Ocean Engineering 38 (2011) 20002014

Contents lists available at SciVerse ScienceDirect

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

Experimental and numerical investigation of the slowly varying wave

exciting drift forces on a restrained body in bi-chromatic waves
Nuno Fonseca a,n, Joa~ o Pessoa a, Spyros Mavrakos b, Marc Le Boulluec c
a

Centre for Marine Tech. and Eng., Technical Univ. of Lisbon, Portugal
Dep. Naval Arch. and Mar. Eng., National Tech. Univ. of Athens, Greece
c
French Res. Institute for Exploitation of the Sea (Ifremer), France
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 6 August 2010
Accepted 10 September 2011
Editor-in-Chief: A.I. Incecik
Available online 28 October 2011

The paper presents an experimental and numerical investigation of the rst order and slowly varying
wave exciting forces on a body of simple geometry which is restrained from moving. Both monochromatic waves and three sets of bi-chromatic waves corresponding to three difference frequencies were
tested. The depth effects on the second order forces are assessed by repeating the wave conditions for
deep water, intermediate water depth and shallow water. The wave exciting mean drift forces and
second order slowly varying forces were successfully measured and identied. However, since the
magnitude of these forces is small compared to the global forces measured, there is some dispersion in
the second order results.
The experimental data is compared with three state of the art numerical methods which are able to
compute the rst and second order wave exciting forces. Two of the methods use Greens function
panel methods (Wamit and HydroStar) and the third is based on an analytical solution (Diffrac-R). All
methods compute the full second order solution, including the effects related to the quadratic
interactions of rst order quantities and to the second order diffraction potentials.
& 2011 Elsevier Ltd. All rights reserved.

Keywords:
Wave exciting forces
Slowly varying forces
Second order forces
Bi-chromatic waves
Difference frequency forces
Water depth effects
Experimental work
Potential ow methods

1. Introduction
A oating or xed marine structure subjected to multi-frequency incident wave eld will experience second order wave
exciting forces. These are related to the interactions between
pairs of harmonic wave components composing the wave eld.
Within a frequency domain approach, these forces can be decomposed into three components namely: a steady force, a difference
frequency component and a sum frequency component. The rst
represents a mean value which is frequency dependent, but time
independent. The other components change harmonically in time,
one with the sum frequency of the pair of harmonic waves and
the other with the difference frequency.
Sum and difference frequency effects are important for different problems. Sum frequency second order forces occur at
relatively high frequencies and they may excite the natural
periods of oating structures moored with taut cables, like the
tension leg platforms. In this case the natural period of the
vertical motions is small, the associated wave damping is low,
therefore the sum frequency forces may induce resonant vertical
motions which results on undesirable dynamic effects (springing

Corresponding author.
E-mail address: nunofonseca@ist.utl.pt (N. Fonseca).

0029-8018/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.oceaneng.2011.09.017

loads). These effects have been identied at the full scale (e.g.
Standing et al., 1993) and conrmed at the model scale (see for
example Herfjord and Nielson, 1986).
In the case of the difference frequency second order forces,
they result on the slowly varying wave drift excitation in irregular
seas, which are important, for example, for oating moored
structures. For many oating structures the mooring system is
compliant with the rst order wave exciting forces, since the
natural period of the moored oater is large compared to the
wave period. However the slowly varying drift forces have longer
periods and therefore they may excite the oater and mooring
system at their natural frequency resulting in large horizontal
motions of the oater and tensions on the mooring lines (see for
example Pinkster, 1976, 1980; Emmerhoff and Sclavounos, 1996).
Semi-submersibles may also be excited in the pitch motion by
slow drift forces because they have a long natural period of pitch.
Stansberg (2007) investigated experimentally the slow drift pitch
motions of a semi-submersible.
Many offshore oating systems use dynamic positioning (DP)
to limit the horizontal excursions of the moored structures. In this
case the dynamic position assistance is used to attenuate the slow
drift oscillations, while the static mooring system opposes the
steady forces. The most efcient DP systems incorporate numerical models to estimate the slow drift motions of the oating
structure.

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

Over the recent years much research work has been focused on
the hydrodynamics of oating structures in shallow water. This is
due mainly to the development of offshore Liqueed Natural Gas
(LNG) terminals. In shallow water the wave exciting drift forces
increase in magnitude (Fonseca et al., 2008) and, on the other
hand, these structures operate often in close proximity with other
oating bodies like buoys, supply vessels, or ofoading shuttle
tankers. For these reasons an accurate prediction of wave exciting
mean and slowly varying drift forces in shallow water is important. van der Valk and Watson (2005) presented a comprehensive
set of model test results for different arrangements of an LNG
carrier and oating production barge in moderate multi directional wave climates. The authors concluded that a good prediction and control of the motions during ofoading operations from
a oating platform to a LNG carrier is important.
The previous paragraphs demonstrate that wave exciting drift
forces induce signicant dynamic responses on oating structures,
therefore it is important understand them and also to develop and
validate numerical methods for their accurate prediction. The focus
of the present study is on the slowly varying wave exciting drift
forces. The related transfer function, also called quadratic transfer
function (QTF), represents the amplitude of the second order
harmonic force for all combinations of pairs of incident harmonic
waves. Usually it is represented by a N  N matrix where N is the
number of wave frequencies considered. The main diagonal corresponds to zero difference frequency, meaning mean drift forces in
monochromatic waves. The usual procedure for calculation of slow
drift forces is to simplify the QTF by representing the difference
frequency components in terms of the zero difference results. In this
way the second order problem is much simplied as well as the
computational effort. However this approximation has some limitations. For the slow drift oscillations it may be important to consider
correctly the difference frequency components, for example when
the difference frequency is close to a motion resonant frequency and
the amplied motions contribute signicantly to the quadratic drift
effects.
1.1. Review of numerical methods
The calculation of mean drift forces is usually carried out using
one of two methods, the so-called far-eld solution, or the neareld solution. Maruo (1960) applied the far-eld method to
represent the horizontal drift forces on two dimensional and
three dimensional bodies considering incident harmonic waves
and deep water. The solution is based on the assessment of the
time rate of change of the momentum in the uid within a control
volume bounded by the free surface, the body, a vertical cylindrical surface with radius tending asymptotically to innity and
the ocean bottom. The force is obtained by averaging the former
over one wave cycle. Newman (1967) applied the same method to
derive the yaw drift moment in deep water, whereas Faltinsen
and Michelsen (1974) modied Newmans (1967) formulation in
conjunction with a three-dimensional sinksource technique
and showed that his expressions were equally valid for innite
and nite water depth. The application of the far-eld method for
evaluating the mean second-order heave force and the roll and
pitch moments leads to expressions that involve innite integrals
on the free surface supplemented by corresponding ones on the
sea bottom, in case of nite depth waters. Lee and Newman
(1971) circumvented this difculty in the case of a submerged
body in deep water by expressing the free-surface integrals in
terms of Kochin functions evaluated on the mean position of the
bodys wetted surface. Mavrakos (1988) implemented the fareld formulation to present expressions for the mean vertical
drift force and the pitching moment on free-surface piercing
arbitrarily shaped vertical bodies of revolution oating in water

