1216 TheoreticalManualSEAWAY
1216 TheoreticalManualSEAWAY
1216 TheoreticalManualSEAWAY
SEAWAY
(Release 4.19, 12-02-2001)
Abstract
This report describes in detail the theoretical backgrounds of the six degrees of
freedom ship motions program SEAWAY.
SEAWAY is a frequency-domain ship motions computer code, based on the linear
strip theory, to calculate the wave-induced loads, motions, added resistance and
internal loads for six degrees of freedom of displacement ships, yachts and barges,
sailing in regular and irregular waves. When not taking into account interaction
e¤ects between the two individual hulls, these calculations can be carried out for
twin-hull ships, such as semi-submersibles and catamarans, too. This potential theory
program is suitable for deep water as well as for shallow water. Viscous roll damping,
bilge keels, anti-roll tanks, free surface e¤ects and linear springs can be added. A
dedicated editor takes care for a simple input of data.
This report and other information on the strip theory program SEAWAY can be
found on the Internet at web site http://dutw189.wbmt.tudelft.nl/~johan, which can
also be reached by a link to this site from http://www.shipmotions.nl.
Aditional information can be obtained by e-mail (J.M.J.Journee@wbmt.tudelft.nl)
from the author.
The last revision of this report is dated: 3 July 2001. Remarks and errata are
very welcome!
Contents
1 Introduction 1
3 Conformal Mapping 35
3.1 Lewis Conformal Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.1.1 Boundaries of Lewis Forms . . . . . . . . . . . . . . . . . . . . . . . 38
3.1.2 Acceptable Lewis Forms . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Extended Lewis Conformal Mapping . . . . . . . . . . . . . . . . . . . . . 40
3.3 Close-Fit Conformal Mapping . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.4 Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3
4 CONTENTS
Introduction
This report aims being a guide and an aid for those who want to study the theoretical
backgrounds and the algorithms of a ship motions computer program based on the strip
theory.
The present report describes in detail the theoretical backgrounds and the algorithms of
a six degrees of freedom ship motions personal computer program, named SEAWAY. A
User Manual is given by [Journée, 2001a]. Extensive veri…cations and validations of this
program have been presented by [Journée, 2001b].
This program, based on the ordinary and the modi…ed strip theory, calculates the wave-
induced loads and motions with six degrees of freedom of mono-hull ships and barges, sailing
in a seaway. When not taking into account interaction e¤ects between the two individual
hulls, these calculations can be carried out for twin-hull ships, such as semi-submersibles
and catamarans, too.
In the past a preliminary description of all algorithms, used in strip theory based ship
motions calculations, has been given by the author, see [Journée, 1992]. Since then, this
program has been extended and adapted considerably, so a revised report is presented here.
Chapter 1, this introduction, gives a short survey of the contents of all chapters in this
report.
Chapter 2 gives a general description of the various strip theory approaches. A general
description of the potential ‡ow theory is given. The derivations of the hydromechanical
forces and moments, the wave potential and the wave and di¤raction forces and moments
have been described.
The equations of motion are given with solid mass and inertia terms and hydromechanical
forces and moments in the left hand side and the wave exciting forces and moments in the
right hand side.
The principal assumptions are a linear relation between forces and motions and the va-
lidity of obtaining the total forces by a simple integration over the ship length of the
two-dimensional cross sectional forces.
This includes for all motions a forward speed e¤ect caused by the potential mass, as it has
been de…ned by [Korvin-Kroukovsky and Jacobs, 1957] for the heave and pitch motions.
This approach is called the ”Ordinary Strip Theory Method”. Also an inclusion of the
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
1
2 CHAPTER 1. INTRODUCTION
forward speed e¤ect caused by the potential damping, as for instance given by [Tasai, 1969],
is given. This approach is called the ”Modi…ed Strip Theory Method”.
The inclusion of so-called ”End-Terms” has been described.
Chapter 3 describes several conformal mapping methods. For the determination of the
two-dimensional hydrodynamic potential coe¢cients for sway, heave and roll motions of
ship-like cross sections, these cross sections are conformal mapped to the unit circle. The
advantage of conformal mapping is that the velocity potential of the ‡uid around an arbi-
trary shape of a cross section in a complex plane can be derived from the more convenient
circular section in another complex plane. In this manner hydrodynamic problems can be
solved directly with the coe¢cients of the mapping function.
The close-…t multi-parameter conformal mapping method is given. A very simple and
straight on iterative least squares method, used to determine the conformal mapping co-
e¢cients, has been described. Two special cases of multi-parameter conformal mapping
have been described too: the well known classic Lewis transformation ([Lewis, 1929]) with
two parameters and an Extended-Lewis transformation with three parameters, as given by
[Athanassoulis and Loukakis, 1985].
Chapter 4 describes the determination of the two-dimensional potential mass and damping
coe¢cients for the six modes of motions at in…nite and …nite water depths.
At in…nite water depths, the principle of the calculation of these potential coe¢cients is
based on work of [Ursell, 1949] for circular cylinders and [Frank, 1967] for any arbitrary
symmetric cross section.
Starting from the velocity potentials and the conjugate stream functions of the ‡uid with
an in…nite depth as have been given by [Tasai, 1959], [Tasai, 1960], [Tasai, 1961] and
[Jong, 1973] and using the multi-parameter conformal mapping technique, the calcula-
tion routines of the two-dimensional hydrodynamic potential coe¢cients of ship-like cross
sections are given for the sway, heave and roll motions.
For shallow water, the method of [Keil, 1974] - based on a variation of the theory of
[Ursell, 1949] - has been given.
The pulsating sources method of [Frank, 1967] for deep water has been described too.
Because of using the strip theory approach here, the pitch and yaw coe¢cients follow
from the moment about the ship’s centre of gravity of the heave and sway coe¢cients,
respectively.
Approximations are given for the surge coe¢cients.
Chapter 5 gives some corrections on the hydrodynamic damping due to viscous e¤ects. The
surge damping coe¢cient is corrected for viscous e¤ects by an empirical method, based on
a simple still water resistance curve as published by [Troost, 1955].
The analysis of free-rolling model experiments and two (semi-)empirical methods pub-
lished by [Miller, 1974] and [Ikeda et al., 1978], to determine a viscous correction of the
roll damping coe¢cients are described in detail.
Chapter 6 describes the determination of the hydromechanical forces and moments in the
left hand side of the six equations of motion of a sailing ship in deep water for both the
ordinary and the modi…ed strip theory method.
Chapter 7 describes the wave exciting forces and moments in the right hand side of the six
equations of motion of a sailing ship in water with an arbitrarily depth, using the relative
motion concept for both the ordinary and the modi…ed strip theory method.
3
First, the classical approach has been described. Then, an alternative approach - based on
di¤raction of waves - with equivalent accelerations an velocities of the water particles has
been described.
Chapter 8 describes the solution of the equations of motion. The determination of the fre-
quency characteristics of the absolute displacements, rotations, velocities and accelerations
and the vertical relative displacements. The use of a wave potential valid for any arbitrary
water depth makes the calculation method, with deep water coe¢cients, suitable for ships
sailing with keel clearances down to about 50 percent of the ship’s draft.
Chapter 9 describes some anti-rolling devices. A description is given of an inclusion of
passive free-surface tanks as de…ned by the experiments of [Bosch and Vugts, 1966] and
by the theory of [Verhagen and van Wijngaarden, 1965]. Active …n stabilizers and active
rudder stabilizers have been described too.
Chapter 10 describes the inclusion of linear spring terms to simulate the behavior of an-
chored or moored ships.
Chapter 11 describes two methods to determine the transfer functions of the added resis-
tances due to waves. The …rst method is a radiated wave energy method, as published by
[Gerritsma and Beukelman, 1972]. The second method is an integrated pressure method,
as published by [Boese, 1970].
Chapter 12 describes the determination of the frequency characteristics of the lateral and
vertical shear forces and bending moments and the torsional moments in a way as presented
by [Fukuda, 1962] for the vertical mode. Still water phenomena are described too.
Chapter 13 describes the statistics in irregular waves, by using the superposition principle.
Three examples of normalized wave spectra are given: the somewhat wide wave spec-
trum of Neumann, an average wave spectrum of Bretschneider and the more narrow Mean
JONSWAP wave spectrum.
A description is given of the calculation procedure of the energy spectra and the statistics
of the ship motions for six degrees of freedom, the added resistances, the vertical relative
motions and the mechanic loads on the ship in waves coming from any direction.
For the calculation of the probability of exceeding a threshold value by the motions, the
Rayleigh probability density function has been used.
The static and dynamic swell up of the waves, of importance when calculating the proba-
bility of shipping green water, are de…ned according to [Tasaki, 1963].
Bow slamming phenomena are de…ned by both the relative bow velocity criterium of
[Ochi, 1964] and the peak bottom impact pressure criterium of [Conolly, 1974].
Chapter 14 describes the additions to the algorithms in case of twin- hull ships, such as
semi-submersibles and catamarans. For interaction e¤ects between the two individual hulls
will not be accounted here.
Chapter 15 shows some typical numerical recipes, as used in program SEAWAY.
4 CHAPTER 1. INTRODUCTION
.
Chapter 2
The ship is considered to be a rigid body ‡oating in an ideal ‡uid: homogeneous, incom-
pressible, free of surface tension, irrotational and without viscosity. It is assumed that the
problem of the motions of this ‡oating body in waves is linear or can be linearized. As a
result of this, only the external loads on the underwater part of the ship are considered
and the e¤ect of the above water part is fully neglected.
The incorporation of seakeeping theories in ship design has been discussed clearly by
[Faltinsen and Svensen, 1990]. An overview of seakeeping theories for ships were presented
and it was concluded that - nevertheless some limitations - strip theories are the most
successful and practical tools for the calculation of the wave induced motions of the ship,
at least in an early design stage of a ship.
The strip theory solves the three-dimensional problem of the hydromechanical and exciting
wave forces and moments on the ship by integrating the two-dimensional potential solutions
over the ship’s length. Interactions between the cross sections are ignored for the zero-speed
case. So each cross section of the ship is considered to be part of an in…nitely long cylinder.
The strip theory is a slender body theory, so one should expect less accurate predictions
for ships with low length to breadth ratios. However, experiments showed that the strip
theory appears to be remarkably e¤ective for predicting the motions of ships with length
to breadth ratios down to about 3.0, or even sometimes lower.
The strip theory is based on the potential ‡ow theory. This holds that viscous e¤ects are
neglected, which can deliver serious problems when predicting roll motions at resonance
frequencies. In practice, for viscous roll damping e¤ects can be accounted fairly by empirical
formulas.
Because of the way that the forced motion problems are solved generally in the strip theory,
substantial disagreements can be found between the calculated results and the experimental
data of the wave loads at low frequencies of encounter in following waves. In practice, these
”near zero frequency of encounter problems” can be solved by forcing the wave loads to go
to zero arti…cially.
For high-speed vessels and for large ship motions, as appear in extreme sea states, the strip
theory can deliver less accurate results. Then the so-called ”end-terms” can be important
too.
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
5
6 CHAPTER 2. STRIP THEORY METHOD
The strip theory accounts for the interaction with the forward speed in a very simple
way. The e¤ect of the steady wave system around the ship is neglected and the free surface
conditions are simpli…ed, so that the unsteady waves generated by the ship are propagating
in directions perpendicular to the centre plane of the ship. In reality the wave systems
around the ship are far more complex. For high-speed vessels, unsteady divergent wave
systems become important. This e¤ect is neglected in the strip theory.
The strip theory is based on linearity. This means that the ship motions are supposed to
be small, relative to the cross sectional dimensions of the ship. Only hydrodynamic e¤ects
of the hull below the still water level are accounted for. So when parts of the ship go out
of or into the water or when green water is shipped, inaccuracies can be expected. Also,
the strip theory does not distinguish between alternative above water hull forms.
Because of the added resistance of a ship due to the waves is proportional to the relative
motions squared, its inaccuracy will be gained strongly by inaccuracies in the predicted
motions.
Nevertheless these limitations, seakeeping prediction methods based upon the strip theory
provide a su¢ciently good basis for optimization studies at an early design stage of the
ship. At a more detailed design stage, it can be considered to carry out additional model
experiments to investigate for instance added resistance or extreme event phenomena, such
as shipping green water and slamming.
2.1 De…nitions
Figure 2.1 shows a harmonic wave as seen from two di¤erent perspectives. Figure 2.1-a
shows what one would observe in a snapshot photo made looking at the side of a (trans-
parent) wave ‡ume; the wave pro…le is shown as a function of distance x along the ‡ume
at a …xed instant in time. Figure 2.1-b is a time record of the water level observed at one
location along the ‡ume; it looks similar in many ways to the other …gure, but time t has
replaced x on the horizontal axis.
Notice that the origin of the coordinate system is at the still water level with the positive
z-axis directed upward; most relevant values of z will be negative. The still water level
is the average water level or the level of the water if no waves were present. The x-axis
is positive in the direction of wave propagation. The water depth, h, (a positive value) is
measured between the sea bed (z = ¡h) and the still water level (z = 0).
2.1. DEFINITIONS 7
The highest point of the wave is called its crest and the lowest point on its surface is the
trough. If the wave is described by a harmonic wave, then its amplitude ³ a is the distance
from the still water level to the crest, or to the trough for that matter. The subscript ”a”
denotes amplitude, here.
The horizontal distance (measured in the direction of wave propagation) between any two
successive wave crests is the wave length, ¸. The distance along the time axis is the wave
period, T: The ratio of wave height to wave length is often referred to as the dimensionless
wave steepness, 2³ a =¸:
Since the distance between any two corresponding points on successive sine waves is the
same, wave lengths and periods are usually actually measured between two consecutive
upward (or downward) crossings of the still water level. Such points are also called zero-
crossings, and are easier to detect in a wave record.
Since sine or cosine waves are expressed in terms of angular arguments, the wave length
and period are converted to angles using:
2¼
k¸ = 2¼ or: k=
¸
2¼
!T = 2¼ or: != (2.1)
T
in which k is the wave number (rad/m) and ! is the circular wave frequency (rad/s).
Obviously, the wave form moves one wave length during one period so that its speed or
phase velocity, c, is given by:
¸ !
c= = (2.2)
T k
Suppose now a sailing ship in waves, with coordinate systems as given in …gure 2.2.
A right-handed coordinate system S(x0 ; y0 ; z0 ) is …xed in space. The (x0 ; y0 )-plane lies in
the still water surface, x0 is directed as the wave propagation and z0 is directed upwards.
Another right-handed coordinate system O(x; y; z) is moving forward with a constant ship
speed V . The directions of the axes are: x in the direction of the forward speed V , y in the
lateral port side direction and z upwards. The ship is supposed to carry out oscillations
around this moving O(x; y; z) coordinate system. The origin O lies above or under the
time-averaged position of the centre of gravity G. The (x; y)-plane lies in the still water
surface.
A third right-handed coordinate system G(xb ; yb ; zb ) is connected to the ship with G at
the ship’s centre of gravity. The directions of the axes are: xb in the longitudinal forward
direction, yb in the lateral port side direction and zb upwards. In still water, the (xb ; yb )-
plane is parallel to the still water surface.
If the wave moves in the positive x0 -direction (de…ned in a direction with an angle ¹ relative
to the ship’s speed vector, V ), the wave pro…le - the form of the water surface - can now
be expressed as a function of both x0 and t as follows:
The right-handed coordinate system O(x; y; z) is moving with the ship’s speed V , which
yields:
x0 = V t cos ¹ + x cos ¹ + y sin ¹ (2.4)
From the relation between the frequency of encounter ! e and the wave frequency !:
! e = ! ¡ kV cos ¹ (2.5)
follows:
³ = ³ a cos(! e t ¡ kx cos ¹ ¡ ky sin ¹) (2.6)
The resulting six ship motions in the O(x; y; z) system are de…ned by three translations
of the ship’s centre of gravity in the direction of the x-, y- and z-axes and three rotations
about them:
surge : x = xa cos(! et + "x³ )
sway : y = ya cos(! e t + "y³ )
heave : z = za cos(! e t + "z³ )
roll : Á = Áa cos(! e t + "Á³ )
pitch : µ = µa cos(! e t + "µ³ )
yaw : à = à a cos(! e t + "ó ) (2.7)
The phase lags of these motions are related to the harmonic wave elevation at the origin
of the O(x; y; x) system, the average position of the ship’s centre of gravity:
wave: ³ = ³ a cos(! e t) (2.8)
The harmonic velocities and accelerations in the O(x; y; z) system are found by taking the
derivatives of the displacements, for instance for surge:
surge displacement : x = xa cos(! e t + "x³ )
2.2. INCIDENT WAVE POTENTIAL 9
@ 2 ©w @ 2 ©w @ 2 ©w
r2 ©w = + + =0 (2.15)
@x2 @y 2 @z 2
Water particles move here in the x-z plane only, so in the equations above:
@©w @v @ 2 ©w
v= = 0 and = =0 (2.16)
@y @y @y 2
Taking this into account, a substitution of equation 2.10 in equation 2.15 yields a homo-
geneous solution of this equation:
d2 P (z)
¡ k 2 P (z) = 0 (2.17)
dz 2
with as solution for P (z):
@©w
=0 for: z = ¡h (2.20)
@z
Substituting this boundary condition in equation 2.19 provides:
C ¡ +k(h+z) ¢
P (z) = e + e¡k(h+z)
2
= C cosh k (h + z) (2.24)
@©w 1 ¡ 2 ¢ p
+ u + v 2 + w2 + + gz = 0 (2.26)
@t 2 ½
In two dimensions, v = 0 and since the waves have a small steepness (u and w are small),
this equation becomes:
@©w p
+ + gz = 0 (2.27)
@t ½
At the free surface this condition becomes:
@©w p0
+ + g³ = 0 for: z = ³ (2.28)
@t ½
The constant value p0 =½ can be included in @©w =@t; this will not in‡uence the velocities
being obtained from the potential ©w . With this the equation becomes:
12 CHAPTER 2. STRIP THEORY METHOD
@©w
+ g³ = 0 for: z = ³ (2.29)
@t
The potential at the free surface can be expanded in a Taylor series, keeping in mind that
the vertical displacement ³ is relatively small:
½ ¾
@©w (x; z; t)
f©w (x; z; t)gz=³ = f©w (x; z; t)gz=0 + ³ ¢ + ::::::
@z z=0
½ ¾ ½ ¾
@©w (x; z; t) @©w (x; z; t)
= + O("2 ) (2.30)
@t z=³ @t z=0
which yields for the linearized form of the free surface dynamic boundary condition:
@©w
+ g³ = 0 for: z = 0 (2.31)
@t
With this, the wave pro…le becomes:
1 @©w
³ =¡ ¢ for: z = 0 (2.32)
g @t
!C
³= ¢ cosh kh ¢ cos (kx ¡ !t) (2.33)
g
or:
!C
³ = ³ a ¢ cos (kx ¡ !t) with: ³ a = ¢ cosh kh (2.34)
g
With this the corresponding wave potential, depending on the water depth h, is given by
the relation:
³ a g cosh k(h + z)
©w = ¢ ¢ sin(kx ¡ !t) (2.35)
! cosh kh
or when !t is the …rst of the sine function arguments, as generally will be used in ship
motion equations:
¡³ a g cosh k(h + z)
©w = ¢ ¢ sin(!t ¡ kx) (2.36)
! cosh kh
In deep water, the expression for the wave potential reduces to:
¡³ a g kz
©w = ¢ e ¢ sin(!t ¡ kx) (deep water) (2.37)
!
2.2. INCIDENT WAVE POTENTIAL 13
dz @³ @³ dx
= + ¢ for the wave surface: z = ³
dt @t @x dt
@³ d³
= +u¢ (2.38)
@t dx
The second term in this expression is a product of two values, which are both small because
of the assumed small wave steepness. This product becomes even smaller (second order)
and can be ignored, see …gure 2.4.
@ 2 ©w @³
2
+g =0 for z = 0 (2.41)
@t @t
or after re-arranging terms:
@³ 1 @ 2 ©w
+ ¢ =0 for z = 0 (2.42)
@t g @t2
Together with equation 2.39 this delivers the free surface kinematic boundary condition or
the Cauchy-Poisson condition:
14 CHAPTER 2. STRIP THEORY METHOD
@z 1 @ 2 ©w
+ ¢ =0 for: z = 0 (2.43)
@t g @t2
! 2 = k g ¢ tanh kh (2.44)
In many situations, ! or T will be know; one must determine k or ¸: Since k appears in a
nonlinear way in 2.44, that equation will generally have to be solved iteratively.
In deep water (tanh kh = 1), equation 2.44 degenerates to a quite simple form which can
be used without di¢culty:
¡³ a g cosh k(h + zb )
©w = ¢ sin(! e t ¡ kxb cos ¹ ¡ kyb sin ¹) (2.47)
! cosh(kh)
©(x; y; z; t) = ©r + ©w + ©d (2.48)
in which:
2.3. FLOATING RIGID BODY IN WAVES 15
©r = the radiation potential for the oscillatory motion of the body in still water
©w = the incident undisturbed wave potential
©d = the di¤raction potential of the waves about the restrained body
The direction cosines are called generalized, because f1 , f2 and f3 have been normal-
ized (the sum of their squares is equal to 1) and used to obtain f4 , f5 and f6 .
Note: The subscripts 1; 2; :::6 are used here to indicate the mode of the motion. Also
displacements are often indicated in literature in the same way: x1 ; x2 ; :::x6 .
5. Radiation Condition
The radiation condition states that when the distance R of a water particle to the
oscillating body tends to in…nity, the potential value tends to zero:
lim © = 0 (2.59)
R!1
in which ~n is the outward normal vector on surface dS and ~r is the position vector of
surface dS in the O(x; y; z) coordinate system.
The pressure p - via the linearized Bernoulli equation - is determined from the velocity
potentials by:
@©
p = ¡½ ¡ ½gz
@t
µ ¶
@©r @©w @©d
= ¡½ + + ¡ ½gz (2.62)
@t @t @t
18 CHAPTER 2. STRIP THEORY METHOD
which can obviously be split into four separate parts, so that the hydromechanical forces
F~ and moments M~ can be split into four parts too:
ZZ µ ¶
~ @©r @©w @©d
F = ½ + + + gz ~n ¢ dS
@t @t @t
S
ZZ µ ¶
~ = ½ @©r @©w @©d
M + + + gz (~r £ ~n) ¢ dS (2.63)
@t @t @t
S
or:
~ = M
M ~r +M
~w +M
~d +M
~s (2.64)
in which the space and time dependent potential term, ©j (x; y; z; t) in direction j, is now
written in terms of a separate space dependent potential, Áj (x; y; z) in direction j, multi-
plied by an oscillatory velocity, vj (t) in direction j.
This allows the normal velocity on the surface of the body to be written as:
@ X
6
@©r
= ©j
@n @n j=1
X6 ½ ¾
@Áj
= ¢vj (2.66)
j=1
@n
and the generalized direction cosines are given by:
@Áj
fj = (2.67)
@n
With this the radiation terms in the hydrodynamic force becomes:
Z Z µ ¶
~ @©r
F = ½ ~n ¢ dS
@t
S
Z Z Ã X 6
!
@
= ½ Á vj ~n ¢ dS (2.68)
@t j=1 j
S
2.3. FLOATING RIGID BODY IN WAVES 19
The components of these radiation forces and moments are de…ned by:
with:
Z Z Ã !
@ X
6
~r = ½
X Á vj fk ¢ dS
k
@t j=1 j
S
Z Z Ã !
@ X
6
@Ák
= ½ Á vj ¢dS for: k = 1; :::6 (2.71)
@t j=1 j @n
S
with: Z Z
dvj @Ák
Xrkj = ½ Áj ¢dS (2.73)
dt @n
S
This radiation force or moment Xrkj in the direction k is caused by a forced harmonic
oscillation of the body in the direction j. This is generally true for all j and k in the range
from 1 to 6. When j = k; the force or moment is caused by a motion in that same direction.
When j 6= k; the force in one direction results from the motion in another direction. This
introduces what is called coupling between the forces and moments (or motions).
The above equation expresses the force and moment components, Xrkj in terms of still
unknown potentials, Áj ; not everything is solved yet! A solution for this will be found later
in this chapter.
Oscillatory Motion
Now an oscillatory motion is de…ned; suppose a motion (in a complex notation) given by:
s_ j = vj = ¡i!saj e¡i!t
dvj
sÄj = = ¡i! 2 saj e¡i!t (2.75)
dt
20 CHAPTER 2. STRIP THEORY METHOD
The hydrodynamic forces and moments can be split into a load in-phase with the acceler-
ation and a load in-phase with the velocity:
Xrkj = ¡Mkj sÄj ¡ Nkj s_ j
¡ ¢
= saj ! 2 Mkj + isaj !Nkj e¡i!t
0 1
Z Z
@Á
= @(¡saj ! 2 ½ Áj k dS)A e¡i!t (2.76)
@n
S
So in case of an oscillation of the body in the direction j with a velocity potential Áj , the
hydrodynamic mass and damping (coupling) coe¢cients are de…ned by:
8 9 8 9
< Z Z @Á = < Z Z
@Á =
Mkj = ¡< ½ Áj k ¢dS and Nkj = ¡= ½! Áj k ¢dS (2.77)
: @n ; : @n ;
S S
In case of an oscillation of the body in the direction k with a velocity potential Ák , the
hydrodynamic mass and damping (coupling) coe¢cients are de…ned by:
8 9 8 9
< Z Z @Áj = < Z Z
@Áj =
Mjk = ¡< ½ Ák ¢dS and Njk = ¡= ½! Ák ¢dS (2.78)
: @n ; : @n ;
S S
This theorem is generally valid for all kinds of potentials; it is not necessary that they
full…l the Laplace equation.
In Green’s theorem, S ¤ is a closed surface with a volume V ¤ . This volume is bounded
by the wall of an imaginary vertical circular cylinder with a very large radius R, the sea
bottom at z = ¡h, the water surface at z = ³ and the wetted surface of the ‡oating body,
S; see …gure 2.6.
Both of the above radiation potentials Áj and Ák must ful…ll the Laplace equation (r2 Áj =
r2 Ák = 0). So the left hand side of the above equation becomes zero which yields for the
right hand side of this equation:
Z Z Z Z
@Ák ¤
@Áj
Áj ¢dS = Ák ¢dS ¤ (2.80)
@n @n
S¤ S¤
@Á
k tanh kh ¢ Á = for: z = 0 (2.82)
@z
This implies that at the free surface of the ‡uid one can write:
When taking also the boundary condition at the sea bed and the radiation condition on
the wall of the cylinder in …gure 2.6:
@Á
= 0 (for: z = ¡h) and lim Á = 0 (2.83)
@n R!1
into account, the integral equation over the surface S ¤ reduces to:
ZZ ZZ
@Á @Áj
Áj k ¢dS = Ák ¢dS (2.84)
@n @n
S S
Potential Coe¢cients
The previous section provides - for the zero forward ship speed case - symmetry in the
coe¢cients matrices with respect to their diagonals so that:
Because of the symmetry of a ship some coe¢cients are zero and the two matrices with
hydrodynamic coe¢cients for a ship become:
0 1
M11 0 M13 0 M15 0
B 0 M22 0 M24 0 M26 C
B C
B M31 0 M33 0 M35 0 C
Hydrodynamic mass matrix: B C (2.86)
B 0 M42 0 M44 0 M46 C
B C
@ M51 0 M53 0 M55 0 A
0 M62 0 M64 0 M66
0 1
N11 0 N13 0 N15 0
B 0 N22 0 N24 0 N26 C
B C
B N31 0 N33 0 N35 0 C
Hydrodynamic damping matrix: B C (2.87)
B 0 N42 0 N44 0 N46 C
B C
@ N51 0 N53 0 N55 0 A
0 N62 0 N64 0 N66
For clarity, the symmetry of terms about the diagonal in these matrices (for example
that M13 = M31 for zero forward speed) has not been included here. The terms on the
diagonals (such as Mnn for example) are the primary coe¢cients relating properties such
as hydrodynamic mass in one direction to the inertia forces in that same direction. O¤-
diagonal terms (such as M13 ) represent hydrodynamic mass only which is associated with
an inertia dependent force in one direction caused by a motion component in another.
Forward speed has an e¤ect on the velocity potentials itself, but is not discussed here. This
e¤ect is quite completely explained by [Timman and Newman, 1962].
and: Z Z µ ¶
~d = ½
~w + M @©w @©d
M + (~r £ ~n)dS (2.89)
@t @t
S
The principles of linear superposition allow the determination of these forces on a restrained
body with zero forward speed; @©=@n = 0. This simpli…es the boundary condition on the
surface of the body to:
@© @©w @©d
= + = 0 (2.90)
@n @n @n
The space and time dependent potentials, ©w (x; y; z; t) and ©d (x; y; z; t), are written now
in terms of isolated space dependent potentials, Áw (x; y; z) and Ád (x; y; z), multiplied by a
normalized oscillatory velocity, v(t) = 1 ¢ e¡i!t :
This elimination of the di¤raction potential results into the so-called Haskind relations:
Z Z µ ¶
¡i!t @Ák @Áw
Xwk = ¡i½e Áw + Ák dS for: k = 1; :::6 (2.96)
@n @n
S
This limiters the problem of the di¤raction potential because the expression for Xwk de-
pends only on the wave potential Áw and the radiation potential Ák .
These relations, found by [Haskind, 1957], are very important; they underlie the relative
motion (displacement - velocity - acceleration) hypothesis, as used in strip theory. These
relations are valid only for a ‡oating body with a zero time-averaged speed in all directions.
[Newman, 1962] however, has generalized the Haskind relations for a body with a constant
forward speed. He derived equations which di¤er only slightly from those found by Haskind.
According to Newman’s approach the wave potential has to be de…ned in the moving
O(x; y; z) system. The radiation potential has to be determined for the constant forward
speed case, taking an opposite sign into account.
The corresponding wave potential for deep water, as given in the previous section, now
becomes:
¡³ a g kz
©w = ¢ e ¢ sin(!t ¡ kx cos ¹ ¡ ky sin ¹)
!
¡i³ a g kz ik(x cos ¹+y sin ¹) ¡i!t
= ¢e ¢e e (2.97)
!
so that the isolated space dependent term is given by:
¡i³ a g kz ik(x cos ¹+y sin ¹)
Áw = ¢e ¢e (2.98)
!
24 CHAPTER 2. STRIP THEORY METHOD
The …rst term in this expression for the wave forces and moments is the so-called Froude-
Krilov force or moment, which is the wave load caused by the undisturbed incident wave.
The second term is caused by the disturbance caused by the presence of the (…xed) body.
or more generally:
ZZ
Xsk = ½g zfk ¢ dS for: k = 1; :::6 (2.102)
S
in which the Xsk are the components of these hydrostatic forces and moments.
The moments of inertia here are often expressed in terms of the radii of inertia and the
solid mass of the structure. Since Archimedes law is valid for a ‡oating structure:
Ixx = kxx2 ¢ ½r
Iyy = kyy 2 ¢ ½r
Izz = kzz 2 ¢ ½r (2.104)
When the distribution of the solid mass of a ship is unknown, the radii of inertia can be
approximated by: 8
< kxx t 0:30 ¢ B to 0:40 ¢ B
for ships: kyy t 0:22 ¢ L to 0:28 ¢ L (2.105)
:
kzz t 0:22 ¢ L to 0:28 ¢ L
in which L is the length and B is the breadth of the ship. Often, the (generally small)
coupling terms, Ixz = Izx , are simply neglected.
Bureau Veritas proposes for the radius of inertia for roll of the ship’s solid mass:
à µ ¶2 !
2 ¢ KG
kxx = 0:289 ¢ B ¢ 1:0 + (2.106)
B
d ³ ~´ ~ = d H
³ ´
F~ = mU and M ~ (2.107)
dt dt
in which:
Two important assumptions are made for the loads in the right hand side of these equations:
a. The so-called hydromechanical forces and moments are induced by the harmonic
oscillations of the rigid body, moving in the undisturbed surface of the ‡uid.
26 CHAPTER 2. STRIP THEORY METHOD
b. The so-called wave exciting forces and moments are produced by waves coming
in on the restrained body.
Since the system is linear, these loads are added to obtain the total loads. Thus, after
assuming small motions, symmetry of the body and that the x-, y- and z-axes are principal
axes, one can write the following six motion equations for the ship:
d
Surge: (½r ¢ x)
_ = ½r ¢ xÄ = Xh1 + Xw1
dt
d
Sway: (½r ¢ y)
_ = ½r ¢ yÄ = Xh2 + Xw2
dt
d
Heave: (½r ¢ z)
_ = ½r ¢ zÄ = Xh3 + Xw3
dt
d ³ ´
Roll: Ixx ¢ Á_ ¡ Ixz ¢ Ã_ = Ixx ¢ Á Ä ¡ Ixz ¢ Ã
Ä = Xh4 + Xw4
dt
d ³ _
´
Pitch: Iyy ¢ µ = Ixx ¢ ĵ = Xh5 + Xw5
dt
d ³ _ _
´
Ä ¡ Izx ¢ Á
Ä
Yaw: Izz ¢ Ã ¡ Izx ¢ Á = Izz ¢ Ã = Xh6 + Xw6
dt
(2.108)
in which:
½ = density of water
r = volume of displacement of the ship
Iij = solid mass moment of inertia of the ship
Xh1 , Xh2 , Xh3 = hydromechanical forces in the x-, y- and z-directions respectively
Xh4 , Xh5 , Xh6 = hydromechanical moments about the x-, y- and z-axes respectively
Xw1 , Xw2 , Xw3 = exciting wave forces in the x-, y- and z-directions respectively
Xw4 , Xw5 , Xw6 = exciting wave moments about the x-, y- and z-axes respectively
Generally, a ship has a vertical-longitudinal plane of symmetry, so that its motions can be
split into symmetric and anti-symmetric components. Surge, heave and pitch motions are
symmetric motions, that is to say that a point to starboard has the same motion as the
mirrored point to port side. It is obvious that the remaining motions sway, roll and yaw
are anti-symmetric motions. Symmetric and anti-symmetric motions are not coupled; they
don’t have any e¤ect on each other. For instance, a vertical force acting at the center of
gravity can cause surge, heave and pitch motions, but will not result in sway, roll or yaw
motions.
Because of this symmetry and anti-symmetry, two sets of three coupled equations of motion
can be distinguished for ships:
9
Surge : ½r ¢ xÄ ¡Xh1 = Xw1 =
Heave : ½r ¢ zÄ ¡Xh3 = Xw3 symmetric motions (2.109)
Ä ;
Pitch : Ixx ¢ µ ¡Xh5 = Xw5
9
Sway : ½r ¢ yÄ ¡Ixz ¢ ÃÄ ¡Xh = Xw =
2 2
Ä
Roll : Ixx ¢ Á ¡Xh4 = Xw4 anti-symmetric motions (2.110)
Ä Ä ;
Yaw : Izz ¢ Ã ¡Izx ¢ Á ¡Xh = Xw 6 6
2.5. STRIP THEORY APPROACHES 27
Note that this distinction between symmetric and anti-symmetric motions disappears when
the ship is anchored. Then, for instance, the pitch motions can generate roll motions via
the anchor lines.
