients for multi-step bottom, section 10.

masses 23, 33, and 34 for high frequencies although in the latter
case some difference in local extrema is also traced for low
frequencies, especially in the 3-step case.
At rst sight, the inuence of the number of steps is
exceptionally small for the heave-heave damping coefcient n33
but this effect is partly due to a different plot scale where large
non-dimensional values were to be shown at very low frequencies
and in fact the differences are rather close to those for m33.
Data for section 22 are qualitatively similar although numerical
absolute values can be very different. The asymmetric
coefcients i.e. those with the indices 23 and 34 are changing
their signs with the frequency and their absolute values are
varying from zero to quite signicant quantities even compared to
some more usual (i.e. observed also in symmetric cases)
coefcients.
The peculiarity of section 10 is that it has a at bottom
faced to the stepped region on the uid boundary, and a
somewhat stronger inuence of the shape of the latter could
be expected. Although in general this is the case, it does not
happen in all modes. For instance, the behaviour of the
coefcients with the indices 33 is very similar to that for another
two sections save for a very small difference between the 3- and
4-step cases visible on the plot for m33. At the same time the
inuence of the bottom shape is somewhat stronger for the added
mass m22 and especially m34 and m44. Of special interest is the
difference between results corresponding to 8 steps and to the
 ARTICLE IN PRESS
956 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958

0.7 0.05

0.04 3 Steps
0.6 4 Steps
3 Steps 0.03 8 Steps
4 Steps Ramp
8 Steps
0.5 Ramp 0.02
22 0.01

23
0.4 0

-0.01
0.3
-0.02

0.2 -0.03
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
-0.08 0.8

-0.1 3 Steps
4 Steps
0.6 8 Steps
-0.12 Ramp

-0.14
0.4
24

-0.16 3 Steps 33
4 Steps
8 Steps
-0.18 Ramp
0.2
-0.2

-0.22 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.01 0.09

0.005 3 Steps
0.08 4 Steps
8 Steps
Ramp
0
34

0.07
44

-0.005
3 Steps
4 Steps
8 Steps 0.06
-0.01 Ramp

-0.015 0.05
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 12. Added-mass coefcients for multi-step bottom, section 22.

continuous ramp. Also, a large difference is observed for m44 in the
3- and 4-step cases.
Plots of the damping coefcients show powerful and, qualita-
tively, a more or less similar inuence of the seabed shape for the
indices 22, 24, and 44, especially at lower frequencies. But this
inuence is expectedly weak in heave damping which is mostly
governed by the contours width at the free surface and conditions
at innity.
5. Conclusions
The algorithm for computing hydrodynamic characteristics
of oscillating contours in shallow water developed earlier by
the authors was applied to investigation of the inuence of
some details of the lower rigid boundary shape approximately
simulating the haven approach channel: natural transition from
shallower to deeper part of the seabed was simulated by a
rectilinear ramp and also by 3, 4, and 8 rectangular steps.
Numerical investigation carried out for 3 characteristic ship
sections showed that depending on the sections shape, oscillation
frequency and mode, the inuence of the local bottom congura-
tion can vary from negligible to very signicant. In general, near-
midship sections with at bottom are more sensitive to the
seabed congurations details than typical bow and stern sections.
Inuence of the number of approximating steps is sometimes
rather systematic but sometimes of a quite unexpected character.
Numerical results presented in this article can be used for
qualitative estimation of possible errors caused by a too
approximate bathymetry and the main practical recommendation
is that the seabed surface should be approximated as accurately
as possible.
 ARTICLE IN PRESS
S. Sutulo et al. / Ocean Engineering 37 (2010) 947958 957

0.4 0.06
3 Steps
4 Steps
0.04 8 Steps
0.3 Ramp

0.02
0.2
22

23
0
3 Steps
4 Steps
0.1 8 Steps
Ramp -0.02

0 -0.04
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0 5

3 Steps 4
-0.02 4 Steps
8 Steps
Ramp 3 Steps
3 4 Steps
8 Steps
24

33
-0.04 Ramp

-0.06
1

-0.08 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency
0.01 0.025

0.005 0.02

0 0.015
34

44

-0.005 0.01

3 Steps 3 Steps
4 Steps 4 Steps
-0.01 8 Steps
0.005 8 Steps
Ramp Ramp

-0.015 0
0 0.5 1 1.5 0 0.5 1 1.5
Dimensionless Frequency Dimensionless Frequency

Fig. 13. Damping coefcients for multi-step bottom, section 22.