2001

of nite depth. The analytic representation of the velocity potential that can be established for this particular type of bodies in
terms of Fourier series (see Kokkinowrachos et al., 1986), enables
the evaluation of the innite free-surface and sea-bottom integrals in closed form. For arbitrarily shaped bodies however, the
momentum method cannot give the vertical drift forces, neither
second order pressures on the body.
Pinkster and Van Oortmersen (1977) introduced the near eld
method to evaluate the mean drift loads on arbitrarily shaped large
volume oating bodies in all six degrees of freedom. The method
requires that all pressure contributions giving rise to second order
terms with respect to the wave amplitude have to be correctly
integrated over the instantaneous wetted surface of the body. The
near eld formulation is not restricted to the evaluation of mean
forces only. It can be also applied for evaluating the second order
sum- and difference-frequency wave load components. However,
numerical schemes implementing panel formulations is much more
sensitive to the meshing of the bodys wetted surface, particularly
near the free surface and close to sharp corners.
While the mean drift forces are calculated from quadratic
products of rst order quantities, the oscillatory component, or
slow drift force, requires the calculation of second order potentials. This means that the hydrodynamic problem must be solved
correctly up to the second order in the wave elevation. In that
context, Pinkster (1976, 1980) implemented the near eld solution to calculate the time dependent slowly-varying differencefrequency drift loads. In doing this, he accounted for all quadratic
terms originated from the rst-order potentials together with the
second-order contributions arising from the undisturbed secondorder incident wave eld. Contributions due to second-order
diffraction potential were omitted in his formulation. For their
evaluation, two methods have been proposed in the literature
classied as indirect and direct.
The rst circumvent the direct calculation of the second-order
diffracted wave eld by introducing an assisting radiation potential (see for example Lighthill, 1979; Molin, 1979; Eatock Taylor
and Hung,1987; Abul-Azm and Williams, 1988; Mavrakos and
Peponis, 1992). These methods have been proven efcient in
calculating the hydrodynamic loading. However, they cannot
provide results for the second-order pressure wave eld and the
free surface elevation around the structure. The second class of
solution methodologies, the so-called direct methods, allows the
derivation of the second-order diffraction potential. Examples of
works in that context are those due to Loken (1986), Kim and Yue
(1989, 1990), Scolan and Molin (1989), Chau and Eatock Taylor
(1992), Zaraphonitis and Papanikolaou (1993), Huang and Eatock
Taylor (1996), Eatock Taylor and Huang (1997), Malenica et al.
(1999), Newman (2005), Mavrakos and Chatjigeorgiou (2006) and
Chatjigeorgiou and Mavrakos, 2007.
More recently the so-called middle-eld formulation was
proposed to overcome the convergence problems of the near eld
method (Chen, 2005; Rezend et al., 2007). In this case the control
surface is located at a nite distance from the body.
1.2. Review of experimental work
Experimental data from model tests is usually very useful to get
insight into the psychics of the hydrodynamic problems. Additionally, experimental data is often needed to validate the numerical
methods. In the case of the second order drift forces on marine
structures, there are not many experimental results available in the
literature. Kobayashi et al. (1985) present the horizontal mean wave
drift force on a tension leg platform (TLP) in regular waves. Standing
(1991) carried out model tests with a moored tanker in bi-chromatic
waves and compares the measured slowly varying wave exciting surge force with numerical predictions. More recently some

2002

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

research effort has focused on the hydrodynamic multi-body interactions motivated by the LNG ofoading from a oating platform to
a shuttle tanker. Kashiwagi et al. (2005) carried out an experimental
program with two side-by-side ship models where the mean drift
forces in regular waves were measured. Their higher order boundary
element method results compare well with the experiments.
Ikoma et al. (2000) investigated experimentally the slowly
varying wave drift forces on a very large oating structure in
irregular short crested waves. Standing et al. (1993) presented full
scale measurements on a Semi-Submersible Support Vessel operating in the North Sea including the low frequency vertical and
horizontal motions. It was concluded that the second order wave
exciting forces contribute strongly to the slow drift motions. The
numerical predictions compare well with the heave motions, but
they often underestimated the horizontal motions and also the
roll and pitch motions.
Regarding high frequency second order loading, Petrauskas
and Liu (1987) investigated the springing loads on a TLP platform
and their results include experimental data of the sum frequency
forces on a vertical cylinder in regular waves. The numerical
results follow approximately the tendency of the experiments,
although the later show some dispersion. Tests were also carried
out with a model of a TLP and the authors identied large
springing forces on the tendons generated by the resonant pitch
period. DeBoom et al. (1984) measured the high frequency tether
forces on a TLP in regular waves. The springing forces occurred at
twice the regular wave frequency.

1.3. Objectives of the present work

From the review presented in the former paragraphs it is clear
that the available experimental data regarding wave exciting drift
forces, and in particular regarding the slowly varying component,
is scarce. For this reason an experimental program was carried
out to obtain the mean and slowly varying wave exciting drift
forces on a body of simple geometry. The objectives consisted on
obtaining insight into the physics of these second order forces and
also obtain experimental data appropriate for the validation of
theoretical and numerical methods. The tests were performed on
a restrained body in monochromatic waves and bi-chromatic
waves. An important part of the study is to access the depth
effects on the drift forces, so the tests were carried out for
3 different water depths corresponding to shallow water, intermediate water depth and deep water.
The experimental results are compared with the predictions
from three different numerical methods. All methods are based on
the complete solution of the second order wave exciting problem
in bi-chromatic waves, however the rst applies the near eld
approach (Lee and Newman, 2004), the second applies the new
middle eld approach (Chen, 2005; Chen and Rezende, 2009) and
the third is based on an analytical solution (Mavrakos, 1988;
Mavrakos and Peponis, 1992).

2. Theory
This section introduces the general formulation of the rst
order and the second order hydrodynamic problem for the wavebody interaction. Details of the formulation can be found in
several publications, as referred in this text.
The uid is assumed non-viscous and the boundary value
problem is formulated in terms of potential ow within a uid
domain with closed boundaries including the wetted body surface,
the free surface, vertical walls far away from the body extending
from the free surface to the sea bottom and the sea bottom itself.

The uid ow, represented by the velocity potential, must satisfy the
Laplace equation and specic conditions on the boundaries.
The space and time dependent nonlinear velocity potential is
expanded into a perturbation series with respect to the wave
slope parameter e (e 51)

F eF1 e2 F2   

The kinematic body and the free surface boundary conditions

are also expanded in terms of Taylor series with respect to their
mean positions. This way the boundary conditions are represented in terms of their known mean positions.
Finally, the expansions series are truncated above the second
order by assuming sufciently small wave amplitudes and body
motions. The rst and second order terms are grouped separately
and two problems are formulated, namely the rst order and the
second order problems. The rst order problem is related to the
harmonic incident wave with frequency oj and the unknown is
the rst order velocity potential complex amplitude f(1).
The second order problem is due to interactions between pairs
of incident harmonic waves with arbitrary different frequencies,
o1, o2 and the oscillating (or xed) body. In the presence of two
plane incident waves of frequencies o1, o2, the rst order
velocity potential can be written in the form:
8
9
2
<X
=
1
F1 x,t Re
fj xeiot
2
:
;
j1

The related second order quantities change harmonically with

frequencies given by the sum and also the difference between the
frequencies of the two incident waves. The corresponding second
order potential is expressed by

F2 x,t Re

2 X
2
X

ff xeio t f xeio


j1l1

where o  oj  ol and o oj ol. The sum- and differencefrequency potentials in (3) f and f  , respectively, represent the
unknowns of the corresponding boundary value problems.
The solution to second order is obtained in three stages. First
the linear boundary value problem is solved to obtain the rst
order velocity potentials and motions. Then the rst order results
are used as input to solve the second order problem. Finally,
responses like the second order drift forces (and moments) and
motions can be computed. A brief presentation of these three
stages is given in the following paragraphs.
2.1. Linear hydrodynamic problem
Incident monochromatic waves are considered and within a
linear approach all time dependent quantities change harmonically in time with the frequency of the incident waves (zero speed
case). A xed Cartesian co-ordinate system is dened with the
vertical z-axis pointing upwards and zero incidence of the waves
when they propagate along the direction of the positive x-axis.
The problem is formulated in terms of a velocity potential
which is linearly decomposed into three independent components namely: incident, scattered and radiation potentials:
1
1
f1 f1
I fS fR