The coupled surge, heave and pitch equations of motion are:
9
(½r + a11 ) ¢ xÄ +b11 ¢ x_ +c11 ¢ x >
>
>
+a13 ¢ zÄ +b13 ¢ z_ +c13 ¢ z >
>
>
+a15 ¢ ĵ +b15 ¢ µ_ +c15 ¢ µ = Xw1 (surge) >
>
>
>
>
>
>
>
a31 ¢ xÄ +b31 ¢ x_ +c31 ¢ x >
>
=
symmetric
+(½r + a33 ) ¢ zÄ +b33 ¢ z_ +c33 ¢ z (2.111)
> motions
+a35 ¢ ĵ +b35 ¢ µ_ +c35 ¢ µ = Xw3 (heave) > >
>
>
>
>
>
>
>
a51 ¢ xÄ +b51 ¢ x_ +c51 ¢ x >
>
>
>
+a53 ¢ zÄ +b53 ¢ z_ +c53 ¢ z >
>
;
+(+Iyy + a55 ) ¢ µÄ +b55 ¢ µ_ +c55 ¢ µ = Xw5 (pitch)
The coupled sway, roll and yaw equations of motion are:
9
(½r + a22 ) ¢ yÄ +b22 ¢ y_ +c22 ¢ y >
>
Ä +b24 ¢ Á_ +c24 ¢ Á >
>
+a24 ¢ Á >
>
>
+a26 ¢ Ã +b26 ¢ Ã +c26 ¢ Ã = Xw2 (sway) >
Ä _ >
>
>
>
>
>
>
a42 ¢ yÄ +b42 ¢ y_ +c42 ¢ y >
>
=
Ä _ anti-symmetric
+(+Ixx + a44 ) ¢ Á +b44 ¢ Á +c44 ¢ Á (2.112)
> motions
Ä +b46 ¢ Ã_ +c46 ¢ Ã = Xw4
+(¡Ixz + a46 ) ¢ Ã (roll) >
>
>
>
>
>
>
>
>
a62 ¢ yÄ +b62 ¢ y_ +c62 ¢ y >
>
Ä _ >
>
+(¡Izx + a64 ) ¢ Á +b64 ¢ Á +c64 ¢ Á >
>
>
;
Ä _
+(+Izz + a66 ) ¢ Ã +b66 ¢ Ã +c66 ¢ Ã = Xw6 (yaw)
In many applications, Ixz = Izx is not known or small; hence their terms are often omitted.
After the determination of the in and out of phase terms of the hydromechanical and the
wave loads, these equations can be solved with a numerical method.
strip theory supposes that the fore and aft side of the body (such as a pontoon) does not
produce waves in the x direction. For the zero forward speed case, interactions between
the cross sections are ignored as well.
Fundamentally, strip theory is valid for long and slender bodies only. In spite of this
restriction, experiments have shown that strip theory can be applied successfully for ‡oating
bodies with a length to breadth ratio larger than three, (L=B = 3), at least from a practical
point of view.
Z Z
surge: Xh1 = Xh0 1 ¢ dxb Xw1 = Xw0 1 ¢ dxb
L L
Z Z
sway: Xh2 = Xh0 2 ¢ dxb Xw2 = Xw0 2 ¢ dxb
L L
Z Z
heave: Xh3 = Xh0 3 ¢ dxb Xw3 = Xw0 3 ¢ dxb
L L
Z Z
roll: Xh4 = Xh0 4 ¢ dxb Xw4 = Xw0 4 ¢ dxb
L L
Z Z
pitch: Xh5 = ¡ Xh0 3 ¢ xb ¢ dxb Xw5 = ¡ Xw0 3 ¢ xb ¢ dxb
L L
2.5. STRIP THEORY APPROACHES 29
Z Z
yaw: Xh6 = Xh0 2 ¢ xb ¢ dxb Xw6 = Xw0 2 ¢ xb ¢ dxb
L L
(2.113)
in which:
The appearance of two-dimensional surge forces seems strange here. It is strange! A more
or less empirical method is used in SEAWAY for the surge motion, by de…ning an equivalent
longitudinal cross section which is swaying. Then, the 2-D hydrodynamic sway coe¢cients
of this equivalent cross section are translated to 2-D hydrodynamic surge coe¢cients by
an empirical method based on theoretical results from three-dimensional calculations and
these coe¢cients are used to determine 2-D loads. In this way, all sets of six surge loads can
be treated in the same numerical way in SEAWAY for the determination of the 3-D loads.
Inaccuracies of the hydromechanical coe¢cients of (slender) ships are of minor importance,
because these coe¢cients are relatively small.
Note how in the strip theory the pitch and yaw moments are derived from the 2-D heave
and sway forces, respectively, while the roll moments are obtained directly.
The equations of motions are de…ned in the moving axis system with the origin at the
average position of the center of gravity, G. All two-dimensional potential coe¢cients
have been de…ned so far in an axis system with the origin, O, in the water plane; the
hydromechanical and exciting wave moments have to be corrected for the distance OG.
The e¤ect of this operator can be understood easily when one realizes that in that earth-
bound coordinate system the sailing ship penetrates through a ”virtual vertical disk”.
When a ship sails with speed V and a trim angle, µ, through still water, the relative
vertical velocity of a water particle with respect to the bottom of the sailing ship becomes
V ¢ µ.
Two di¤erent types of strip theory methods are discussed here:
D n 0 ¤
o ¤
Xh¤j = Mjj ¢ ³_ hj + Njj
0
¢ ³_ hj + Xrs
0
Dt j
D n 0 ¤
o ¤
Xw¤ j = Mjj ¢ ³_ wj + Njj
0
¢ ³_ wj + Xf0 kj (2.117)
Dt
This in the …rst formulation of the strip theory that can be found in the literature. It
contains a more or less intuitive approach to the forward speed problem, as published
in detail by [Korvin-Kroukovsky and Jacobs, 1957] and others.
0 0
In the equations above, Mjj and Njj are the 2-D potential mass and damping coe¢cients.
0
Xrsj is the two-dimensional quasi-static restoring spring term, generally present for heave,
roll and pitch only. Xf0 kj is the two-dimensional Froude-Krilov force or moment which is
calculated by an integration of the directional pressure gradient in the undisturbed wave
over the cross sectional area of the hull.
Equivalent directional components of the orbital acceleration and velocity, derived from
these Froude-Krilov loads, are used to calculate the di¤raction parts of the total wave
forces and moments.
From a theoretical point of view, one should prefer the use of the modi…ed method, but it
appeared from user’s experience that for ships with moderate forward speed (F n 5 0:30),
the ordinary method provides a better …t with experimental data. Thus from a practical
point of view, the use of the ordinary method is advised generally.
2.5.3 End-Terms
From the previous, it is obvious that in the equations of motion longitudinal derivatives of
the two-dimensional potential mass Mij0 and damping Nij0 will appear. These derivatives
2.5. STRIP THEORY APPROACHES 31
have to be determined numerically over the whole ship length in such a manner that the
following relation is ful…lled:
xbZ
(L)+" xZb (0) xZb (L) xbZ
(L)+"
df (xb ) df (xb ) df(xb ) df (xb )
dxb = dxb + dxb + dxb
dxb dxb dxb dxb
xb (0)¡" xb (0)¡" xb (0) xb (L)
xZb (L)
df(xb )
= f (0) + dxb ¡ f (L)
dxb
xb (0)
= 0 (2.119)
0 0
with " ¿ L, while f (xb ) is equal to the local values of Mjj (xb ) or Njj (xb ); see …gure 2.8.
The numerical integrations of the derivatives are carried out in the region xb (0) · xb ·
xb (L) only. So, the additional so-called ”end terms” are de…ned by f(0) and f (L).
Because the integration of the derivatives has to be carried out in the region xb (0) ¡ " ·
xb · xb (L) + ", some algebra provides the integral and the …rst and second order moments
(with respect to G) over the whole ship length:
xbZ
(L)+"
df (xb )
¢ dxb = 0
dxb
xb (0)¡"
xbZ
(L)+" xZb (L)
df (xb )
¢ xb ¢ dxb = ¡ f(xb ) ¢ dxb
dxb
xb (0)¡" xb (0)
32 CHAPTER 2. STRIP THEORY METHOD
xbZ
(L)+" xZb (L)
df (xb ) 2
¢ xb ¢ dxb = ¡2 f(xb ) ¢ xb ¢ dxb (2.120)
dxb
xb (0)¡" xb (0)
Note that these expressions are valid for the integrations of the potential coe¢cients over
the full ship length only. They can not be used for calculating local hydromechanical loads.
Also for the wave loads, these expressions can not be used, because there these derivatives
are multiplied with xb -depending orbital motions.
point source of unit strength at the midpoint of each line element. Wave systems
were de…ned in such a way that all required boundary conditions were ful…lled. The
linearized Bernoulli equation provides the pressures after which the potential coe¢-
cients were obtained from the in-phase and out-of-phase components of the resultant
hydrodynamic loads.
The 2-D potential pitch and yaw (moment) coe¢cients follow from the previous heave and
sway coe¢cients and the arm, i.e., the distance of the cross section to the center of gravity
G.
A more or less empiric procedure has followed by the author for the surge motion. An equiv-
alent longitudinal cross section has been de…ned. For each frequency, the two-dimensional
potential hydrodynamic sway coe¢cient of this equivalent cross section is translated to
two-dimensional potential hydrodynamic surge coe¢cients by an empiric method, which is
based on theoretical results of three-dimensional calculations.
The 3-D coe¢cients follow from an integration of these 2-D coe¢cients over the ship’s
length. Viscous terms can be added for surge and roll.
34 CHAPTER 2. STRIP THEORY METHOD
.
Chapter 3
Conformal Mapping
Ursell’s derivation of potential coe¢cients is valid for circular cross sections only. For the
determination of the two-dimensional added mass and damping in the sway, heave and
roll mode of the motions of ship-like cross sections by Ursell’s method, the cross sections
have to be mapped conformally to the unit circle. The advantage of conformal mapping is
that the velocity potential of the ‡uid around an arbitrarily shape of a cross section in a
complex plane can be derived from the more convenient circular section in another complex
plane. In this manner hydrodynamic problems can be solved directly with the coe¢cients
of the mapping function.
The general transformation formula is given by:
X
N
¡ ¢
z = Ms a¡(2n¡1 ³ 2n¡1
n=0
with:
From this follows the relation between the coordinates in the z-plane (the ship’s cross
section) and the variables in the ³-plane (the circular cross section):
N
X © ª
x = ¡Ms (¡1)n a2n¡1 e¡(2n¡1)® sin ((2n ¡ 1)µ)
n=0
XN
© ª
y = +Ms (¡1)n a2n¡1 e¡(2n¡1)® cos ((2n ¡ 1)µ)
n=0
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
35
36 CHAPTER 3. CONFORMAL MAPPING
The contour of the - by conformal mapping approximated - ship’s cross section follows
from putting ® = 0 in the previous relations:
X
N
x0 = ¡Ms f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g
n=0
N
X
y0 = +Ms f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g
n=0
The breadth on the waterline of the approximate ship’s cross section is de…ned by:
N
X
b0 = 2Ms ¸a with: ¸a = a2n¡1
n=0
In here a¡1 =+1 and the conformal mapping coe¢cients a1 and a3 are called Lewis coe¢-
cients, while Ms is the scale factor.
Then:
¡ ¢
x = Ms ¢ e® sin µ + a1 e¡® sin µ ¡ a3 e¡3® sin 3µ
¡ ¢
y = Ms ¢ e® cos µ ¡ a1 e¡® cos µ + a3 e¡3® cos 3µ
Now the coe¢cients a1 and a3 and the scale factor Ms will be determined in such a manner
that the sectional breadth, draft and area of the approximate cross section and of the
actual cross section are identical.
The half breadth to draft ratio H0 is given by:
Bs =2 1 + a1 + a3
H0 = =
Ds 1 ¡ a1 + a3
An integration of the Lewis form delivers the sectional area coe¢cient ¾ s :
As ¼ 1 ¡ a21 ¡ 3a23
¾s = = ¢
Bs Ds 4 (1 + a3 )2 ¡ a21
These ranges of the half breadth to draft ratio H0 and the area coe¢cient ¾ s for the
di¤erent typical Lewis forms are shown in …gure 3.3.
bounded by a lower limit to omit re-entrant Lewis forms and by an upper limit to omit
non-symmetric Lewis forms:
µ ¶
3¼ ¼ 1
for H0 · 1:0 : (2 ¡ H0 ) < ¾s < 10 + H0 +
32 32 H0
µ ¶ µ ¶
3¼ 1 ¼ 1
for H0 ¸ 1:0 : 2¡ < ¾s < 10 + H0 +
32 H0 32 H0
If a value of ¾ s is outside of this range it has to be set to the value of the nearest border
of this range, to calculate the Lewis coe¢cients.
Numerical problems, for instance with bulbous or aft cross sections of a ship, are avoided
when the following requirements are ful…lled:
Bs Bs
> °Ds and Ds > °
2 2
with for instance ° = 0:01.
Z = Ms (a¡1 ³ + a1 ³ ¡1 + a3 ³ ¡3 + a5 ³ ¡5 )
Now the coe¢cients a1 , a3 and a5 and the scale factor Ms can be determined in such
a manner that, except the sectional breadth, draft and area, also the centroids of the
approximate cross section and the actual cross section of the ship have an equal position.
The half beam to draft ratio is given by:
Bs =2 1 + a1 + a3 + a5
H0 = =
Ds 1 ¡ a1 + a3 ¡ a5
An integration of the approximate form results into the sectional area coe¢cient:
For the relative distance of the centroid to the keel point a more complex expression has
been obtained by [Athanassoulis and Loukakis, 1985]:
P
3 P
3 P
3
(Aijk a2i¡1 a2j¡1 a2k¡1 )
KB i=0 j=0 k=0
· = = 1¡
Ds P
3
H0 ¾ s a32i¡1
i=0
in which:
µ ¶
1 1 ¡ 2k 1 ¡ 2k 1 ¡ 2k 1 ¡ 2k
Aijk = ¡ + +
4 3 ¡ 2(i + j + k) 1 ¡ 2(i ¡ j + k) 1 ¡ 2(i + j ¡ k) 1 ¡ 2(¡i + j + k)
The following requirements should be ful…lled when also bulbous cross sections are allowed:
² re-entrant forms are avoided when both the following requirements are ful…lled:
Taking these restrictions into account, the equations above can be solved in an iterative
manner.
The scale factor Ms and the conformal mapping coe¢cients a2n¡1 , with a maximum value
of n varying from N = 2 until N = 10, have been determined successfully from the o¤sets
of various cross sections in such a manner that the mean squares of the deviations of the
actual cross section from the approximate described cross section is minimized.
The general transformation formula is given by:
N n
X o
Z = Ms a2n¡1 ³ ¡(2n¡1)
n=0
N
X
x0 = ¡Ms f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g
n=0
N
X
y0 = +Ms f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g
n=0
Bs =2 Ds
Ms = or: Ms =
P
N P
N
fa2n¡1 g f(¡1)n a2n¡1 g
n=0 n=0
The procedure starts with initial values for [Ms ¢ a2n¡1 ]. The initial values of Ms , a1 and a3
are obtained with the Lewis method as has been described before, while the initial values of
a5 until a2N¡1 are set to zero. With these [Ms ¢a2n¡1 ] values, a µi -value is determined for each
o¤set in such a manner that the actual o¤set (xi ; yi ) lies on the normal of the approximate
contour of the cross section in (x0i ; y0i ). Now µi has to be determined. Therefore a function
F (µi ), will be de…ned by the distance from the o¤set (xi ; yi ) to the normal of the contour
to the actual cross section through (x0i ; y0i ), see …gure 3.4.
Then new values of [Ms ¢ a2n¡1 ] have to be obtained in such a manner that E is minimized.
This means that each of the derivatives of this equation with respect to each coe¢cients
[Ms ¢ a2n¡1 ] is zero, so:
@E
=0 for: j = 0; :::N
@fMs a2j¡1 g
This yields N+1 equations:
(
XI N
X
¡ sin ((2j ¡ 1)µi ) f(¡1)n[Ms ¢ a2n¡1 ] sin ((2n ¡ 1)µi )g +
i=0 n=0
)
N
X
¡ cos ((2j ¡ 1)µi ) f(¡1)n [Ms ¢ a2n¡1 ] cos ((2n ¡ 1)µi )g =
n=0
I
X
= fxi sin ((2j ¡ 1)µi ) ¡ yi cos ((2j ¡ 1)µi )g for: j = 0; :::N
i=0
To obtain the exact breadth and draft, the last two equations are replaced by the equations
for the breadth at the water line and the draft:
(N )
X I
X
f(¡1)n [Ms ¢ a2n¡1 ]g fcos ((2j ¡ 2n)µi )g =
n=0 i=0
I
X
= f¡xi sin ((2j ¡ 1)µi ) + yi cos ((2j ¡ 1)µi )g for: j = 0; :::N ¡ 2
i=0
N
X
f[Ms ¢ a2n¡1 ]g = Bs =2 for: j = N ¡ 1
n=0
N
X
f(¡1)n [Ms ¢ a2n¡1 ]g = Ds for: j = N
n=0
These N + 1 equations can be solved numerically, so that new values for [Ms ¢ a2n¡1 ] will be
obtained. These new values are used instead of the initial values to obtain new µi -values of
the I + 1 o¤sets again, etc. This procedure will be repeated several times and stops when
the di¤erence between the numerical E-values of two subsequent calculations becomes less
than a certain threshold value ¢E, depending on the dimensions of the cross section; for
instance: ³ ´2
p
¢E = (I + 1) 0:00005 b2max + d2max
in which:
bmax = maximum half breadth of the cross section
dmax = maximum height of the cross section
Because a¡1 = +1 the scale factor Ms is equal to the …nal solution of the …rst coe¢cient
(n = 0). The N other coe¢cients a2n¡1 can be found by dividing the …nal solutions of
[Ms ¢ a2n¡1 ] by this Ms -value.
Reference is also given here to a report of [Jong, 1973]. In that report several other,
suitable but more complex, methods are described to determine the scale factor Ms and
the conformal mapping coe¢cients a2n¡1 from the o¤sets of a cross section.
Attention has to be paid to divergence in the calculation routines and re-entrant forms.
In these cases the number N will be increased until the divergence or re-entrance vanish.
In the most worse case a ”maximum” value of N will be attained without success.. One
can then switch to Lewis coe¢cients with an area coe¢cient of the cross section set to the
nearest border of the valid Lewis form area.
3.4 Comparisons
A …rst example has been given here for the amidships cross section of a container vessel,
with a breadth of 25.40 meter and a draft of 9.00 meter, with o¤sets as tabled below:
i Ds ¡ y i xi
(¡) (m) (m)
0 0 .0 0 0 0 .0 0 0
1 0 .1 3 5 4 .9 5 0
2 0 .2 7 0 9 .9 0 0
3 0 .5 0 0 1 0 .9 6 0
4 1 .0 0 0 1 1 .7 4 0
5 2 .0 0 0 1 2 .4 4 0
6 3 .0 5 0 1 2 .7 0 0
7 6 .0 0 0 1 2 .7 0 0
8 9 .0 0 0 1 2 .7 0 0
3.4. COMPARISONS 45
For the least squares method in the conformal mapping method, 33 new o¤sets at equidis-
tant length intervals on the contour of this cross section can be determined by a second
degree interpolation routine. The calculated data of the two-parameter Lewis and the N-
parameter Close-Fit conformal mapping of this amidships cross section are tabled below.
The last line lists the RMS-values for the deviations of the 33 equidistant points on the
approximate contour of this cross section..
L ew is
C o n fo rm al N -Pa ra m e ter C lo se F it C o n fo rm a l M a p p in g
M a p p in g
N (¡) 2 2 3 4 5 6 7 8 9 10
2N -1 (¡) 3 3 5 7 9 11 13 15 17 19
Ms (m) 1 2.24 00 12 .2 45 7 12 .2 84 1 1 2 .3 1 9 3 1 2 .3 1 8 6 1 2 .3 1 8 3 1 2 .3 1 9 1 1 2 .3 1 9 0 1 2 .3 1 9 5 1 2 .3 1 9 4
Another example is given in …gure 3.5, which shows the di¤erences between a Lewis trans-
formation and a 10-parameter close-…t conformal mapping of a rectangular cross section
with a breadth of 20.00 meter and a draft of 10.00 meter.
.
Chapter 4
2-D Potential Coe¢cients
This chapter described the various methods, used in the SEAWAY program, to obtain the
2-D potential coe¢cients:
² the theory of Tasai for deep water, based on Ursell’s potential theory for circular
cylinders and N-parameter conformal mapping
² the theory of Keil for shallow and deep water, based on a variation of Ursell’s potential
theory for circular cylinders and Lewis conformal mapping
² the theory of Frank for deep water, using pulsating sources on the cross sectional
contour.
During the ship motions calculations di¤erent coordinate systems, as shown in …gure 2.2,
will be used. The two-dimensional hydrodynamic potential coe¢cients have been de…ned
here with respect to the O(x; y; z) coordinate system for the moving ship in still water.
However, in this section deviating axes systems are used for the determination of the two-
dimensional hydrodynamic potential coe¢cients for sway, heave and roll motions. This
holds for the sway and roll coupling coe¢cients a change of sign. The signs of the uncoupled
sway, heave and roll coe¢cients do not change.
For each cross section, the following two-dimensional hydrodynamic coe¢cients have to be
obtained:
0 0
M22 and N22 = 2-D potential mass and damping coe¢cients of sway
0 0
M24 and N24 = 2-D potential mass and damping coupling coe¢cients of roll into sway
0 0
M33 and N33 = 2-D potential mass and damping coe¢cients of heave
0 0
M44 and N44 = 2-D potential mass and damping coe¢cients of roll
0 0
M42 and N42 = 2-D potential mass and damping coupling coe¢cients of sway into roll
The 2-D potential pitch and yaw (moment) coe¢cients follow from the previous heave and
sway coe¢cients and the arm, i.e., the distance of the cross section to the center of gravity
G.
0
Finally, an approximation is given for the determination of the surge coe¢cients M11 and
0
N11 .
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
47
48 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
The cylinder is forced to carry out a simple harmonic vertical motion about its initial
position with a frequency of oscillation ! and a small amplitude of displacement ya :
y = ya cos(!t + ±)
y_ = ¡!ya sin(!t + ±)
yÄ = ¡! 2 ya sin(!t + ±)
This forced vertical oscillation of the cylinder causes a surface disturbance of the ‡uid.
4.1. THEORY OF TASAI 49
Because the cylinder is supposed to be in…nitely long, the generated waves will be two-
dimensional. These waves travel away from the cylinder and a stationary state is rapidly
attained.
Two kinds of waves will be produced:
² A standing wave system, denoted here by subscript A. The amplitudes of these waves
decrease strongly with the distance to the cylinder.
² A regular progressive wave system, denoted here by subscript B. These waves dis-
sipate energy. At a distance of a few wave lengths from the cylinder, the waves on
each side can be described by a single regular wave train.. The wave amplitude at
in…nity ´ a is proportional to the amplitude of oscillation of the cylinder ya , provided
that this amplitude is su¢ciently small compared with the radius of the cylinder and
the wave length is not much smaller than the diameter of the cylinder.
The two-dimensional velocity potential of the ‡uid has to ful…l the following requirements:
@2© @2©
r2 © = + 2 =0
@x2 @y
2. Because the heave motion of the ‡uid is symmetrical about the y-axis, this velocity
potential has the following relation:
©(¡x; y) = ©(+x; y)
!2 @© Bs
©+ =0 for: jxj ¸ and y = 0
g @y 2
In consequence of the conformal mapping, this free surface condition can be written as:
»b X ©
N
ª @© ¼
© (2n ¡ 1)a2n¡1 e¡(2n¡1)® § = 0 for: ® ¸ 0 and µ = §
¾ a n=0 @µ 2
in which:
»b !2 ! 2 b0
= Ms or: » b = (non-dimensional frequency squared)
¾a g 2g
From the de…nition of the velocity potential follows the boundary condition on the surface
of the cylinder for ® = 0:
@©0 (µ) @y0
= y_
@n @n
in which n is the outward normal of the cylinder surface.
50 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Using the stream function ª, this boundary condition on the surface of the cylinder (®=0)
reduces to:
¡@ª0 (µ) @y0
= y_
@µ @®
N
X
= ¡yM
_ s f(¡1)n (2n ¡ 1)a2n¡1 cos ((2n ¡ 1)µ)g
n=0
Integration results into the following requirement for the stream function on the surface of
the cylinder:
N
X
ª0 (µ) = yM_ s f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g
n=0
1 X
N
2x0
h(µ) = = ¡ f(¡1)n a2n¡1 sin ((2n ¡ 1)µ)g
b0 ¾ a n=0
C(t) = 0
So:
b0
ª0 (µ) = ¡y_ h(µ)
2
For the standing wave system a velocity potential and a stream function satisfying to the
equation of Laplace, the symmetrical motion of the ‡uid and the free surface condition has
to be found.
The following set of velocity potentials, as they are given by [Tasai, 1959], [Tasai, 1960]
and [Jong, 1973], ful…l these requirements:
à 1 !
g´ a X © ª X 1
© ª
©a = P2m ÁA2m (®; µ) cos(!t) + Q2m ÁA2m (®; µ) sin(!t)
¼! m=1 m=1
in which:
in which:
while:
!2
º= (wave number for deep water)
g
Changing the parameters delivers:
g´a ¡ ¢
©B = ÁBc (®; µ) cos(!t) + ÁBs (®; µ) sin(!t)
¼!
The conjugate stream function is given by:
g´ a ¡ ¢
ªB = Ã Bc (x; y) cos(!t) + Ã Bs (x; y) sin(!t)
¼!
in which:
à Bc = ¼e¡ºy sin(ºjxj)
Z1
¡ºy º cos(ky) + k sin(ky) ¡kjxj
à Bs = ¡¼e cos(ºx) + e dk
k2 + º 2
0
Power series expansions, as given by [Porter, 1960], can be used instead of these last inte-
grals over k. The summation in these expansions converge much faster than the numeric
integration procedure. This has been shown before for the sway case.
The total velocity potential and stream function to describe the waves generated by a
heaving cylinder are:
© = ©A + ©B
ª = ªA + ªB
So the velocity potential and the conjugate stream function are expressed by:
ÃÃ !
g´ a X1
© ª
©(®; µ) = ÁBc (®; µ) + P2m ÁA2m (®; µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÁBs (®; µ) + Q2m ÁA2m (®; µ) sin(!t)
m=1
ÃÃ !
g´ a
1
X © ª
ª(®; µ) = à Bc (®; µ) + P2m à A2m (®; µ) cos(!t)
¼! m=1
à ! !
X
1
© ª
+ à Bs (®; µ) + Q2m à A2m (®; µ) sin(!t)
m=1
When putting ® = 0, the stream function is equal to the expression found before from the
boundary condition on the surface of the cylinder:
ÃÃ !
g´ a X1
© ª
ª0 (µ) = à B0c (µ) + P2m à A02m (µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÃB0s (µ) + Q2m à A02m (µ) sin(!t)
m=1
b0
= ¡y_ h(µ)
2
in which:
In this expression à B0c (µ) and à B0s (µ) are the values of ÃBc (®; µ) and ÃBs (®; µ) at the
surface of the cylinder, so for ® = 0.
So for each µ, the following equation has been obtained:
à !
X1
© ª
à B0c (µ) + P2m à A02m (µ) cos(!t)
m=1
à !
X1
© ª ¼!b0
+ à B0s (µ) + Q2m à A02m (µ) sin(!t) = ¡y_ h(µ)
m=1
2g´ a
4.1. THEORY OF TASAI 53
in which:
ya ya
A0 = ¼» sin ± and B0 = ¼» cos ±
´a b ´a b
This results for each µ into a set of two equations:
1
X © ª
à B0c (µ) + P2m à A02m (µ) = h(µ)A0
m=1
X1
© ª
à B0s (µ) + Q2m à A02m µ) = h(µ)B0
m=1
When putting µ = ¼=2, so at the intersection of the surface of the cylinder with the free
surface of the ‡uid where h(µ)=1, we obtain the coe¢cients A0 and B0 :
1
X © ª
A0 = à B0c (¼=2) + P2m à A02m (¼=2)
m=1
X1
© ª
B0 = à B0s (¼=2) + Q2m à A02m (¼=2)
m=1
in which:
XN ½ ¾
»b m 2n ¡ 1
à A02m (¼=2) = (¡1) a2n¡1
¾a n=0
2m + 2n ¡ 1
A substitution of A0 and B0 into the set of two equations for each µ, results for any µ-value
less than ¼=2 into a set of two equations with the unknown parameters P2m and Q2m .
So:
X
1
à B0c (µ) ¡ h(µ)à B0c (¼=2) = ff2m (µ)P2m g
m=1
X1
à B0s (µ) ¡ h(µ)à B0s (¼=2) = ff2m (µ)Q2m g
m=1
in which:
f2m (µ) = ¡Ã A02m (µ) + h(µ)Ã A02m (¼=2)
The series in these two sets of equations converges uniformly with an increasing value of
m. For practical reasons the maximum value of m is limited to M.
Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series. For a lot
of µ-values, the best …t values of P2m and Q2m are supposed to be those found by means
of a least squares method. Note that at least M values of µ, less than ¼=2, are required to
solve these equations.
Another favorable method is to multiply both sides of the equations with ¢µ. Then the
summations over µ can be replaced by integrations.
54 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Herewith, two sets of M equations have been obtained, one set for P2m and one set for
Q2m :
8 9
XM < Z¼=2 = Z¼=2
¡ ¢
P2m f2m (µ)f2n (µ)dµ = Ã B0c (µ) ¡ h(µ)Ã B0c (¼=2) f2n (µ)dµ n = 1; :::M
: ;
m=1 0 0
8 9
M <
X Z¼=2 = Z¼=2
¡ ¢
Q2m f2m (µ)f2n (µ)dµ = Ã B0s (µ) ¡ h(µ)Ã B0s (¼=2) f2n (µ)dµ n = 1; :::M
: ;
m=1 0 0
Now the P2m and Q2m series can be solved by a numerical method.
With these P2m and Q2m values, the coe¢cients A0 and B0 are known too.
From the de…nition of these coe¢cients follows the amplitude ratio of the radiated waves
and the forced heave oscillation:
´a ¼» b
=p
ya A20 + B02
With the solved P2m and Q2m values, the velocity potential on the surface of the cylinder
(®=0) is known too:
ÃÃ !
g´ a XM
© ª
©0 (µ) = ÁB0c (µ) + P2m ÁA02m (µ) cos(!t)
¼! m=1
à ! !
XM
© ª
+ ÁB0s (µ) + Q2m ÁA02m (µ) sin(!t)
m=1
in which:
In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®; µ) and ÁBs (®; µ) at the
surface of the cylinder, so for ® = 0.
Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the
linearized equation of Bernoulli:
@©0 (µ)
p(µ) = ¡½
@tÃÃ !
¡½g´ a XM
© ª
= ÁB0s (µ) + Q2m ÁA02m (µ) cos(!t)
¼ m=1
à ! !
XM
© ª
¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t)
m=1
Heave Coe¢cients
The two-dimensional hydrodynamic vertical force, acting on the cylinder in the direction
of the y-axis, can be found by integrating the vertical component of the hydrodynamic
pressure on the surface of the cylinder:
Z
+¼=2
dx0
Fy0 = ¡ p(µ) ds
ds
¡¼=2
Z¼=2
dx0
= ¡2 p(µ) dµ
dµ
0
With this the two-dimensional hydrodynamic vertical force due to heave oscillations can
be written as follows:
½gb0 ´ a
Fy0 = (M0 cos(!t) ¡ N0 sin(!t))
¼
in which:
Z¼=2 N
X
1
M0 = ¡ ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 cos ((2n ¡ 1)µ)g dµ
¾a n=0
0
( N ½ ¾)
1 XM X (2n ¡ 1)2
¡ (¡1)m Q2m a2n¡1
¾a m=1 n=0
(2m)2 ¡ (2n ¡ 1)2
à ( )!
¼»
N
X N
X ¡m
+ 2b Q2 + (¡1)m Q2m f(2n ¡ 1)a2n¡1 a2m+2n¡1 g
4¾ a m=1 n=0
and N0 as obtained from this expression for M0 , by replacing there ÁB0s (µ) by ÁB0c (µ) and
Q2m by P2m .
For the determination of M0 and N0 it is advised: M = N .
These expressions coincide with those given by [Tasai, 1960].
With:
½gb0 ´a
Fy0 = (M0 cos(!t + ± ¡ ±) ¡ N0 sin(!t + ± ¡ ±))
¼
and:
´a ´a
sin ± = A0 cos ± = B0
ya ¼» b ya ¼» b
the two-dimensional hydrodynamic vertical force can be resolved into components in phase
and out phase with the vertical displacement of the cylinder:
½gb0 ´ 2a
Fy0 = ((M0 B0 + N0 A0 ) cos(!t + ±) + (M0 A0 ¡ N0 B0 ) sin(!t + ±))
¼ 2 » b ya
This hydrodynamic vertical force can also be written as:
Fy0 = ¡M33
0 0
yÄ ¡ N33 y_
0
= M33 0
! 2 ya cos(!t + ±) + N33 !ya sin(!t + ±)
in which:
56 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
0
M33 = 2-D hydrodynamic mass coe¢cient of heave
0
N33 = 2-D hydrodynamic damping coe¢cient of heave
When using also the amplitude ratio of the radiated waves and the forced heave oscillation,
found before, the two-dimensional hydrodynamic mass and damping coe¢cients of heave
are given by:
0 ½b20 M0 B0 + N0 A0 0 ½b20 M0 A0 ¡ N0 B0
M33 = ¢ and N33 = ¢ ¢!
2 A20 + B02 2 A20 + B02
The signs of these two coe¢cients are proper in both, the axes system of Tasai and the
ship motions O(x; y; z) coordinate system.
The energy delivered by the exciting forces should be equal to the energy radiated by the
waves.
So:
T osc
Z
1 0 ½g´ 2a c
N33 y_ ¢ ydt
_ =
Tosc 2
0
¼2
M0 A0 ¡ N0 B0 =
2
4.1. THEORY OF TASAI 57
The cylinder is forced to carry out a simple harmonic lateral motion about its initial
position with a frequency of oscillation ! and a small amplitude of displacement xa :
x = xa cos(!t + ")
This forced lateral oscillation of the cylinder causes a surface disturbance of the ‡uid.
Because the cylinder is supposed to be in…nitely long, the generated waves will be two-
dimensional. These waves travel away from the cylinder and a stationary state is rapidly
attained.
Two kinds of waves will be produced:
² A standing wave system, denoted here by subscript A. The amplitudes of these waves
decrease strongly with the distance to the cylinder.
58 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
² A regular progressive wave system, denoted here by subscript B. These waves dis-
sipate energy. At a distance of a few wave lengths from the cylinder, the waves on
each side can be described by a single regular wave train.. The wave amplitude at
in…nity ´ a is proportional to the amplitude of oscillation of the cylinder xa , provided
that this amplitude is su¢ciently small compared with the radius of the cylinder and
the wave length is not much smaller than the diameter of the cylinder.
The two-dimensional velocity potential of the ‡uid has to ful…l the following three require-
ments:
@2© @2©
r2 © = + 2 =0
@x2 @y
2. Because the sway motion of the ‡uid is not symmetrical about the y-axis, this velocity
potential has the following anti-symmetric relation:
©(¡x; y) = ¡©(+x; y)
!2 @© Bs
©+ =0 for: jxj ¸ and y = 0
g @y 2
»b X ©
N
ª @© ¼
© (2n ¡ 1)a2n¡1 e¡(2n¡1)® § = 0 for: ® ¸ 0 and µ = §
¾ a n=0 @µ 2
in which:
»b !2 ! 2 b0
= Ms or: » b = (non-demensional frequency squared)
¾a g 2g
From the de…nition of the velocity potential follows the boundary condition on the surface
S of the cylinder for ® = 0:
@©0 (µ) @x0
= x_
@n @n
in which n is the outward normal of the cylinder surface S.
Using the stream function ª this boundary condition on the surface of the cylinder (® = 0)
reduces to:
@ª0 (µ) @x0
= ¡x_
@µ @®
XN
= ¡xM
_ s f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g
n=0
4.1. THEORY OF TASAI 59
Integration results into the following requirement for the stream function on the surface of
the cylinder:
XN
ª0 (µ) = xM
_ s f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g + C(t)
n=0
1 X
N
2y0
g(µ) = = f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g
b0 ¾ a n=0
in which:
in which:
in…nity. For this, a function describing a two-dimensional horizontal doublet at the origin
O is chosen.
[Tasai, 1961] and [Jong, 1973] gave the velocity potential of the progressive wave system
as:
g´ ¡ ¢
©B = a ÁBc (x; y) cos(!t) + ÁBs (x; y) sin(!t)
¼!
in which:
while:
j = +1 for x > 0
j = ¡1 for x < 0
!2
º = (wave number for deep water)
g
Changing the parameters delivers:
g´ ¡ ¢
©B = a ÁBc (®; µ) cos(!t) + ÁBs (®; µ) sin(!t)
¼!