Acknowledgments Bertram, V., 2000. Practical Ship Hydrodynamics. Butterworth Heinemann,

The study was carried out within the framework of the
research project PTDC/ECM/65806/2006 Dynamics and Hydro-
dynamics of Ships in Approaching Fairways nanced by Fundac-
a~ o para a Ciencia e a Tecnologia (FCT), Portugal. The rst author
was supported by FCT, Grant SFRH/BPD/26722/2006. The authors
are thankful to Mr. Antonio Pac- o for his considerable aid at editing
numerous plots.
References
Andersen, P., 1979. Ship motions and sea loads in restricted water depth. Ocean
Engineering 6, 557569.
Beck, R.F., Tuck, E.O., 1972. Computation of shallow water ship motions. In:
Proceedings of the Ninth Symposium Naval Hydrodynamics, Frontier
Problems, pp. 15431587.
Oxford.
Beukelman, W., Journee, J.M.J., 2001. Hydrodynamic transverse loads on ships in
deep and shallow water. In: Proceedings of the 22nd International Conference
HADMAR 2001, 14 October 2001, Varna, Bulgaria, pp. 5586.
Gray, W.O., Waters, J.K., Blume, A.L., Landsburg, A.C., 2003. Channel design and
vessel maneuverability: next steps. Marine Technology 40, 93105.
ITTC, 1983. Summary of Results Obtained with Computer Programs to Predict Ship
Motions in Six Degree of Freedom and Related Responses. Report on 15th and
16th ITTC Seakeeping Committee Comparative Study on Ship Motion Program
(19761981), 53 pp.
Kobayashi, E., 1995. The development of practical simulation system to evaluate
ship maneuverability in shallow water. In: Proceedings of the Sixth Interna-
tional Symposium on Practical Design of Ships and Mobile Units, Seoul, Korea,
September 1722, 1995, pp. 1.7121.723.
Landsburg, A.C., Barr, R.A., Dagget, L., Hwang, W.Y., Jakobsen, B., Morris, M., Vest, L.,
2005. Critical needs for ship maneuverability: lessons from the Houston ship
channel full-scale maneuvering trials. Marine Technology 42, 1120.
Soding, H., 2005. Discussion to the Paper (vol. 51, no. 2) by Serge Sutulo and Carlos
Guedes Soares A boundary integral equation method for computing inertial
and damping characteristics of arbitrary contours in deep uid. Ship
Technology Research 52 (1), 52.
 ARTICLE IN PRESS
958 S. Sutulo et al. / Ocean Engineering 37 (2010) 947958
Sutulo, S., Guedes Soares, C., 2004. A boundary integral equation method for
computing inertial and damping characteristics of arbitrary contours in deep
uid. Ship Technology Research 51, 6993.
Sutulo, S., Rodrigues, J.M., Guedes Soares, C., 2009. Computation of inertial and
damping characteristics of ship sections in shallow water. Ocean Engineering
36, 10981111.
Svendsen, I.A., Madsen, P.A., 1981. The dynamics of wave induced ship motions in
shallow water. Ocean Engineering 8, 443479.
Van Oortmerssen, G., 1976. The motions of a ship in shallow water. Ocean
Engineering 3, 221255.
Yeung, R.W.-C., 1973. A singularity-distribution method for free-surface ow problems
with an oscillating body. Ph.D. Dissertation, University of California, Berkeley.
Yeung, R.W.-C., 1982. Numerical methods in free-surface ows. Annual Review of
Fluid Mechanics 14, 395442.
Zhao, Y.-X., 1986. Computations of manoeuvring motions in shallow water (Jahrbuch
der Schiffsbautechnischen Gesellschaft), Bd. 80, pp. 261275 (in German).