The radiation potential can be further decomposed into com1

ponents from all six degrees of rigid body rst order motions xj

j1
R io

6
X

j1
x1
j
j

j1

The potentials must satisfy the Laplace equation in the uid

domain, the rst order free surface boundary condition and the

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

rst order body boundary conditions

2

r j

1
I,S,R

@2 j1
I,S,R
@t

@j1
I,S,R
@z

0,

z0

The second order potentials must satisfy also the Laplace

equation in the uid domain, the bottom condition and appropriate radiation conditions. Additionally, the second order incident potential must comply with the following free surface
condition (Kim and Yue, 1990; Lee, 1995):
o 7 j2
I g
2

1
@fS

@n
@j

1
j

@n

1
@fI

@n

on S0

nj , j 1, . . ., 6 on S0

where @ =@n denotes the derivative in the direction of the inward

unit normal vector n to the mean wetted surface S0 of the body
and nj are its generalized components dened as (n1, n2, n3) n
and (n4, n5, n6)r  n where r is the position vector with respect
to the origin of the co-ordinate system of body. Appropriate
radiation and sea bottom conditions also need to be specied.
The former equations show that the radiation and diffracted
potentials can be obtained solving two independent boundary
value problems.
2.2. Second order problem
For the second order problem one considers two incident
harmonic waves with arbitrary frequencies oj and ol and the
related velocity potential is bi-harmonic as given by (3). Besides
Laplaces equation, bottom and radiation conditions, the second
order potential must comply with the free surface and body
boundary conditions. Following Lee (1995) the conditions are:
2

@2 f
@t 2

@f
@z

QF,

on z 0

10

on S0

11

where QF and QB are dened, respectively, as free surface forcing

and body boundary forcing. Since the former functions were
obtained from Taylor expansions about the mean positions of
the boundaries, Eqs. (10) and (11) are applied on these known
positions.
QF represents the second order terms of the expansion (see
Newman, 1974), it is expressed in terms of the rst order
potentials of Eq. (2), and it is possible to decompose the forcing
function into quadratic components related to interactions
between: two incident harmonic potentials, incident and body
disturbance potentials and two body disturbance potentials.
The second order forcing function on the body includes
quadratic terms that are function of the rst order solution and
terms that depend linearly of the second order motions (Ogilvie,
1983).
The second order potential can also be decomposed into
incident, radiation and scattering components:

@j2
I
Q II7 ,
@z

z0

14

where Q II7 is related to the quadratic interactions between two

incident harmonic waves.
The second order radiation potential satises the same free
surface and body conditions as the rst order problem, respectively, Eqs. (7) and (9).
The second order scattering potential satises the following
conditions:
o 7 j2
S g

@j2
7
7
S
Q IB
Q BB
,
@z

@j2
S
Q B7 ,
@n

on S0

z0

15

16

7
7
where Q IB
and Q BB
are the free surface forcing related to bichromatic quadratic interactions between incident wave and
body disturbance and two body disturbance wave systems,
respectively. Q B7 is the forcing function on the body boundary
and it includes the derivative of the second order incident
potential and quadratic terms related to the rst order motions,
rst order uid velocities and also second order velocities of
the body.

@f
Q B,
@n

2003

6
X

Consider two co-ordinate systems: the rst, X0 (x0 ,y0 ,z0 ), is

xed to the body, and the second, X(x,y,z), is xed in space and
coincides with the rst when the body is at rest. The vector
!
representing the displacement a of a point of the instantaneous
hull wetted surface Sw in the space xed system is given by
!

!
s O  r
a !

17

!
where s is the translation vector of the origin of the body xed
!
!
co-ordinate system X0 (x0 ,y0 ,z0 ), O is the rotation vector and r is
the position vector of a point on Sw with respect to the body xed
reference system.
The hydrodynamic forces and moments on the body are
obtained by direct integration of pressure over the hull wetted
surface:
Z
!
!!
F
18
P a n dS
Sw

!
M

! ! !
P a a  n dS

19

Sw

2
2
2
fI fR fS

7
j2
R io

2.3. Hydrodynamic forces by the near eld method

12

1
x2
j jj

13

j1
2

The rst term in (12), fI , is the potential that satises the

second order free surface condition of the incident waves, the
2
second fR represents the potential linearly proportional to the
2
second order motions, see Eq. (13), and the third fS includes all
remaining components basically related to the quadratic interactions between pairs of incident potentials and body disturbance
potentials.

!
where n is the unit normal vector to the instantaneous wetted
surface of the body. Bernoullis equation gives the pressure P


@F 1
rFUrF gz
20
P r
2
@t
!
!
Since the exact wetted hull is not known a priori, P, a and n
are expanded in a Taylor series about the mean body position S0.
Terms up to the second order are retained in the resulting force
and moment equations. First order terms are the linear hydrodynamic and hydrostatic forces. The remaining terms are the
second order forces and moments.

2004

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

The linear forces in harmonic waves are proportional to the

wave amplitude and given by
0
1
Z
6
X
!1
!1
1
1
1 1 A!
@
n dS F H
ior
jI jS
jj xj
21
F
S0

j1

Resulting, respectively, in1the linear Froude Krilov, scattering

!
and radiation forces, and F H is the rst order hydrostatic force
linearly proportional to the rst order motion amplitudes. For the
sake of simplicity, only the equations for the forces are presented
here, although the reasoning can be generalized for the moments.
The second order wave exciting forces are proportional to the
product of the two wave amplitudes and given by
Z
Z
!2
1
@F1 !1 !0
!0
F
 r 9rF1 j2 n dSr
O  n dS
2 S0
S0 @t
Z
Z
0
@F1 !1 !0
!0
r
a n dSrg R2 !
x Urz n dS
S0 @t
S0
Z
Z
q
1
!0
1
rg zR2 n dS rg xr 2 1n2z dl
2
S0
wl
Z
@F2 !0
n dS
r
22
S0 @t
R(2) is the rotation matrix used to represent the second order
effects related to the rst order angular motions.
The rst term in Eq. (22) represents the uid velocity squared
in the Bernoullis equation, the second term represents the
angular rotation of the rst order hydrodynamic forces, the third
components includes the second order effect of the rst order
displacements on the hull pressures, the fourth term accounts for
the second order rotational motions effects on the hydrostatic
forces, the fth component represents a second order correction
due to the second order rotational motions on the vertical
hydrostatic force, the sixth term is an integral over the mean
waterline and represents force contribution from the rst order
pressure acting on the additional hull wetted surface due to the
relative free surface elevation, and nally the last component
accounts for the second order velocity potential.