The conjugate stream function is given by:
g´ ¡ ¢
ªB = a à Bc (x; y) cos(!t) + à Bs (x; y) sin(!t)
¼!
in which:
à Bc = ¼e¡ºy cos(ºx)
Z1
º sin(ky) ¡ k cos(ky) ¡kjxj y
à Bs = ¼e¡ºy sin(ºjxj) + 2 2
e dk ¡
k +º º(x + y 2 )
2
0
in which:
³ p ´ X1
Q = ° + ln º x2 + y 2 + fpn cos(n¯)g
n=1
X
1
S = ¯+ fpn sin(n¯)g
n=1
µ ¶
x
¯ = arctan
y
³ p ´n
2
º x +y 2
pn =
n ¢ n!
° = 0:5772156649:::::: (Euler constant)
The summation in these expansions converge much faster than the numeric integration
procedure. A suitable maximum value of n should be chosen.
The total velocity potential and stream function to describe the waves generated by a
swaying cylinder are:
© = ©A + ©B
ª = ªA + ªB
So the velocity potential and the conjugate stream function are expressed by:
ÃÃ !
g´ a X1
© ª
©(®; µ) = ÁBc (®; µ) + P2m ÁA2m (®; µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÁBs (®; µ) + Q2m ÁA2m (®; µ) sin(!t)
m=1
ÃÃ !
g´ a
1
X © ª
ª(®; µ) = à Bc (®; µ) + P2m à A2m (®; µ) cos(!t)
¼! m=1
à ! !
X
1
© ª
+ à Bs (®; µ) + Q2m à A2m (®; µ) sin(!t)
m=1
When putting ® = 0, the stream function is equal to the expression found before from the
boundary condition on the surface of the cylinder:
ÃÃ !
g´ a X1
© ª
ª0 (µ) = à B0c (µ) + P2m à A02m (µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÃB0s (µ) + Q2m à A02m (µ) sin(!t)
m=1
b0
= x_ g(µ) + C(t)
2
in which:
à A02m (µ) = ¡ cos ((2m + 1)µ)
N ½ ¾
»b X n 2n ¡ 1
+ (¡1) a2n¡1 cos ((2m + 2n)µ)
¾ a n=0 2m + 2n
62 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
In this expression ªB0c (µ) and ªB0s (µ) are the values of ªBc (®; µ) and ªBs (®; µ) at the
surface of the cylinder, so for ® = 0.
So for each µ-value, the following equation has been obtained:
à !
1
X © ª
à B0c (µ) + P2m à A02m (µ) cos(!t)+
m=1
à !
X1
© ª ¼!b0
à B0s (µ) + Q2m à A02m (µ) sin(!t) = x_ g(µ) + C ¤ (t)
m=1
2g´ a
When putting µ = ¼=2, so at the intersection of the surface of the cylinder with the free
surface of the ‡uid where g(µ) = 0, we obtain the constant C ¤ (t):
à !
1
X © ª
C ¤ (t) = à B0c (¼=2) + P2m à A02m (¼=2) cos(!t) +
m=1
à !
X1
© ª
à B0s (¼=2) + Q2m à A02m (¼=2) sin(!t)
m=1
in which:
XN ½ ¾
»b m 2n ¡ 1
ªA02m (¼=2) = (¡1) a2n¡1
¾a n=0
2m + 2n
A substitution of C ¤ (t) in the equation for each µ-value, results for any µ-value less than
¼=2 into the following equation:
à !
1
X © ¡ ¢ª
à B0c (µ) ¡ à B0c (¼=2) + P2m à A02m (µ) ¡ à A02m (¼=2) cos(!t)+
m=1
à !
X1
© ¡ ¢ª ¼! b0
à B0s (µ) ¡ à B0s (¼=2) + Q2m à A02m (µ) ¡ à A02m (¼=2) sin(!t) = x_ g(µ)
m=1
g´a 2
in which:
xa xa
P0 = ¡ ¼» sin " and Q0 = ¡ ¼» cos "
´a b ´a b
This delivers for any µ-value less than ¼=2 two sets of equations with the unknown para-
meters P2m and Q2m .
So:
1
X
à B0c (µ) ¡ à B0c (¼=2) = g(µ)P0 + ff2m (µ)P2m g
m=1
X1
à B0s (µ) ¡ à B0s (¼=2) = g(µ)Q0 + ff2m (µ)Q2m g
m=1
4.1. THEORY OF TASAI 63
in which:
f2m (µ) = ¡Ã A02m (µ) + Ã A02m (¼=2)
These equations can also be written as:
X
1
à B0c (µ) ¡ à B0c (¼=2) = ff2m (µ)P2m g
m=0
X1
à B0s (µ) ¡ à B0s (¼=2) = ff2m (µ)Q2m g
m=0
in which:
The series in these two sets of equations converges uniformly with an increasing value of
m. For practical reasons the maximum value of m is limited to M.
Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series. The best …t
values of P2m and Q2m are supposed to be those found by means of a least squares method.
Note that at least M + 1 values of µ, less than ¼=2, are required to solve these equations.
Another favorable method is to multiply both sides of the equations with ¢µ. Then the
summations over µ can be replaced by integrations.
Herewith, two sets of M + 1 equations have been obtained, one set for P2m and one set for
Q2m :
8 9
M
X < Z¼=2 = Z¼=2
¡ ¢
P2m f2m (µ)f2n (µ)dµ = Ã B0c (µ) ¡ Ã B0c (¼=2) f2n (µ)dµ n = 0; :::M
: ;
m=0 0 0
8 9
M <
X Z¼=2
= Z¼=2
¡ ¢
Q2m f2m (µ)f2n (µ)dµ = Ã B0s (µ) ¡ Ã B0s (¼=2) f2n(µ)dµ n = 0; :::M
: ;
m=0 0 0
Now the P2m and Q2m series can be solved with a numerical method. Then P0 and Q0 are
known now and from the de…nition of these coe¢cients follows the amplitude ratio of the
radiated waves and the forced sway oscillation:
´a ¼» b
=p
xa P02 + Q20
With the solved P2m and Q2m values, the velocity potential on the surface of the cylinder
(® = 0) is known too:
ÃÃ !
g´ a X1
© ª
©0 (µ) = ( ÁB0c (µ) + P2m ÁA02m (µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÁB0s (µ) + Q2m ÁA02m (µ) sin(!t)
m=1
64 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
in which:
In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®; µ) and ÁBs (®; µ) at the
surface of the cylinder, so for ® = 0.
Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the
linearized equation of Bernoulli:
@©0 (µ)
p(µ) = ¡½
@tÃÃ !
¡½g´a X1
© ª
= ÁB0s (µ) + Q2m ÁA02m (µ) cos(!t)
¼ m=1
à ! !
1
X© ª
¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t)
m=1
Sway Coe¢cients
The two-dimensional hydrodynamic lateral force, acting on the cylinder in the direction
of the x¡axis, can be found by integrating the lateral component of the hydrodynamic
pressure on the surface S of the cylinder:
Z¼=2
¡dy0
Fx0 = ¡ (p(+µ) ¡ p(¡µ)) ds
ds
0
Z¼=2
dy0
= 2 p(µ) dµ
dµ
0
With this the two-dimensional hydrodynamic lateral force due to sway oscillations can be
written as follows:
¡½gb0 ´ a
Fx0 = (M0 cos(!t) ¡ N0 sin(!t))
¼
in which:
Z¼=2 X
N
1
M0 = ¡ ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g dµ
¾a n=0
0
N¡1
X
¼
+ f(¡1)m Q2m (2m + 1)a2m+1 g
4¾ a
m=1
( N ½ ¾)
»
M
X N X
X (2n ¡ 1)(2i ¡ 1)
+ b2 (¡1)m Q2m a2n¡1 a2i¡1
¾a m=1 n=0 i=0
(2m + 2i)2 ¡ (2n ¡ 1)2
4.1. THEORY OF TASAI 65
and N0 as obtained from this expression for M0 , by replacing there ÁB0s (µ) by ÁB0c (µ) and
Q2m by P2m .
For the determination of M0 and N0 it is required: M = N .
With:
¡½gb0 ´ a
Fx0 = (M0 cos(!t + " ¡ ") ¡ N0 sin(!t + " ¡ "))
¼
and:
¡´ a P0 ¡´a Q0
sin " = cos " =
xa ¼» b xa ¼» b
the two-dimensional hydrodynamic lateral force can be resolved into components in phase
and out phase with the lateral displacement of the cylinder:
½gb0 ´ 2a
Fx0 = ((M0 Q0 + N0 P0 ) cos(!t + ") + (M0 P0 ¡ N0 Q0 ) sin(!t + "))
¼ 2 » b xa
This hydrodynamic lateral force can also be written as:
Fx0 = ¡M220 0
xÄ ¡ N22 x_
0 2 0
= M22 ! xa cos(!t + ") + N22 !xa sin(!t + ")
in which:
0
M22 = 2-D hydrodynamic mass coe¢cient of sway
0
N22 = 2-D hydrodynamic damping coe¢cient of sway
When using also the amplitude ratio of the radiated waves and the forced sway oscillation,
found before, the two-dimensional hydrodynamic mass and damping coe¢cients of sway
are given by:
0 ½b20 M0 Q0 + N0 P0 0 ½b20 M0 P0 ¡ N0 Q0
M22 = ¢ and N22 = ¢ ¢!
2 P02 + Q20 2 P02 + Q20
The signs of these two coe¢cients are proper in both, the axes system of Tasai and the
ship motions O(x; y; z) coordinate system.
The energy delivered by the exciting forces should be equal to the energy radiated by the
waves.
So:
T osc
Z
1 0 ½g´ 2a c
N22 x_ ¢ xdt
_ =
Tosc 2
0
0
When comparing this expression for N22 with the expression found before, the following
energy balance relation is found:
¼2
M0 P0 ¡ N0 Q0 =
2
With this the two-dimensional hydrodynamic roll moment due to sway oscillations can be
written as follows:
½gb20 ´ a
MR0 = (YR cos(!t) ¡ XR sin(!t))
¼
in which:
Z¼=2 X N
N X
1 © ª
YR = ÁB0s (µ) (¡1)n+i (2i ¡ 1) ¢ a2n¡1 a2i¡1 sin ((2n ¡ 2i)µ) dµ
2¾ 2a n=0 i=0
0
( N ½ ¾ )
1 X
M XN X
m (2i ¡ 1)(2n ¡ 2i)
+ 2 (¡1) Q2m ¢ a2n¡1 a2i¡1
2¾ a m=1 n=0 i=0
(2m + 1)2 ¡ (2n ¡ 2i)2
¼» b X
N
¡ 3 f(¡1)m Q2m ¢
8¾ a m=1
à N n¡m ½ ¾
X X (¡2m + 2n ¡ 2i ¡ 1)(2i ¡ 1)
¢ a2n¡1 a2i¡1 a¡2m+2n¡2i¡1
n=m i=0
2n ¡ 2i
N¡m
X X N ½ ¾!)
(¡2m ¡ 2n + 2i ¡ 1)(2i ¡ 1)
+ ¢ a2n¡1 a2i¡1 a¡2m¡2n+2i¡1
n=0 i=m+n
2n ¡ 2i
and XR as obtained from this expression for YR , by replacing there ÁB0s (µ) by ÁB0c (µ) and
Q2m by P2m .
With:
½gb20 ´ a
MR0 = (YR cos(!t + " ¡ ") ¡ XR sin(!t + " ¡ "))
¼
and:
¡´ a ¡´ a
sin " = P0 cos " = Q0
xa ¼» b xa ¼» b
4.1. THEORY OF TASAI 67
the two-dimensional hydrodynamic roll moment can be resolved into components in phase
and out phase with the lateral displacement of the cylinder:
¡½gb20 ´2a
MR0 = 2 ((YR Q0 + XR P0 ) cos(!t + ") + (YR P0 ¡ XR Q0 ) sin(!t + "))
¼ » b xa
This hydrodynamic roll moment can also be written as:
MR0 = ¡M420 0
xÄ ¡ N42 x_
0 2 0
= M42 ! xa cos(!t + ") + N42 !xa sin(!t + ")
in which:
0
M22 = 2-D hydrodynamic mass coupling coe¢cient of sway into roll
0
N22 = 2-D hydrodynamic damping coupling coe¢cient of sway into roll
When using also the amplitude ratio of the radiated waves and the forced sway oscillation,
found before, the two-dimensional hydrodynamic mass and damping coupling coe¢cients
of sway into roll in Tasai’s axes system are given by:
0 ¡½b30 YR Q0 + XR P0 0 ¡½b30 YR P0 ¡ XR Q0
M42 = ¢ and N42 = ¢ ¢!
2 P02 + Q20 2 P02 + Q20
In the ship motions O(x; y; z) coordinate system these two coupling coe¢cients will change
sign.
68 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
The cylinder is forced to carry out a simple harmonic roll motion about the origin O with
a frequency of oscillation ! and a small amplitude of displacement ¯ a :
¯ = ¯ a cos(!t + °)
This forced angular oscillation of the cylinder causes a surface disturbance of the ‡uid.
Because the cylinder is supposed to be in…nitely long, the generated waves will be two-
dimensional. These waves travel away from the cylinder and a stationary state is rapidly
attained.
Two kinds of waves will be produced:
² A standing wave system, denoted here by subscript A. The amplitudes of these waves
decrease strongly with the distance to the cylinder.
4.1. THEORY OF TASAI 69
² A regular progressive wave system, denoted here by subscript B. These waves dis-
sipate energy. At a distance of a few wave lengths from the cylinder, the waves on
each side can be described by a single regular wave train.. The wave amplitude at
in…nity ´ a is proportional to the amplitude of oscillation of the cylinder ¯ a , provided
that this amplitude is su¢ciently small compared with the radius of the cylinder and
the wave length is not much smaller than the diameter of the cylinder.
The two-dimensional velocity potential of the ‡uid has to ful…l the following three require-
ments:
@2© @2©
r2 © = + 2 =0
@x2 @y
2. Because the roll motion of the ‡uid is not symmetrical about the y-axis, this velocity
potential has the following relation:
©(¡x; y) = ¡©(+x; y)
!2 @© Bs
©+ =0 for: jxj ¸ and y = 0
g @y 2
»b X ©
N
ª @© ¼
© (2n ¡ 1)a2n¡1 e¡(2n¡1)® § = 0 for: ® ¸ 0 and µ = §
¾ a n=0 @µ 2
in which:
»b !2 ! 2 b0
= Ms or: » b = (non-dimensional frequency squared)
¾a g 2g
From the de…nition of the velocity potential follows the boundary condition on the surface
S of the cylinder for ® = 0:
@©0 (µ) @r0
= r0 ¯_
@n @s
in which n is the outward normal of the cylinder surface S and r0 is the radius from the
origin to the surface of the cylinder.
Using the stream function ª this boundary condition on the surface of the cylinder (® = 0)
reduces to: µ 2 ¶
2
¡@ª0 (µ) @ x + y
= ¯_ 0 0
@s @s 2
Integration results into the following requirement for the stream function on the surface of
the cylinder:
¯_
ª0 (µ) = ¡ (x20 + y02 ) + C(t)
2
70 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
x20 + y02
¹(µ) = ¡ b 0 ¢2
2
à !2
1 X
N
n
= ¡ f(¡1) a2n¡1 sin ((2n ¡ 1)µ)g
¾ a n=0
à !2
1 X
N
+ + f(¡1)n a2n¡1 cos ((2n ¡ 1)µ)g
¾ a n=0
in which:
in which:
In these expressions the magnitudes of the P2m and the Q2m series follow from the boundary
conditions, as will be explained further on.
Another requirement is a diverging wave train for ® goes to in…nity. Therefore it is nec-
essary to add a stream function, satisfying the equation of Laplace, the non-symmetrical
motion of the ‡uid and the free surface condition, representing such a train of waves at
in…nity. For this, a function describing a two-dimensional horizontal doublet at the origin
O is chosen.
[Tasai, 1961] and [Jong, 1973] gave the velocity potential of the progressive wave system
as:
g´ ¡ ¢
©B = a ÁBc (x; y) cos(!t) + ÁBs (x; y) sin(!t)
¼!
in which:
while:
j = +1 for x > 0
j = ¡1 for x < 0
!2
º = (wave number for deep water)
g
Changing the parameters delivers:
g´ ¡ ¢
©B = a ÁBc (®; µ) cos(!t) + ÁBs (®; µ) sin(!t)
¼!
The conjugate stream function is given by:
g´ ¡ ¢
ªB = a à Bc (x; y) cos(!t) + à Bs (x; y) sin(!t)
¼!
in which:
à Bc = ¼e¡ºy cos(ºx)
Z1
º sin(ky) ¡ k cos(ky) ¡kjxj y
à Bs = ¼e¡ºy sin(ºjxj) + 2 2
e dk ¡
k +º º(x + y 2 )
2
0
The total velocity potential and stream function to describe the waves generated by a
swaying cylinder are:
© = ©A + ©B
ª = ªA + ªB
So the velocity potential and the conjugate stream function are expressed by:
ÃÃ !
g´ a X1
© ª
©(®; µ) = ÁBc (®; µ) + P2m ÁA2m (®; µ) cos(!t)
¼! m=1
à ! !
1
X© ª
+ ÁBs (®; µ) + Q2m ÁA2m (®; µ) sin(!t)
m=1
ÃÃ !
g´ a
1
X © ª
ª(®; µ) = à Bc (®; µ) + P2m à A2m (®; µ) cos(!t)
¼! m=1
à ! !
1
X © ª
+ à Bs (®; µ) + Q2m à A2m (®; µ) sin(!t)
m=1
When putting ® = 0, the stream function is equal to the expression found before from the
boundary condition on the surface of the cylinder:
ÃÃ !
g´ a X1
© ª
ª0 (µ) = à B0c (µ) + P2m à A02m (µ) cos(!t)
¼! m=1
à ! !
X1
© ª
+ ÃB0s (µ) + Q2m à A02m (µ) sin(!t)
m=1
b0
= ¡Â_ ¹(µ) + C(t)
4
in which:
à A02m (µ) = ¡ cos ((2m + 1)µ)
N ½ ¾
»b X n 2n ¡ 1
+ (¡1) a2n¡1 cos ((2m + 2n)µ)
¾ a n=0 2m + 2n
In this expression ªB0c (µ) and ªB0s (µ) are the values of ªBc (®; µ) and ªBs (®; µ) at the
surface of the cylinder, so for ® = 0.
So for each µ-value, the following equation has been obtained:
à !
X
1
© ª
à B0c (µ) + P2m à A02m (µ) cos(!t)+
m=1
à !
X
1
© ª ¼!b0
à B0s (µ) + Q2m à A02m (µ) sin(!t) = ¡Â_ ¹(µ) + C ¤ (t)
m=1
4g´ a
When putting µ = ¼=2, so at the intersection of the surface of the cylinder with the free
surface of the ‡uid where ¹(µ) = 1, we obtain the constant C ¤ (t):
1
X © ª
¤
C (t) = à B0c (¼=2) + P2m à A02m (¼=2) cos(!t)
m=1
4.1. THEORY OF TASAI 73
à !
1
X © ª
+ à B0s (¼=2) + Q2m à A02m (¼=2) sin(!t)
m=1
¼!b0
+Â_
4g´ a
in which:
»b XN
2n ¡ 1
m
ªA02m (¼=2) = (¡1) a2n¡1
¾a n=0
2m + 2n
A substitution of C ¤ (t) in the equation for each µ-value, results for any µ-value less than
¼=2 into the following equation:
à !
1
X © ¡ ¢ª
à B0c (µ) ¡ à B0c (¼=2) + P2m à A02m (µ) ¡ à A02m (¼=2) cos(!t)
m=1
à !
1
X © ¡ ¢ª
+ à B0s (µ) ¡ à B0s (¼=2) + Q2m à A02m (µ) ¡ à A02m (¼=2) sin(!t)
m=1
¼!b0
= ¡Â_ (¹(µ) ¡ 1)
4g´ a
in which:
¼Âa ¼Âa
P0 = » sin ° and Q0 = » cos °
2´ a b 2´ a b
This delivers for any µ-value less than ¼=2 two sets of equations with the unknown para-
meters P2m and Q2m .
So:
1
X
à B0c (µ) ¡ à B0c (¼=2) = (¹(µ) ¡ 1) P0 + ff2m (µ)P2m g
m=1
X1
à B0s (µ) ¡ à B0s (¼=2) = (¹(µ) ¡ 1) Q0 + ff2m (µ)Q2m g
m=1
in which:
f2m (µ) = ¡Ã A0c (µ) + Ã A0c (¼=2)
These equations can also be written as:
1
X
à B0c (µ) ¡ à B0c (¼=2) = ff2m (µ)P2m g
m=0
X1
à B0s (µ) ¡ à B0s (¼=2) = ff2m (µ)Q2m g
m=0
74 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
in which:
The series in these two sets of equations converges uniformly with an increasing value of
m. For practical reasons the maximum value of m is limited to M.
Each µ-value less than ¼=2 will deliver an equation for the P2m and Q2m series. The best …t
values of P2m and Q2m are supposed to be those found by means of a least squares method.
Note that at least M +1 values of µ, less than ¼=2, are required to solve these equations.
Another favorable method is to multiply both sides of the equations with ¢µ. Then the
summations over µ can be replaced by integrations.
Herewith, two sets of M + 1 equations have been obtained, one set for P2m and one set for
Q2m :
8 9
XM < Z¼=2 = Z¼=2
¡ ¢
P2m f2m (µ)f2n (µ)dµ = Ã B0c (µ) ¡ Ã B0c (¼=2) f2n (µ)dµ n = 0; :::M
: ;
m=0 0 0
8 9
M <
X Z¼=2
= Z¼=2
¡ ¢
Q2m f2m (µ)f2n (µ)dµ = Ã B0s (µ) ¡ Ã B0s (¼=2) f2n(µ)dµ n = 0; :::M
: ;
m=0 0 0
Now the P2m and Q2m series can be solved with a numerical method. Then P0 and Q0 are
known now and from the de…nition of these coe¢cients follows the amplitude ratio of the
radiated waves and the forced sway oscillation:
´a ¼»
= p b
Âa 2 P02 + Q20
With the solved P2m and Q2m values, the velocity potential on the surface of the cylinder
(®=0) is known too:
ÃÃ !
g´ a XM
© ª
©0 (µ) = ÁB0c (µ) + P2m ÁA02m (µ) cos(!t)
¼! m=1
à ! !
X©
M
ª
+ ÁB0s (µ) + Q2m ÁA02m (µ) sin(!t)
m=1
in which:
In this expression ÁB0c (µ) and ÁB0s (µ) are the values of ÁBc (®; µ) and ÁBs (®; µ) at the
surface of the cylinder, so for ® = 0.
4.1. THEORY OF TASAI 75
Now the hydrodynamic pressure on the surface of the cylinder can be obtained from the
linearized equation of Bernoulli:
@©0 (µ)
p(µ) = ¡½
@tÃÃ !
¡½g´ a XM
© ª
= ÁB0s (µ) + Q2m ÁA02m (µ) cos(!t)
¼ m=1
à ! !
XM
© ª
¡ ÁB0c (µ) + P2m ÁA02m (µ) sin(!t)
m=1
Roll Coe¢cients
The two-dimensional hydrodynamic moment acting on the cylinder in the clockwise direc-
tion can be found by integrating the roll component of the hydrodynamic pressure on the
surface S of the cylinder:
Z¼=2 µ ¶
+dx0 ¡dy0
MR0 = (p(+µ) ¡ p(¡µ)) ¡x0 + y0 ds
ds ds
0
Z¼=2 µ ¶
dx0 dy0
= ¡2 p(µ) x0 + y0 dµ
dµ dµ
0
With this the two-dimensional hydrodynamic moment due to roll oscillations can be written
as follows:
½gb20 ´ a
MR0 = (YR cos(!t) ¡ XR sin(!t))
¼
in which:
Z¼=2 X N
N X
1 © ª
YR = ÁB0s (µ) (¡1)n+i (2i ¡ 1) ¢ a2n¡1 a2i¡1 sin ((2n ¡ 2i)µ) dµ
2¾ 2a n=0 i=0
0
( N ½ ¾)
1
M
X N X
X (2i ¡ 1)(2n ¡ 2i)
+ 2 (¡1)m Q2m ¢ a2n¡1 a2i¡1
2¾ a m=1 n=0 i=0
(2m + 1)2 ¡ (2n ¡ 2i)2
¼» b XN
¡ f(¡1)m Q2m ¢
8¾ 3a
m=1
Ã
X X½
N n¡m
(¡2m + 2n ¡ 2i ¡ 1)(2i ¡ 1)
¾
¢ a2n¡1 a2i¡1 a¡2m+2n¡2i¡1
n=m i=0
2n ¡ 2i
N¡m
X X N ½ ¾!)
(¡2m ¡ 2n + 2i ¡ 1)(2i ¡ 1)
+ ¢ a2n¡1 a2i¡1 a¡2m¡2n+2i¡1
n=0 i=m+n
2n ¡ 2i
and XR as obtained from this expression for YR , by replacing there ÁB0s (µ) by ÁB0c (µ) and
Q2m by P2m .
76 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
These expressions are similar to those found before for the hydrodynamic roll moment due
to sway oscillations.
With:
0 ½gb20 ´ a
MR = (YR cos(!t + ° ¡ °) ¡ XR sin(!t + ° ¡ °))
¼
and:
2´ a 2´ a
sin ° = P0 cos ° = Q0
Âa ¼» b Âa ¼» b
the two-dimensional hydrodynamic roll moment can be resolved into components in phase
and out phase with the angular displacement of the cylinder:
2½gb20 ´ 2a
MR0 = ((YR Q0 + XR P0 ) cos(!t + °) + (YR P0 ¡ XR Q0 ) sin(!t + °))
¼ 2 » b Âa
This hydrodynamic roll moment can also be written as:
M 0 = ¡M 0 Ǟ ¡ N 0 ¯_
R 44 44
= 0
M44 ! 2 ¯ a 0
cos(!t + °) + N44 !¯ a sin(!t + °)
in which:
0
M44 = 2-D hydrodynamic mass moment of inertia coe¢cient of roll
0
N44 = 2-D hydrodynamic damping coe¢cient of roll
When using also the amplitude ratio of the radiated waves and the forced roll oscillation,
found before, the two-dimensional hydrodynamic mass and damping coe¢cients of roll in
Tasai’s axes system are given by:
0 ½b40 YR Q0 + XR P0 0 ½b40 YR P0 ¡ XR Q0
M44 = ¢ and N44 = ¢ ¢!
8 P02 + Q20 8 P02 + Q20
The signs of these two coe¢cients are proper in both, the axes system of Tasai and the
ship motions O(x; y; z) coordinate system.
The energy delivered by the exciting moments should be equal to the energy radiated by
the waves.
So:
T osc
Z 2
1 0 _ _ = ½g´ a c
N44 ¯ ¢ ¯dt
Tosc 2
0
in which Tosc is the period of oscillation.
With the relation for the wave speed c = g=!, follows the relation between the two-
dimensional roll damping coe¢cient and the amplitude ratio of the radiated waves and the
forced roll oscillation: µ ¶2
0 ½g 2 ´ a
N44 = 3
! ¯a
With this amplitude ratio the two-dimensional hydrodynamic damping coe¢cient of roll
is also given by:
0 ½¼2 b40 1
N44 = ¢ 2 ¢!
64 P0 + Q20
0
When comparing this expression for N44 with the expression found before, the following
energy balance relation is found:
¼2
YR P0 ¡ XR Q0 =
8
4.1. THEORY OF TASAI 77
With this the two-dimensional hydrodynamic angular force due to roll oscillations can be
written as follows:
¡½gb0 ´ a
Fx0 = (M0 cos(!t) ¡ N0 sin(!t))
¼
in which:
Z¼=2 N
X
1
M0 = ¡ ÁB0s (µ) f(¡1)n (2n ¡ 1)a2n¡1 sin ((2n ¡ 1)µ)g dµ
¾a n=0
0
N¡1
X
¼
+ f(¡1)m Q2m (2m + 1)a2m+1 g
4¾ a
m=1
( N ½ ¾)
»
M
X N X
X (2n ¡ 1)(2i ¡ 1)
+ b2 (¡1)m Q2m a2n¡1 a2i¡1
¾a m=1 n=0 i=0
(2m + 2i)2 ¡ (2n ¡ 1)2
and N0 as obtained from this expression for M0 , by replacing there ÁB0s (µ) by ÁB0c (µ) and
Q2m by P2m .
For the determination of M0 and N0 it is required: M = N .
These expressions are similar to those found before for the hydrodynamic lateral force due
to sway oscillations.
With:
¡½gb0 ´ a
Fx0 = (M0 cos(!t + ° ¡ °) ¡ N0 sin(!t + ° ¡ °))
¼
and:
2´ a 2´ a
sin ° = P0 cos ° = Q0
Âa ¼» b Âa ¼» b
the two-dimensional hydrodynamic angular force can be resolved into components in phase
and out phase with the angular displacement of the cylinder:
¡2½gb0 ´ 2a
Fx0 = ((M0 Q0 + N0 P0 ) cos(!t + °) + (M0 P0 ¡ N0 Q0 ) sin(!t + °))
¼ 2 » b Âa
This hydrodynamic angular force can also be written as:
0 Ǟ 0 _
Fx0 = ¡M24 ¡ N24 ¯
0 2 0
= M24 ! ¯ a cos(!t + °) + N24 !¯ a sin(!t + °)
in which:
78 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
0
M24 = 2-D hydrodynamic mass coupling coe¢cient of roll into sway
0
N24 = 2-D hydrodynamic damping coupling coe¢cient of roll into sway
When using also the amplitude ratio of the radiated waves and the forced roll oscillation,
found before, the two-dimensional hydrodynamic mass and damping coupling coe¢cients
of roll into sway are given by:
0 ¡½b30 M0 Q0 + N0 P0 0 ¡½b30 M0 P0 ¡ N0 Q0
M24 = ¢ and N24 = ¢ ¢!
8 P02 + Q20 8 P02 + Q20
In the ship motions O(x; y; z) coordinate system these two coupling coe¢cients will change
sign.
4.1. THEORY OF TASAI 79
Zero-Frequency Coe¢cients
The two-dimensional hydrodynamic mass coe¢cient of sway of a Lewis cross section is
given by [Tasai, 1961]:
µ ¶2
0 ½¼ Ds ¡ ¢
M22 (! = 0) = (1 ¡ a1 )2 + 3a23
2 1 ¡ a1 + a3
The two-dimensional hydrodynamic mass coupling coe¢cient of sway into roll of a Lewis
cross section is given by [Grim, 1955]:
0 0 16 Ds
M42 (! = 0) = ¡M22 (! = 0) ¢ ¢ ¢
3¼ 1 ¡ a1 + a3
¡ ¢
a1 1 ¡ a1 + 45 a3 ¡ a1 a3 + 35 a23 + 45 a3 ¡ 12 2
a
7 3
(1 ¡ a1 )2 + 3a23
0
In Tasai’s axes system M42 will change sign.
The two-dimensional hydrodynamic mass moment of inertia coe¢cient of roll of a Lewis
cross section is given by [Grim, 1955]:
µ ¶4
0 16½ Bs
M44 (! = 0) = ¢ ¢
¼ 2(1 + a1 + a3 )
µ ¶
2 2 8 16 2
a1 (1 + a3 ) + a1 a3 (1 + a3 ) + a3
9 9
The two-dimensional hydrodynamic mass coupling coe¢cient of roll into sway of a Lewis
cross section is given by [Grim, 1955]:
0 0 ¼ 1 ¡ a1 + a3
M24 (! = 0) = ¡M44 (! = 0) ¢ ¢ ¢
6 Ds
a1 (1 ¡ a1 )(1 + a3 ) + 35 a1 a3 (1 + a3 ) + 45 a3 (1 ¡ a1 ) ¡ 12 2
a
7 3
a21 (1 + a3 )2 + 89 a1 a3 (1 + a3 ) + 16 a2
9 3
0
In Tasai’s axes system M24 will change sign.
All potential damping values for zero frequency will be zero:
0
N22 (! = 0) = 0
0
N42 (! = 0) = 0
0
N33 (! = 0) = 0
0
N44 (! = 0) = 0
0
N24 (! = 0) = 0
80 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
4.2.1 Notation
His notations have been maintained here as far as possible and references are given here
to the formula numbers in his report.
a = Lewis coe¢cient
Aindex = source strength
A¹index = amplitude ratio
b = Lewis coe¢cient
B = breadth of body
c = wave velocity
Cindex = non-dimensional force or moment
Eindex = non-dimensional exciting force or moment
Findex = hydrodynamic force
g = acceleration of gravity
G = function (real part)
h = water depth
H = function (imaginary part)
Hindices = …ctive moment arm
HT = water depth - draft ratio
I" = hydrodynamic moment of inertia
kx = wave number in x-direction
ky = wave number in y-direction
m" = hydrodynamic mass
Mindices = hydrodynamic moment
Nindex = hydrodynamic damping coe¢cient
p p = pressure
r = y2 + z2 = polar coordinates
t = time or integer value
T = draft
U¹ = velocity amplitude of horizontal oscillation
V¹ = velocity amplitude of vertical oscillation
Ax = cross sectional area
x; y; z = space-bound coordinates
Yindices = transfer functions
° = Euler constant (= 0.57722)
" = phase lag
³¹ = wave amplitude
µ = polar coordinate or pitch amplitude
82 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
¸ = wave length
¹ = wave direction
º = ! 2 =g = wave number at deep water
º 0 = 2¼=¸ = wave number
½ = density of water
' = roll angle
'indices = part of potential
©indices = time-dependent potential
à indices = part of stream function
ªindices = time-dependent stream function
! = circular frequency of oscillation
E = related to excitation
H = horizontal or related to horizontal motions
j = imaginary part
n = ordering of potential parts
Q = related to transverse motions
r = real part
R = related to roll motions
V = vertical or related to vertical motions
W = related to waves
!2 2¼ 2¼h
º= = tanh = º 0 tanh [º 0 h]
g ¸ ¸
The ‡uid is supposed to be incompressible and inviscid. The ‡ow caused by the oscillating
body in the surface of this ‡uid can be described by a potential ‡ow. The problem will
be linearized, i.e., contributions of second and higher order in the de…nition of the bound-
ary conditions will be ignored. Physically, this yields an assumption of small amplitude
motions.
The space-bound axes system of the sectional contour is given in …gure 4.5-a.
Velocities are positive if they are directed in the positive coordinate direction:
@© @©
= vy = vz
@y @z
The value of the stream function increases when - going in the positive direction - the ‡ow
goes in the negative y-direction:
@© @ª
ª1 < ª2 ! =+
@y @z
4.2. THEORY OF KEIL 83
@© @ª
ª3 > ª4 ! =¡
@z @y
@© @ª
=¡
@n @s
@© @ª
=+
@s @n
For the imaginary parts, i and j have been used: i for geometrical variables (potential and
stream function) and j for functions of time.
1. The ‡uid is incompressible and the velocity potential must satisfy to the Continuity
Condition and the Equation of Laplace:
@2© @2©
r2 © = + 2 =0 (Keil-1)
@y 2 @z
2. The linearized free surface condition follows from the condition that the pressure at
the free surface is not time-depending:
µ ¶
@p @© @ 2 © B
=½ g ¡ 2 =0 for: jyj ¸ and z = 0 (Keil-2/1)
@t @z @t 2
from which follows:
!2 @© B
©+ =0 for: jyj ¸ and z = 0
g @z 2
or:
@© B
º© + =0 for: jyj ¸ and z = 0 (Keil-2/2)
@z 2
3. The sea bottom is impervious, so the vertical ‡uid velocity at z = h is zero:
@©
=0 for: z = h (Keil-3)
@z
4. The harmonic oscillating cylinder produces a regular progressive wave system, trav-
elling away from the cylinder, which ful…ls the Sommerfeld Radiation Condition:
½p µ ¶¾
@
lim jyj Re (©) ¡ º 0 Im (©) =0 (Keil-4)
y!1 @y
5. The oscillating cylinder is impervious too; thus at the surface of the body is the ‡uid
velocity equal to the body velocity, see …gure 4.5-b. This yields that the boundary
conditions on the surface of the body are given by:
· ¸
@©
= [vn ]body
@n body
· ¸
@© dy @© dz
= ¢ ¡ ¢
@z ds @y ds body
· ¸
dª
= ¡ (Keil-5)
ds body
a. The hydromechanical loads, which have to be obtained for the vertically oscillating
cylinder in still water with a vertical velocity equal to:
V = V¹ ¢ ejwt
or:
ªbody (y; z; t) = ¡V¹ ¢ ejwt ¢ ybody + C (Keil-5a/2)
b. The wave loads, which have to be obtained for the restrained cylinder in regular waves
from the incoming undisturbed wave potential ©W and the di¤raction potential ©S :
· ¸
@©W @©S
+ = 0
@n @n body
· ¸
@©W dy @©W dz @©S dy @©S dz
= ¢ ¡ ¢ + ¢ ¡ ¢
@z ds @y ds @z ds @y ds body
· ¸
dªW dªS
= ¡ + (Keil-6/1)
ds ds body
or:
The stream function of an incoming wave - which travels in the negative y-direction,
so ¹ = 900 - is given in Appendix 1 of this section by:
¹
³!
ªW = j ¢ ej(wt+º 0 y) ¢ (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) (Keil-A1)
º
Because only vertical forces have to be determined, only the in y-symmetric part
of the potential and stream functions have to be considered. From this follows the
boundary condition on the surface of the body for beam waves, so wave direction
¹ = 900 :
out of the body per unit of length so that totally no ‡uid of the incoming wave enters
into the body. This function can be used as an approximation of the problem:
2 y1 3
h i Z Zz1
@ªW @ªW 5
~ S (x1 ; y1 ; z1 ; t)
ª =4 dy ¡ dz = (Keil-6b)
body @z @y
0 z0 body
¹
³!
= ¢ ejwt ¢ cos(º 0 x cos ¹) ¢
º
fsin ¹ sin(º 0 y1 sin ¹) (sinh [º 0 z1 ] ¡ tanh [º 0 h] cosh [º 0 z1 ])
Zy1
2
+º 0 (1 ¡ sin ¹) cos(º 0 y sin ¹) (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) dyg
0
Potentials
3-D Radiation Potential Suppose a three-dimensional oscillating cylindrical body in
previously still water. To …nd the potential of the resulting ‡uid motions, this body will
be replaced by an oscillating pressure p at the free surface. The unknown amplitude p¹ of
this pressure has to follow from the boundary conditions.
This pressure is not supposed to act over the full breadth of the body; it is supposed to
act - over the full length L of the body - only over a small distance ¢y=2 to both sides of
y = 0, so:
Z +¢y=2
+L=2 Z
P¹ = p¹(x; y; z=z0 ) dydx
¡L=2 ¡¢y=2
Z
+L=2
0
= P¹ (x)dx < 1 (Keil-7a)
¡L=2
Boundary condition (Keil-7) can be ful…lled by a pressure amplitude p¹(x; y; z=z0 ) which
is found by a superposition of an in…nite number of harmonic pressures. From equation
(Keil-7a) follows that the pressure amplitude p¹(x) can be integrated, so a Fourier series
expansion follows from:
Z1 Z+1
1
p¹(x) = p¹(») cos [kx(x ¡ »)] d» dkx
¼
0 ¡1
4.2. THEORY OF KEIL 87
Because the pressure amplitude p¹ depends on two variables, the Fourier series expansion
has to be two-dimensional:
Z1 Z1 Z+1 Z+1
1
p¹(x; y; z=z0 ) = 2 p¹(»; ´) cos [ky (y ¡ ´)] cos [kx (x ¡ »)] d´ d» dky dkx
¼
0 0 ¡1 ¡1
in which kx is the wave number in the x-direction and ky is the wave number in the
y-direction.
According to equation (Keil-7), the pressure amplitude p¹ disappears for jyj > ¢y=2 and
jxj > L=2, so for this pressure expression remains:
Z1 Z1 +L=2
Z +¢y=2
Z
1
p¹(x; y; z=z0 ) = 2 p¹(»; ´) cos [ky (y ¡ ´)] cos [kx (x ¡ »)] d´ d» dky dkx
¼
0 0 ¡L=2 ¡¢y=2
It is assumed that the value of ¢y is small. This means that ´ remains small too, thus one
can safely suppose that:
cos [ky (y ¡ ´)] t cos(ky y)
which results in:
Z1 Z1 +L=2
Z +¢y=2
Z
1
p¹(x; y; z=z0 ) = p¹(»; ´) d´ cos [kx(x ¡ »)] d» cos(ky y) dky dkx
¼2
0 0 ¡L=2 ¡¢y=2
Z1 Z1 Z
+L=2
1 0
= cos(ky y) P¹ (») cos [kx (x ¡ »)] d» dky dkx (Keil-7b)
¼2
0 0 ¡L=2
This pressure de…nition leads - as a start - to the following initial de…nition of the radiation
potential:
Z1 Z1 Z1
j!t
©0r (x; y; z; t) = e C(kx ; ky ) cos [kx (x ¡ »)] cos(ky y) ¢
0 0 0
£p ¤
cosh kx2 + ky2 ¢ (z ¡ h)
¢ £p ¤ dkx dky d» (Keil-8)
sinh kx2 + ky2 ¢ h
in which the function C(kx ; ky ) is still unknown.
This expression (Keil-8) for the radiation potential ful…ls the Equation of Laplace:
@ 2© @ 2© @ 2©
+ 2 + 2 =0
@x2 @y @z
Now, the pressure at the free surface p1 can be obtained from an integration of the - with
the Bernoulli Equation obtained - derivative to the time of the pressure:
Z1 Z1 Z1
@p1
= ej!t ½ C(kx ; ky ) cos [kx (x ¡ »)] cos(ky y) ¢
@t
0 0 0
2
p £p ¤
! ¡ g kx2 + ky2 ¢ tanh kx2 + ky2 ¢ h
¢ £p ¤ dkx dky d» (Keil-8a)
tanh kx2 + ky2 ¢ h
88 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Z1 Z1 Z1
½
p¹1 (x; y; z=z0 ) = C(kx; ky ) cos [kx (x ¡ »)] cos(ky y) ¢ (Keil-8b)
!
0 0 0
2
p £p ¤
! ¡g kx2 + ky2 ¢ tanh kx2 + ky2 ¢ h
¢ £p ¤ dkx dky d»
tanh kx2 + ky2 ¢ h
If this pressure amplitude p¹1 is supposed to be equal to the amplitude p¹, then combining
equations (Keil-7b) and (Keil-8b) provides the unknown function C(kx ; ky ):
Z1 Z1 Z
+L=2
1 0
p¹(x; y; z=z0 ) = cos(ky y) P¹ (») cos [kx (x ¡ »)] d» dky dkx
¼2
0 0 ¡L=2
Z1 p £p ¤
½g º ¡ kx2 + ky2 ¢ tanh kx2 + ky2 ¢ h
C(kx ; ky ) cos [kx (x ¡ »)] ¢ £p ¤ d» =
! tanh kx2 + ky2 ¢ h
0
Z
+L=2
1 0
= 2 P¹ (») cos [kx (x ¡ »)] d»
¼
¡L=2
or:
Z1 ¤ £p
kx2 + ky2 ¢ h
tanh
C(kx ; ky ) cos [kx (x ¡ »)] d» = p £p ¤¢
º ¡ kx2 + ky2 ¢ tanh kx2 + ky2 ¢ h
0
Z
+L=2
! 0
¢ P¹ (») cos [kx (x ¡ »)] d» (Keil-9)
½g¼ 2
¡L=2
When de…ning:
4.2. THEORY OF KEIL 89
! P¹ (»)
0
A0 (») =
½g¼ 2
and substituting equation (Keil-9) in equation (Keil-8) provides the radiation potential:
Z
+L=2 Z1
j!t
©0r (x; y; z; t) = e A0 (») cos [kx (x ¡ »)] ¢ (Keil-10)
¡L=2 0
Z1 £p ¤
cosh kx2 + ky2 ¢ (z ¡ h) cos(ky y)
¢ £p ¤ p £p ¤ dkx dky d»
º cosh kx2 + ky2 ¢ h ¡ kx2 + ky2 ¢ sinh kx2 + ky2 ¢ h
0
This potential ful…lls both the radiation conditions at in…nity and the boundary conditions
at the free surface.
Z1
cosh [k(z ¡ h)]
©0r (y; z; t) = ej!t A0 cos(ky) dk (Keil-11)
º cosh [kh] ¡ k sinh [kh]
0
To ful…l also the Sommerfeld Radiation Condition in (Keil-4), a term has to be added to
equation (Keil-11). For this, use will be made here of the value of the potential given in
equation (Keil-11) at a large distance from the body:
Z1
j!t cosh [k(z ¡ h)]
©0r (y; z; t) = e A0 cos(ky) dk
º cosh [kh] ¡ k sinh [kh]
0
81 9
< Z Z1 =
1
= ej!t A0 F (k)e+iky dk + F (k)e¡iky dk (K-11)
2: ;
0 0
with:
The location of the singular point follows from the denominator in the expression
(K-11a) for F (k):
º cosh [º 0 h] ¡ º 0 sinh [º 0 h] = 0
Because limR!1 [JIII ] = 0 and JIV disappears too for a large y; the singular point
itself delivers a contribution only:
Z1
cosh [º 0 h] cosh [º 0 (z ¡ h)] +iº 0 y
F (k)e+iky dk = ¡i¼ e
º 0 h + sinh [º 0 h] cosh [º 0 h]
0
4.2. THEORY OF KEIL 91
Because limR!1 [JV I ] = 0 and JV II disappears too for a large y; the singular point
itself delivers a contribution only:
JV = +i¼ ¢ Residue(º 0 )
cosh [º 0 h] cosh [º 0 (z ¡ h)] ¡iº 0 y
= ¡i¼ e
º 0 h + sinh [º 0 h] cosh [º 0 h]
This means that equation (Keil-12) describes a ‡ow, consisting of waves with an amplitude:
¹³ = A0 !¼ cosh2 [º 0 h]
(Keil-12c)
g º 0 h + sinh [º 0 h] cosh [º 0 h]
travelling away from both sides of the cylinder.
From the orthogonality condition:
@© @ª
=+
@y @z
follows the stream function:
For an in…nite water depth, equations (Keil-12) and (Keil-13) reduces to:
81 9
<Z e¡kz =
©0r1 (y; z; t) + j©0j1 (y; z; t) = ej!t A0 cos(ky) dk + j¼e¡ºz cos(ºy)
: º¡k ;
0
(Keil-12a)
81 9
<Z e¡kz =
ª0r1 (y; z; t) + jª0j1 (y; z; t) = ej!t A0 ¡ºz
sin(ky) dk + j¼e sin(ºy)
: º¡k ;
0
(Keil-13a)
Alternative Derivation Assuming that the real part of the potential at an in…nite water
depth, ©r1 ,is known, another derivation of the 2-D potential is given by [Porter, 1960].
The additional potential for a restricted water depth, ©rad , will be determined in such a
way that it ful…lls the free surface condition and - together with ©r1 - also the boundary
condition at the sea bed.
As an extension of the additional real potential will be chosen:
Z1
©0rad (y; z; t) = ej!t A0 fC1 (k) sinh [kz] + C2 (k) cosh [k(z ¡ h)]g cos(ky)dk
0
(Keil-14)
From the free surface condition (Keil-2) follows for jyj > B=2:
Z1
ej!t A0 fºC2 (k) cosh [kh] + kC1 (k) ¡ kC2 (k) sinh [kh]g cos(ky)dk = 0
0
is known:
Z1
1
f (k) = g(») cos(k») d»
¼
0
The still unknown function C1 (k) follows from the boundary condition at the sea bed:
· ¸
@©0r1 @©0rad
+ =0=
@z @z z=h
8 1 9
< Z ke¡kh Z1 =
= ej!t A0 ¡ cos(ky) dk + kC1 (k) cosh [kh] cos(ky) dk
: º¡k ;
0 0
So:
e¡kh
C1 (k) =
(º ¡ k) cosh [kh]
¡ke¡kh
C2 (k) =
(º ¡ k) fº cosh [kh] ¡ k sinh [kh]g cosh [kh]
With this, the extension of the additional real potential, as given in equation (Keil-12b),
becomes:
Z1
j!t e¡kh º sinh [kz] ¡ k cosh [kz]
©0rad (y; z; t) = e A0 ¢ cos(ky) dk
º ¡ k º cosh [kh] ¡ k sinh [kh]
0
(Keil-14a)
2-D Multi-Potential The free surface conditions can not be ful…lled with the potential
(Keil-12b) - and the stream function (Keil-13b) respectively - only. Additional potentials
©n are required which ful…ll the boundary conditions (Keil-1) through (Keil-4) and together
with ©0 also ful…ll the boundary conditions (Keil-5) or (Keil-6):
0
1
X 0
©(y; z; t) = A0 ©0 (y; z; t) + An ©n (y; z; t) (Keil-15)
n=1
h 0 0 0
i
= A0 ©0r1(y; z; t) + ©0rad (y; z; t) + j©0j (y; z; t)
1
X h 0 0 0
i
+ An ©nr1(y; z; t) + ©nrad (y; z; t) + j©nj (y; z; t)
n=1
Use will be made here of multi-potentials given by [Grim, 1956] and [Grim, 1957], of which
- using the Sommerfeld Radiation Condition - the real part of the additional potential
©nrad and the imaginary part potential ©nj will be determined. This results in:
Z1
©nr1 (y; z; t) = +ej!t An (k + º)k 2(n¡1) e¡kz cos(ky) dk
0
4.2. THEORY OF KEIL 95
(Keil-16)
Z1
º sinh [kz] ¡ k cosh [kz]
©nrad (y; z; t) = +ej!t An (k + º)k 2(n¡1) e¡kh cos(ky) dk
º cosh [kh] ¡ k sinh [kh]
0
(Keil-16a)
¼º 2n
0 cosh [º 0 (z ¡ h)]
©nj (y; z; t) = ¡ej!t An ¢ cos(º 0 y)
cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h]
(Keil-16b)
Z1
j!t
ªnr1 (y; z; t) = +e An (k + º)k 2(n¡1) e¡kz sin(ky) dk
0
(Keil-17)
Z1
º cosh [kz] ¡ k sinh [kz]
ªnrad (y; z; t) = ¡ej!t An (k + º)k 2(n¡1) e¡kh sin(ky) dk
º cosh [kh] ¡ k sinh [kh]
0
(Keil-17a)
¼º 2n
0 sinh [º 0 (z ¡ h)]
ªnj (y; z; t) = +ej!t An ¢ sin(º 0 y)
cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h]
(Keil-17b)
Total Potentials Only the complex constant An with (0 6 n 6 1) in the potential has
to be determined:
1
X h 0 0 0
i
©(y; z; t) = [Anr + jAnj ] ©nr1 (y; z; t) + ©nrad (y; z; t) + j©nj (y; z; t)
n=0
X1 n h 0 i h io
0 0 0 0 0
= Anr ©nr1 + ©nrad ¡ Anj ©nj + j Anj (©nr1 + ©nrad ) + Anr ©nj
n=0
1
X © £ ¤ª
j!t
= e Anr ['nr1 + 'nrad ] ¡ Anj 'nj + j Anj ('nr1 + 'nrad ) + Anr 'nj
n=0
(Keil-18)
and
1
X h 0 0 0
i
ª(y; z; t) = [Anr + jAnj ] ªnr1(y; z; t) + ªnrad (y; z; t) + jªnj (y; z; t)
n=0
1 n
X h 0 i h io
0 0 0 0 0
= Anr ªnr1 + ªnrad ¡ Anj ªnj + j Anj (ªnr1 + ªnrad ) + Anr ªnj
n=0
96 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
1
X © £ ¤ª
j!t
= e Anr [à nr1 + à nrad ] ¡ Anj à nj + j Anj (à nr1 + à nrad ) + Anr à nj
n=0
(Keil-19)
The coe¢cients Anr and Anj with (0 6 n 6 1) have to be determined in such a way
that the instantaneous boundary conditions on the body surface have been ful…lled. These
coe¢cients are dimensional and it is very practical to determine them for the amplitude of
the ‡ow velocity V¹ ; also if they then have the dimension [L2n+1 ]:
Ar + jAj ³ ´
Ar + jAj = V¹ = ¹ A0 + jA0
V
V¹ r j
0 0
Then, An 'n and An à n have the dimensions of a length [L].
(" #
º
1
X
m
'0r1 = e¡ºz ° + ln(ºr) + Re f(z + iy)m g cos(ºy)
m=1
m ¢ m!
" # )
y X1
ºm
+ arctan + Im f(z + iy)m g sin(ºy) (Keil-21)
z m=1 m ¢ m!
(" #
X1
º m
à 0r1 = e¡ºz ° + ln(ºr) + Re f(z + iy)m g sin(ºy)
m=1
m ¢ m!
" # )
y X1
ºm
¡ arctan + Im f(z + iy)m g cos(ºy) (Keil-21a)
z m=1 m ¢ m!
XM
(m ¡ 1)! m ¡ºz y
'0r1 = m r 2m
Re f(z + iy) g + ¼e sin(ºy)
m=1
º jyj
XM
(m ¡ 1)! y
à 0r1 = m 2m
Im f(z + iy)m g ¡ ¼e¡ºz cos(ºy)
m=1
º r jyj
(m¡1)!
Mind you that º m r2m
f(z + iy)m g is semi–convergent.
© ª º © ª
'nr1 = (¡1)n (2n ¡ 1)! Re (y + iz)¡2n + Im (y + iz)¡(2n¡1)
2n ¡ 1
(Keil-22)
© ª º © ª
à nr1 = (¡1)n (2n ¡ 1)! Im (y + iz)¡2n ¡ Re (y + iz)¡(2n¡1)
2n ¡ 1
(Keil-22a)
For the expansion of the remaining potential parts use has been made of the following
relations as derived in Appendix 2:
X1
k 2t © ª
cosh [kz] cos(ky) = Re (z + iy)2t
t=0
(2t)!
X1
k 2t © ª
sinh [kz] sin(ky) = Im (z + iy)2t
t=0
(2t)!
X1
k 2t+1 © ª
sinh [kz] cos(ky) = Re (z + iy)2t+1
t=0
(2t + 1)!
X1
k 2t+1 © ª
cosh [kz] sin(ky) = Im (z + iy)2t+1
t=0
(2t + 1)!
It is obvious that:
1
X G(2t + 1) © © ª © ªª
ª0rad = ¡ (2t + 1) Im (z + iy)2t ¡ º Im (z + iy)2t+1 (Keil-23a)
t=0
(2t + 1)!
The function:
Z1
k t e¡kh
G(t) = dk
(k ¡ º) [º cosh [kh] ¡ k sinh [kh]]
0
Z1
(k 2 ¡ º 2 ) k 2(n¡1) e¡kh
'nrad = dk ¢
(k ¡ º) (º cosh [kh] ¡ k sinh [kh])
0
( )
X1
k 2t+1 © ª X1
k 2t © ª
¢ º Re (z + iy)2t+1 ¡ k Re (z + iy)2t
t=0
(2t + 1)! t=0
(2t)!
81
X 1
1 <Z k 2t+2n+1 e¡kh
= dk
(2t + 1)! : (k ¡ º) (º cosh [kh] ¡ k sinh [kh])
t=0 0
9
Z1 2t+2n¡1 ¡kh =
k e
¡º 2 dk ¢
(k ¡ º) (º cosh [kh] ¡ k sinh [kh]) ;
0
© © ª © ªª
¢ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t
X 1
G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1)
= ¢
t=0
(2t + 1)!
© © ª © ªª
¢ º Re (z + iy)2t+1 ¡ (2t + 1) Re (z + iy)2t (Keil-24)
It is clear that:
1
X G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1)
à nrad = ¢
t=0
(2t + 1)!
© © ª © ªª
¢ (2t + 1) Im (z + iy)2t ¡ º Im (z + iy)2t+1 (Keil-24a)
cosh2 (º 0 h)
'oj = +¼ ¢ (Keil-25)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t © ª X1
º 2t+1 © ª
0 0
¢ Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1
t=0
(2t)! t=0
(2t + 1)!
4.2. THEORY OF KEIL 99
cosh2 (º 0 h)
à oj = ¡¼ ¢ (Keil-25a)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t © ª X1
º 2t+1 © ª
0 0
¢ Im (z + iy)2t ¡ tanh [º 0 h] Im (z + iy)2t+1
t=0
(2t)! t=0
(2t + 1)!
º 2n
0
'nj = ¡¼ ¢ (Keil-26)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t © ª X1
º 2t+1 © ª
0 0
¢ Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1
t=0
(2t)! t=0
(2t + 1)!
¡ ¢
= ¡º 2(n¡1)
0 º 20 ¡ º 2 ¢ 'oj
º 2n
0
à nj = +¼ ¢ (Keil-26a)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t © ª X1
º 2t+1 © ª
0 2t 0 2t+1
¢ Im (z + iy) ¡ tanh [º 0 h] Im (z + iy)
t=0
(2t)! t=0
(2t + 1)!
2(n¡1) ¡ 2 ¢
= ¡º 0 º 0 ¡ º 2 ¢ à oj
Function G(t)
The function:
Z1
k t e¡kh £ 1¡t ¤
G(t) = dk with unit: L
(k ¡ º) (º cosh [kh] ¡ k sinh [kh])
0
has two singular points: k = º and k = º 0 . Thus, it is not possible to solve this integral
directly.
First, this integral will be normalized:
0
G (t) = G(t) ¢ ht¡1
Z1
ut e¡u
= du
(u ¡ ºh) (ºh cosh [u] ¡ u sinh [u])
0
A substitution of:
&p
w = u + iv = 2 ¢ ei¾
2
provides:
I
wt e¡w
J = dw
(w ¡ ºh) (ºh cosh [w] ¡ w sinh [w])
Z1 Z Z Z Z
= ...du + ...dw + ...dw + ...dw + ...dw
0 I II III IV
100 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Z1
= ...du + JI + JII + JIII + JIV
0
= 0 (Keil-27a)
JI and JII are imaginary because they are residues and JIII = 0 for R ! 1.
So, it remains:
Z1
ut e¡u
du = ¡ Re fJIV g
(u ¡ ºh) (ºh cosh [u] ¡ u sinh [u])
0
8 9
<Z t ¡w
we =
= ¡ Re dw
: (w ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) ;
IV
0
G (t) =
8 9
<Z t ¡w
w e (w~ ¡ ºh) (ºh cosh [w] ~ ¡ w~ sinh [w])
~ =
¡ Re dw
: (w ¡ ºh) (w~ ¡ ºh) (ºh cosh [w] ¡ w sinh [w]) (ºh cosh [w]
~ ¡ w~ sinh [w])
~ ;
IV
(Keil-28)
4.2. THEORY OF KEIL 101
t¼
= ¡ cos ¢
0 4 ©¡ ¢ ¡ ¢ ª 1
2 (ºh)2 1 ¡ tan t¼4 ©(cos & + e¡& ) + 1 + ª tan t¼
4
sin &
B ¡4ºh& cos & + tan t¼4 ¡sin & C
Z1 B
Bµ & ¶t ©¡ ¢ ¢ ª C
C
B +& 2 1 + tan t¼4 (cos & ¡ e¡& ) ¡ 1 ¡ tan t¼4 sin & C
B p C d&
B 2 f& 2 ¡ 2ºh (& ¡ ºh)g ¢ C
0 B C
@ ¢ fcosh [&] (2º 2 h2 + & 2 ¡ 2ºh& tanh [&]) A
+ (2º 2 h2 ¡ & 2 ) cos & + 2ºh& sin &g
Because t is always odd:
Z1 µ ¶t
0 t¼ 4 (ºh)2 sin & + 2& 2 (cos & ¡ e¡& ) ¡ 4ºh& (cos & + sin &) &
G (t) = ¡ cos p d&
4 denominator 2
0
for t = 1, 5, 9, ...
Z1 µ ¶
t¼ 4 (ºh)2 (cos & + e¡& ) ¡ 2& 2 sin & ¡ 4ºh& (cos & ¡ sin &) & t
= ¡ cos p d&
4 denominator 2
0
for t = 3, 7, 11, ...
0 0
For t > 1 the function G (t) becomes …nite. However, G (1) does not converge for ºh ! 0;
the integral increases monotone with decreasing ºh. This will be investigated …rst.
Z1
0 ue¡u
G (1) = G(1) = du (Keil-28a)
(u ¡ ºh) (ºh cosh [u] ¡ u sinh [u])
0
This integral converges fast for small ºh values. This will be approximated by:
Z1
ue¡u
lim G1 (1) = du (Keil-29)
ºh!0 (u ¡ ºh) (ºh ¡ u2 )
0
This can be written as:
8
< Z1
1 e¡u
lim G1 (1) = lim du
ºh!0 ºh!0 : 1 ¡ ºh u ¡ ºh
0
Z1
1 e¡u
+ p ³p ´ p du
2 ºh ºh ¡ 1 u ¡ ºh
0
9
Z1 ¡u =
1 e
+ p ³p ´ p du
2 ºh ºh + 1 u + ºh ;
0
From:
Z1 ( )
e¡u
1
X am
du = ¡e¡a ¢ ° + ln jaj +
u¡a m=1
m ¢ m!
0
102 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
follows:
( " #
¡e¡ºh X1
(ºh)m
lim G1 (1) = lim ° + ln (ºh) +
ºh!0 ºh!0 1 ¡ ºh m=1
m ¢ m!
p " m
#
e¡ ºh 1 X1
(ºh) 2
+ p ³ p ´ ° + 2 ln (ºh) +
2 ºh 1 ¡ ºh m=1
m ¢ m!
p " #9
(ºh) =
m
e ºh
1
1
X 2
¡ p ³ p ´ ° + 2 ln (ºh) + (¡1)m
2 ºh 1 + ºh m=1
m ¢ m! ;
( " #
¡e¡ºh X1
(ºh)m
= lim ° + ln (ºh) +
ºh!0 1 ¡ ºh m=1
m ¢ m!
³ p ´ ¡pºh " #
m
1 + ºh e 1 p X1
(ºh) 2
+ p ° + ln (ºh) + ºh +
2 ºh (1 ¡ ºh) 2 m=2
m ¢ m!
³ p ´ p " #9
1 ¡ ºh e ºh (ºh) 2 =
m
1 p X1
¡ p ° + ln (ºh) ¡ ºh + (¡1)m
2 ºh (1 ¡ ºh) 2 m=2
m ¢ m! ;
( " #
¡e¡ºh X1
(ºh)m
= lim ° + ln (ºh) +
ºh!0 1 ¡ ºh m=1
m ¢ m!
2 hp i 3
1 4 hp i sinh ºh · 1
¸
+ cosh ºh ¡ p 5 ° + ln (ºh)
1 ¡ ºh ºh 2
hp i hp i
cosh ºh sinh ºh
+ ¡
1 ¡ ºh " 1 ¡ ºh
m¡1 m
1 p X1
(ºh) 2 p X1
(ºh) 2
¡ ºh ¡ ºh
+ e +e
2 (1 ¡ ºh) m=2
m ¢ m! m=2
m ¢ m!
m¡1 m
#)
p X1
(ºh) 2 p 1
X (ºh) 2
¡e ºh (¡1)m + e ºh (¡1)m
m=2
m ¢ m! m=2
m ¢ m!
= 1 ¡ ° ¡ ln (ºh) (Keil-29a)
Hydrodynamic Loads
The hydrodynamic loads can be found from an integration of the pressures on the hull of
the oscillating body in (previously) still water. With a known potential, these pressures
can be found from the linear part of the instationary pressures as follows from the Bernoulli
equation:
@©
pdyn = p ¡ pstat = ¡½
@t
4.2. THEORY OF KEIL 103
= ¡j!½©
The potential is in-phase with the oscillation velocity. To obtain the phase of the pressures
with respect to the oscillatory motion a phase shift of -900 is required, which means a
multiplication with -j. Then the pressure is:
pdyn = ¡½!©
The hydrodynamic force on the body is equal to the integrated pressure on the body. This
is - in the two-dimensional case - a force per unit length.
The vertical force becomes:
Z
FV = FV r + jFV j = pdy
S
h 0 0
i
¹
= ½! V FV r + jFV j (Keil-30)
1
X Z h 0 ³ 0 ´i
0 0
¹
= ¡½! V ej!t
Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dy
n=0
S
The real part of this force is equal to the hydrodynamic mass coe¢cient times the oscillatory
acceleration, from which the hydrodynamic mass coe¢cient follows:
FV r F¹V r
m" = = ¹
b !V
or non-dimensional:
m" F¹V r
CV = ¼ 2 = ¼ 2 ¹ (Keil-30a)
½8B ½ 8 B !V
The imaginary part of the force must be equal to the hydrodynamic damping coe¢cient
times the oscillatory velocity, from which the damping coe¢cient follows:
¯ ¯
¯F¹V j ¯
NV = ¹
V
Instead of this coe¢cient, generally the ratio between amplitude of the radiated wave ¹³
and the oscillatory motion z¹ will be used. The energy balance provides:
2
³¹ !2 NV
A¹2V = 2 = ¢
z¹ ½g 2 ¢ cgroup
!º 0 sinh [2º 0 h]
= ¢ NV
½g 2º 0 h + sinh [2º 0 h]
º2 cosh2 [º 0 h]
= ¢ F¹V j (Keil-30b)
½! V¹ º 0 h + sinh [º 0 h] cosh [º 0 h]
In deep water becomes the hydrodynamic mass for º ! 0 in…nite, because potential
(Keil-21) becomes:
104 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
0
m
lim CV = lim ¼ 2
º!0 º!0 ½ B
" 8 #
8 X1
1
= 2
¡° ¡ ln (ºr) +
¼ n=1
n (4n2 ¡ 1)
The hydrodynamic mass for º ! 0 in shallow water remains …nite. Because the multi-
potentials - just as in deep water - provide …nite contributions, the radiation potential
has to be discussed only, which is decisive (in…nite mass) in deep water. the borderline
change-over ”deep to shallow” water provides for this radiation potential:
¹³ !¼ cosh2 [º 0 h]
A¹V = = A0 ¢
z¹ g¹z º 0 h + sinh [º 0 h] cosh [º 0 h]
º¼ cosh2 [º 0 h]
= A0 ¢
!¹z º 0 h + sinh [º 0 h] cosh [º 0 h]
A0 cosh2 [º 0 h]
= ¹ ¼º
V º 0 h + sinh [º 0 h] cosh [º 0 h]
0 º 0 sinh [º 0 h] cosh [º 0 h]
= A0 ¼
º 0 h + sinh [º 0 h] cosh [º 0 h]
Because:
0
A¼
lim 0 = 1
º!0 B
it follows:
º 0 sinh [º 0 h] cosh [º 0 h]
lim A¹V = B
º!0 º 0 h + sinh [º 0 h] cosh [º 0 h]
4.2. THEORY OF KEIL 105
and:
dA¹V dA¹V
lim = lim
º!0 dº º 0 !0 dº 0
B 1
= lim
2 º 0 !0 º 0 h + sinh [º 0 h] cosh [º 0 h]
= 1
F¹V j
0
F¹V i
"HT =1 = arctan ¹ = arctan ¹ 0 = 0
FV r FV r
This condition is ful…lled only when F¹V j = 0 or F¹V r = m" =½ = 1. However, F¹V j is …nite:
0 0 0
0 F¹V j NV g 2 ¹2 A¹2V
F¹V j = = = A =
½! V¹ ½! !4 V º2
Because limº 0 !0 A¹V = BV follows limº 0 !0 F¹V j = B 2 . The term F¹V r = m" =½ has to be
0 0
in…nite.
The …nite value of the hydrodynamic mass at shallow water is physically hard to interpret.
A full explanation is not given here. However, it has been shown here that the result makes
some sense. At shallow water can the wave (even p in an incompressible ‡uid) not travel
with an in…nite velocity; its maximum velocity is gh In case of long waves at shallow
water, the energy has the same velocity. From that can be concluded that at low decreasing
frequencies the damping part in the hydrodynamic force will increase. This means that:
F¹V i
0
Wave Loads
The wave forces on the restrained body in waves consists of forces F1 in the undisturbed
incoming waves (Froude-Krylov hypothesis), and the forces caused by the disturbance of
the waves by the body, one part F2 in phase with the accelerations of the water particles
and another part F3 in phase with the velocity of the water particles:
FE = F1 + F2 + jF3
106 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
These forces will be determined from the undisturbed wave potential ©W and the dis-
turbance potential ©S . As mentioned before, for ¹ 6= 900 only an approximation will be
found.
FE = F1 + F2 + jF3 =
Z
= ¡½! [©W + ©S ] dy
S
Z ½¹
j!t ³! ¡jº 0 x cos ¹
= ¡½!e e (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹)
º
S
1 h
)
X 0 0
³ 0 0
´i
+V¹ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dy (Keil-31)
n=0
In here:
Z
½! 2 ³¹ j(!t¡º 0 x cos ¹)
F1 = ¡ e f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹)g dy
º
S
1 h
X i
0 0
F2 ¹
= ¡½! V ej!t
Anr 'nr ¡ Anj 'nj dy
n=0
X1 h i
0 0
F3 = ¡½! V¹ ej!t Anr 'nj + Anj 'nr dy
n=0
Z
F¹1 1
E¹1 = ¹ = ¡ B f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) cos (º 0 y sin ¹)g dy
½g ³B
S
º Xh 0 i
1
F¹2 0
E¹2 = ¹ = ¡ A '
nr nr ¡ Anj nj dy
'
½g ³B B n=0
º Xh 0 i
1
F¹3 0
E¹3 = ¹ = ¡B Anr 'nj + Anj 'nr dy
½g ³B n=0
In case of ¹ = 900 , so beam waves, the theory of [Haskind, 1957] can be used too to
determine the amplitudes E¹1 , E¹2 and E¹3 . When ©W = 'W ej!t is the potential of the
incoming wave and ©S = 'S ej!t is the potential of the disturbance by the body at a large
distance from the body with a velocity amplitude V¹ = 1, then:
Zh µ ¶
@' @'
FE = ¡½!e j!t
'w ¡' w dz (Keil-32)
@y @y
0
From the previous subsections follows the asymptotic expression for the disturbance po-
tential in still water with V¹ = 1:
¼ cosh2 [º 0 h]
'y!1 = (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y) ¢
º 0 h + sinh [º 0 h] cosh [º 0 h]
" 1 ³
#
0 0 ¡ 2 ¢ X 0 0
´
2(n¡1)
¢ A0r + jA0j ¡ º 0 ¡ º 2 Anr + jAnj º 0
n=1
2¼ cosh2 [º 0 h]
FE = ¡½g ¹³º 0 ej!t ¢
º 0 h + sinh [º 0 h] cosh [º 0 h]
Zh
¢ (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z])2 dz ¢
"0 1 ³
#
0 0 ¡ ¢X 0 0
´
¢ A0r + jA0j ¡ º 20 ¡ º 2
Anr + jAnj º 2(n¡1)
0
n=1
" 1 ³
#
0 0 ¡ ¢X 0 0
´
2(n¡1)
= ¡½g ¹³¼ej!t 2
A0r + jA0j ¡ º 0 ¡ º 2
Anr + jAnj º 0
n=1
Non-dimensional:
© ª " #
Re F¹E ¼ 0 ¡ 2 ¢X1
0 2(n¡1)
E¹1 + E
¹2 = ¹ =¡ A0r ¡ º 0 ¡ º 2
Anr º 0
½g ³B B n=1
(Keil-32a)
© ª " #
Im F¹E ¼ 0 ¡ 2 1
¢X 0 2(n¡1)
¹3
E = =¡ A0j ¡ º 0 ¡ º 2
Anj º 0
¹
½g ³B B n=1
Solution
The Lewis transformation of a cross section is given by:
Ship
Ship Lewis Form Lewis Form
The yet unknown complex coe¢cients An (0 6 n 6 1), the source strengths of the by the
‡ow generated singularity, can be determined by substituting the stream function (Keil-
19) and the coordinates of the cross section in the relevant boundary conditions (Keil-5a)
through (Keil-6b).
h i h i
ª (y; z)body ; t = ªbody (y; z)body ; t (Keil-34)
To determine the unknowns An , an equal number of equations has to be formulated.