3. Numerical solution
3.1. Wamit
Within Wamit, which is a commercial code, the boundary
value problem is solved by the integral equation method, meaning that the Greens theorem is used to derive integral equations
for the velocity potential on the mean body wetted surface. In
order to calculate the second order exciting forces, it necessary to
evaluate the uid velocities on the body surface (see Eq. (20)) and
on the free surface (for the free surface forcing function). The uid
velocity is calculated by the source formulation, which also
results on surface integral equations including Greens functions
in the unknown source strengths.
The Green function satises the radiation condition and all
boundary conditions with the exception of the kinematic boundary
condition on the body, therefore the problem is to be solved only
across the mean wet surface of the body. The body surface is
represented by a set of quadrilateral panels, the potentials (and
source strengths) are assumed constant over each panel and the
integral equations are enforced at the centroid of the panels. This
way the integral equations are reduced to a set of linear equations.
The uid in the closed domain dened by the body surface and
the interior free surface will have a set of eigen frequencies which
lead to incorrect numerical solution of the boundary value
problem. These are called the irregular frequencies. Wamit uses

the extended boundary integral equation method to remove the

irregular frequencies. Basically the interior free surface is also
meshed and an articial velocity potential is enforced at this
surface (see Lee et al., 1996).
Regarding the second order boundary value problem, we have
a set of boundary conditions for the radiation and scattering
potentials similar to the rst order problem, however one needs
to consider the free surface forcing in the incident and scattering
free surface conditions (Eqs. (12) and (13)) and the body forcing in
the scattering body boundary condition. The radiation boundary
value problem is solved by the same methods as the rst order
problem, since the only difference is the frequency of the forced
motion which given by the sum or difference frequency.
The second order scattering boundary value problem is also
solved with the integral equation method and using the same
Green functions as for the solution of the rst order problem.
However, since the Green function does not satises the forcing
functions on the free surface boundary conditions, one needs to
distribute unknown source potentials also on the free surface
around the body. The free surface integral is calculated separately
in two domains divided by a partition circle, and inner domain
and an outer domain. The integration in the inner domain is done
numerically and this domain is further divided in two parts
separated by an inner circle. Inside the circle the free surface is
discretized with quadrilateral panels as the body surface. In the
region between the inner circle and the partition circle, the
integration is carried out using Gaussian quadrature rule methods
(see Lee and Zhu, 1993). In the outer domain, between the
partition circle and innite, the incident and the body potentials
are approximated by FourierBessel expansions and the integrations are carried out analytically.
3.2. HydroStar
The HydroStar, which is also a commercial software, implements a solution to solve the rst order 3D boundary value
problem which is in the basic aspects similar to the one described
in the former paragraphs: it is a Greens function 3D diffraction/
radiation method where the mean hull wetted surface is represented by quadrilateral at panels. The irregular frequencies are
removed by extending the domain of integration to the inside free
surface. Regarding the second order wave exciting drift forces,
three formulations are implemented, namely: the far eld
approach (Maruo, 1960; Newman, 1967), the near eld solution
(Pinkster et al., 1979) and the middle eld approach (Chen, 2004,
2005; Chen and Rezende, 2009). The complete quadratic transfer
functions and the vertical force components are given only by the
near eld and middle eld solutions, therefore these two
approaches are used for the present investigation.
The near eld approach is based on the second order direct
pressure integration on the body surface and the related forces
can be decomposed into several components representing quadratic interactions of rst order quantities and a component
related to the second order potentials (see Eq. (22)). The second
order diffraction potential is taken into account through the
Haskind theorem method (Chen and Rezende, 2009).
The convergence of the results from the near eld method is
difcult in some cases, especially for geometries with sharp corners,
therefore the middle-eld approach was developed to improve the
numerical accuracy of the predictions. The middle-eld formulation
was developed from the most general near-eld formulation by
making use of two variants of Stokes theorems. The second order
loads are computed over a control surface surrounding the body at a
nite distance. In this way, the solution provides the advantages of
both classical approaches: good numerical precision as the far-eld
formulation and the possibility of computing all the components of

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

the drift loads and the complete quadratic transfer function as the
near-eld formulation.
3.3. Diffrac-R
The method developed by Mavrakos and Peponis (1992) to
predict the sum- and difference-frequency second-order forces on
vertical axisymmetric bodies restrained in regular waves does not
solve directly for the second-order potential. It exploits the
Greens second identity together with the use of an assisting
rst-order radiation potential (Molin, 1979) to recast the expression of the second-order force due to the second-order potential,
i.e. the last term in Eq. (22), in a summation of terms involving
rst-order quantities and the second-order undisturbed incident
wave potential alone. The assisting rst-order radiation potential
7
Ck7 x,t Refck7 xeio t g is associated with a unit forced oscillation of the body in otherwise still water in the k-th mode of
motion at frequency o 7 oj 7 ol. The complex potential func7
tion ck fulll the Laplace equation and satises the usual freesurface, sea-bottom, radiation and kinematic body boundary
conditions, see Eqs. (6)(9), at the sum- or difference-frequency.
The required rst-order assisting potentials are calculated using
the method of matched axisymmetric eigenfunction expansions
that allows analytic representations of the velocity potential
around the body (Kokkinowrachos et al., 1986).
Following the methodology outlined above, it can be shown
that the integral of the second-order potential over the mean
wetted surface of the body, i.e. the last term in (22), can be
reduced to



Z
Z
@F2 !0
7
2 !0
r
n dS Re
io 7 r f n dS eio t
23
S0 @t
S0
where
Z

!0

f2 n

S0

dS

Z
S0

ck7

@fI
1
dS
g
@n

SF

7
7
Q IB
Q BB
ck dS

24

The rst term in (24) includes products of the normal derivative of the bi-chromatic second-order undisturbed incident wave
potential with the assisting rst-order radiation potential. The
integrand in the last term includes products of the free surface
7
7
second-order forcing functions Q IB
and Q BB
, see Eq. (16), with
the assisting rst-order radiation potential. The integration has to

Table 1
Main particulars of the restrained body.
Cylinder diameter, R (cm)
Bilge radius, r (cm)
Draught, T (cm)
Depth, D (cm)
Underwater volume, V (m3)

65
10
20
40
0.062

2005

be carried out over the entire free surface. By carrying out rst the
integration over the azimuthal angle y, an one dimensional
integration over the radial co-ordinate from the external radius
of the body to a large radius has to be performed. The integrand
converges far away from the body rather slowly. To enforce faster
convergence, we lter out the oscillatory behavior by averaging
the integrand over a period. Convergence is for all practical
purposes achieved after 10 to 12 periods.
As far as the calculation of the mean drift loads is concerned, the
Diffrac-R software uses the far-led method for the prediction of
both their horizontal and vertical components. The latter includes
evaluation of integrals over the entire free-surface, which, however,
in the case of vertical axisymmetric bodies can be treated analytically in closed form (Mavrakos, 1988).

4. Experimental program
4.1. Experimental setup
The experimental tests were carried out with a restrained
vertical cylinder with a shallow draft. Table 1 presents the main
characteristics of the body. Fig. 1 shows the side view lines of the
model with the dimensions and a photo with the tested cylinder.
The cylinder has a rounded bottom, with a bilge radius between
the at bottom and the vertical sides of 10 cm. The objective of
the rounded bottom was to minimize the viscous effects related
to vortex shedding from the transition between the bottom and
sides associated to the oscillatory ow.
The tests were carried out at Danish Hydraulics Institute,
namely at the shallow water basin, for the water depths of
40 cm and 55 cm, and at the offshore basin for the water depth
of 300 cm. The shallow water basin (25 m  35 m  0.20.8 m)
includes a segmented piston 3D wave maker equipped with an
active wave absorption system to eliminate the reected waves. A
beach is located 20 m away from the wave maker. The offshore
basin (20 m  30 m  3 m) uses 60 individually controlled aps to
generate 2D or 3D waves and includes a parabolic beach on the
opposite side. Current and wind can be simulated in both basins.
The model was kept in place by a triangular shaped rigid rig xed
to the bottom of the tank, as presented in Fig. 2.
Regarding the instrumentation, a force transducer, connecting
the model to a triangular rig, measures horizontal and vertical
wave exciting forces. Eight wave gauges measure the wave
elevation, where seven are in front of the model and in line with
the wave propagation and one is placed sideway with respect to
the model (see Fig. 3).
The tests with the restrained model were carried out in
monochromatic waves, bi-chromatic waves and irregular seastates. Only the rst two are presented and discussed here. Three
sets of bi-chromatic waves are considered, namely corresponding
to the difference of frequencies between the two harmonic waves

Fig. 1. Geometry of the vertical cylinder (left side) and a photo of the tested body.

2006

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

Fig. 2. Triangular rig to x the model during the tests.