Because Lewis forms are used only here, a simple approach is possible.
All stream function parts and boundary conditions can be given as a Fourier series:
1
X
Ãn = fcnm sin [2mµ] + dnm cos [(2m + 1) µ]g
m=0
or with
2µ 2 X
1
(2m + 1)2
cos [(2m + 1) µ] = 1 ¡ + £ ¤ sin (2kµ)
¼ ¼ k=1 k 4k 2 ¡ (2m + 1)2
by:
1
X
à n = an0 + anm sin [2mµ]
m=1
0
The solution of the by equating coe¢cients generated equations provide the unknowns An .
4.2. THEORY OF KEIL 109
1. For sway:
· ¸
@©
= [vn ]body
@n body
· ¸
@© dy @© dz
= ¢ ¡ ¢
@z ds @y ds body
· ¸
dª
= ¡
ds body
· ¸
dz
= ¡U¹ e j!t
ds body
or:
¹ j!t [dz]
dªbody = Ue body
2. For roll:
· ¸
@©
= [vn ]body
@n body
· ¸
@© dy @© dz
= ¢ ¡ ¢
@z ds @y ds body
· ¸
dª
= ¡
ds body
· ¸
j!t dr
= ¡¹'!e r
ds body
or:
'!ej!t [r dr]body
dªbody = ¡¹
from which follows:
¹ ! j!t £ 2
' ¤
ªbody (y; z; t) = ¡ e y + z 2 body + C (Keil-35b)
2
110 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
For the restrained body in waves, only the force in the horizontal direction and the moment
about the longitudinal axis of the body will be calculated. One gets in beam waves only
the in y point-symmetric part of the potential and the in y symmetric part of the stream
function of the wave (see Appendix 1), respectively:
¹³!
[ªs (y; z; t)]body = ¢ ejwt ¢ [(sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) ¢ cos(º 0 y)]body (Keil-36a)
º
and in oblique waves:
h i ¹
³!
ª~ s (x; y1 ; z1 ; t) = ¢ ejwt ¢ sin(º 0 x cos ¹) ¢ (Keil-36b)
body º
[sin ¹ cos(º 0 y1 sin ¹) ¢ (sinh [º 0 z1 ] ¡ tanh [º 0 h] cosh [º 0 z1 ])
3
Zy1
¡º 0 (1 ¡ sin2 ¹) sin(º 0 y sin ¹) ¢ (sinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]) dy 5
0 body
Potentials
2-D Radiation Potential In a similar way as equation (Keil-10) for heave, the three
dimensional radiation potential for sway and roll can be derived as:
Z
+L=2 Z1
©0r (x; y; z; t) = ej!t A0 (») cos [kx (x ¡ »)] ¢ (Keil-37)
¡L=2 0
Z1 p 2 £p ¤
kx + ky2 ¢ cosh kx2 + ky2 ¢ (z ¡ h)
¢ £p ¤ p £p ¤ sin(ky y)dkx dky d»
º cosh kx2 + ky2 ¢ h ¡ kx2 + ky2 ¢ sinh kx2 + ky2 ¢ h
0
With the Sommerfeld radiation Condition (Keil-4) and Appendix 3 the total radiation
potential becomes:
Z1
©nr1 (y; z; t) = ¡ej!t An (k + º)k 2n¡1 e¡kz sin(ky) dk
0
(Keil-41)
Z1
º sinh [kz] ¡ k cosh [kz]
©nrad (y; z; t) = ¡ej!t An (k + º)k 2n¡1 e¡kh sin(ky) dk
º cosh [kh] ¡ k sinh [kh]
0
(Keil-41a)
¼º 2n+1
0 cosh [º 0 (z ¡ h)]
©nj (y; z; t) = +ej!t An ¢ sin(º 0 y)
cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h]
(Keil-41b)
Z1
ªnr1 (y; z; t) = +ej!t An (k + º)k 2n¡1 e¡kz cos(ky) dk
0
112 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
(Keil-42)
Z1
º cosh [kz] ¡ k sinh [kz]
ªnrad (y; z; t) = ¡ej!t An (k + º)k 2n¡1 e¡kh cos(ky) dk
º cosh [kh] ¡ k sinh [kh]
0
(Keil-42a)
¼º 2n+1
0 sinh [º 0 (z ¡ h)]
ªnj (y; z; t) = +ej!t An ¢ cos(º 0 y)
cosh [º 0 h] º 0 h + sinh [º 0 h] cosh [º 0 h]
(Keil-42b)
Total Potentials With exception of the complex constant An with (0 6 n 6 1), the
potential is known:
X
1
© ª
j!t
©(y; z; t) = e Anr ['nr1 + 'nrad ] ¡ Anj 'nj + j [Anj ('nr1 + 'nrad )] + Anr 'nj
n=0
(Keil-43a)
and
1
X © ª
j!t
ª(y; z; t) = e Anr [à nr1 + à nrad ] ¡ Anj à nj + j [Anj (à nr1 + à nrad )] + Anr à nj
n=0
(Keil-43b)
with:
h 0 0
i
Ar + jAj = U¹ Ar + jAj for sway
h 0 0
i
¹ ! Ar + jAj
' for roll
0 0 0
An has the dimension [L2n+2 ]. Then, An 'n and An à n have for sway the dimension [L] and
for roll the dimension [L2 ].
0
The determination of the coe¢cients An follow from the conditions at the body contour.
4.2. THEORY OF KEIL 113
( " #
y ºm
1
X
'0r1 = ¡ 2 + ºe¡ºz sin(ºy) ° + ln(ºr) + Re f(z + iy)m g
y + z2 m=1
m ¢ m!
" #)
y X ºm
1
¡ cos(ºy) arctan + Im f(z + iy)m g (Keil-44)
z m=1 m ¢ m!
( " #
z X1
º m
à 0r1 = + 2 ¡ ºe¡ºz cos(ºy) ° + ln(ºr) + Re f(z + iy)m g
y + z2 m=1
m ¢ m!
" #)
y X ºm
1
+ sin(ºy) arctan + Im f(z + iy)m g (Keil-44a)
z m=1 m ¢ m!
y XM
(m ¡ 1)! y
'0r1 = ¡ 2 2
+º m 2m
Im f(z + iy)m g ¡ ¼ºe¡ºz cos(ºy)
y +z m=1
º r jyj
z XM
(m ¡ 1)! y
à 0r1 = + 2 2
¡º m 2m
Re f(z + iy)m g ¡ ¼ºe¡ºz sin(ºy)
y +z m=1
º r jyj
(m¡1)!
Mind you that º m r2m
f(z + iy)m g is semi–convergent.
n © ª º © ªo
n+1 ¡(2n+1) ¡2n
'nr1 = (¡1) (2n)! Re (y + iz) + Im (y + iz)
2n
(Keil-45)
n © ª º © ªo
à nr1 = (¡1)n+1 (2n)! Im (y + iz)¡(2n+1) ¡ Re (y + iz)¡2n
2n
(Keil-45a)
1 ½
X ¾
G(2t + 3) © 2t+1
ª G(2t + 1) © 2t
ª
'0rad = Im (z + iy) ¡º Im (z + iy)
t=0
(2t + 1)! (2t)!
(Keil-46)
1 ½
X ¾
G(2t + 3) © ª G(2t + 1) © ª
ª0rad = Re (z + iy)2t+1 ¡ º Re (z + iy) 2t
t=0
(2t + 1)! (2t)!
(Keil-46a)
1 ½
X G(2t + 2n + 3) ¡ º 2 ¢ G(2t + 2n + 1) © ª
'nrad = Im (z + iy)2t+1
t=0
(2t + 1)!
114 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
¾
G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) © 2t
ª
¡º Im (z + iy)
(2t)!
(Keil-47)
1 ½
X G(2t + 2n + 3) ¡ º 2 ¢ G(2t + 2n + 1) © ª
à nrad = Re (z + iy)2t+1
t=0
(2t + 1)!
¾
G(2t + 2n + 1) ¡ º 2 ¢ G(2t + 2n ¡ 1) © 2t
ª
¡º Re (z + iy)
(2t)!
(Keil-47a)
cosh2 (º 0 h)
'oj = +¼º 0 ¢ (Keil-48)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t+1 © ª X1
º 2t © ª
0 0
¢ Im (z + iy)2t+1 ¡ tanh [º 0 h] Im (z + iy)2t
t=0
(2t + 1)! t=0
(2t)!
cosh2 (º 0 h)
à oj = +¼º 0 (Keil-48a)
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t+1 © ª X1
º 2t © ª
0 0
¢ Re (z + iy)2t+1 ¡ tanh [º 0 h] Re (z + iy)2t
t=0
(2t + 1)! t=0
(2t)!
2(n¡1) ¡ ¢
'nj = +º 0 º 20 ¡ º 2 ¢ 'oj (Keil-49)
2(n¡1) ¡ 2 ¢
à nj = +º 0 º 0 ¡ º 2 ¢ à oj (Keil-49a)
Zero-Frequency Potential
[Grim, 1956] and [Grim, 1957] give for the horizontal motions at zero frequency the complex
potential:
1
X n
' + ià = An (y + iz)¡(2n+1) + (Keil-50)
n=0
)
X1 n o
(y + iz + i2mh)¡(2n+1) + (y + iz ¡ i2mh)¡(2n+1)
m=1
( µ ¶
1
X 1
X 2n + p ¡ ¡i2µ
p ¢p
¡i(2n+1)µ
' + ià = An e (¡1) ae + be¡i4µ
n=0 p=0
p
1 n
X
+i2 (i2mH)¡(2n+1) ¢
m=1
µ ¶ µ +iµ ¶ ))
X
1
2p + 2n + 1 ¡iµ
e + ae + be¡i3µ 2p+1
¢ (¡1)p
p=0
2p + 1 2mH
4.2. THEORY OF KEIL 115
(1 ( p µ ¶µ ¶ )
1
X X p
X 2n + p p p¡l l ¡i(2n+2p+2l+1)µ
= An (¡1) a be
n=0 p=0 l=0
p l
" 2p+1 l
1 X
X 1 XX
+2 (¡1)p+n (2mH)¡2(p+n+1) (Keil-50a)
m=1 p=0 l=0 k=0
µ ¶µ ¶µ ¶ ¸¾
2n + 2p + 1 2p + 1 l l¡k k i(2p¡2l¡2k+1)µ
a b e
2p + 1 l º0
These sums converge as long as:
1+a+b B
2HT > = (Keil-51a)
1¡a+b 2T
Hydrodynamic Loads
The hydrodynamic force at sway oscillations in still water becomes:
¹ j!t ¢
FQ = FQr + jFQj = ¡½! Ue (Keil-52)
1 Z h
X ³ 0 ´i
0 0 0
¢ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dz
n=0 S
0 0
Of course, the coe¢cients Anr and Anj of sway and roll will di¤er.
Fictive moment levers are de…ned by:
M ¹ Qr ¹
¹ Qr =
H ¹ Qj = MQj
H
¹ !m"
U ¹ Q
UN
¹ Rr I "!2 '
¹ ¹ Rj = NR ! '
¹B
H = ¹ H ¹
FRr 2FRj
m" F¹Qr
CH = = ¹ ¼T2
½ ¼2 T 2 ½! U 2
¹³ 2 º 2
cosh2 [º 0 h]
A¹2Q = 2 = ¢ F¹Qj
y¹ ½! U¹ º 0 h + sinh [º 0 h] cosh [º 0 h]
¹ Qr
H M ¹ Qr
=
T T ¢ F¹Qr
¹ Qj
H M ¹ Qj
= (Keil-54)
T T ¢ F¹Qj
I" M ¹ Rr
CR = ¼ 4 =
½8T ¹ ¼8 T 4
½! 2 '
2
³¹ 4º 2 cosh2 [º 0 h]
A¹2R = 2 = ¢ ¹ Rj
M
¹ 2 B4
' ½! 2 '
¹ B 2 º 0 h + sinh [º 0 h] cosh [º 0 h]
¹ Rr
H M ¹ Rr
=
T T ¢ F¹Rr
¹
HRj M ¹ Rj
= (Keil-54a)
T T ¢ F¹Rj
Wave Loads
The wave loads are separated in contributions of the undisturbed wave and di¤raction:
FE = F1 + F2 + jF3 =
4.2. THEORY OF KEIL 117
Z
= ¡½! [©W + ©S ] dz
S
Z ½ ¹
j!t ³!
= ¡½!e ¡ e¡jº 0 x cos ¹ (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)
º
S
1 h
)
X 0 0
³ 0 0
´i
+U¹ Anr 'nr ¡ Anj 'nj + j Anr 'nj + Anj 'nr dz (Keil-55)
n=0
ME = M1 + M2 + jM3 =
Z ½ ¹
j!t ³!
= ¡½!e ¡ e¡jº 0 x cos ¹ (cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)
º
S
1 h
)
X 0 0
³ 0 0
´i
+U¹ A 'nr ¡ A 'nj + j A 'nj + A 'nr
nr nj nr [ydy + zdz]
nj (Keil-56)
n=0
Z
½! 2 ¹³ j(!t¡º 0 x cos ¹)
F1 = + e f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g dz
º
S
1 h
X i
0 0
F2 ¹
= ¡½! U ej!t
Anr 'nr ¡ Anj 'nj dz
n=0
X1 h i
0 0
F2 = ¡½! U¹ ej!t Anr 'nj + Anj 'nr dz
n=0
Z
½! 2 ¹³ j(!t¡º 0 x cos ¹)
M1 = + e f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g ¢
º
S
¢ [ydy + zdz]
1 h
X i
0 0
M2 ¹
= ¡½! U ej!t
Anr 'nr ¡ Anj 'nj [ydy + zdz]
n=0
1 h
X i
0 0
M2 ¹
= ¡½! U ej!t
Anr 'nj + Anj 'nr [ydy + zdz]
n=0
Dimensionless:
Z
¹1 = F¹1 tanh [º 0 h]
E =+ f(cosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]) sin (º 0 y sin ¹)g dz
½g ¹³º 0 Ax ºAx
S
X1 h i
¹2 = F¹2 tanh [º 0 h] 0 0
E = ¡ Anr 'nr ¡ Anj 'nj dz
½g ¹³º 0 Ax Ax n=0
tanh [º 0 h] X h 0 i
1
¹3 F¹3 0
E = = ¡ A ' + Anj nr dz
' (Keil-55a)
½g ¹³º 0 Ax Ax nr nj
n=0
118 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
¹W r
H M¹1 + M¹2 ¹Wj
H M¹3
= ¡ ¢ = (Keil-56a)
T T ¢ F¹1 + F¹2 T T ¢ F¹3
The Haskind-Newman relations are valid too here:
© ª " #
Re F¹E ¼ 0 ¡ 2 ¢X1
0 2(n¡1)
¹1 + E¹2 2
E = ¹ 0 Ax = + Ax A0r + º 0 ¡ º
½g ³º
Anr º 0
n=1
(Keil-55b)
© ª " #
Im F¹E ¼ 0 ¡ 2 1
¢X 0 2(n¡1)
E¹3 = =+ A0j + º 0 ¡ º 2
Anj º 0
½g ¹³º 0 Ax Ax n=1
Solution
To determine the unknowns An , an equal number of equations has to be formulated.
Because Lewis forms are used only here, a simple approach is possible.
All stream function parts and boundary conditions can be given as a Fourier series:
1
X
Ãn = fcnm sin [(2m + 1) µ] + dnm cos [2mµ]g
m=0
or with
16 X
1
m2
cos [2mµ] = 1 + sin [(2k + 1) µ]
¼ k=0 (2k + 2m + 1) (2k ¡ 2m + 1) (2k + 1)
in:
1
X
Ãn = anm sin [(2m + 1) µ]
m=0
0
The solution of the by equating coe¢cients generated equations provide the unknowns An .
4.2. THEORY OF KEIL 119
4.2.5 Appendices
Appendix 1: Undisturbed Wave Potential
The general expression of the complex potential of a shallow wave, travelling in the negative
y-direction, is:
¹
³!
©W = cosh [º 0 (z ¡ h)] cos (º 0 y + !t)
º cosh [º 0 h]
¹
³!
= ej!t ¢ cosh [º 0 (z ¡ h)] ejº 0 y
º cosh [º 0 h]
¹
³!
= ej!t fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g ejº 0 y
º
¹
³!
ªW = ¡ sinh [º 0 (z ¡ h)] sin (º 0 y + !t)
º cosh [º 0 h]
¹
³!
= j ej!t ¢ sinh [º 0 (z ¡ h)] ejº 0 y
º cosh [º 0 h]
¹³!
= j ej!t fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g ejº 0 y
º
For the vertical motions is the in y symmetrical part of the potential signi…cant. For the
horizontal motions is the in y point symmetrical part - multiplied with j, so a phase shift
of 900 - important.
¹
³!
©W V = + ej!t fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g cos (º 0 y) (Keil-A2)
º
¹
³!
ªW V = ¡ ej!t fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g sin (º 0 y) (Keil-A2a)
º
¹
³!
©W H = ¡ ej!t fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g sin (º 0 y) (Keil-A3)
º
¹
³!
ªW H = ¡ ej!t fsinh [º 0 z] ¡ tanh [º 0 h] cosh [º 0 z]g cos (º 0 y) (Keil-A3a)
º
When the wave travels in the xw -direction, the potential becomes:
120 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
¹
³!
©W = ej(!t¡º 0 xw ) fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g
º
With:
xw = x cos ¹ ¡ y sin ¹
yw = x sin ¹ + y cos ¹
¹
³!
©W V = + ej(!t¡º 0 x cos ¹) fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g cos (º 0 y sin ¹) (Keil-A4)
º
and for the horizontal motions in:
¹
³!
©W H = ¡ ej(!t¡º 0 x cos ¹) fcosh [º 0 z] ¡ tanh [º 0 h] sinh [º 0 z]g sin (º 0 y sin ¹) (Keil-A5)
º
p y
z § iy = y 2 + z 2 ¢ e§i arctan z
= r ¢ e§i¯
the body away) is given - is also possible in another way. This approach is based on work
carried out by Rayleigh and is given in the literature by [Lamb, 1932] for an in…nite water
depth.
In this approach, an viscous force ½¹w will be included in the Euler equations, ¹ is the
dynamic viscosity and w is the velocity. Because the ‡uid is assumed to be non-viscous,
in a later stage this dynamic viscosity ¹ will be set to zero.
From the Euler equation follows with this viscosity force the with time changing pressure
change:
· ¸
@p @© @© @ 2 ©
= ½ lim g ¡¹ ¡ 2
@t ¹!0 @z @t @t
From this follows the approach as given in a subsection before for the two-dimensional
radiation potential:
The …rst integral leads to equation (Keil-12a) and the second integral can be expanded as
follows:
©0rad +j©0jad =
³ ´
Z1 !¹
º ¡ j g sinh [kz] ¡ k cosh [kz]
e¡kh
= lim ¢ ³ ´ cos(ky) dk
¹!0 º ¡ j !¹g
¡k º ¡ j !¹
cosh [kh] ¡ k sinh [kh]
0 g
("µ ¶ © 2t+1 ª © ª#
X1
!¹ Re (z + iy) Re (z + iy)2t
= lim º¡j ¡ ¢
¹!0
t=0
g (2t + 1)! (2t)!
9
Z1 ¡kh 2t+1 =
e k
¢ ³ ´ h³ ´ i dk
º ¡ j !¹
¡ k º ¡ j !¹
cosh [kh] ¡ k sinh [kh] ;
0 g g
("µ ¶ © 2t+1 ª © ª#
X 1
!¹ Re (z + iy) Re (z + iy)2t
= ¡ lim º¡j ¡ ¢
¹!0
t=0
g (2t + 1)! (2t)!
¢ [G (2t + 1) + jH (2t + 1)]g
The function:
4.2. THEORY OF KEIL 123
Z1
e¡kh kt
G (t) + iH (t) = lim ³ ´ ¢ h³ ´ i dk
¹!0
0
k ¡ º ¡ i !¹
g
º ¡ i !¹
g
cosh [kh] ¡ k sinh [kh]
will be normalized as done before:
0 0
G (t) + iH (t) = ht¡1 [G (t) + iH (t)]
Z1
ut e¡u
= lim ³ ´ h³ ´ i du
¹!0 !¹h !¹h
0
u ¡ ºh ¡ i g
ºh ¡ i g
cosh [hu] ¡ u sinh [hu]
This is a complex integral and must be solved in the complex domain with w = u + iv .
The integrand has singularity for:
!¹h
w1 = ºh ¡ i and w2
g
where w2 is the solution of the equation:
µ ¶
!¹h
ºh ¡ i cosh [w] ¡ w sinh [w] = 0
g
see …gure 4.8-a.
8 R 9
I ¡w t <Z Z Z =
e w
lim dw = lim ...du + ...dw + ...dw
¹!0 (w ¡ w1 ) (w1 cosh w ¡ w sinh w) ¹!0 : ;
0 I II
8 R 9
<Z =
= lim ...du + JI + JII
¹!0 : ;
0
= 0
124 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Thus:
Z1
0 0 ut e¡u
G (t) + iH (t) = lim ³ ´ h³ ´ i du
¹!0
0
u ¡ ºh ¡ i !¹h
g
ºh ¡ i !¹h
g
cosh [hu] ¡ u sinh [hu]
= lim fJI + JII g
¹!0 R!1
Because:
lim JI = 0
R!1
follows:
Z
0 0 wt e¡w
G (t) + iH (t) = ¡ lim dw
¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w])
II
will be calculated as done before. This integral has no singularity and the boundary ¹ ! 0
can be passed before integration; see …gure 4.8-b.
The imaginary part of this integral:
8 9
< Z t ¡w =
0 we
H (t) = ¡ Im lim dw
:¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) ;
II
can be calculated numerically in an analog way. It is also possible to solve this integral
independently by using another integral path:
I
wt e¡w
lim dw = lim fJII + JIII + JIV g
¹!0 (w ¡ w1 ) (w1 cosh [w] ¡ w sinh [w]) ¹!0
Re fJII g = ¡ Re fJIII g
Im fJII g = + Im fJIII g
Then it follows:
4.2. THEORY OF KEIL 125
½ ¾
0
H (t) = ¡ Im lim JII
¹!0
( © ª © ª)
1
X Re (z + iy)2t Re (z + iy)2t+1
©0jad = H (2t + 1) ¡º
t=0
(2t)! (2t + 1)!
X1 ½ 2 ¾
º 2t
0 cosh [º 0 h]
= ¼ ¡ º 2t ¢
t=0
º 0 h + sinh [º 0 h] cosh [º 0 h]
( © ª © ª)
Re (z + iy)2t Re (z + iy)2t+1
¢ ¡º
(2t)! (2t + 1)!
¼ cosh2 [º 0 h]
= ¢
º 0 h + sinh [º 0 h] cosh [º 0 h]
(1 )
X º 2t © ª X1 2t+1
º0 © ª
¢ 0
Re (z + iy)2t ¡ tanh [º 0 h] Re (z + iy)2t+1
t=0
(2t)! t=0
(2t + 1)!
¡¼e¡ºz cos (ºy)
4.3.1 Notation
º = ! 2 =g = wave number
ºk = k-th irregular wave number for ajoint interior problem
»j = abscissa of the j-th input point
½ = density of ‡uid
©(m) = velocity potential for th m-th mode of oscillation
! = radian frequency of oscillation
!k = k-th irregular frequency for adjoint interior problem
(k-th eigen-frequency)
4. The condition of the normal velocity component of the ‡uid at the surface of the
oscillating cylinder being equal to the normal component of the forced velocity of
the cylinder. i.e., if vn is the component of the forced velocity of the cylinder in the
direction of the outgoing unit normal vector ~n, then
~ (m) = vn
~n ¢ r© (Frank-5)
this kinematic boundary condition on the oscillating body surface the kinematic
boundary condition being satis…ed at the mean (rest) position of the cylindrical
surface.
5. The radiation condition that the disturbed surface of the ‡uid takes the form of
regular progressive outgoing gravity waves at large distances from the cylinder.
According to Wehausen and Laitone, the complex potential at z of a pulsating point source
of unit strength at the point ³ in the lower half plane is:
8 9
< Z1 ¡ik(z¡b³) =
1 e
G¤ (z; ³; t) = ln(z ¡ ³) ¡ ln(z ¡ b
³) + 2P V dk cos !t
2¼ : º¡k ;
0
n o
b
¡ e¡iº(z¡³) sin !t
(Frank-6)
b !2
z = x + iy ³ = » + i´ ³ = » ¡ i´ º=
g
Letting:
8 9
< Z1 ¡ik(z¡b
³) =
1 e
G(z; ³) = Re ln(z ¡ ³) ¡ ln(z ¡ b
³) + 2P V dk
2¼ : º¡k ;
0
n o
¡iº(z¡b
³)
¡i Re e (Frank-8)
then:
© ª
H(x; y; »; ´; t) = Re G(z; ³) ¢ e¡i!t (Frank-9)
130 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
Equation (Frank-9) satis…es the radiation condition and also the equations (Frank-1)
through (Frank-4).
Another expression satisfying all these conditions is:
¼ © ª
H(x; y; »; ´; t ¡ ) = Re iG(z; ³) ¢ e¡i!t (Frank-10)
2!
Since the problem is linear, a superposition of equations (Frank-9) and (Frank-10) results
in the velocity potential:
8 9
<Z =
(m) ¡i!t
© (x; y; t) = Re Q(s) ¢ G(z; ³) ¢ e ¢ ds (Frank-11)
: ;
C0
where C0 is the submerged contour of the cylindrical cross section at its mean (rest) position
and Q(s) represents the complex source density as a function of the position along C0 .
Application of the kinematic boundary condition on the oscillating cylinder at z yields:
8 9
< Z =
~
Re (~n ¢ r) Q(s) ¢ G(z; ³) ¢ ds = 0
: ;
C0
(Frank-12)
8 9
< Z =
Im ~
(~n ¢ r) Q(s) ¢ G(z; ³) ¢ ds = A(m) ¢ ! ¢ n(m)
: ;
C0
where A(m) denotes the amplitude of oscillation and n(m) the direction cosine of the normal
velocity at z on the cylinder. Both A(m) and n(m) depend on the mode of motion of the
cylinder, as will be shown in the following section.
The fact that Q(s) is complex implies that equations (Frank-12) represent a set of coupled
integral equations for the real functions RefQ(s)g and ImfQ(s)g. The solution of these
integral equations and the evaluation of the kernel and potential integrals are described in
the following section and in Appendices B and C, respectively.
The coordinates, length and angle associated with the j-th segment are identi…ed by the
subscript j, whereas the corresponding quantities for the re‡ected image in the third quad-
rant are denoted by the subscript ¡j, so that by symmetry » ¡j = ¡» j and ´ ¡j = ¡´j for
1 6 j 6 N + 1.
Potentials and pressures are to be evaluated at the midpoint of each segment. The coordi-
nates of the midpoint of the i-th segment are:
» i + » i+1 ´i + ´i+1
xi = and yi = (Frank-13)
2 2
for 1 6 i 6 N.
The length of the i-th segment is:
q¡ ¢2 ¡ ¢2
jsi j = » i+1 ¡ » i + ´ i+1 ¡ ´ i (Frank-14)
while the angle made by the i-th segment with the positive x-axis is given by:
½ ¾
´ i+1 ¡ ´ i
®i = arctan (Frank-15)
» i+1 ¡ » i
The outgoing unit vector normal to the cross section at the i-th midpoint (xi ; yi ) is:
where ~i and ~j are unit vectors in the directions of increasing x and y, respectively.
The cylinder is forced into simple harmonic motion with radian frequency !, according to
the displacement equation:
S (m) = A(m) ¢ cos !t (Frank-17)
for m = 2, 3 or 4, corresponding to sway, heave or roll, respectively. The rolling motions
are about an axis through a point (0; y0 ) in the symmetry plane of the cylinder.
In the translational modes, any point on the cylinder moves with the velocity:
The normal components of the velocity vi (m) = ~ni ¢~v(m) at the midpoint of the i-th segment
(xi ; yi ) are:
De…ning:
vi (m)
ni (m) =
A(m) ! sin !t
then, consistent with the previously mentioned notation, the direction cosines for the three
modes of motion are:
Equations (Frank-22) illustrate that heaving is symmetrical, i.e., n¡i (3) = ni (3) . Swaying
and rolling, however, are anti-symmetrical modes, i.e., n¡i (2) = ¡ni (2) and n¡i (4) = ¡ni (4) .
Equations (Frank-12) are applied at the midpoints of each of the N segments and it is
assumed that over an individual segment the complex source strength Q(s) remains con-
stant, although it varies from segment to segment. With these stipulations, the set of
coupled integral equations (Frank-12) becomes a set of 2N linear algebraic equations in
the unknowns:
© ª
Re Q(m) ¢ (sj ) = Qj (m)
© ª
Im Q(m) ¢ (sj ) = QN+j (m)
4.3. THEORY OF FRANK 133
where pa (m) and pv (m) are the hydrodynamic pressures in-phase with the displacement and
in-phase with the velocity, respectively and ½ denotes the density of the ‡uid.
As indicated by the notation of equations (Frank-24) and (Frank-25), the pressure as well
as the potential is a function of the oscillation frequency !.
The hydrodynamic force or moment (when m = 4) per unit length along the cylinder,
necessary to sustain the oscillations, is the integral of p(m) ¢n(m) over the submerged contour
of the cross section C0 . It is assumed that the pressure at the i-th midpoint is the mean
pressure for the i-th segment, so that the integration reduces to a summation, whence:
N
X
(m)
M (!) = 2 pa (m) (xi ; yi ; !) ¢ ni (m) ¢ jsi j (Frank-26)
i=1
N
X
(m)
N (!) = 2 pv (m) (xi ; yi ; !) ¢ ni (m) ¢ jsi j (Frank-27)
i=1
on the surface of the ‡uid outside the cylinder, whereas for extremely large frequencies,
i.e., when ! ! 1, the free-surface condition becomes the ”impulsive” surface condition:
©=0 (Frank-50)
on y = 0 outside the cylinder.
Equations (Frank-2), (Frank-4) and (Frank-5) remain valid for both asymptotic cases. The
radiation condition is replaced by a condition of boundedness at in…nity.
Therefore, there is a Neumann problem for the case ! ! 0 and a mixed problem when
! ! 1. The appropriate complex potentials for a source of unit strength at a point ³ in
the lower half plane are:
1 n o
G0 (z; ³) = ln(z ¡ ³) + ln(z ¡ b
³) + K0 (Frank-51)
2¼
and:
1 n o
G1 (z; ³) = ln(z ¡ ³) ¡ ln(z ¡ b³) + K1 (Frank-52)
2¼
for the Neuman and mixed problems, respectively, where K0 and K1 are constants not
yet speci…ed.
Let:
which - after application at the N segmental midpoint - yields a set of N linear algebraic
equations in the N unknown source strengths Qj .
It remains to be shown whether these two problems are, in the language of potential theory,
well posed, i.e., whether the solutions to these problems lead to unique forces or moments.
The mixed problem raises no di¢culty, since as z ! 1, G1 (z; ³) ! 0. In fact K1 = 0,
which can be inferred from the pulsating source-potential equation (Frank-8) by letting
º ! 1.
Considering the Neumann problem, note that the constant K0 in the Green’s function
equation (Frank-51) yields by integration an additive constant K to the potential. However,
for a completely submerged cylinder the cross sectional contour C0 is a simply closed curve,
so that the contribution of K in integrating the product of the pressure with the direction
cosine of the body-surface velocity vanishes. For partially submerged bodies C0 is no longer
(m) (m)
closed. But since n¡i = ¡ni for m being even,
4.3. THEORY OF FRANK 135
Z
K n(m) ds = 0
C0
For a rectangular cylinder with beam B and draft T , the irregular wave numbers may be
easily obtained by separation of variables in the Laplace equation. Separating variables
gives the eigen-frequencies:
µ ¶ · ¸
k¼x k¼y
à k = Bk sin sinh for: k = 1; 2; 3; ...... etc.
B B
where Bk are Fourier coe¢cients to be determined from an appropriate boundary condition.
Applying the free surface condition (Frank-2) on y = T for 0 < x < B, the eigen-wave
numbers (or irregular wave numbers):
· ¸
k¼ k¼T
ºk = coth (Frank-28)
B B
136 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
are obtained for k = 1, 2, 3, ..., etc. In particular, the lowest such irregular wave number
is given by:
· ¸
¼ ¼T
k1 = coth (Frank-29)
B B
Keeping T …xed in equation (Frank-29) but letting B vary and setting b = ¼=B, then from
the Taylor expansion:
" #
1 bT (bT )3
b coth [bT ] = b + ¡ + ......
bT 3 45
1
º1 = (Frank-30)
T
a relation that John proved for general shapes complying with the restrictions previously
outlined. For a beam-to-draft ratio of B=T = 2:5: º 1 = 1:48, while for B=T = 2: º 1 = 1:71.
At an irregular frequency the matrix of in‡uence coe¢cients of equations (Frank-23) be-
comes singular as the number of de…ning points per cross section increases without limit,
i,e., as N ! 1. In practice, with …nite N , the determinant of this matrix becomes very
small, not only at the irregular frequency but also at an interval about this frequency. This
interval can be reduced by increasing the number of de…ning points N for the cross section.
Most surface vessels have nearly constant draft over the length of the ship and the maximum
beam occurs at or near amidships, where the cross section is usually almost rectangular, so
that for most surface ships the …rst irregular frequency ! 1 is less for the midsection than
for any other cross section.
For a ship with a 7:1 length-to-beam and a 5:2 beam-to-draft, the …rst irregular wave
encounter frequency - in non-dimensional form with L denoting the ship length - occurs
at:
s
L
!1 t 5:09
g
2000 1250
surge/sway
1000
1500
Damping (ton/s)
750
Mass (ton)
heave
1000
heave
500
surge/sway
500
250
0 0
0 1 2 3 0 1 2 3
Frequency (rad/s Frequency (rad/s
)
)
Figure 4.10: E¤ect of ”Lid-Method” on Irregular Frequencies
4.3.6 Appendices
Appendix A: Principle Value Integrals
The real and imaginary parts of the principle value integral:
Z1 b
e¡ik(z¡³)
PV dk
º¡k
0
Z1 b Z1 b
e¡ik(z¡³) e¡ik(z¡³) b
PV dk = dk ¡ i¼e¡iº(z¡³) (Frank-31)
º¡k º ¡k
0 y0
where the path of integration is the positive real axis indented into the upper half plane
about k = º.
Note that º = ! 2 =g > 0, Im fzg <³0 and´Im f³g 6 0.
The transformation ! = i (k ¡ º) z ¡ b ³ converts the contour integral on the right hand
side of equation (Frank-31) to:
Z1 b Z1
e¡ik(z¡³) b e¡w
dk = ¡e¡iº(z¡³) dw
º¡k w
y 0 ¡iº(z¡b
³)
h i µ 2¼i x ¡ » > 0
¶
¡iº(z¡b
= ¡e E1 ¡iº(z ¡ b
³)
³) + for
0 x¡» < 0
n h i
= ¡e¡iº(z¡³) ° + ln ¡iº(z ¡ b
³) (Frank-32)
138 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
h in 9
1 (¡1)
X
n
¡iº(z ¡ b
³) = µ ¶
2¼i x¡» > 0
+ + for
n ¢ n! ; 0 x¡» < 0
n=1
where ° = 0:5772::: is the well known Euler-Mascheroni constant and the value of E1 is
de…ned by Abramowitz and Stegun.
Setting:
n 2 o3
¯
¯
¯
¯ Im ¡iº(z ¡ b
³)
b
r = ¯¡iº(z ¡ ³)¯ and µ = arctan 4 n o5 + ¼
b
Re ¡iº(z ¡ ³)
Z1 b
e¡ik(z¡³)
PV dk = eº(y+´) [cos º (x ¡ ») ¡ i sin º (x ¡ »)] ¢
º¡k
0
(" #
X1
rn cos (nµ)
¢ ° + ln r + (Frank-33)
n=1
n ¢ n!
"µ ¶ X 1
#)
µ x¡» > 0 rn sin (nµ)
+i for +
µ ¡ 2¼ x¡» < 0 n=1
n ¢ n!