Table 2
Characteristics of the monochromatic and bi-chromatic waves tested.
Monochromatic waves
Wave index

T (s)

A (cm)

1
2
3
4
5
6
7
8
9
10

0.70
0.90
1.10
1.15
1.35
1.75
1.90
2.00
2.20
1.55

1.1
1.6
2.2
2.3
2.9
4.0
4.4
4.7
4.8
3.4

Bi-chromatic wavesdw 0.5 (rad/s)

Fig. 3. Plan view of the wave gauges.

of 0.5, 1.5 and 4.0 rad/s. Table 2 presents the characteristics of the
waves tested, where, T, T1 and T2 are the wave periods and A, A1
and A2 are the wave amplitudes. The former wave conditions
were tested for three water depths, namely 40 cm, 55 cm and
300 cm. These correspond to shallow water conditions, intermediate water depth and deep water.

Wave index

T1 (s)

T2 (s)

A1 (cm)

A2 (cm)

1
2
3
4
5
6
7
8
9

0.72
0.93
1.15
1.55
1.65
1.77
2.02
2.23
2.40

0.68
0.87
1.06
1.38
1.46
1.55
1.74
1.90
2.01

1.1
1.7
2.3
3.4
3.7
4.0
4.8
4.8
4.8

1.0
1.5
2.0
3.0
3.2
3.4
4.0
4.4
4.7

Bi-chromatic wavesdw 1.5 (rad/s)

4.2. Analysis of the experimental data

The wave elevations and force transducer data was acquired at
a sample rate of 80 Hz. Afterwards, the time signals were
analyzed to identify the harmonic content, namely, the mean
values and all rst and second order harmonics. Two procedures
were used; (a) the standard Fourier analysis and (b) tting a
second order expansion model to the measured signal with a least
squared procedure. In bi-chromatic waves the wave exciting
forces are approximated by (components above the 2nd order
are neglected)
s1
c1
s1
Ft F F c1
1 coso1 t F 1 sino1 t F 2 coso2 t F 2 sino2 t
s2
c2
F c2
1 cos2o1 t F 1 sin2o1 t F 2 cos2o2 t
c
F s2
cosfo1 o2 tg F s sinfo1 o2 tg
2 sin2o2 t F

F c cosfo2 o1 tg F s sinfo2 o1 tg

25

where o1 and o2 are the two incident wave frequencies (rst

harmonics) and the exciting force is represented by the sum of a
mean value, F, and six harmonic components. Each harmonic
component is represented by a sum of cosine and sine parts, thus
allowing the identication of the phase angles. This way the

Wave index

T1 (s)

T2 (s)

A1 (cm)

A2 (cm)

10
11
12
13
14
15
16
17
18

0.76
1.01
1.27
1.55
1.83
2.00
2.31
2.46
2.72

0.65
0.81
0.97
1.13
1.27
1.36
1.49
1.55
1.65

1.2
1.9
2.6
3.4
4.2
4.7
4.8
4.8
4.8

0.9
1.4
1.8
2.3
2.7
2.9
3.3
3.4
3.7

Bi-chromatic wavesdw 4.0 (rad/s)

Wave index

T1 (s)

T2 (s)

A1 (cm)

A2 (cm)

19
20
21
22
23

1.11
1.35
1.55
1.70
1.90

0.65
0.73
0.78
0.82
0.86

2.2
2.9
3.4
3.9
4.4

0.9
1.1
1.3
1.4
1.5

exciting force is represented by the sum of: the rst order

harmonics at the incident wave frequencies, F 1
and F 1
1
2 , the
second order components at double the wave frequencies,

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

()
F 2
and F 2
, and the
1
2 , the second order sum frequency, F
second order difference frequency components, F(  ).
Fig. 4 shows a typical time record of the wave elevation (Z) and
corresponding horizontal (Fx) and vertical (Fy) wave exciting
forces on the body measured for a bi-chromatic incident wave
with frequencies of 4.05 and 4.55 rad/s. One clearly observes
modulated amplitude on the wave elevation, which is reected
also on the exciting forces. The period of this modulation is 12.57 s
which corresponds to the difference frequency of 0.5 rad/s. The
other second order effects are much more difcult to identify from
the observation of the time record, but they can be identied by the
harmonic analysis of the signal.
Fig. 5 presents the harmonic analysis result for the former
measurements. The black line stands for the Fourier analysis
results, while the small circle symbols with different colors, and
related vertical segment pointers, represent the different harmonic amplitudes obtained by the tting method. The rst conclusion is that one is able to identify all the expected harmonic

2007

components on the incident wave elevation and on the exciting

forces as well. The order of magnitude of the mean drift force and
slow drift force is much smaller than that of the rst order force,
therefore the accuracy of the measurements is important. The
Fourier analysis identies same energy, on the incident wave and
on the exciting forces, at a frequency slightly lower (3.6 rad/s) and
at a frequency slightly higher (5.1 rad/s) than the two rst
harmonics. This is not a second order effect, but rather a third
order effect related to the interaction between second order and
rst order harmonics (2o1  o2 and 2o2  o1).

5. Experimental and numerical results

5.1. Numerical models
Fig. 6 shows the meshes for the Wamit and the Hydrostar
calculations. Regarding Wamit, one quarter of the body is represented
numerical fitting

experimental data

(m)

0.1
0

Fx (N)

-0.1
150
0
-150
Fz (N)

200
0
-200
10

20

30

40

50

60

70

80

Time (s)
Fig. 4. Time record of the wave elevation, surge force and heave exciting force. Bi-chromatic wave with periods of 1.55 s and 1.38 s and corresponding amplitudes of
3.4 cm and 3.0 cm.

fourier

w1

w2

2w1

2w2

w2-w1

w1+w2

steady drift

0.04

(m)

0.03
0.02
0.01
0

Fx (N)

60
40
20
0
80

Fz (N)

60
40
20
0
0

5
w [rad/s]

10

Fig. 5. Harmonic analysis of the wave elevation, surge and heave exciting forces. Bi-chromatic wave with periods of 1.55 s and 1.38 s and corresponding amplitudes of
3.4 cm and 3.0 cm.

2008

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

by 860 low order panels with cosine spacing on the vertical direction
to obtain a ner mesh near the free surface, which is important for
the accurate second order pressure integration. The interior free
surface is automatically meshed to remove the irregular frequencies.
The convergence analysis was performed based on the results of the
mean drift forces calculated by the pressure integration method.
Previous numerical tests show that the mean drift forces by the direct
method need much more elaborate meshes to converge than the rst
order results, or the drift forces by the momentum method. The free
surface in the vicinity of the body is meshed with 175 quadrilateral
panels for one quadrant of the annulus between the body and a circle
with a radius of 0.6 m. Outside this circle the free surface integrations
are carried out analytically.

The Hydrostar mesh is composed of 1494 quadrilateral low

order panels distributed over the wetted body surface, with
constant steps on the vertical wall, on the horizontal bottom
and on the rounded bilge. An interior free surface was automatically generated for removal of irregular frequencies. HydroStar
middle eld option uses an additional cylindrical control surface
surrounding the body, that extends from the bottom to the free
surface, closing the domain. In this case the radius of the control
surface is around 2.5 times the body radius.
Diffrac-R is based on an analytical solution for axis-symmetrical bodies and therefore uses the information regarding the
radius of the horizontal cross sections along the vertical axis.
5.2. First order wave exciting forces
This section presents the rst order wave exciting forces as
function of harmonic wave period. The rst harmonic amplitudes
are non-dimensionalized as
F 1

nd

26

rgAL2

where r, g and A are, respectively, the uid density, acceleration

of gravity and rst harmonic incident wave amplitude. L is a
characteristic length here dened as 1 m.
The graphs of Fig. 7 shows the surge exciting forces for the three
tested water depths, namely 40 cm, 55 cm and 300 cm. Fig. 8

Fig. 6. Wamit mesh (left side) and HydroStar middle eld mesh (right side).

Surge first order force

(h=300cm)

F 1

Surge first order force

(h=55cm)

Surge first order force

(h=40cm)

0.3

0.3

0.3

0.2

0.2

0.2

0.1

0.1

0.1

0
0.5

1.5
T (s)

2.5

0
0.5

1.5
T (s)

2.5

0
0.5

Wamit
experimental - regular

HydroStar
experimental - dw=0.5 T1

Diffrac-R
experimental - dw=0.5 T2

experimental - dw=1.5 T1
experimental - dw=4 T2

experimental - dw=1.5 T2

experimental - dw=4 T1

1.5
T (s)

2.5

Fig. 7. First order wave exciting surge force.

Heave first order force

(h=300 cm)

Heave first order force

(h=55 cm)

Heave first order force

(h=40 cm)
0.3

0.2

0.2

0.2

(1)
F3nd

(1)
F3nd

0.3

(1)
F3nd

0.3

0.1

0
0.5

0.1

0.1

1.5
T (s)

2.5

0
0.5

Wamit
experimental - regular
experimental - dw=1.5 T1
experimental - dw=4 T2

1.5
T (s)

2.5

HydroStar
experimental - dw=0.5 T1
experimental - dw=1.5 T2

0
0.5

1.5
T (s)

2.5

Diffrac-R
experimental - dw=0.5 T2
experimental - dw=4 T1

Fig. 8. First order wave exciting heave force. (For interpretation of the references to color in this gure, the reader is referred to the web version of this article.)

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

presents similar results but in this case for the heave exciting force.
Each graph includes experimental data (symbols) and numerical
results (lines). There are different sets of experimental results,
represented by different symbols, which correspond to the monochromatic waves and bi-chromatic waves with 3 difference frequencies. Each difference frequency tests considers two harmonic
incident waves and, in the graphs, T1 corresponds to the lower
period and T2 to the higher period. Regarding the numerical results,
the black, blue and red lines correspond, respectively, to Wamit,
HydroStar and Diffrac-R.
The experimental data shows a slightly smaller dispersion of
results for the monochromatic waves than for the bi-chromatic,
but overall the agreement between different sets of results is
good. The agreement between the experimental data and the
numerical results is very good indicating that the potential ow
methods are able to estimate very well the rst order wave
exciting forces. The results from Diffrac-R tend to be slightly
larger than the others for long periods. Both the experimental and
numerical surge forces increase as the water depth decreases,
especially for the medium to longer period range.

5.3. Steady drift forces

The horizontal and vertical mean drift forces, respectively in
the surge and heave directions, were measured in monochromatic
waves for the wave period range of interest. Figs. 9 and 10 present
0
0
the nondimensional force coefcients, F 1nd
and F 3nd
, as function of
the wave period. Each gure includes three graphs corresponding

2009

to three water depths, namely 300 cm, 55 cm and 40 cm. The

forces are non-dimensionalized as
F 0

nd

F 0

rgA2 L

27

where the nomenclature has been dene before. The symbols in

the graphs represent the experimental data and the lines with
different colors represent numerical predictions by different
methods. As dened before, Wamit and Diffract-C results are
represented, respectively, by the black and red lines. Regarding
HydroStar, the second order forces are calculated with the near
eld method (pink line) and with the middle eld method (blue
line). This correspondence is used along the text.
The experimental data shows a consistent behavior for different water depths and correct tendencies along the frequency
range, which is an indication that the data is correct. However
there is some dispersion, which is associated to the experimental
uncertainty since the second order forces are very small compared to the global forces measured. The surge forces tend for a
constant value for short periods and to zero for long periods. On
the contrary, the heave forces go to zero for short periods and to a
constant value for long periods. One observes a signicant
increase in the mean drift forces as the water depth decreases,
especially for the medium to long period range, where the
shallow water effects are more important.
Regarding the numerical results, there is an overall very good
agreement between the four methods. The Wamit surge drift
force for short periods gives slightly lower results than the other

Fig. 9. Surge steady drift force in monochromatic waves for three water depths. (For interpretation of the references to color in this gure, the reader is referred to the web
version of this article.)

Fig. 10. Heave steady drift force in monochromatic waves for three water depths. (For interpretation of the references to color in this gure, the reader is referred to the
web version of this article.)

2010

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

methods, while Diffrac-R heave force results for the shallower

water depth (40 cm) are very slightly larger than the others. The
comparison between experimental data and numerical results
leads to the conclusion that the numerical methods are able to
compute quite accurately the mean surge and heave drift forces in
monochromatic waves, including the depth effects.
5.4. Slowly varying wave exciting drift forces
This section presents the slowly varying wave exciting drift
forces in bi-chromatic waves, meaning that we have two incident
regular waves (harmonic waves for the numerical calculations)
with different periods and propagating in the same direction.
Surge and heave forces are considered and, again, three water
depths. Before the wave induced forces are discussed, it is
interesting to analyze the experimental slowly varying incident
wave amplitudes. The graphs of Fig. 11 show the ratio between
the slowly varying wave amplitude and the mean incident rst
harmonic wave amplitude
A

nd

A
A1 A2 =2

28

where A(  ) is the slowly varying wave amplitude, and A1, A2 are

the rst harmonic amplitudes of the two regular incident waves.
The results are presented as function of the mean period of the
two incident regular waves
T T 1 T 2 =2

29

The left graph corresponds to the bi-chromatic difference frequency of 0.50 rad/s and the right graph to the difference frequency

of 1.5 rad/s. Each graph includes the wave characteristics for the
three water depths. The blue lines represent linear ts to each group
of experimental data and they are used to make clear the different
tendencies.
One observes that the slowly varying wave amplitude is very
small for deep water and it remains constant along the mean
wave period range. This is expected since the wave steepnesses
used are small and approximately constant along the wave period
range. However the slowly varying wave amplitude increases
very much as the water depth decreases, especially for the long
period range where the shallow water effects are stronger. The
largest slowly varying amplitudes are around 40% of the mean
rst harmonic amplitude, which occurs for the shallow water
depth. The slowly varying amplitudes tend to increase as the
difference frequency increase.
Regarding the wave exciting slow drift forces, the graphs in
Figs. 1217 present the nondimensional forces as function of the
mean period of the two incident waves. Each graph includes
results for pairs of regular waves corresponding to one difference
frequency. Three difference frequencies are considered, namely
0.5, 1.5 and 4.0 rad/s. The second order forces are nondinensionalized as
F 

nd

F 
rgA1 A2 L

30

where in addition to the nomenclature already dened, A1 and A2

are the rst harmonic amplitudes of the two incident regular
waves.
One observes, as before for the mean drift forces, some
dispersion of the experimental data which seems to increase for

Slowly varying wave amplitude

(dw=0.5 rad/s)

Slowly varying wave amplitude

(dw=1.5 rad/s)
0.60

0.60
h = 40cm

h = 40cm

h = 55cm

h = 55cm

h = 300cm

0.40

h = 300cm

And

A nd

()

()

0.40

0.20

0.20

0.00
0.5

1.0

1.5
2.0
(T1+T2 )/2 (s)

2.5

0.00
0.5

1.0

1.5
2.0
(T1+T2 )/2 (s)

2.5

Fig. 11. Normalized slowly varying wave amplitudes in bi-chromatic waves as function of the mean wave period.

Fig. 12. Surge slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 0.5 rad/s.

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

Fig. 13. Surge slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 1.5 rad/s.

Fig. 14. Surge slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 4.0 rad/s.

Fig. 15. Heave slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 0.5 rad/s.

Fig. 16. Heave slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 1.5 rad/s.