Separating equation (Frank-33) into its real and imaginary parts yields:
Z1
ek(y+´) cos k (x ¡ »)
PV dk = eº(y+´) [C (r; µ) cos º (x ¡ »)]
º ¡k
0
+S (r; µ) sin º (x ¡ »)
(Frank-34)
Z1
ek(y+´) sin k (x ¡ »)
PV dk = eº(y+´) [C (r; µ) sin º (x ¡ »)]
º¡k
0
¡S (r; µ) cos º (x ¡ »)
provided that:
1
X rn cos (nµ)
C (r; µ) = ° + ln r +
n ¢ n!
n=1
X1 µ ¶
rn sin (nµ) µ x¡» > 0
S (r; µ) = µ + + for
n=1
n ¢ n! µ ¡ 2¼ x¡» < 0
4.3. THEORY OF FRANK 139
8 2 0 1
>
< Z ³ ´ Z1 ¡ik(z¡b³)
~ 6 1 1 e
ln(z ¡ ³) ¡ ln(z ¡ b
(m)
Iij = Re (~ni ¢ r) 4 @ ³) + P V dk A ds
>
: 2¼ ¼ º¡k
sj 0
(Frank-35)
0 3 9 1
Z Z1 >
=
m @ 1 ³ b
´ 1 e¡ik(z+³) A 7
¡ (¡1) ln(z + ³) ¡ ln(z + ³) + P V dk ds5
2¼ ¼ º¡k >
;
s¡j 0
z=zi
and:
8 2 3 9
>
< Z Z >
=
(m)
Jij = Re ~ 6
(~ni ¢ r) 4
b
e¡iº(z¡³) ds ¡ (¡1)m
7
e¡iº(z+³) ds5 (Frank-36)
>
: >
;
sj s¡j
z=zi
Considering the term containing ln (z ¡ ³), it is evident that the kernel integral is singular
when i = j, so that the indicated di¤erentiation cannot be performed under the integral
sign. However, in that case one may proceed as follows:
since:
³ = » + i´
d³ = d» + id´
= ds cos ®j + ids sin ®j
= ei®j ds
8 2 3 9 8 2 3 9
>
> Z >
> >
> ³
Z >
>
< = < d
j+1
=
~ 6
Re (~nj ¢ r)
7
4 ln (z ¡ ³) ds5 = Re ¡iei®j 6
4 e¡i®j 7
d³ ln (z ¡ ³)5
>
> >
> >
> dz >
>
: sj ; : ³j ;
z=zj z=zj
8 2 3 9
>
> ³ j+1
Z >
>
< d =
6 7
= Re ¡i 4 d³ ln (z ¡ ³) d³ 5
>
> dz >
>
: ³j ;
z=zj
140 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
0
Setting ³ = z ¡ ³, the last integral becomes:
8 2 3 9
>
> z¡³ j
Z >
>
< d 6 0 07
= ¡ ¢ ¡ ¢
Re ¡i 4 ln ³ d³ 5 = arg zj ¡ ³ j ¡ arg zj ¡ ³ j+1
>
> dz >
>
: z¡³ j+1 ;
z=zj
= ¼ (Frank-37)
8 2 3 9
>
< Z >
=
~ 6 7
(L1 ) = Re (~ni ¢ r) 4 ln (z ¡ ³) ds5
>
: >
;
sj
z=zi
v
u ¡ ¢ ¡ ¢
u xi ¡ » j 2 + yi ¡ ´ j 2
= sin (®i ¡ ®j ) ln ¡t ¢2 ¡ ¢2 (Frank-38)
xi ¡ » j+1 + yi ¡ ´ j+1
· ¸
yi ¡ ´ j yi ¡ ´ j+1
+ cos (®i ¡ ®j ) arctan ¡ arctan
xi ¡ » j xi ¡ » j+1
³ ´
b
For the integral containing the ln z ¡ ³ term, ds = ei®j db ³, so that:
8 2 3 9
>
< Z ³ ´ >
=
(L2 ) = Re ~ 6
(~ni ¢ r) 4 ln z ¡ b
7
³ ds5
>
: >
;
sj
z=zi
v
u ¡ ¢ ¡ ¢
u xi ¡ » j 2 + yi + ´j 2
t
= sin (®i + ®j ) ln ¡ ¢2 ¡ ¢2 (Frank-39)
xi ¡ » j+1 + yi + ´j+1
· ¸
yi + ´j yi + ´ j+1
+ cos (®i + ®j ) arctan ¡ arctan
xi ¡ » j xi ¡ » j+1
8 2 3 9
>
< Z Z1 b >
=
~ 6 e¡ik(zi ¡³) 7
(L5) = Re (~ni ¢ r) 4 ds ¢ P V dk 5
>
: º ¡k >
;
sj 0
z=zi
8 b
9
>
<
³j
Z Z1 ¡ik(zi ¡b
³)
>
=
d e
= Re ¡iei(®i +®j ) db
³ PV dk
>
: db³ º ¡k >
;
b
³ j+1 0
2 ¡ ¢
Z1
ek(yi +´j ) cos k xi ¡ » j
= + sin (®i + ®j ) 4P V dk
º¡k
0
4.3. THEORY OF FRANK 141
¡ ¢ 3
Z1 k(yi +´ j+1 )
e cos k xi ¡ » j+1
¡P V dk 5
º¡k
0
2 ¡ ¢
Z1
ek(yi +´j ) sin k xi ¡ » j
¡ cos (®i + ®j ) 4P V dk (Frank-40)
º ¡k
0
¡ ¢ 3
Z1 k(yi +´ j+1 )
e sin k xi ¡ » j+1
¡P V dk 5
º¡k
0
The …rst integral on the right hand side of equation (Frank-36) becomes:
8 2 3 9
>
< Z >
=
(L7) = Re ~ 6
(~ni ¢ r) 4 e ³) 7
¡iº(z¡b
ds5 (Frank-41)
>
: >
;
sj
z=z
£ º(yi +´ ) ¡ i ¢ ¡ ¢¤
= ¡ sin (®i + ®j ) e j cos º xi ¡ » j ¡ eº(yi +´j+1 ) cos º xi ¡ » j+1
£ ¡ ¢ ¡ ¢¤
+ cos (®i + ®j ) eº(yi +´j ) sin º xi ¡ » j ¡ eº(yi +´j+1 ) sin º xi ¡ » j+1
The kernel integrals over the image segments are obtained from equations (Frank-38)
through (Frank-41) by replacing » j , » j+1 and ®j with » ¡j = ¡» j , » ¡(j+1) = ¡» j+1 and
®¡j = ¡®j , respectively.
8 2 3
>
< Z Z1
b
1 X
N ¡ik(z ¡³)
4ln(zi ¡ ³) ¡ ln(zi ¡ b e i
©(m) (xi ; yi ; t) = Qj Re ³) + 2P V dk 5 ds
2¼ j=1 >
: º¡k
sj 0
2 3 9
Z Z1 ¡ik(zi +³) >
=
m 4 b e 5
¡ (¡1) ln(zi + ³) ¡ ln(zi + ³) + 2P V dk ds
º¡k >
;
s¡j 0
8 9
XN >
<Z Z >
= cos !t
¡iº(zi ¡b
³) m ¡iº(zi +³)
¨ QN+j Re e ds ¡ (¡1) e ds ¢
>
: >
; sin !t
j=1 sj s¡j
(Frank-42)
8 9
>
<Z > ·
= ¡ ¢ q¡ ¢2 ¡ ¢2
Re ln(zi ¡ ³)ds = + cos ®j xi ¡ » j ln xi ¡ » j + yi ¡ ´j + » j ¡ » j¡1
>
: >
;
sj
142 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
¡ ¢ q¡ ¢2 ¡ ¢2
¡ xi ¡ » j+1 ln xi ¡ » j+1 + yi ¡ ´ j+1
¸
¡ ¢ yi ¡ ´ j ¡ ¢ yi ¡ ´j+1
¡ yi ¡ ´j arctan + yi ¡ ´ j+1 arctan
xi ¡ » j xi ¡ » j+1
· q
¡ ¢ ¡ ¢2 ¡ ¢2
+ sin ®j yi ¡ ´ j ln xi ¡ » j + yi ¡ ´ j + ´j ¡ ´j¡1
¡ ¢ q¡ ¢2 ¡ ¢2
¡ yi ¡ ´j+1 ln xi ¡ » j+1 + yi ¡ ´ j+1 (Frank-43)
¸
¡ ¢ yi ¡ ´ j ¡ ¢ yi ¡ ´ j+1
+ xi ¡ » j arctan ¡ xi ¡ » j+1 arctan
xi ¡ » j xi ¡ » j+1
b
³ j+1 b
³ j+1
Z Z1 ¡ik(z¡b
³) Z1 Z
e dk b
ds ¢ P V dk = e i®j
¢ PV e¡ik(z¡³) db
³
º¡k º¡k
b
³j 0 0 b
³j
b
³ j+1
Z1 Z
e¡ikz b
= e i®j
¢ PV dk eik³ db
³
º¡k
0 b
³j
Z1 b b
i®j e¡ikz eik³ j+1 ¡ eik³ j
= ¡e ¢ PV dk
º ¡k k
0
where the change of integration is permissible since only one integral requires a principle
value interpretation.
After dividing by º and multiplying by º ¡ k +k under the integral sign, the last expression
becomes:
21 3
i®jZ ikb
³ j+1 ikb
³j Z1 ¡ik(z¡b³ j+1 ) Z1 ¡ik(z¡b³ j )
ie 4 e ¡e e e
¡ e¡ikz dk + P V dk ¡ P V dk 5
º k º¡k º¡k
0 0 0
(Frank-44)
Regarding the …rst integral in equation (Frank-44) as a function of z:
Z1 ikb
³ j+1 b
¡ikz e ¡ eik³ j
F (z) = e dk (Frank-45)
k
0
Di¤erentiating equation (Frank-45) with respect to z gives:
81 9
<Z Z1 =
0 b b
F (z) = ¡i e¡ik(z¡³ j+1 ) dk ¡ e¡ik(z¡³ j ) dk
: ;
0 0
4.3. THEORY OF FRANK 143
1 1
= ¡
z ¡b
³j z ¡b
³ j+1
So:
³ ´ ³ ´
F (z) = ln z ¡ b
³ j ¡ ln z ¡ b
³ j+1 + º (Frank-46)
8 9
>
<Z Z1 ¡ik(zi ¡b
³)
>
=
e
(K5) = Re ds ¢ P V dk
>
: º¡k >
;
sj 0
½
iei®j h
= Re ¡ ln(zi ¡ b ³ j ) ¡ ln(zi ¡ b ³ j+1 )
º
39
Z1 ¡ik(zi ¡b³ j+1 ) Z1 ¡ik(zi ¡b³ j ) =
e e 5
+P V dk ¡ P V dk
º¡k º¡k ;
0 0
8 v
2 u ¡
< ¢ ¡ ¢
1 u xi ¡ » j 2 + yi + ´ j 2
= sin ®j 4ln t ¡ ¢2 ¡ ¢2 (Frank-47)
º: xi ¡ » j+1 + yi + ´ j+1
¡ ¢ ¡ ¢ 3
Z1 k(yi +´ j+1 ) Z1 k(yi +´j )
e cos k xi ¡ » j+1 e cos k xi ¡ » j
+P V dk ¡ P V dk 5
º ¡k º¡k
0 0
·
yi + ´ j yi + ´j+1
+ cos ®j arctan ¡ arctan
xi ¡ » j xi ¡ » j+1
Z k(yi +´j )
1 ¡ ¢ Z1 k(yi +´j+1 ) ¡ ¢ 39 =
e sin k xi ¡ » j e sin k xi ¡ » j+1
+P V dk ¡ P V dk 5
º ¡k º¡k ;
0 0
The integration of the potential component in-phase with the velocity over sj gives:
8 9
>
<Z >
=
¡iº(zi ¡b
³)
(K7) = Re e ds (Frank-48)
>
: >
;
sj
1 © º(yi +´j ) £ ¡ ¢ ¤ £ ¡ ¢ ¤ª
= e sin º xi ¡ » j ¡ ®j ¡ eº(yi +´j+1 ) sin º xi ¡ » j+1 ¡ ®j
º
144 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
By using a Lewis transformation of this equivalent longitudinal section to the unit circle,
¤ ¤
the two-dimensional potential mass M11 and damping N11 can be calculated in an analog
manner as has been described for the two-dimensional potential mass and damping of sway,
0 0
M22 and N22 .
With these two-dimensional values, the total potential mass and damping of surge are
de…ned by:
¤
M11 = B ¢ M11
¤
N11 = B ¢ N11
When using this formula for surge, the total potential mass of surge is de…ned by:
¤
M11 (! = 0) = B ¢ M11 (! = 0)
M11 (S&K) = ® ¢ ½r
So, the potential mass and damping of surge are de…ned by:
¤
M11 = B ¢ M11 ¢¯
¤
N11 = B ¢ N11
To obtain a uniform approach during all ship motions calculations, the cross sectional two-
dimensional values of the hydrodynamic mass and damping have to be obtained. Based
on the results of numerical 3-D studies with a Wigley hull form, a proportionality of both
the two-dimensional hydrodynamic mass and damping with the absolute values of the
derivatives of the cross sectional areas Ax in the xb -direction is assumed:
0
j dAx
dxb
j 0
j dAx
dxb
j
M11 =R ¢ M11 N11 =R ¢ N11
j dAx
dxb
jdxb j dAx
dxb
jdxb
L L
146 CHAPTER 4. 2-D POTENTIAL COEFFICIENTS
200
300 3000 3750
Sway - Roll
200 2000 3500 Roll - Sway
100 Midship section ' '
M 24 = M 42
of a containership
100 1000 3250
0 0 0 3000
0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5
100 100
With the exception of the roll motions, the three results are very close. The roll motion de-
viation, predicted with the Lewis conformal mapping method, is caused by the description
of the ”bilge” by the simple Lewis transformation.
A disadvantage of Frank’s method can be the large computing time, when compared with
Ursell-Tasai’s method. Generally, it is advised to use Ursell-Tasai’s method with 10 para-
meter close-…t conformal mapping. For submerged sections, bulbous sections and sections
with an area coe¢cient, ¾ s , less than 0.5, Frank’s pulsating source method should be used.
Chapter 5
Viscous Damping
The strip theory is based on the potential ‡ow theory. This holds that viscous e¤ects are
neglected, which can deliver serious problems when predicting roll motions at resonance
frequencies. In practice, viscous roll damping e¤ects can be accounted for by empirical for-
mulas. For surge and roll, additional damping coe¢cients have to be introduced. Because
of these additional contributions to the damping are from a viscous origin mainly, it is not
possible to calculate the total damping in a pure theoretical way.
147
148 CHAPTER 5. VISCOUS DAMPING
Then the total surge damping coe¢cient at forward ship speed V becomes:
2
B11t = 2Ct ¢ ½r 3 V
with:
2
A44 = a44 + OG ¢ a42 + OG ¢ a24 + OG ¢ a22 (potential mass coe¢cient)
2
B44 = b44 + OG ¢ b42 + OG ¢ b24 + OG ¢ b22 (potential damping coe¢ciernt)
B44V = b44V (viscous damping coe¢cient)
C44 = ½gr ¢ GM (restoring term coe¢cient)
while for zero forward speed:
a42 = a24 and b42 = b24
These equations can be rewritten as:
Ä + 2º ¢ Á_ + ! 2 ¢ Á = 0
Á 0
in which:
B44 + B44v
2º = (quotient of damping and moment of inertia)
Ixx + A44
C44
! 20 = (natural roll frequency squared)
Ixx + A44
The non-dimensional roll damping coe¢cient, ·, is given by:
º B44 + B44v
·= = p
! 0 2 (Ixx + A44 ) ¢ C44
This damping coe¢cient is written as a fraction betweenpthe actual damping coe¢cient,
B44 + B44v , and the critical damping coe¢cient, B44cr = 2 (Ixx + A44 ) ¢ C44 ; so for critical
damping: ·cr = 1.
Herewith, the equation of motion can be re-written as:
Ä + 2·! 0 ¢ Á_ + ! 2 ¢ Á = 0
Á 0
Suppose the vessel is de‡ected to an initial heel angle, Áa , in still water and then released.
The solution of the equation of motion of this decay becomes:
µ ¶
¡ºt º
Á = Áa e cos ! Á t + sin ! Á t
!Á
Then the logarithmic decrement of the motion is:
½ ¾
Á(t)
ºTÁ = ·! 0 TÁ = ln
Á(t + TÁ )
Because ! 2Á = ! 20 ¡ º 2 for the natural frequency oscillation and the damping is small so
that º 2 ¿ ! 20 , one can neglect º 2 here and use ! Á ¼ ! 0 ; this leads to:
! 0 TÁ ¼ ! Á TÁ = 2¼
The non-dimensional total roll damping is given now by:
½ ¾
1 Á(t) !0
·= ln = (B44 + B44v ) ¢
2¼ Á(t + TÁ ) 2C44
The non-potential part of the total roll damping coe¢cient follows from the average value
of · by:
2C44
B44v = · ¢ ¡ B44
!0
150 CHAPTER 5. VISCOUS DAMPING
The decay coe¢cient · can therefore be estimated from the decaying oscillation by deter-
mining the ratio between any pair of successive (double) amplitudes. When the damping
is very small and the oscillation decays very slowly, several estimates of the decay can be
obtained from a single record. It is obvious that for a linear system a constant ·-value
should be found in relation to Áa .
Note that these decay tests provide no information about the relation between the potential
coe¢cients and the frequency of oscillation. Indeed, this is impossible since decay tests are
carried out at only one frequency: the natural frequency. These experiments deliver no
information on the relation with the frequency of oscillation.
The method is not really practical when º is much greater than about 0.2 and is in any
case strictly valid for small values of º only. Luckily, this is generally the case.
Be aware that this damping coe¢cient is determined by assuming an uncoupled roll motion
(no other motions involved). Strictly, this damping coe¢cient is not valid for the actual
coupled motions of a ship which will be moving in all directions simultaneously.
The successively found values for ·, plotted on base of the average roll amplitude, will
often have a non-linear behavior as illustrated in the next …gure.
while sometimes even a cubic roll damping coe¢cient, ·3 ¢ Á2a , has to be added to this
formula.
This non-linear behavior holds that during frequency domain calculations, the damping
term is depending on the - so far unknown - solution for the transfer function of roll:
Áa =³ a . With a known wave amplitude, ³ a , this problem can be solved in an iterative
manner. A less accurate method is to use a …xed Áa .
Generally, CV = 1:0 but for slender ships, like frigates, a suitable value for CV seems to
be: q
CV = 4:85 ¡ 3:00 ¢ GM F ull Scale
ZT Á ZT Á
(1) _ = B (2) ¢ jÁj
B44V ¢ Á_ ¢ Ádt 44V
_ Á_ ¢ Ádt
_
0 0
or:
(1)8 (2)
B44V =Áa ! ¢ B44V
3¼
For the estimation of the non-potential parts of the roll damping, use has been made of
work published by [Ikeda et al., 1978]. A few subordinate parts have been modi…ed and
this empiric method is called here the ”Ikeda method”.
The Ikeda method estimates the following linear components of the roll damping coe¢cient
of a ship:
B44S = a correction on the potential roll damping coe¢cient due to forward speed
B44F = the frictional roll damping coe¢cient
B44E = the eddy making roll damping coe¢cient
B44L = the lift roll damping coe¢cient
B44K = the bilge keel roll damping coe¢cient
So, the additional - mainly viscous - roll damping coe¢cient B44V is given by:
Ikeda, Himeno and Tanaka claim fairly good agreements between their prediction method
and experimental results. They conclude that the method can be used safely for ordinary
ship forms, which conclusion has been con…rmed by the author too. But for unusual ship
forms, for very full ship forms and for ships with a very large breadth to draft ratio the
method is not always accurate su¢ciently.
For numerical reasons three restrictions have been made here during the calculations:
- if, local, ¾ s > 0:999 then ¾ s = 0:999
- if, local, OG < ¡Ds ¾ s then OG = ¡Ds ¾ s
- if a calculated component of the viscous roll damping coe¢cient becomes less than
zero, this component will be set to zero.
154 CHAPTER 5. VISCOUS DAMPING
Notation
In this description of the Ikeda method the notation of the authors (Ikeda, Himeno and
Tanaka) is maintained as far as possible:
with:
Sf = L(1:7D + CB B)
The relation between the density, kinematic viscosity and temperture of fresh water and
sea water are given in …gure 5.4.
1030 2.0
Sea Water
1020
Kinematic Viscosity (m s)
1.5
2
Density (kg/m )
3
Sea Water
1010
Fresh Water
1.0
1000
Fresh Water
990 0.5
0 10 20 30 0 10 20 30
0 0
Temperature ( C) Temperature ( C)
Figure 5.4: Relation Between Density, Kinematic Viscosity and Temperature of Water
When eliminating the temperature of water, the kinematic viscosity can be expressed into
the density of water by the following relation in the kg-m-s system:
25 25
20 20
Kinematic Viscosity *10 (m s)
2
Temperature ( C)
Temperature ( C)
7
7
15 15
0
0
10 10
5 5
The …rst part in this expression represents the laminar ‡ow case. The second part has
been ignored by Ikeda, but has been included here.
Using this, the quadratic roll damping coe¢cient due to skin friction at zero forward ship
speed is expressed as:
(2) 1
B44F = ½rf3 Sf Cf
0 2
This frictional roll damping component increases slightly with forward speed. Semi-
theoretically, Tamiya deduced a modi…cation coe¢cient for the e¤ect of forward speed
on the friction component.
Accurate enough from a practical point of view, this results into the following formula for
the speed dependent frictional damping coe¢cient:
µ ¶
(2) 1 3 V
B44F = ½rf Sf Cf ¢ 1:0 + 4:1 ¢
2 !L
Then, the equivalent linear roll damping coe¢cient due to skin friction is expressed as:
µ ¶
8 1 3 V
B44F = Á ! ¢ ½rf Sf Cf ¢ 1:0 + 4:1 ¢
3¼ a 2 !L
Ikeda con…rmed the use of his formula for the three-dimensional turbulent boundary layer
over the hull of an oscillating ellipsoid in roll motion.
5.2. ROLL DAMPING 157
The value of ° around a cross section is approximated by the potential ‡ow theory for a
rotating Lewis form cylinder in an in…nite ‡uid.
An estimation of the local maximum distance between the roll axis and the hull surface,
rmax , has to be made.
Values of rmax (Ã) have to be calculated for:
0:5
à = à 1 = 0:0 and à = Ã2 = ³ ´
a1 (1+a3 )
cos 4a3
With these two results a value rmax and a value à follow from the conditions:
- for rmax (à 1 ) > rmax (à 2 ) : rmax = rmax (à 1 ) and à = à 1
- for rmax (à 1 ) < rmax (à 2 ) : rmax = rmax (à 2 ) and à = à 2
The relative velocity ratio ° on a cross section is obtained by:
p µ ¶
f3 ¼ 2Ms p 2
°= ³ ´ p ¢ rmax + a +b 2
2Ds ¾ s + OG H ¾ H
Ds 0 s
with:
Bs
H0 =
2Ds
As
¾s =
Bs Ds
Bs Ds
Ms = =
2(1 + a1 + a3 ) 1 ¡ a1 + a3
158 CHAPTER 5. VISCOUS DAMPING
With this a quadratic sectional eddy making damping coe¢cient for zero forward speed
follows from:
µ ¶2
(2) 0 1 4 rmax
B44E = ½D Cp ¢
0 2 s Ds
õ ¶µ ¶ µ ¶2 !
f1 rb OG f1 rb f1 rb
1¡ 1+ ¡ + f2 H0 ¡
Ds Ds Ds Ds
with:
The approximations of the local radius of the bilge circle rb are given as:
r
H0 (¾ s ¡ 1) Bs
rb = 2Ds for: rb < Ds and rb <
¼¡4 2
rb = Ds for: H0 > 1 and rb > Ds
Bs
rb = for: H0 < 1 and rb > H0 Ds
2
For three-dimensional ship forms the zero forward speed eddy making quadratic roll damp-
ing coe¢cient is found by an integration over the ship length:
Z
(2) (2) 0
B44E = B44E dxb
0 0
L
This eddy making roll damping decreases rapidly with the forward speed to a non-linear
correction for the lift force on a ship with a small angle of attack. Ikeda has analyzed this
forward speed e¤ect by experiments and the result has been given in an empirical formula.
With this the equivalent linear eddy making damping coe¢cient at forward speed is given
by:
8 (2) 1
B44E = ¢Áa !B44e0 ¢
3¼ 1 + K2
with:
V
K=
0:04 ¢ !L
CL
kN =
® µ ¶
2¼D B
= + Â 4:1 ¢ ¡ 0:045
L L
in which the coe¢cient  is given by Ikeda in relation to the amidships section coe¢cient
CM :
with:
hk
CD = 22:5 ¢ + 2:40
¼rk Áa fk
fk = 1:0 + 0:3 ¢ e¡160¢(1:0¡¾s )
The approximation of the local distance between the roll axis and the bilge keel rk is given
as: s
µ ¶ µ ¶2
rb 2 OG rb
rk = Ds H0 ¡ 0:293 ¢ + 1:0 + ¡ 0:293 ¢
Ds Ds Ds
The approximation of the local radius of the bilge circle rb in here is given before.
160 CHAPTER 5. VISCOUS DAMPING
Assuming a pressure distribution on the hull caused by the bilge keels, a quadratic sectional
roll damping coe¢cient can be de…ned:
Zh k
(2) 0 1
B44K = ½rk2 fk2 Cp lm dh
S 2
0
Ikeda carried out experiments to measure the pressure on the hull surface created by bilge
keels. He found that the coe¢cient Cp+ the pressure on the front face of the bilge keel does
not depend on the period parameter, while the coe¢cient Cp¡ of the pressure on the back
face of the bilge keel and the length of the negative pressure region depend on the period
parameter.
Ikeda de…nes an equivalent length of a constant negative pressure region S0 over the height
of the bilge keels, which is …tted to the following empirical formula:
S0 = 0:3 ¢ ¼fk rk Áa + 1:95 ¢ hk
The pressure coe¢cients on the front face of the bilge keel, Cp+ , and on the back face of
the bilge keel, Cp¡ , are given by:
hk
Cp+ = 1:20 and Cp¡ = ¡22:5 ¢ ¡ 1:20
¼fk rk Áa
and the sectional pressure moment is given by:
Zh k
¡ ¢
Cp lm ¢ dh = Ds2 ¡A ¢ Cp¡ + B ¢ Cp+
0
with:
A = (m3 + m4 )m8 ¡ m27
m34 (1 ¡ m1 )2 (2m3 ¡ m2 )
B = + + m1 (m3 m5 + m4 m6 )
3(H0 ¡ 0:215 ¢ m1 ) 6(1 ¡ 0:215 ¢ m1 )
rb
m1 =
Ds
¡OG
m2 =
Ds
m3 = 1:0 ¡ m1 ¡ m2
m4 = H0 ¡ m1
0:414 ¢ H0 + 0:0651 ¢ m21 ¡ (0:382 ¢ H0 + 0:0106)m1
m5 =
(H0 ¡ 0:215 ¢ m1 )(1 ¡ 0:215 ¢ m1 )
0:414 ¢ H0 + 0:0651 ¢ m21 ¡ (0:382 + 0:0106 ¢ H0 )m1
m6 =
(H0 ¡ 0:215 ¢ m1 )(1 ¡ 0:215 ¢ m1 )
S0
m7 = ¡ 0:25 ¢ ¼m1 for: S0 > 0:25 ¢ ¼rb
Ds
= 0:0 for: S0 < 0:25 ¢ ¼rb
m8 = m7 + 0:414 ¢ m1 for: S0 > 0:25 ¢ ¼rb
µ µ ¶¶
S0
= m7 + 1:414 ¢ m1 1 ¡ cos for: S0 < 0:25 ¢ ¼rb
rb
5.2. ROLL DAMPING 161
The equivalent linear total bilge keel damping coe¢cient can be obtained now by integrating
the two sectional roll damping coe¢cients over the length of the bilge keels and linearizing
the result: Z ³ ´
8 (2) 0 (2) 0
B44K = Áa ! B44K + B44K dxb
3¼ N S
Lk
Experiments of Ikeda have shown that the e¤ect of forward ship speed on this roll damping
coe¢cient can be ignored.
It may be noted that for full scale ships, because of the higher Reynolds number, the
frictional part of the roll damping is expected to be smaller than showed above.
162 CHAPTER 5. VISCOUS DAMPING
.
Chapter 6
Hydromechanical Loads
With an approach as mentioned before, a description will be given here of the determination
of the hydromechanical forces and moments for all six modes of motions.
In the ”Ordinary Strip Theory”, as published by [Korvin-Kroukovsky and Jacobs, 1957]
and others, the uncoupled two-dimensional potential hydromechanical loads in the direction
j are de…ned by:
D n 0 _¤ o ¤
Xh0 j = Mjj ¢ ³ hj + Njj0
¢ ³_ hj + XRS
0
(6.1)
Dt j
In the ”Modi…ed Strip Theory”, as published for instance by [Tasai, 1969] and others, these
loads become: ½µ ¶ ¾
D i 0 ¤
0
Xhj = Mjj ¡ Njj ¢ ³_ hj + XRS
0 0
Dt !e j
When assuming that the cross sectional hydromechanical force hold at a plane through
the local centroid of the cross section b, parallel to (xb ; yb ), equivalent longitudinal motions
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
163
164 CHAPTER 6. HYDROMECHANICAL LOADS
of the water particles, relative to the cross section of an oscillating ship in still water, are
de…ned by:
³ ¤h1 = ¡x + bG ¢ µ
¤ @bG
³_ h1 = ¡x_ + bG ¢ µ_ ¡ V ¢µ
@xb
¼ ¡x_ + bG ¢ µ_
¤ @bG _ @ 2 bG
³Äh1 = ¡Äx + bG ¢ ĵ ¡ 2V ¢µ+V2 ¢µ
@xb @x2b
x + bG ¢ µÄ
¼ ¡Ä
In here, bG is the vertical distance of the centre of gravity of the ship G above the centroid
b of the local submerged sectional area.
According to the ”Ordinary Strip Theory” the two-dimensional potential hydromechanical
force on a surging cross section in still water is de…ned by:
D n 0 _¤ o ¤
Xh0 1 = M11 ¢ ³ h1 + N110
¢ ³_ h1
Dt µ ¶
0
Ä ¤ dM11 ¤
0 0
= M11 ¢ ³ h1 + N11 ¡ V ¢ ¢ ³_ h1
dxb
According to the ”Modi…ed Strip Theory” this hydromechanical force becomes:
½µ ¶ ¾
0 D 0 i 0 _ ¤
Xh1 = M11 ¡ ¢ N11 ¢ ³ h1
Dt !e
µ 0
¶ µ 0
¶
V dN11 Ä ¤ dM11 ¤
= 0
M11 + 2 ¢ 0
¢ ³ h1 + N11 ¡ V ¢ ¢ ³_ h1
! e dxb dxb
This results into the following coupled surge equation:
with:
Z
0
a11 = + M11 ¢ dxb
L
¯ ¯
¯ Z ¯
¯V dN 0 ¯
+¯¯ 2 11
¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM11
b11 = + N11 ¡ V ¢ ¢ dxb + b11V
dxb
L
c11 = 0
a13 = 0
6.2. HYDROMECHANICAL FORCES FOR SWAY 165
b13 = 0
c13 = 0 Z
0
a15 = ¡ M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V 0
dN11 ¯
¯
+¯ 2 ¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM11
b15 = ¡ N11 ¡V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG
dxb
L
c15 = 0
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
A small viscous surge damping coe¢cient b11V , derived from the still water resistance
approximation of [Troost, 1955], has been added here.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled surge equation become:
Z
0
a11 = + M11 ¢ dxb
L
Z
0
b11 = + N11 ¢ dxb + b11V
L
c11 = 0
a13 = 0
b13 = 0
c13 = 0 Z
0
a15 = ¡ M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0 ¯
+¯¯ 2 11
¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM11
b15 = ¡ N11 ¡V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG
dxb
L
c15 = 0
The lateral and roll motions of the water particles, relative to the cross section of an
oscillating ship in still water, are de…ned by:
³ ¤h2 = ¡y ¡ xb ¢ Ã ¡ OG ¢ Á ³ ¤h4 = ¡Á
¤ ¤
³_ h2 = ¡y_ ¡ xb ¢ Ã_ + V ¢ Ã ¡ OG ¢ Á_ ³_ h4 = ¡Á_
¤ ¤
³Äh2 = ¡Äy ¡ xb ¢ ÃÄ + 2V ¢ Ã_ ¡ OG ¢ ÁÄ ³Ä = ¡Á
h4
Ä
with:
Z
0
a22 = + M22 ¢ dxb
L
¯ ¯
¯ Z ¯
¯V dN 0 ¯
+¯¯ 2 22
¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM22
b22 = + N22 ¡ V ¢ ¢ dxb
dxb
L
6.2. HYDROMECHANICAL FORCES FOR SWAY 167
c22 = 0 Z Z
0 0
a24 = + M24 ¢ dxb + OG M22 ¢ dxb
L L
¯ ¯
¯ Z Z ¯
¯V dN 0
V dN 0 ¯
+¯¯ 2 24
¢ dxb + 2 ¢ OG 22
¢ dxb ¯¯
¯ !e dxb !e dxb ¯
L L
Z µ 0
¶ Z µ 0
¶
0 dM24 0 dM22
b24 = + N24 ¡ V ¢ ¢ dxb + OG N22 ¡ V ¢ ¢ dxb
dxb dxb
L L
c24 = 0
Z Z µ 0
¶
0 V 0 dM22
a26 = + M22 ¢ xb ¢ dxb + 2 N22 ¡ V ¢ ¢ dxb
!e dxb
L L
¯ ¯
¯ Z Z ¯
¯V V dN22 0 ¯
¯ 0
+¯ 2 N22 ¢ dxb + 2 ¢ xb ¢ dxb ¯¯
¯ !e !e dxb ¯
L L
Z µ 0
¶ Z
0 dM22 0
b26 = + N22 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M22 ¢ dxb
dxb
L L
¯ ¯
¯ 2Z ¯
¯V dN 0 ¯
+¯¯ 2 22
¢ dxb ¯¯
¯ !e dxb ¯
L
c26 = 0
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled sway equation become:
Z
0
a22 = + M22 ¢ dxb
L
Z
0
b22 = + N22 ¢ dxb
L
c22 = 0 Z Z
0 0
a24 = + M24 ¢ dxb + OG M22 ¢ dxb
L
Z ZL
0 0
b24 = + N24 ¢ dxb + OG N22 ¢ dxb
L L
c24 = 0 Z Z
0 V 0
a26 = + M22 ¢ xb ¢ dxb + 2 N22 ¢ dxb
!e
L
Z ZL
0 0
b26 = + N22 ¢ xb ¢ dxb ¡ V M22 ¢ dxb
L L
168 CHAPTER 6. HYDROMECHANICAL LOADS
c26 = 0
The vertical motions of the water particles, relative to the cross section of an oscillating
ship in still water, are de…ned by:
³ ¤h3 = ¡z + xb ¢ µ
¤
³_ h3 = ¡z_ + xb ¢ µ_ ¡ V ¢ µ
ij ¤ = ¡Ä z + xb ¢ ĵ ¡ 2V ¢ µ_
h3
with:
a31 = 0
b31 = 0
c31 = 0 Z
0
a33 = + M33 ¢ dxb
L
6.3. HYDROMECHANICAL FORCES FOR HEAVE 169
¯ ¯
¯ Z ¯
¯V dN 0 ¯
+¯¯ 2 33
¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM33
b33 = + N33 ¡ V ¢ ¢ dxb
dxb
L
Z
c33 = +2½g yw ¢ dxb
L
Z Z µ 0
¶
0 V 0 dM33
a35 = ¡ M33 ¢ xb ¢ dxb ¡ 2 ¡V ¢
N33 ¢ dxb
!e dxb
L L
¯ ¯
¯ Z Z ¯
¯ ¡V V dN 0 ¯
+¯¯ 2 0
N33 ¢ dxb ¡ 2 33
¢ xb ¢ dxb ¯¯
¯ !e !e dxb ¯
L L
Z µ 0
¶ Z
0 dM33 0
b35 = ¡ N33 ¡ V ¢ ¢ xb ¢ dxb + 2V M33 ¢ dxb
dxb
L L
¯ ¯
¯ 2Z ¯
¯V dN33 0 ¯
¯
+¯ 2 ¢ dxb ¯¯
¯ !e dxb ¯
L
Z
c35 = ¡2½g yw ¢ xb ¢ dxb
L
The ”Modi…ed Strip Theory Method” includes the outlined terms. When ignoring the
outlined terms the ”Ordinary Strip Theory” is presented.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled heave equation become:
a31 = 0
b31 = 0
c31 = 0 Z
0
a33 = + M33 ¢ dxb
L
Z
0
b33 = + N33 ¢ dxb
L
Z
c33 = +2½g yw ¢ dxb
L
Z Z
0 V 0
a35 = ¡ ¢ xb ¢ dxb ¡ 2 N33
M33 ¢ dxb
!e
L
Z ZL
0 0
b35 = ¡ N33 ¢ xb ¢ dxb + V M33 ¢ dxb
L L
170 CHAPTER 6. HYDROMECHANICAL LOADS
Z
c35 = ¡2½g yw ¢ xb ¢ dxb
L
So no terms are added for the ”Modi…ed Strip Theory”.
with:
Z
0
a42 = + M42 ¢ dxb + OG ¢ a22
L
¯ ¯
¯ Z ¯
¯V dN420 ¯
¯
+¯ 2 ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM42
b42 = + N42 ¡ V ¢ ¢ dxb + OG ¢ b22
dxb
L
Z OG ¢ c22
c42 = 0 + Z
0 0
a44 = + M44 ¢ dxb + OG M42 ¢ dxb + OG ¢ a24
L L
¯ ¯
¯ Z Z ¯
¯V dN44 0
V dN420 ¯
¯
+¯ 2 ¢ dxb + 2 ¢ OG ¢ dxb ¯¯
¯ !e dxb !e dxb ¯
L L
Z µ 0
¶ Z µ 0
¶
0 dM44 0 dM42
b44 = + N24 ¡ V ¢ ¢ dxb + OG N42 ¡ V ¢ ¢ dxb + b44V + OG ¢ b24
dxb dxb
L L
Z µ 3 ¶
yw As
c44 = +2½g ¡ ¢ bG dxb
3 2
L
= +½gr ¢ GM
Z Z µ 0
¶
0 V 0 dM42
a46 = + M42 ¢ xb ¢ dxb + 2 N42 ¡ V ¢ ¢ dxb + OG ¢ a26
!e dxb
L L
¯ ¯
¯ Z Z ¯
¯V V dN 0 ¯
+¯¯ 2 N42 0
¢ dxb + 2 42
¢ xb ¢ dxb ¯¯
¯ !e !e dxb ¯
L L
Z µ 0
¶ Z
0 dM42 0
b46 = + N42 ¡V ¢ ¢ xb ¢ dxb ¡ 2V M42 ¢ dxb + OG ¢ b26
dxb
L L
¯ ¯
¯ 2Z ¯
¯V dN420 ¯
¯
+¯ 2 ¢ dxb ¯¯
¯ !e dxb ¯
L
c46 = 0 + OG ¢ c26
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory Method” is presented.