2011

2012

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

Fig. 17. Heave slowly varying wave exciting drift force in bi-chromatic waves. Three water depths and difference frequency of 4.0 rad/s.

longer wave periods and for shallower water depths. However it

is clear that the slowly varying wave exciting drift forces were
successfully identied and measured in this experimental program. The numerical results follow, in general, the force magnitudes and tendencies of the experimental data. Therefore it can be
said that the numerical models are able to compute correctly, at
least qualitatively, the slowly varying drift forces. We observe a
signicant increase of the force coefcients as the water depth
decreases. At the long periods range, there is an increase of the surge
forces of around 400% from the deep water to the shallow water.
The increase is even larger for the heave second order force.
When we observe and compare the rst order results
(Figs. 7 and 8), the mean drift forces (Figs. 9 and 10) and the
slow drift forces (Figs. 1217), we conclude that the second order
mean drift forces increase signicantly more than the rst order
forces as the water depth decreases, and the slowly varying forces
increase much more than the other two. This behavior is justied
by the fact that second order effects in the incident wave eld
increase as the water depth decreases. In fact, the harmonic
analysis of the bi-chromatic incident waves demonstrated that
the slowly varying incident wave increases very much for shallow
water (Fig. 11).
Returning to the graphs of Figs. 1217, one observes that the
agreement between different numerical methods is not as good as
for the rst order forces and steady drift forces. The agreement
worsens as the difference frequency increases. The HydroStar
near eld and HydroStar middle eld methods give nearly the
same results, which can be explained by the fact that the second
order forces by both methods are calculated from the same rst
order quantities. The middle eld method has some advantage
when the body has sharp corners since in this case the near eld
method may encounter numerical difculties. The body under
investigation does not have sharp corners. The near eld method
also needs a more rened mesh close to the waterline to carry out
the integrations accurately at this part of the hull.
Diffrac-R analytical method gives qualitatively similar results
to the other methods for the surge force, and in particular the
agreement with Wamit is good. Regarding Wamit and HydroStar
results, they compare well for the lower difference frequency, but
not for the other two cases. Given the dispersion of the experimental data, it is not possible to conclude which of the methods is
more accurate, except in the case of the heave force for the larger
difference frequency where the HydroStar shows better results.
Figs. 13 and 14 show signicant dispersion of the surge slow
drift force numerical results from the three codes and this
dispersion increases with the increase of the difference frequency.
It is expected that the differences, between codes, of the second
order results are approximately double the differences of rst

order results, and this partly explains the former observation. But
in fact the exact reason for the discrepancies was not determined.
In deep analysis of the different second order components would
be necessary to draw denite conclusions, however the commercial codes do not allow for the detailed analysis of partial results.
In fact, the slow drift forces in the case of a xed body include
three components, namely: a quadratic term arising from the
uid velocities squared in Bernoullis equation, an integral over
the mean waterline accounting for the effect that the rst order
pressure has when it acts on the additional hull wetted surface
due to the relative free surface elevation, as well as the forces
associated to the second order velocity potential. It would therefore be interesting to conclude on which of the components
originates the identied discrepancies. It is however believed
that the discrepancies can be primarily traced back to the fact that
for the evaluation of the wave loads due to second order
potentials, different methods have been used to properly account
for the second order free surface effects (direct evaluation of the
second-order pressure itself along with indirect methods that
evaluate second-order integrated quantities only).

6. Conclusions
The paper presents an experimental study of the rst order
and slowly varying wave exciting forces on a body of simple
geometry in monochromatic and bi-chromatic. Three water
depths were tested. The wave exciting mean drift forces and
second order slowly varying forces were successfully measured
and identied. However, since the magnitude of these forces is
small compared to the global forces measured, there is some
dispersion in the second order results.
The experimental data clearly shows that the second order
forces increase signicantly as the water depth decreases. The
rst order forces increase for shallow water as well, however the
rate of increase of the second order forces is much higher. This
leads to the conclusion that second order forces may be more
important in shallow water than in deep water. The magnitude of
the slowly varying second order drift forces are very much related
to the magnitude of the second order characteristics of the bichromatic incident wave. In fact, the second order effects on the
wave eld itself increase as the water depth decreases and this
induces the increase on the second order exciting forces.
The experimental data is compared with three state of the art
numerical methods which are able to compute the second order
loads. Two of the methods use Greens function panel methods
(Wamit and HydroStar) and the third is based on an analytical
solution (Diffrac-R). All methods take into account the full second

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

order solution, meaning the quadratic interactions of rst order

quantities and the second order diffraction potentials. Regarding
the rst order exciting forces, all methods give very similar
results, with differences up to 3%. The agreement with the
experimental data is very good. The mean drift forces in regular
waves are very well predicted by the three numerical methods as
well. All numerical results are in very good agreement between
each other.
With respect to the slowly varying drift forces, the agreement
between the three numerical methods is good for the lower wave
frequency difference and it worsens as the difference frequency
increases. Furthermore, the numerical methods are able to qualitatively represent the measured slowly varying wave exciting
drift forces, namely the order of magnitude of the forces, the
tendencies along the mean wave period range and the increase of
these forces with the decreasing water depth. The predicted
forces are within the range of uncertainty of the experimental
data. Comparing the different numerical methods, it is not
possible to say that one agrees better with the experimental data
than the others. Overall, it can be said that the tested numerical
methods are appropriate for the calculation of the mean and
slowly varying wave exciting drift forces.

Acknowledgments
The experimental work has been supported by European
Communitys Sixth Framework Programme through the grant to
the budget of the Integrated Infrastructure Initiative HYDRALAB
III within the Transnational Access Activities, Contract no.
022441.
The analysis of experimental data and numerical calculations
have been partially performed in the scope of the project Design
of mooring systems for oating wave energy converters, which is
nanced by Fundac- a~ o para a Ciencia e a Tecnologia (FCT) under
contract PTDC/EME-MFE/103524/2008.

References
Abul-Azm, A.G., Williams, A.N., 1988. Second-order diffraction loads on truncated
cylinders. J. Waterw. Port Coastal Ocean Div. 14, 436454.
Chatjigeorgiou, I.K., Mavrakos, S.A., 2007. Second-order sum-frequency wave
diffraction by a truncated surface-piercing cylinder in bi-chromatic waves.
J. Mar. Sci. Technol. 12, 218231.
Chau, F.P., Eatock Taylor, R., 1992. Second-order wave diffraction by a vertical
cylinder. J. Fluid Mech. 240, 571599.
Chen, X.B., 2004. Hydrodynamics in offshore and naval applicationspart I. In:
Proceedings of the Keynote Lecture at the Sixth International Conference of
HydroDynamics, Perth, Australia.
Chen, X.B., 2005. Hydrodynamic analysis for offshore LNG terminals. In: Proceedings of the Second Offshore Hydrodynamics Symposium, Rio de Janeiro.
Chen, X.B., Rezende, F., 2009. Efcient computations of second-order lowfrequency wave load. In: Proceedings of the 28th International Conference
on Offshore Mechanics and Arctic Engineering (OMAE2009), Paper
OMAE2009-79522, ASME, New York, USA.
DeBoom, W.C., Pinkster, J.A., Tan, P.S.G., 1984. Motion and tether force prediction
for a TLP. J. Waterw. Port Coastal Ocean Eng., ASCE 110 (4), 472486.
Eatock Taylor, R., Hung, S.M., 1987. Second order diffraction forces on a vertical
cylinder in regular waves. Appl. Ocean Res. 9, 1930.
Eatock Taylor, R., Huang, J.B., 1997. Semi-analytical formulation for second-order
diffraction by a vertical cylinder in bichromatic waves. J. Fluid Struct. 11,
465484.
Emmerhoff, O.J., Sclavounos, P.D., 1996. The simulation of slow-drift motions of
offshore structures. Appl. Ocean Res. 18, 5564.
Faltinsen, O.M., Michelsen, F.C., 1974. Motions of large structures in waves at zero
Froude number. In: Proceedings of the International Symposium on Dynamics
of Marine Vehicles and Structures in Waves, University College London, UK,
pp. 91106.
Fonseca, N., Pessoa, J., Guedes Soares, C., 2008. Calculation of second order drift
forces on a FLNG accounting for difference frequency components. In:
Proceedings 27th International Conference on Offshore Mechanics and Arctic
Engineering (OMAE2008), Paper OMAE2008-57942, ASME, New York, USA.

2013

Herfjord, K., Nielson, F.G., 1986. Nonlinear wave forces on a xed vertical cylinder
due to the sum frequency of waves in irregular seas. Appl. Ocean Res. 8,
821.
Huang, J.B., Eatock Taylor, R., 1996. Semi-analytical solution for second-order
wave diffraction by a truncated circular cylinder in monochromatic waves.
J. Fluid Mech. 319, 171196.
Ikoma, T., Maeda, H. Rheem, C.-K., 2000. Slowly varying wave drifting force on a
very large oating structure in short crested waves. In: Proceedings of the
Oceans Conference Record (IEEE), vol. 1, pp. 533539.
Kim, M.H., Yue, D.K.P., 1989. The complete second order diffraction solution for an
axisymmetric body. Part I. Monochromatic incident waves. J. Fluid Mech. 200,
235264.
Kim, M.H., Yue, D.K.P., 1990. The complete second-order diffraction solution for an
axisymmetric body. Part II: bichromatic incident waves and body motions.
J. Fluid Mech. 201, 557593.
Kashiwagi, M., Endo, K., Yamaguchi, H., 2005. Wave drift forces and moments on
two ships arranged side by side in waves. Ocean Eng. 32, 529555.
Kobayashi, M., Shimada, K., Fujchira, T., 1985, Study on the dynamic response of a
TLP in waves. In: Proceedings of the Fourth International Offshore Mechanics
and Arctic Engineering Symposium, ASME, vol. 1, pp. 2935.
Kokkinowrachos, K., Mavrakos, S.A., Asorakos, S., 1986. Behavior of vertical bodies
of revolution in waves. Ocean Eng. 13, 505538.
Lee, C.-H., Zhu, X., 1993. Second-order diffraction and radiation solutions on
oating bodies. In: Proceedings of the Eight International Workshop on Water
Waves and Floating Bodies, St. Johns, Newfoundland, Canada.
Lee, C.-H., Newman, J.N., Zhu, X., 1996. An extended boundary-integral-equation
method for the removal of irregular-frequency effects. Int. J. Num. Methods
Fluids 23, 637660.
Lee, C.-H., Newman, J.N., 2004. Computation of wave effects using the panel
method. In: Chakrabarti, S. (Ed.), Numerical Models in Fluid-Structure Interaction, Preprint, WIT Press, Southhampton.
Lee, C.-H., 1995. WAMIT Theory Manual. MIT.
Lee, C.M., Newman, J.N., 1971. The vertical man force and moment of submerged
bodies under waves. J. Ship Res. 15, 231245.
Lighthill, J., 1979. Waves and hydrodynamic loading. In: Proceedings of the
Second International Conference on Behaviour of Offshore Structures,
vol. 1, pp. 140.
Loken, A.E., 1986. Three-Dimensional Second-Order Hydrodynamic Effects on
Ocean Structures in Waves, Report UR-86-54, Dept. of Marine Technology,
University of Trondheim, Norway.
Malenica, S., Eatock Taylor, R., Huang, J.B., 1999. Second-order water wave
diffraction by an array of vertical cylinders. J. Fluid Mech. 390, 349373.
Maruo, H., 1960. The drift of a body oating in waves. J. Ship Res. 4 (3), 110.
Mavrakos, S.A., 1988. The vertical drift force and pitch moment on axisymmetric
bodies in regular waves. Appl. Ocean Res. 10, 207218.
Mavrakos, S.A., Peponis, V., 1992. Second-order sum and difference wave loads on
axisymmetric bodies restrained in irregular waves. In: Proceedings of the
Second International Offshore Polar Engineering Conference, San Francisco
USA, vol. III, pp. 546553.
Mavrakos, S.A., Chatjigeorgiou, I.K., 2006. Second-order diffraction by a bottom
seated compound cylinder. J. Fluids Struct. 22, 463492.
Molin, B., 1979. Second order diffraction loads upon three-dimensional bodies.
Appl. Ocean Res. 1, 197202.
Newman, J.N., 1967. The drift force and moment on ships in waves. J. Ship Res. 11,
5160.
Newman, J.N., 1974, Second order slowly varying forces on vessels in irregular
waves. In: Proceedings of the Symposium on Dynamics of Marine Vehicles and
Structures in Waves, London.
Newman, J.N., 2005. Second-order diffraction in short waves. J. Ship Res. 49,
263273.
Ogilvie, T.F., 1983. Second-order hydrodynamic effects on ocean platforms. In:
Proceedings of the International Workshop on Ships & Platform Motions,
Berkeley, USA, pp. 205265.
Petrauskas, C., Liu, S.V., 1987. Springing force response of a tension leg platform.
In: Proceedings of the Ninth Annual Offshore Technology Conference, OTC
5458, Houston, Texas, pp. 333341.
Pinkster, J.A.., 1976. Low frequency second order wave forces on vessels moored
at sea. In: Proceedings 11th Symposium on Naval Hydrodynamics, pp.
603615.
Pinkster, J.A., Van Oortmersen, G., 1977, Computation of the rst and second order
wave forces on oscillating bodies in regular waves. In: Proceedings of the
Second International Conference on Numerical Ship Hydrodynamics, Berkeley,
USA, pp. 121.
Pinkster, J.A., 1979. Mean and low frequency wave drifting forces on oating
structures. Ocean Engineering 6 (6), 593615.
Pinkster, J.A., 1980. Low-Frequency Second Order Wave Exciting Force on Floating
Structures, Netherlands Ship Model Basin, Publication No. 650.
Rezend, F., Li, X., Chen, X.-B., 2007. Second Order Loads on LNG Terminals in
Multi-Directional Sea in Water of Finite Depth, OMAE 2007, San Diego,
paper29231.
Scolan, Y.-M., Molin, B., 1989, Second-order deformation of the free surface around
a cylinder, Part II. In: Proceedings Fourth International Workshop on Water
Waves and Floating Bodies, Oystese, Norway.
Stansberg C.-T., 2007. Slow-drift pitch motions and air-gap observed from model
testing with moored semi-submersibles. In: Proceedings of the 26th OMAE,
San Diego, CA, United States, vol. 1, pp. 659668.

2014

N. Fonseca et al. / Ocean Engineering 38 (2011) 20002014

Standing, R.G., 1991. Uncertainties in estimating second-order low-frequency

wave forces and responses. In: Proceeding of the International Conference
Offshore Mechanics and Arctic Engineering.
Standing, R.G., Brendling, W.J., Jackson, G.E., 1993. Summary Report on an
Investigation into the Correlation Between Full-Scale Measured and Predicted
Motions of the SSSV Uncle John, Health and Safety ExecutiveOffshore
Technology Report OTH 92 363, ISBN 0118821369.

van der Valk, C.A.C., Watson, A., 2005. Mooring Of LNG carriers to a weathervaning
oaterside-by-side or stern-by-bow. In: Proceedings of the 2005 OTC,
Houston, Paper 17154.
Zaraphonitis, G.N., Papanikolaou, A.D., 1993. Second order theory and calculations
of motions and loads of arbitrarily shaped 3D bodies in waves. Mar. Struct. 6,
165185.