A viscous roll damping coe¢cient b44V , derived for instance with the empiric method of
[Ikeda et al., 1978], has been added here.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled roll equation become:
Z
0
a42 = + M42 ¢ dxb + OG ¢ a22
L
172 CHAPTER 6. HYDROMECHANICAL LOADS
Z
0
b42 = + N42 ¢ dxb + OG ¢ b22
L
c42 = 0 Z Z
0 0
a44 = + M44 ¢ dxb + OG M42 ¢ dxb + OG ¢ a24
L
Z ZL
0 0
b44 = + N44 ¢ dxb + OG N42 ¢ dxb + b44V + OG ¢ b24
L L
c44 = +½gr ¢ GM
Z Z
0 V 0
a46 = + M42 ¢ xb ¢ dxb + 2 N42 ¢ dxb + OG ¢ a26
!e
L
Z ZL
0 0
b46 = + N42 ¢ xb ¢ dxb ¡ V M42 ¢ dxb + OG ¢ b26
L L
c46 = 0
with:
Z
0
a51 = ¡ M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0 ¯
+¯¯ 2 11
¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM11
b51 = ¡ N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG
dxb
L
c51 = 0 Z
0
a53 = ¡ M33 ¢ xb ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0 ¯
+¯¯ 2 33
¢ xb ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM33
b53 = ¡ N33 ¡ V ¢ ¢ xb ¢ dxb
dxb
L
Z
c53 = ¡2½g yw ¢ xb ¢ dxb
L
Z
0 2
a55 = + M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯V dN 0
2 ¯
+¯¯ 2 11
¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z Z µ 0
¶
0 2 V 0 dM33
+ M33 ¢ xb ¢ dxb + 2 N33 ¡ V ¢ ¢ xb ¢ dxb
!e dxb
L L
¯ ¯
¯ Z Z ¯
¯V V dN 0 ¯
+¯¯ 2 N33 ¢ xb ¢ dxb + 2
0 33
¢ xb ¢ dxb ¯¯
2
¯ !e !e dxb ¯
L L
Z µ 0
¶
0 dM11 2 2
b55 = + N11 ¡ V ¢ ¢ bG ¢ dxb + b11V ¢ BG
dxb
L
Z µ 0
¶ Z
0 dM33 2 0
+ N33 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M33 ¢ xb ¢ dxb
dxb
L L
¯ ¯
¯ Z ¯
¯ ¡V 2 dN330 ¯
¯
+¯ 2 ¢ xb ¢ dxb ¯¯
¯ !e dxb ¯
L
Z
c55 = +2½g yw ¢ x2b ¢ dxb
L
174 CHAPTER 6. HYDROMECHANICAL LOADS
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled pitch equation become:
Z
0
a51 = ¡ M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0 ¯
+¯¯ 2 11
¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM11
b51 = ¡ N11 ¡ V ¢ ¢ bG ¢ dxb ¡ b11V ¢ BG
dxb
L
c51 = 0 Z
0
a53 = ¡ M33 ¢ xb ¢ dxb
L
¯ ¯
¯ Z ¯
¯V ¯
+¯¯ 2 N33 0
¢ dxb ¯¯
¯ !e ¯
L
Z Z
0 0
b53 = ¡ N33 ¢ xb ¢ dxb ¡ V M33 ¢ dxb
L L
Z
c53 = ¡2½g yw ¢ xb ¢ dxb
L
Z
0 2
a55 = + M11 ¢ bG ¢ dxb
L
¯ ¯
¯ Z ¯
¯V dN 0
2 ¯
+¯¯ 2 11
¢ bG ¢ dxb ¯¯
¯ !e dxb ¯
L
Z Z Z
0 2 V 0 V2 0
+ M33 ¢ xb ¢ dxb + 2 N33 ¢ xb ¢ dxb + 2 M33 ¢ dxb
!e !e
L L L
¯ ¯
¯ Z ¯
¯ ¡V ¯
+¯¯ 2 N33 ¢ xb ¢ dxb ¯¯
0
¯ !e ¯
L
Z µ 0
¶ Z
0 dM11 2 2 0
b55 = + N11 ¡ V ¢ ¢ bG ¢ dxb + b11V ¢ BG + N33 ¢ x2b ¢ dxb
dxb
L L
¯ ¯
¯ 2Z ¯
¯V ¯
+¯¯ 2 N330
¢ dxb ¯¯
¯ !e ¯
L
Z
c55 = +2½g yw ¢ x2b ¢ dxb
L
6.6. HYDROMECHANICAL MOMENTS FOR YAW 175
with:
Z
0
a62 = + M22 ¢ xb ¢ dxb
L
¯ ¯
¯ Z ¯
¯V dN 0 ¯
+¯¯ 2 22
¢ xb ¢ dxb ¯¯
¯ !e dxb ¯
L
Z µ 0
¶
0 dM22
b62 = + N22 ¡ V ¢ ¢ xb ¢ dxb
dxb
L
176 CHAPTER 6. HYDROMECHANICAL LOADS
c62 = 0 Z Z
0 0
a64 = + M24 ¢ xb ¢ dxb + OG M22 ¢ xb ¢ dxb
L L
¯ ¯
¯ Z Z ¯
¯V dN 0
V dN 0 ¯
+¯¯ 2 24
¢ xb ¢ dxb + 2 ¢ OG 22
¢ xb ¢ dxb ¯¯
¯ !e dxb !e dxb ¯
L L
Z µ 0
¶ Z µ 0
¶
0 dM24 0 dM22
b64 = + N24 ¡ V ¢ ¢ xb ¢ dxb + OG N22 ¡ V ¢ ¢ xb ¢ dxb
dxb dxb
L L
c64 = 0
Z Z µ 0
¶
0 V dM22
a66 = + M22 ¢ x2b ¢ dxb + 2 0
N22 ¡V ¢ ¢ xb ¢ dxb
!e dxb
L L
¯ ¯
¯ Z Z ¯
¯V V 0
dN22 2 ¯
¯ 0
+¯ 2 N22 ¢ xb ¢ dxb + 2 ¢ xb ¢ dxb ¯¯
¯ !e !e dxb ¯
L L
Z µ 0
¶ Z
0 dM22 2 0
b66 = + N22 ¡ V ¢ ¢ xb ¢ dxb ¡ 2V M22 ¢ xb ¢ dxb
dxb
L L
¯ ¯
¯ 2Z ¯
¯V dN 0 ¯
+¯¯ 2 22
¢ xb ¢ dxb ¯¯
¯ !e dxb ¯
L
c66 = 0
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
After simpli…cation, the expressions for the total hydromechanical coe¢cients in the cou-
pled yaw equation become:
Z
0
a62 = + M22 ¢ xb ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V ¯
+¯¯ 2 N22 ¢ dxb ¯¯
0
¯ !e ¯
L
Z Z
0 0
b62 = + N22 ¢ xb ¢ dxb + V M22 ¢ dxb
L L
c62 = 0 Z Z
0 0
a64 = + M24 ¢ xb ¢ dxb + OG M22 ¢ xb ¢ dxb
L L
¯ ¯
¯ Z Z ¯
¯ ¡V V ¯
+¯¯ 2 N24 ¢ dxb ¡ 2 ¢ OG N22 ¢ dxb ¯¯
0 0
¯ !e !e ¯
L L
6.6. HYDROMECHANICAL MOMENTS FOR YAW 177
Z Z Z Z
0 0 0 0
b64 = + N24 ¢ xb ¢ dxb + V M24 ¢ dxb + OG N22 ¢ xb ¢ dxb + V ¢ OG M22 ¢ dxb
L L L L
c64 = 0
Z Z Z
0 V V2
a66 = + M22 x2b
¢ ¢ dxb + 2 0
N22 ¢ xb ¢ dxb + 2 0
M22 ¢ dxb
!e !e
L L L
¯ ¯
¯ Z ¯
¯ ¡V ¯
+¯¯ 2 N220
¢ xb ¢ dxb ¯¯
¯ !e ¯
L
Z
0
b66 = + N22 ¢ x2b ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V 2 ¯
+¯¯ 2 0
N22 ¢ dxb ¯¯
¯ !e ¯
L
c66 = 0
178 CHAPTER 6. HYDROMECHANICAL LOADS
.
Chapter 7
Exciting Wave Loads
The …rst order wave potential in a ‡uid - with any arbitrary water depth h - is given by:
¡g cosh k(h + zb )
©w = ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
! cosh(kh)
in an axes system with the centre of gravity in the waterline.
The velocities and accelerations in the direction j of the water particles have to be de…ned.
The local relative orbital velocities of the water particles in a certain direction follow from
the derivative in that direction of the wave potential. The orbital accelerations of the water
particles can be obtained from these velocities by:
n o ½ ¾
ij 0 = D ³_ 0 with:
D
=
@
¡V ¢
@
for: j = 1; 2; 3; 4
wj
Dt wj Dt @t @xb
With this, the relative velocities and accelerations in the di¤erent directions can be found:
² Surge direction:
0 @©w
³_ w1 =
@xb
cosh k(h + zb )
= +! ¢ ¢ cos ¹ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
0
ij 0 @ ³_ w1
w1 =
@t
cosh k(h + zb )
= ¡! 2 ¢ ¢ cos ¹ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
² Sway direction:
0 @©w
³_ w2 =
@yb
cosh k(h + zb )
= +! ¢ ¢ sin ¹ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
179
180 CHAPTER 7. EXCITING WAVE LOADS
0
0 @ ³_ w2
³Äw2 =
@t
0 cosh k(h + zb )
³Äw2 = ¡! 2 ¢ ¢ sin ¹ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
² Heave direction:
0 @©w
³_ w3 =
@zb
sinh k(h + zb )
= ¡! ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
0
ij 0 @ ³_ w3
w3 =
@t
sinh k(h + zb )
= ¡! 2 ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
² Roll direction:
0 0
_³ 0 @ ³_ w2 @ ³_ w3
w4 = ¡ =0
@zb @yb
ij 0 = 0
w4
of which the zero solution is obvious, because the potential ‡uid is free of rotation.
The pressure in the ‡uid follows from the linearized equation of Bernoulli:
cosh k(h + zb )
p = ¡½gzb + ½g ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
cosh(kh)
@p @p @p
= p0 + ¢ dxb + ¢ dyb + ¢ dzb
@xb @yb @zb
with the following expressions for the pressure gradients:
@p cosh k(h + zb )
= +½kg cos ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
@xb cosh(kh)
cosh k(h + zb )
= +½! 2 cos ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
@p cosh k(h + zb )
= +½kg sin ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
@yb cosh(kh)
cosh k(h + zb )
= +½! 2 sin ¹ ¢ ¢ ³ a sin(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
@p sinh k(h + zb )
= ¡½g + ½kg ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
@zb cosh(kh)
sinh k(h + zb )
= ¡½g + ½! ¢ ¢ ³ a cos(!t ¡ kxb cos ¹ ¡ kyb sin ¹)
sinh(kh)
7.1. CLASSICAL APPROACH 181
@p 0
= ¡½ ¢ ³Äw1
@xb
@p 0
= ¡½ ¢ ³Äw2
@yb
@p 0
= ¡½ ¢ (g + ij w3 )
@zb
According to the ”Ordinary Strip Theory” the exciting wave forces for surge on a restrained
cross section of a ship in waves are de…ned by:
D n 0 _¤ o ¤
Xw0 1 = 0
M11 ¢ ³ w1 + N11 ¢ ³_ w1 + XF0 K1
Dt µ ¶
0
Ä ¤ dM11 ¤
0 0
= M11 ¢ ³ w1 + N11 ¡ V ¢ ¢ ³_ w1 + XF0 K1
dxb
The Froude-Krilov force in the surge direction - so the longitudinal force due to the pressure
in the undisturbed ‡uid - is given by:
Z³ Z+yb
@p
XF K1 = ¡ ¢ dyb ¢ dzb
@xb
¡T ¡yb
Z³ Z+yb
0
= ½ ³Äw1 ¢ dyb ¢ dzb
¡T ¡yb
182 CHAPTER 7. EXCITING WAVE LOADS
After neglecting the second order terms, this can be written as:
with:
Z0
sin(¡kyb sin ¹) cosh k(h + zb )
Ach = 2 ¢ ¢ yb ¢ dzb
¡kyb sin ¹ cosh(kh)
¡T
with:
Z0 Z0 Z0
A=2 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb Iy = 2 yb ¢ zb2 ¢ dzb
¡T ¡T ¡T
The acceleration term kg cos ¹ ¢ ³ a in here is the amplitude of the longitudinal component
of the relative orbital acceleration in deep water at zb = 0.
The dominating …rst term in this series consists of a mass and this acceleration.
This mass term ½A is used to obtain from the total Froude-Krilov force an equivalent
longitudinal component of the orbital acceleration of the water particles:
¤
XF K1 = ½A ¢ ij w1
This holds that the equivalent longitudinal components of the orbital acceleration and
velocity are equal to the values at zb = 0 in a wave with a reduced amplitude ³ ¤a1 :
¤
³Äw1 = ¡kg cos ¹ ¢ ³ ¤a1 sin(! e t ¡ kxb cos ¹)
¤ +kg cos ¹ ¤
³_ w1 = ¢ ³ a1 cos(! e t ¡ kxb cos ¹)
!
7.1. CLASSICAL APPROACH 183
with:
Ach
³ ¤a1 = ¢ ³a
A
This equivalent acceleration and velocity will be used in the di¤raction part of the wave
force for surge.
From the previous follows the total wave loads for surge:
Z
¤
Xw1 = + M11 0
¢ ³Äw1 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ V 0
dN11 Ĥ ¯
¯
+¯ ¢ ³ w1 ¢ dxb ¯¯
¯ ! ¢ !e dxb ¯
L
Z ï ¯ !
¯!¯ 0 dM 0
¤
+ ¯ ¯¢N ¡ V ¢ 11
¢ ³_ w1 ¢ dxb
¯ ! e ¯ 11 dxb
L
Z
+ XF0 K1 ¢ dxb
L
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
Z³ Z+yb
0
= ½ ³Äw2 ¢ dyb ¢ dzb
¡T ¡yb
After neglecting the second order terms, this can be written as:
XF K2 = ½Ach ¢ (¡kg sin ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹)
with:
Z0
sin(¡kyb sin ¹) cosh k(h + zb )
Ach = 2 ¢ ¢ yb ¢ dzb
¡kyb sin ¹ cosh(kh)
¡T
When expanding the Froude-Krilov force in deep water with ¸ À 2¼ ¢ yw and ¸ À 2¼ ¢ T
in series, it is found:
µ ¶
1 2
XF K2 = ½ ¢ A + k ¢ Sy + k ¢ Iy + ::: ¢ (¡kg sin ¹) ¢ ³ a sin(! et ¡ kxb cos ¹)
2
with:
Z0 Z0 Z0
A=2 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb Iy = 2 yb ¢ zb2 ¢ dzb
¡T ¡T ¡T
The acceleration term kg sin ¹ ¢ ³ a in here is the amplitude of the lateral component of the
relative orbital acceleration in deep water at zb = 0.
The dominating …rst term in this series consists of a mass and this acceleration.
This mass term ½A is used to obtain from the total Froude-Krilov force an equivalent
lateral component of the orbital acceleration of the water particles:
¤
XF K2 = ½A ¢ ij w2
This holds that the equivalent lateral components of the orbital acceleration and velocity
are equal to the values at zb = 0 in a wave with a reduced amplitude ³ ¤a2 :
ij ¤ = ¡kg sin ¹ ¢ ³ ¤ sin(! e t ¡ kxb cos ¹)
w2 a2
¤ +kg sin ¹ ¤
³_ w2 = ¢ ³ a2 cos(! e t ¡ kxb cos ¹)
!
7.1. CLASSICAL APPROACH 185
with:
Ach
³ ¤a2 = ¢ ³a
A
This equivalent acceleration and velocity will be used in the di¤raction part of the wave
force for sway.
From the previous follows the total wave loads for sway:
Z
¤
Xw2 = + 0
M22 ¢ ³Äw2 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ V dN 0
¤ ¯
+¯¯ 22
¢ ³Äw2 ¢ dxb ¯¯
¯ ! ¢ !e dxb ¯
L
Z ï ¯ !
¯!¯ 0 dM 0
¤
+ ¯ ¯¢N ¡ V ¢ 22
¢ ³_ w2 ¢ dxb
¯ ! e ¯ 22 dxb
L
Z
+ XF0 K2 ¢ dxb
L
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
According to the ”Ordinary Strip Theory” the exciting wave forces for heave on a restrained
cross section of a ship in waves are de…ned by:
D n 0 _¤ o ¤
Xw0 3 = M33 ¢ ³ w3 + N330
¢ ³_ w3 + XF0 K3
Dt
µ 0
¶
¤ dM ¤
= M33 ¢ ij w3 + N33 ¡ V ¢
0 0 33
¢ ³_ w3 + XF0 K3
dxb
The Froude-Krilov force in the heave direction - so the vertical force due to the pressure
in the undisturbed ‡uid - is given by:
Z³ Z+yb
0 @p
XF K3 = ¡ ¢ dyb ¢ dzb
@zb
¡T ¡yb
Z³ Z+yb
0
= ½ (g + ³Äw3 ) ¢ dyb ¢ dzb
¡T ¡yb
After neglecting the second order terms, this force can be written as:
0
XF K3 = 2½gyw ¢ C3 ¢ ³ a cos(! e t ¡ kxb cos ¹)
with:
Z0
sin(¡kyw sin ¹) k sin(¡kyb sin ¹) sinh k(h + zb )
C3 = ¡ ¢ ¢ yb ¢ dzb
¡kyw sin ¹ yw ¡kyb sin ¹ cosh(kh)
¡T
When expanding the Froude-Krilov force in deep water with ¸ À 2¼ ¢ T and in long
waves with ¸ À 2¼ ¢ yw in series, it is found:
Z0
sin(¡kyw sin ¹) k sin(¡kyb sin ¹) sinh k(h + zb )
C3 = ¡ ¢ ¢ yb ¢ dzb
¡kyw sin ¹ yw ¡kyb sin ¹ cosh(kh)
¡T
Z0
k
t 1¡ ekzb ¢ yb ¢ dzb
yw
¡T
Z0 µ ¶
k k2 2
= 1¡ 1 + k ¢ zb + ¢ zb + :::::: ¢ yb ¢ dzb
yw 2
¡T
7.1. CLASSICAL APPROACH 187
µ ¶
A Sy k 2 Iy
= 1 ¡ k¢ +k¢ + ¢ + ::::::
2yw 2yw 2 2yw
µ ¶
A
= 1 ¡ k¢ + ::::::
2yw
= 1 ¡ kT3¤
t exp f¡kT3¤ g
with:
Z0 Z0 Z0
A=2 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb Iy = 2 yb ¢ zb2 ¢ dzb
¡T ¡T ¡T
T3¤ can be considered as the draft at which the pressure in the vertical direction is equal
to the average vertical pressure on the cross section in the ‡uid and can be obtained by.
¡ ln C3
T3¤ =
k
This holds that the equivalent vertical components of the orbital acceleration and velocity
are equal to the values at zb = ¡T3¤ :
¤ ¤
³Äw3 = ¡kg ¢ e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹)
¡ ¤ ¢
= ¡! 2 ¢ e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹)
¤ ¡kg ¡kT3¤
³_ w3 = ¢e ¢ ³ a3 ¢ sin(! e t ¡ kxb cos ¹)
! ¡ ¢
¤
= ¡! ¢ e¡kT3 ¢ ³ a3 ¢ sin(! e t ¡ kxb cos ¹)
¤ kg ¡kT3¤
³_ w3 = 2
¢e ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹)
!
¡ ¢
¤
= e¡kT3 ¢ ³ a3 ¢ cos(! e t ¡ kxb cos ¹)
When expanding the Froude-Krilov force in shallow water with kh ! 0 and in long waves
with ¸ À 2¼ ¢ yw in series, it is found:
Z0
sin(¡kyw sin ¹) k sin(¡kyb sin ¹) sinh k(h + zb )
C3 = ¡ ¢ ¢ yb ¢ dzb
¡kyw sin ¹ yw ¡kyb sin ¹ cosh(kh)
¡T
Z0
k sinh k(h + zb )
t 1¡ ¢ yb ¢ dzb
yw cosh(kh)
¡T
Z0
k k(h + zb ) + k 3 ¢ (::::::) + ::::::
= 1¡ ¢ yb ¢ dzb
yw cosh(kh)
¡T
0 0 1
Z Z0
kh k @ 1
= 1¡ ¢ ¢ yb ¢ dzb + yb ¢ zb ¢ dzb + ::::::A
cosh(kh) yw h
¡T ¡T
188 CHAPTER 7. EXCITING WAVE LOADS
µ ¶
kh A Sy
= 1¡ ¢k¢ + + ::::::
cosh(kh) 2yw h ¢ 2yw
kh ³ zB ´ A
= 1¡ ¢k¢ 1+ + :::::: ¢
cosh(kh) h 2yw
³ ´ µ ¶
kh zB A
= 1¡ ¢ 1+ ¢k¢ + ::::::
cosh(kh) h 2yw
kh ³ zB ´
= 1¡ 1+ ¢ kT3¤
cosh(kh) h
½ ³ ¾
kh zB ´ ¤
t exp ¡ 1+ ¢ kT3
cosh(kh) h
with:
Z0 Z0
A=2 yb ¢ dzb Sy = 2 yb ¢ zb ¢ dzb = zB ¢ A
¡T ¡T
¡ ln C3
T3¤ = ¡ ¢
kh
cosh(kh)
1 + zhB ¢ k
This holds that the equivalent vertical components of the orbital acceleration and velocity
are equal to the values at zb = ¡T3¤ :
It may be noted that this shallow water de…nition for T3¤ is valid in deep water too, because:
kh zB
!1 and !0 for: h ! 1
cosh(kh) h
These equivalent accelerations and velocities will be used to determine the di¤raction part
of the wave forces for heave.
7.1. CLASSICAL APPROACH 189
From the previous follows the total wave loads for heave:
Z
0 ¤
Xw3 = + M33 0
¢ ³Äw3 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ V 0
dN33 Ĥ ¯
¯
+¯ ¢ ³ w3 ¢ dxb ¯¯
¯ ! ¢ !e dxb ¯
L
Z ï ¯ !
¯!¯ dM 0
¤
+ ¯ ¯ ¢ N33 ¡ V ¢
0 33
¢ ³_ w3 ¢ dxb
¯ !e ¯ dxb
L
Z
+ XF0 K3 ¢ dxb
L
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
According to the ”Ordinary Strip Theory” the exciting wave moments for roll on a re-
strained cross section of a ship in waves are de…ned by:
D n 0 _¤ o ¤
Xw0 4 = +XF0 K4 + 0
M42 ¢ ³ w2 + N42 ¢ ³_ w2 + OG ¢ Xw0 2
Dt
µ 0
¶
¤ dM ¤
= XF K4 + M42 ¢ ³Äw2 + N42 ¡ V ¢
0 0 0 42
¢ ³_ w2 + OG ¢ Xw0 2
dxb
According to the ”Modi…ed Strip Theory” these moments become:
½µ ¶ ¾
0 0 D 0 i _ ¤
Xw4 = +XF K4 + M42 ¡ ¢ N42 ¢ ³ w2 + OG ¢ Xw0 2
0
Dt !e
µ 0
¶ µ 0
¶
V dN42 Ä ¤ dM42 ¤
0 0
= XF K4 + M42 + 2 ¢ 0
¢ ³ w2 + N42 ¡ V ¢ ¢ ³_ w2 + OG ¢ Xw0 2
! e dxb dxb
The Froude-Krilov moment in the roll direction - so the roll moment due to the pressure
in the undisturbed ‡uid - is given by:
Z³ Z+ybµ ¶
@p @p
XF K4 = ¡ ¡ ¢ zb + ¢ yb ¢ dyb ¢ dzb
@yb @zb
¡T ¡yb
Z³ Z+yb³ ´
0 0
= ½ Ä Ä
¡³ w2 ¢ zb + (g + ³ w3 ) ¢ yb ¢ dyb ¢ dzb
¡T ¡yb
190 CHAPTER 7. EXCITING WAVE LOADS
After neglecting the second order terms, this can be written as:
µ ¶
CCy CSy
XF K4 = ½ ¢ ¡ ¡ + CIz ¢ (¡k 2 g sin ¹) ¢ ³ a sin(! e t ¡ kxb cos ¹)
k k
with:
sin(¡kyw sin ¹)
¡kyw sin ¹
¡ cos(¡kyw sin ¹)
CCy = 2 ¢ ¢ yw3
(¡kyw sin ¹)2
Z0
sin(¡kyb sin ¹) cosh k(h + zb )
CSy = 2 ¢ ¢ yb ¢ zb ¢ dzb
¡kyb sin ¹ cosh(kh)
¡T
Z0 sin(¡kyb sin ¹)
¡ cos(¡kyb sin ¹) cosh k(h + zb ) 3
¡kyb sin ¹
CIz = 2 ¢ ¢ yb ¢ dzb
(¡kyb sin ¹)2 cosh(kh)
¡T
with:
Xw0 5 = ¡Xw0 1 ¢ bG ¡ Xw0 3 ¢ xb
In here, bG is the vertical distance of the centre of gravity of the ship G above the centroid
b of the local submerged sectional area.
From this follows the total wave loads for pitch:
Z
¤
Xw5 = ¡ M11 0
¢ bG ¢ ³Äw1 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0
¤ ¯
+¯¯ 11
¢ bG ¢ ³Äw1 ¢ dxb ¯¯
¯ ! ¢ !e dxb ¯
L
Ã
Z ¯ ¯ !
¯!¯ 0 0
¡ ¯ ¯¢N11 ¡ V ¢ dM11 ¢ bG ¢ ³_ ¤w ¢ dxb
¯ !e ¯ dxb 1
L
Z
¡ XF0 K1 ¢ bG ¢ dxb
L
Z
¤
¡ 0
M33 ¢ xb ¢ ³Äw3 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ ¡V dN 0
¤ ¯
+¯¯ 33
¢ xb ¢ ij w3 ¢ dxb ¯¯
¯ ! ¢ !e dxb ¯
L
Z ï¯ ¯¯ 0
!
!
¯ ¯¢N 0 ¡ V ¢ dM ¤
¡ ¯ ! e ¯ 33
33
¢ xb ¢ ³_ w3 ¢ dxb
dxb
L
Z
¡ XF0 K3 ¢ xb ¢ dxb
L
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
with:
Xw0 6 = +Xw0 2 ¢ xb
From this follows the total wave loads for yaw:
Z
¤
Xw6 = + M22 0
¢ xb ¢ ij w2 ¢ dxb
L
¯ ¯
¯ Z ¯
¯ V 0
dN22 ¤ ¯
¯
+¯ ¢ xb ¢ ³ w2 ¢ dxb ¯¯
Ä
¯ ! ¢ !e dxb ¯
L
Z ¯ ¯
¯!¯ dM22 0
¤
+ ¯¯ ¯¯ ¢ N220
¡V ¢ ¢ xb ¢ ³_ w2 ¢ dxb
!e dxb
L
Z
+ XF0 K2 ¢ xb ¢ dxb
L
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
7.2. EQUIVALENT MOTIONS OF WATER PARTICLES 193
xja ! 2
P0j = ¡ ¼ ywl sin "Áxj
³ ja g
xja ! 2
Q0j = + ¼ ywl cos "Áxj
³ ja g
The phase lag "Áxj between he velocity potentials and the forced motion is incorporated in
the coe¢cients P0j and Q0j and can be obtained by using:
µ ¶
¡P0j
"Áxj = arctan
+Q0j
This equation will be used further on for obtaining wave load phases.
194 CHAPTER 7. EXCITING WAVE LOADS
Generally, these hydrodynamic loads are expressed in terms of potential mass and damping
coe¢cients:
0
Xhi = ¡Mij xÄj ¡ Nij x_ j
= Mij ! 2 xja cos !t + Nij !xja sin !t
with:
The energy required for this oscillation should be equal to the energy radiated by the
damping waves:
7.2. EQUIVALENT MOTIONS OF WATER PARTICLES 195
ZT ZT
1 0 1 0 1
Xhj ¢ x_ j dt = Njj x_ j ¢ x_ j dt = 2 ¢ ½g³ 2a ¢ cgroup
T T 2
0 0
or:
1 0 0 1 0
Xhja !xja sin "hj = Njj ! 2 x2ja = ½g³ 2a ¢ cgroup
2 2
From the …rst part of this equation follows
0 0
Xhja sin "hj 0 xja
= Njj !
³a ³a
From the second part of this equation follows the amplitude ratio of the oscillatory motions
and the radiated waves:
s
xja 1 2½g ¢ cgroup
= 0
³a ! Njj
Combining these two last equations provides for the out-phase part - so the damping part
- of the oscillatory force:
0
Xhja sin "hj q
0
0
= 2½g ¢ cgroup ¢ Njj for: j = 2; 3; 4
³a
0 0
In here, Xhja sin "hj is the in-phase with the velocity part of the exciting force or moment.
Figure 7.5: Vector Diagrams of Wave Components for Sway and Heave
Heave Mode
The vertical wave force on a restrained cross section in waves is:
0 0
³ 0
´
Xw3 = Xw3a cos !t + "w3
0 0 0 0
= Xw3a cos "w3 cos !t ¡ Xw3a sin "w3 sin !t
of which the amplitude is equal to:
q
0 0
Xw3a = ³ a 2½g ¢ cgroup ¢ N33
0
For the phase lag of this wave force, "w3 , an approximation has to be found.
0
The phase lag of a radiated wave, "wR 3 , at the intersection of the ship’s hull with the
waterline, yb = ywl , is
0
"wR 3 = kywl
0
The phase lag of the wave force, "w3 , has been approximated by this phase:
0 0
"w3 = "wR 3 = kywl
Then, the in-phase and out-phase parts of the wave loads are:
µ q ¶
0 0 0 0 0
XF0 K3 + Xw31 = +Xw3a cos "w3 = + ³ a 2½g ¢ cgroup ¢ N33 cos "w3
µ q ¶
0 0 0 0 0
Xw32 = ¡Xw3a sin "w3 = ¡ ³ a 2½g ¢ cgroup ¢ N33 sin "w3
7.2. EQUIVALENT MOTIONS OF WATER PARTICLES 197
0 0
from which the di¤raction terms, Xw31 and Xw32 follow.
These di¤raction terms can also be written as:
0 0 0
Xw31 = M33 ¢ a
¹3
0 0 0
Xw32 = M33 ¢ v¹3
0 0
in which a
¹3 and v¹3 are the equivalent amplitudes of the acceleration and the velocity of
the water particles around the cross section.
Herewith, the equivalent acceleration and velocity amplitudes of the water particles are:
0
0 Xw31
a
¹3 = 0
M33
0
0 Xw32
v¹3 = 0
N33
Sway Mode
The horizontal wave force on a restrained cross section in beam waves is:
0 0
³ 0
´
Xw2 = Xw2a cos !t + "w2
0 0 0 0
= Xw2a cos "w2 cos !t ¡ Xw2a sin "w2 sin !t
of which the amplitude is equal to:
q
0 0
Xw2a = ³ a 2½g ¢ cgroup ¢ N22
0
For the phase lag of this wave force, "w2 , an approximation has to be found.
0
The phase lag of an incoming undisturbed wave, "wI 2 , at the intersection of the ship’s hull
with the waterline, yb = ywl , is
0
"wI 2 = ¡kywl sin ¹
0 0
if sin ¹ < 0 then: "wI 2 = "WI 2 + ¼
In very short waves - so at high wave frequencies ! ! 1 - the ship’s hull behaves like a
0
vertical wall and all waves will be di¤racted. Then, the phase lag of the wave force, "w2 , is
equal to:
0 0
"w2 (! ! 1) = ¡"wI 2
The acceleration and velocity amplitudes of the water particles in the undisturbed surface
of the incoming waves are:
³ 0´
a2 = ¡kg sin ¹
still water surface
³ 0´ ¡a2
0
kg sin ¹
v2 = =
still water surface ! !
198 CHAPTER 7. EXCITING WAVE LOADS
In very long waves - so at low wave frequencies ! ! 0 - the wave force is dominated by
the Froude-Krylov force and the amplitudes of the water particle motions do not change
0
very much over the draft of the section. Apparently, the phase lag of the wave force, "w2 ,
can be approximated by:
" 0
#
0 XF0 K2 + M22 ¢ (¡kg sin ¹)
"w2 (! ! 0) = ¡ arctan ¡ ¢
N22 ¢ kg sin ¹
0
!
0 0
When plotted against !, the two curves "w2 (! ! 0) and "w2 (! ! 1) will intersect each
0
other. The phase lag of the wave force, "w2 , can now be approximated by the lowest of
these two values:
0 0
"w2 = "w2 (! ! 1)
0 0 0 0
if "w2 (! ! 0) > "w2 (! ! 1) then: "w2 = "w2 (! ! 0)
0 0
Because "w2 (! ! 1) goes to zero in the low frequency region and "w2 (! ! 0) can have
values between 0 and 2¼, one simple precaution has to be taken:
0 ¼ 0 0
if "w2 (! ! 0) > then: "w2 (! ! 0) = "w2 (! ! 0) ¡ 2¼
2
Now the in-phase and out-phase terms of the wave force in beam waves are:
µ q ¶
0 0 0 0 0
XF0 K2 + Xw21 = ¡Xw2a sin "w2 = ¡ ³ a 2½g ¢ cgroup ¢ N22 sin "w2
µ q ¶
0 0 0 0 0
Xw22 = +Xw2a cos "w2 = + ³ a 2½g ¢ cgroup ¢ N22 cos "w2
0 0
from which the di¤raction terms, Xw21 and Xw22 follow.
These terms can also be written as:
0 0 0
Xw21 = M22 ¢ a¹2
0 0 0
Xw22 = N22 ¢ v¹2
0 0
in which a¹2 and v¹2 are the equivalent amplitudes of the acceleration and the velocity of
the water particles around the cross section.
Then - when using an approximation for the in‡uence of the wave direction - the equivalent
acceleration and velocity amplitudes of the water particles are:
0
0 Xw21
a
¹2 = 0 ¢ jsin ¹j
M22
0
0 Xw22
v¹2 = 0 ¢ jsin ¹j
N22
7.2. EQUIVALENT MOTIONS OF WATER PARTICLES 199
Roll Mode
The ‡uid is free of rotation; so the wave moment for roll consists of sway contributions
only. However, the equivalent amplitudes of the acceleration and the velocity of the water
particles will di¤er from those of sway.
From a study on potential coe¢cients, the following relation between sway and roll damping
coe¢cients has been found:
¡ 0 ¢2 ¡ 0 ¢2
0 N24 N42
N44 = 0 = 0
N22 N22
The horizontal wave moment on a restrained cross section in beam waves is:
0 0
³ 0
´
Xw4 = Xw4a cos !t + "w4
of which the amplitude is equal to:
q
0 0
Xw4a = ³ a 2½g ¢ cgroup ¢ N44
s ¡ 0 ¢2
N24
= ³ a 2½g ¢ cgroup ¢ 0
N22
q ¯ 0 ¯
¯N ¯
= ³ a 2½g ¢ cgroup ¢ N22 ¢ 24
0
0
N22
¯ 0 ¯
0
¯N ¯
= Xw2a ¢ 24 0
N22
The in-phase and out-phase parts of the wave moment in beam waves are:
0
³ ´ ¯¯N 0 ¯¯
0
XF0 K4 + Xw41 = XF0 K2 + Xw21 ¢ 24 0
N22
¯ 0 ¯
0 0
¯N ¯
24
Xw42 = Xw22 ¢ 0
N22
0 0
from which the di¤raction terms, Xw41 and Xw42 follow.
These terms can also be written as:
0 0 0
X41 = M24 ¢ a¹24
0 0 0
X42 = N24 ¢ v¹24
0 0
in which a¹24 and v¹24 are the equivalent amplitudes of the acceleration and the velocity of
the water particles around the cross section.
Then - when using an approximation for the in‡uence of the wave direction - the equivalent
acceleration and velocity amplitudes of the water particles are:
0
0 Xw41
a
¹24 = 0 ¢ jsin ¹j
M24
0
0 Xw42
v¹24 = 0 ¢ jsin ¹j
N24
200 CHAPTER 7. EXCITING WAVE LOADS
Surge Mode
The equivalent acceleration and velocity amplitudes of the water particles around the cross
section for surge have been found from:
0
0 a¹2
a
¹1 =
tan ¹
0
0 ¡¹a1
v¹1 =
!
7.3. NUMERICAL COMPARISON 201
Figure 7.7: Comparison of Simple Di¤raction Wave Loads with DELFRAC Data
202 CHAPTER 7. EXCITING WAVE LOADS
.
Chapter 8
Transfer Functions of Motions
After dividing the left and right hand terms by the wave amplitude ³ a , two sets of six
coupled equations of motion are available.
The variables in the coupled equations for the vertical plane motions are:
xa xa
Surge: ³a
¢ cos "x³ and ³a
¢ sin "x³
za za
Heave: ³a
¢ cos "z³ and ³a
¢ sin "z³
µa µa
Pitch: ³a
¢ cos "µ³ and ³a
¢ sin "µ³
The variables in the coupled equations for the horizontal plane motions are:
ya ya
Sway: ³a
¢ cos "y³ and ³a
¢ sin "y³
Áa Áa
Roll: ³a
¢ cos "Á³ and ³a
¢ sin "Á³
Ãa Ãa
Yaw: ³a
¢ cos "ó and ³a
¢ sin "ó
These sets of motions have to be solved by a numerical method. A method which provides
continuous good results, given by [Zwaan, 1977], has been used in the strip theory program
SEAWAY-DELFT.
203
204 CHAPTER 8. TRANSFER FUNCTIONS OF MOTIONS
² the low frequency region (! 2 ¿ c=(m + a)), with motions dominated by the restoring
spring term,
² the natural frequency region (! 2 t c=(m + a)), with motions dominated by the
damping term and
² the high frequency region (! 2 À c=a), with motions dominated by the mass term.
With the six centre of gravity motions, the harmonic motions in the ship-bound xb , yb and
zb directions - or in the earth bound x, y and z directions - in any point P (xb ; yb ; zb ) on
the ship can be calculated.
x, y and z directions - in a point P (xb ; yb ; zb ) on the ship can be obtained from the six
centre of gravity motions as presented below.
The harmonic longitudinal displacement is given by:
xP = x ¡ yb ¢ Ã + zb ¢ µ
= xPa ¢ cos(! e t + "xP ³ )
yP = y + x b ¢ Ã ¡ z b ¢ Á
= yPa ¢ cos(! e t + "yP ³ )
zP = z ¡ xb ¢ µ ¡ yb ¢ Á
= zPa ¢ cos(! e t + "zP ³ )
Some examples of calculated transfer functions of a crude oil carrier and a containership
are given in the …gures 8.2, 8.3 and 8.4.
1.5 1.5
Crude Oil Carrier Crude Oil Carrier
V = 0 kn V = 0 kn
Heave Pitch
RAO Heave (-)
1.0 1.0
RAO Pitch (-)
0
0
µ = 90 µ = 180
0
µ = 180
0.5 0.5
0
µ = 90
0 0
0 0.25 0.50 0.75 1.00 0 0.25 0.50 0.75 1.00
0 0
-90 0
-90 0
µ = 90 µ = 180
-180 -180
Phase εz ζ (deg)
-270 -270 0
Phase εθ ζ
µ = 90
-360 0
-360
µ = 180
-450 -450
-540 -540
-630 -630
-720 -720
0 0.25 0.50 0.75 1.00 0 0.25 0.50 0.75 1.00
Wave Frequency (rad/s) Wave Frequency (rad/s)
Notify the di¤erent speed e¤ects for roll and pitch in …gure 8.4.
1.5 1.5
Crude Oil Carrier Crude Oil Carrier
V = 16 kn V = 16 kn
Heave Pitch
RAO Heave (-)
1.0 1.0
0
µ = 90
0 0
0 0.25 0.50 0.75 1.00 0 0.25 0.50 0.75 1.00
0 0
-90 0
-90 0
µ = 90 µ = 180
-180 -180
Phase εz ζ (deg)
-270 -270
Phase εθ ζ
0 0
µ = 180 µ = 90
-360 -360
-450 -450
-540 -540
-630 -630
-720 -720
0 0.25 0.50 0.75 1.00 0 0.25 0.50 0.75 1.00
15 1.50
RAO of roll RAO of pitch
V = 0 knots
Beam waves Head waves
1.25
V = 20 knots
V = 10 knots V = 10 knots
Non-dim. RAO of pitch (-)
10 1.00
Non-dim. RAO of roll (-)
V = 0 knots
0.75
V = 20 knots
5 0.50
0.25
Containership
Lpp = 175 metre Containership
Lpp = 175 metre
0 0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
wave frequency (rad/s) wave frequency (rad/s)
² Lateral acceleration:
yÄP = yÄ + xb ¢ ÃÄ ¡ zb ¢ Á Ä
= ¡! 2e ¢ yPa ¢ cos(! e t + "yP ³ )
= yÄPa ¢ cos(! et + "yÄP ³ )
² Vertical acceleration:
zÄP = zÄ ¡ xb ¢ ĵ + yb ¢ Á Ä
= ¡! 2e ¢ zPa ¢ cos(! et + "zP ³ )
= zÄPa ¢ cos(! e t + "zÄP ³ )
208 CHAPTER 8. TRANSFER FUNCTIONS OF MOTIONS
² Longitudinal acceleration:
xÄP = xÄ ¡ yb ¢ ÃÄ + zb ¢ ĵ ¡ g ¢ µ
= ¡! 2e ¢ xPa ¢ cos(! e t + "xP ³ ) ¡ g ¢ µa ¢ cos(! et + "µ³ )
= xÄPa ¢ cos(! e t + "xÄP ³ )
² Lateral acceleration:
yÄP = yÄ + xb ¢ ÃÄ ¡ zb ¢ Á Ä +g¢Á
= ¡! 2e ¢ yPa ¢ cos(! e t + "yP ³ ) + g ¢ Áa ¢ cos(! e t + "Á³ )
= yÄPa ¢ cos(! e t + "yÄP ³ )
² Vertical acceleration:
zÄP = zÄ ¡ xb ¢ ĵ + yb ¢ Á Ä
= ¡! 2e ¢ zPa ¢ cos(! et + "zP ³ )
= zÄPa ¢ cos(! e t + "zÄP ³ )
sP = ³ P ¡ z + xb ¢ µ ¡ yb ¢ Á
= sPa ¢ cos(! et + "sP ³ )
with:
³ P = ³ a cos(! e t ¡ kxb cos ¹ ¡ kyb sin ¹)
It may be noted that the sign of the relative motion is chosen here in such a way that a
positive relative displacement implies a decrease of the freeboard.
An oscillating ship will produce waves and these phenomena will change the relative motion.
A dynamical swell up should be taken into account, which is not included in the previous
formulation.
Notify the di¤erent behavior of absolute and relative vertical motioins as given in …gure
8.5.
8.6. VERTICAL RELATIVE VELOCITIES 209
5 5
Containership Containership
Head waves Head waves V = 20 knots
V = 0 knots
2 2
V = 0 knots
RAO tends
to 1.0
1 1
RAO tends
to 1.0
RAO tends RAO tends
to 0.0 to 0.0
0 0
0 0.5 1.0 1.5 0 0.5 1.0 1.5
wave frequency (rad/s) wave frequency (rad/s)
D
s_ P = f³ ¡ z + xb ¢ µ ¡ yb ¢ Ág
Dt P
= ³_ P ¡ z_ + xb ¢ µ_ ¡ V ¢ µ ¡ yb ¢ Á_
.
Chapter 9
Anti-Rolling Devices
Since the disappearence of sails on oceangoing ships, with their stabilising wind e¤ect on
the rolling motions, naval architects have been concerned in reducing the rolling of ships
among waves. With bilge keels they performed a …rst successful attack on the problem
of rolling, but in several cases these bilge keels did not prove to be su¢cient. Since 1880,
numerous other more or less successful ideas have been tested and used.
² Four types of anti-rolling devices and its contribution to the equations of motion are
described here:
² bilge keels
² active …n stabilisers
The active …n and rudder stabilisers are not build into the program SEAWAY yet.
211
212 CHAPTER 9. ANTI-ROLLING DEVICES
pressure on the hull surface created by bilge keels. He found that the coe¢cient Cp+ of
the pressure on the front face of the bilge keel does not depend on the period parameter,
while the coe¢cient Cp¡ of the pressure on the back face of the bilge keel and the length of
the negative pressure region depend on the period parameter. Ikeda de…nes an equivalent
length of a constant negative pressure region S0 over the height of the bilge keels and a
two-dimensional roll damping component can be found.
The total bilge keel damping has been obtained by integrating these two two-dimensional
roll damping components over the length of the bilge keels.
Experiments of Ikeda showed that the e¤ect of forward speed on the roll damping due to
the bilge keels can be ignored.
The equivalent linear total bilge keel damping has been obtained by linearising the result,
as has been shown in a separate chapter.
The natural frequency of the surface waves in a harmonic rolling tank appears as the wave
length ¸ equals twice the breadth b, so: ¸0 = 2b.
With the wave number and the dispersion relation:
2¼ p
k= and ! = kg tanh [kh]
¸
it follows for the natural frequency of surface waves in the tank:
s · ¸
¼g ¼h
!0 = tanh
b b
[Verhagen and van Wijngaarden, 1965] have investigated the shallow water wave loads in
a rolling rectangular container, with the centre of rotation at the bottom of the container.
Their expressions for the internal wave loads are rewritten and modi…ed to be useful for
any arbitrary vertical position of the centre of rotation by [Journée, 1997]. For low and
high frequencies and the frequencies near to the natural frequency, di¤erent approaches
have been used. A calculation routine has been made to connect these regions.
Á = Áa sin(!t)
In the axis-system of …gure 9.1 and after linearisation, the vertical displacement of the
tankbottom is described by:
z = s + yÁ
and after linearisation, the surface elevation of the ‡uid is described by:
z =s+h+³
Relative to the bottom of the tank, the linearised surface elevation of the ‡uid is described
by:
» = h + ³ ¡ yÁ
Using the shallow water theory, the continuity and momentum equations are:
@» @» @v
+v +» = 0
@t @y @y
@v @v @»
+v +g + gÁ = 0
@t @y @y
In these formulations, v denotes the velocity of the ‡uid in the y-direction and the vertical
pressure distribution is assumed to be hydrostatic. Therefore, the acceleration in the z-
direction, introduced by the excitation, must be small with respect to the acceleration of
gravity g, so:
Áa ! 2 b ¿ g
The boundary conditions for v are determined by the velocity produced in the horizontal
direction by the excitation. Between the surface of the ‡uid and the bottom of the tank,
214 CHAPTER 9. ANTI-ROLLING DEVICES
the velocity of the ‡uid v varies between vs and vs = cosh kh with a mean velocity: vs =kh.
However, in very shallow water v does not vary between the bottom and the surface. When
taking the value at the surface, it is required that:
b
v = ¡(s + h)Á_ at: y = §
2
For small values of Áa , the continuity equation and the momentum equation can be given
in a linearised form:
@» @v
+h = 0
@t @y
@v @»
+g + gÁ = 0
@t @y
The solution of the surface elevation » in these equations, satisfying the boundary values
for v, is: n o
b! 0 1 + g(s+h)!2 µ ¶
¼!y
» =h¡ ³ ´ sin Á
¼! cos 2! ¼! b! 0
0
Now, the roll moment follows from the quasi-static moment of the mass of the frozen liquid
½lbh and an integration of » over the breadth of the tank:
µ ¶ Z
+b=2
h
MÁ = ½glbh s + Á + ½gl »y dy
2
¡b=2
This delivers the roll moment amplitude for low and high frequencies at small water depths:
µ ¶
h
MaÁ = ½glbh s + Áa
2
½ ¾ ½ ³ ´ µ ¶ ³ ´¾
3 (s + h) ! 2 !0 3 ¼! !0 2
+½glb 1 + ¢ 2 tan ¡ Áa
g ¼! 2! 0 ¼!
For very low frequencies, so for the limit value ! ! 0, this will result into the static
moment: ½ µ ¶ ¾
h b3
MÁ = ½gl bh s + + Á
2 12
The phase lags between the roll moments and the roll motions have not been obtained
here. However, they can be set to zero for low frequencies and to -¼ for high frequencies:
"MÁ Á = 0 for: ! ¿ ! 0
"MÁ Á = ¡¼ for: ! À ! 0
Verhagen and van Wijngaarden solved the problem by using the approach in gas dynamics
when a column of gas is oscillated at a small amplitude, e.g. by a piston. At frequencies
near to the natural frequency at small water depths, they found a roll moment amplitude,
de…ned by: ( )
3
µ ¶4 r 2 2
lb 4 2Áa h ¼ b (! ¡ ! 0 )
MaÁ = ½g ¢ 1¡
12 ¼ 3b 32gÁa
The phase lags between the roll moment and the roll motion at small water depths are
given by:
¼
"MÁ Á = ¡ +® for: ! < ! 0
2
¼
"MÁ Á = ¡ ¡® for: ! > ! 0
2
with:
8s 9
< ¼ 2 b (! ¡ ! )2 =
0
® = 2 arcsin
: 24gÁa ;
(s )
¼ 2 b (! ¡ ! 0 )2
¡ arcsin
96gÁa ¡ 3¼ 2 b (! ¡ ! 0 )2
Because that the arguments of the square roots in the expression for "MÁ Á have to be
positive, the limits for the frequency ! are at least:
r r
24gÁa 24gÁa
!0 ¡ < ! < ! 0 +
b¼ 2 b¼ 2
and the dimensions of the rectangular free-surface tank are given by:
216 CHAPTER 9. ANTI-ROLLING DEVICES
It is obvious that for an anti-rolling free-surface tank, build into a ship, it holds:
Áa = 'a and !e = !
So it can be written:
Á = Áa cos(! e t + "Á³ )
Kt = Kta cos(! e t + "Á³ + "t )
Then, an additional moment has to be added to the right hand side of the equations of
motion for roll:
Ä + b44
Xtank4 = a44tank ¢ Á ¢ Á_ + c44tank ¢ Á
tank
with:
a44tank = 0
Kta
Áa
¢ sin "t
b44tank =
!e
Kta
c44tank = ¢ cos "t
Áa
218 CHAPTER 9. ANTI-ROLLING DEVICES
This holds that the anti-rolling coe¢cients a44tank , b44tank and c44tank have to be subtracted
from the coe¢cients a44 , b44 and c44 in the left hand side of the equations of motion for
roll.
40 30
Trawler Trawler
30
20
20 15 With tank
10
With tank
10
0 0
0 0.5 1.0 1.5 2.0 2.5 0 1 2 3 4 5 6
circular wave frequency (1/s Significant wave height (m
)
¯ = ¯ a cos(! e t + ²¯Á )
The exciting forces and moments, caused by an oscillating …n pair are given by:
with:
a2¯ = ¡2 sin ° ¢ a¯
b2¯ = ¡2 sin ° ¢ b¯
9.3. ACTIVE FIN STABILISERS 219
c2¯ = ¡2 sin ° ¢ c¯
a4¯ = +2(ybfin cos ° + zbfin sin °) ¢ a¯
b4¯ = +2(ybfin cos ° + zbfin sin °) ¢ b¯
c4¯ = +2(ybfin cos ° + zbfin sin °) ¢ c¯
a6¯ = ¡2xbf in sin ° ¢ a¯
b6¯ = ¡2xbf in sin ° ¢ b¯
c6¯ = ¡2xbf in sin ° ¢ c¯
and:
1 ³ c ´3
f in
a¯ = ½sf in ¢ ¢¼
2 2 Ã !
µ ¶
1 cf in @CL
b¯ = ½V Af in ¢ ¼ + ¢ C(k)
2 2 @® f in
µ ¶
1 2 @CL
c¯ = ½V Af in ¢ ¢ C(k)
2 @® f in
In here:
° = angle of port …n
µ ¶
@CL
= lift curve slope of …n
@® f in
C(k) = circulation delay function
! e cr
k = = reduced frequency
2V
Af in = projected …n area
sf in = span of …n
cf in = mean chord of …n
xbf in = xb -coordinate of the centroid of …n forces
ybf in = yb -coordinate of the centroid of …n forces
zbf in = zb -coordinate of the centroid of …n forces
The nominal lift curve slope of a …n pro…le in a uniform ‡ow is approximated by:
with:
Of normal …ns, the sweep angle of the …n pro…le is zero, so ¤=0 or cos ¤=1.
The …n acts in the boundary layer of the ship, which will reduce the lift. This e¤ect is
translated into a reduced lift curve slope of the …n.
220 CHAPTER 9. ANTI-ROLLING DEVICES
The velocity distribution in the hull boundary layer is estimated by the following two
equations:
r
±
V (±) = V ¢ 7 with: ± < ± BL
±BL
V ¢x
± BL = 0:377 ¢ xf in ¢ Rx¡0:2 with: Rx =
º
in which:
V (±) = ‡ow velocity inside boundary layer
V = forward ship speed
± = normal distance from hull
± BL = thickness of boundary layer
xf in = distance aft of forward perpendicular of …n
Rx = local Reynolds number
º = kinematic density of ‡uid
The kinematic viscosity of seawater can be found from the water temperature T in degrees
centigrade by:
1:78
º ¢ 106 = m2 /s
1:0 + 0:0336 ¢ T + 0:000221 ¢ T 2
It is assumed here that the total lift of the …n can be found from:
sfin
Z
1 1
½CL V 2 (±) ¢ c(±) ¢ d± = ½CLfin ¢ V 2 ¢ Af in
2 2
0
± = ± a cos(! e t + "±Á )
with:
a2± = +a±
b2± = +b±
c2± = +c±
a4± = ¡zbrudder ¢ a±
b4± = ¡zbrudder ¢ b±
c4± = ¡zbrudder ¢ c±
a6± = +xbrudder ¢ a±
b6± = +xbrudder ¢ b±
c6± = +xbrudder ¢ c±
and:
1 ³c ´3
rudder
a± = ½srudder ¢ ¢¼
2 2 µ µ ¶ ¶
1 crudder @CL
b± = ½Vrudder ¢ Arudder ¢ ¢ ¼+ ¢ C(k)
2 2 @® rudder
µ ¶
1 2 @CL
c± = ½V ¢ Arudder ¢ ¢ C(k)
2 rudder @® rudder
In here:
Generally a rudder is not mounted close to the hull, so the e¤ective aspect ratio is equal
to the geometric aspect ratio:
srudder
(ARE )rudder = (AR)rudder =
crudder
Chapter 10
External Linear Springs
The harmonic longitudinal, lateral and vertical displacements of a certain point P on the
ship are given by:
x(P ) = x ¡ yp ¢ Ã + zp ¢ µ
y(P ) = y + xp ¢ Ã ¡ zp ¢ Á
z(P ) = z ¡ xp ¢ µ + yp ¢ Á
The linear spring coe¢cients in the three directions in a certain point P are de…ned by
(Cpx ; Cpy ; Cpz ). The units of these coe¢cients are N/m or kN/m.
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
223
224 CHAPTER 10. EXTERNAL LINEAR SPRINGS
Xs1 = ¡Cpx ¢ (x ¡ yp ¢ Ã + zp ¢ µ)
Xs2 = ¡Cpy ¢ (y + xp ¢ Ã ¡ zp ¢ Á)
Xs3 = ¡Cpz ¢ (z ¡ xp ¢ µ + yp ¢ Á)
Xs4 = ¡Xs2 ¢ zp + Xs3 ¢ yp
Xs5 = +Xs1 ¢ zp ¡ Xs3 ¢ xp
Xs6 = ¡Xs1 ¢ yp + Xs2 ¢ xp
² Surge:
¢c11 = +Cpx
¢c12 = 0
¢c13 = 0
¢c14 = 0
¢c15 = +Cpx ¢ zp
¢c16 = ¡Cpx ¢ yp
² Sway:
¢c21 = 0
¢c22 = +Cpy
¢c23 = 0
¢c24 = ¡Cpy ¢ zp
¢c25 = 0
¢c26 = +Cpy ¢ xp
² Heave:
¢c31 = 0
¢c32 = 0
¢c33 = +Cpz
¢c34 = +Cpz ¢ yp
¢c35 = ¡Cpz ¢ xp
¢c36 = 0
10.3. LINEARIZED MOORING COEFFICIENTS 225
² Roll:
¢c41 = 0
¢c42 = ¡Cpy ¢ zp
¢c43 = +Cpz ¢ yp
¢c44 = +Cpy ¢ zp2 + Cpz ¢ yp2
¢c45 = ¡Cpz ¢ xp ¢ yp
¢c46 = ¡Cpy ¢ xp ¢ zp
² Pitch:
¢c51 = +Cpx ¢ zp
¢c52 = 0
¢c53 = ¡Cpz ¢ xp
¢c54 = ¡Cpz ¢ xp ¢ yp
¢c55 = +Cpx ¢ zp2 + Cpz ¢ x2p
¢c56 = ¡Cpx ¢ yp ¢ zp
² Yaw:
¢c61 = ¡Cpx ¢ yp
¢c62 = +Cpy ¢ xp
¢c63 = 0
¢c64 = ¡Cpy ¢ xp ¢ zp
¢c65 = ¡Cpx ¢ yp ¢ zp
¢c66 = +Cpx ¢ yp2 + Cpy ¢ x2p
A ship moving forward in a wave …eld will generate ”two sets of waves”: waves associated
with forward speed through still water and waves associated with its vertical relative motion
response to waves. Since both wave patterns dissipate energy, it is logical to conclude that
a ship moving through still water will dissipate less energy than one moving through waves.
The extra wave-induced loss of energy can be treated as an added propulsion resistance.
Figure 11.1 shows the resistance in regular waves as a function of the time: a constant
part due the calm water resistance and an oscillating part due to the motions of the ship,
relative to the incoming regular waves. The time-averaged part of the increase of resistance
is called: the added resistance due to waves, Raw .
2500
Still water resistance RSW
+
Mean added resistance RAW
Resistance
2000
Resistance (kN)
1500
1000
500
0
0 10 20 30
Time (s)
Two theoretical methods have been used for the estimation of the time-averaged added
resistance of a ship due to the waves and the resulting ship motions:
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
227
228 CHAPTER 11. ADDED RESISTANCES DUE TO WAVES
Because of the added resistance of a ship due to the waves is proportional to the relative
motions squared, its inaccuracy will be gained strongly by inaccuracies in the predicted
motions.
The transfer function of the mean added resistance is presented as:
00 Raw
Raw =
³ 2a
In a non-dimensional way the transfer function of the mean added resistance is presented
as:
00 Raw
Raw =
½g³ 2a B 2 =L
in which:
L = length between perpendiculars
B = maximum breadth of the waterline
in which:
b033 = hydrodynamic damping coe¢cient of the vertical motion of the cross section
Vz¤ = vertical average velocity of the water particles, relative to the cross sections
Te = period of vertical oscillation of the cross section
The speed dependent hydrodynamic damping coe¢cient for the vertical motion of a cross
section is de…ned here as shown before:
0
dM33
b033 = N33
0
¡V ¢
dxb
The harmonic vertical relative velocity of a point on the ship with respect to the water
particles is de…ned by:
0 D
Vz = ³_ w3 ¡ fz ¡ xb ¢ µ + yb ¢ Ág
Dt
³ ´
0
_ _ _
= ³ w3 ¡ z_ ¡ xb ¢ µ + V ¢ µ + yb ¢ Á
11.2. INTEGRATED PRESSURE METHOD 229
For a cross section of the ship, an equivalent harmonic vertical relative velocity has to be
found.
This equivalent relative velocity is de…ned by:
¤
³ ´
Vz¤ = ³_ w3 ¡ z_ ¡ xb ¢ µ_ + V ¢ µ
= Vz¤a ¢ cos(! e t + "Vz¤ ³ )
With this the radiated energy during one period of oscillation is given by:
Z µ 0
¶
¼ dM33
P = 0
N33 ¡ V ¢ ¢ Vz¤a 2 ¢ dxb
!e dxb
L
To maintain a constant forward ship speed, this energy should be delivered by the ship’s
propulsion plant. A mean added resistance Raw has to be gained.
The energy delivered to the surrounding water is given by:
µ ¶
c
P = Raw ¢ v ¡ ¢ Te
cos ¹
2¼
= Raw ¢
¡k cos ¹
From this the transfer function of the mean added resistance according to Gerritsma and
Beukelman can be found:
Z µ ¶
Raw ¡k cos ¹ 0
0
dM33 Vz¤a 2
= ¢ N33 ¡ V ¢ ¢ ¢ dxb
³ 2a 2! e dxb ³ 2a
L
This method gives good results in head to beam waves. However, in following waves this
method fails.
When the wave speed in following waves approaches the ship speed the frequency of en-
counter in the denominator tends to zero. At these low frequencies, the potential sectional
mass is very high and the potential sectional damping is almost zero. The damping mul-
tiplied with the relative velocity squared in the nominator does not tend to zero, as fast
as the frequency of encounter. This is caused by the presence of a natural frequency for
heave and pitch at this low ! e , so a high motion peak can be expected. This results into
extreme positive and negative added resistances.
with: zx = z ¡ xb µ.
As the mean added resistance during one period will be calculated, the constant term and
the …rst harmonic term can be ignored. So:
µ 2 ¶
¤ ¡³ + zx2 ³ ¢ (³ ¡ zx)
f (xb ; t) = ½g ¢ +
2 tanh [kh]
The vertical relative motion is de…ned by s = ³ ¡ zx , so:
µ 2 ¶
¤ ¡³ + zx2 ³ ¢s
f (xb ; t) = ½g ¢ +
2 tanh [kh]
The average horizontal force on a cross section follows from:
ZTe
f ¤ (xb ) = f ¤ (xb ; t) ¢ dt
0
µ ¶
½g³ 2a zx2a 2sa ¢ cos(¡kxb cos ¹ ¡ "s³ )
= ¢ ¡1 + 2 +
4 ³a ³ a ¢ tanh [kh]
The added resistance due to this force is:
Z µ ¶
¤
dyw
Raw1 = 2 f (xb ) ¢ ¡ ¢ dxb
dxb
L
Z µ ¶
½g³ 2a zx2a 2sa ¢ cos(¡kxb cos ¹ ¡ "s³ ) dyw
= 1¡ 2 ¡ ¢ ¢ dxb
2 ³a ³ a ¢ tanh [kh] dxb
L
For deep water, this part of the mean added resistance reduces to:
Z
¡½g dyw
Raw1 = s2a ¢ ¢ dxb (as given by Boese for deep water)
2 dxb
L
The integrated vertical hydromechanical and wave forces in the shipborne axis system
varies not only in time but also in direction with the pitch angle.
11.3. COMPARISON OF RESULTS 231
.
Chapter 12
The axes system (of which the hydrodynamic sign convention di¤ers from that commonly
used in structural engineering) and the internal load de…nitions are given in …gure 12.1.
To obtain the vertical and lateral shear forces and bending moments and the torsional
moments the following information over a length Lm on the solid mass distribution of the
ship including its cargo is required:
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
233
234 CHAPTER 12. BENDING AND TORSIONAL MOMENTS
m0 (xb ): distribution over the ship length of the solid mass of the ship
per unit length
0
zm (xb ): distribution over the ship length of the vertical zb values of the centre of
gravity of the solid mass of the ship per unit length
0
kxx (xb ): distribution over the ship length of the radius of inertia of the solid mass
of the ship per unit length, about a horizontal longitudinal axis
through the centre of gravity
It is obvious that this integrated mass should be equal to the mass of displacement, calcu-
lated from the underwater hull form:
m = ½r
Both terms will be calculated from independently derived data, so small deviations are
possible. A proportional correction of the masses per unit length m0 (xb ) can be used.
Then m0 (xb ) will be replaced by:
½r
m0 (xb ) ¢
m
The longitudinal position of the centre of gravity is found from the distribution of the mass
per unit length: Z
1
xG = m0 (xb ) ¢ xb ¢ dxb
m
Lm
235
xG = xB
Lm
c(xb ) = ¡c1 ¢ (xb ¡ xA ¡ 0) for: 0 < xb ¡ xA <
µ ¶ 4
Lm Lm 3Lm
c(xb ) = +c1 ¢ xb ¡ xA ¡ for: < xb ¡ xA <
2 4 4
3Lm
c(xb ) = ¡c1 ¢ (xb ¡ xA ¡ Lm ) for: < xb ¡ xA < Lm
4
with:
32 ¢ ½r ¢ (xB ¡ xG )
c1 =
L3m
In here:
For relatively slender bodies, the longitudinal gyradius of the mass can be found from the
distribution of the mass per unit length:
Z
2 1
kyy = m0 (xb ) ¢ x2b ¢ dxb
m
Lm
It can be desirable to change the mass distribution in such a way that a certain required
longitudinal gyradius kyy (new) or kzz (new) will be achieved, without changing the total
mass or the position of its centre of gravity.
236 CHAPTER 12. BENDING AND TORSIONAL MOMENTS
The position in height of the centre of gravity is found from the distribution of the heights
of the centre of gravity of the masses per unit length:
Z
1
zG = m0 (xb ) ¢ zm
0
(xb ) ¢ dxb
m
Lm
It is obvious that this value should be zero. If not so, this value has to be subtracted from
0
zm (xb ).
0 0
So, zm (xb ) will be replaced by zm (xb ) ¡ zG .
237
The transverse radius of inertia kxx is found from the distribution of the radii of inertia of
the masses per unit length:
Z
2 1 2
kxx = m0 (xb ) ¢ kxx
0
(xb ) ¢ dxb
m
Lm
If this value of kxx di¤ers from a required value kxx (new) of the gyradius, a proportional
correction of the longitudinal distribution of the radii of inertia can be used:
0 0 kxx (new)
kxx (xb ; new) = kxx (xb ; old) ¢
kxx (old)
Consider a section of the ship with a length dxb to calculate the shear forces and the
bending and the torsional moments.
When the disk is loaded by a load q(xb ), this implies for the disk:
dQ(xb )
q(xb ) ¢ dxb = ¡dQ(xb ) so: = ¡q(xb )
dxb
dM (xb )
Q(xb ) ¢ dxb = +dM (xb ) so: = +Q(xb )
dxb
in which:
The shear force and the bending moment in a cross section x1 follows from an integration
of the loads from the aftmost part of the ship x0 to this cross section x1 :
Zx1
dQ(xb )
Q(x1 ) = ¡ ¢ dxb
dxb
x0
Zx1
M (x1 ) = ¡ Q(xb ) ¢ dxb
x0
0 1
Zx1 Zxb
@ dQ(xb )
= + ¢ dxb A ¢ dxb
dxb
x0 x0
So, the shear force Q(x1 ) and the bending moment M (x1 ) in a cross section can be expressed
in the load q(xb ) by the following integrals:
Zx1
Q(x1 ) = ¡ q(xb ) ¢ dxb
x0
Zx1
M (x1 ) = + q(xb ) ¢ (x1 ¡ xb ) ¢ dxb
x0
Zx1 Zx1
= + q(xb ) ¢ xb ¢ dxb ¡ x1 ¢ q(xb ) ¢ dxb
x0 x0
For the torsional moment an approach similar to the approach for the shear force can be
used.
The load q(xb ) consists of solid mass and hydromechanical terms. The ordinates of these
terms will di¤er generally, so the numerical integrations of these two terms have to be
carried out separately.
According to Newton’s second law of dynamics, the vertical forces on the unfastened disk
of a ship in still water are given by:
with:
q3sw (xb ) = ½As g ¡ m0 g
So, the vertical shear force Q3sw (x1 ) and the bending moment Q5sw (x1 ) in still water in a
cross section can be obtained from the vertical load q3sw (xb ) by the following integrals:
Zx1
Q3sw (x1 ) = ¡ q3sw (xb ) ¢ dxb
x0
Zx1 Zx1
Q5sw (x1 ) = + q3sw (xb ) ¢ xb ¢ dxb ¡ x1 q3sw (xb ) ¢ dxb
x0 x0
For obtaining the dynamic parts of the vertical shear forces and the vertical bending
moments in regular waves, reference is given to [Fukuda, 1962]. For the lateral mode and
the roll mode a similar procedure can be followed. This will be shown at the following
pages.
with:
a022 = +M22 0
¯ ¯
¯ V dN220 ¯
+¯¯ 2 ¢ ¯
! e dxb ¯
0
dM22
b022 = +N22
0
¡V ¢
dxb
c022 = 0
a024 = +M24 0
+ OG ¢ M22 0
¯ ¯
¯ V dN24 0
V dN22 0 ¯
¯
+¯ 2 ¢ + 2 ¢ OG ¢ ¯
! e dxb !e dxb ¯
0
µ 0
¶
0 0 dM24 0 dM22
b24 = +N24 ¡ V ¢ + OG ¢ N22 ¡ V ¢
dxb dxb
0
c24 = 0
µ 0
¶
0 0 V 0 dM22
a26 = +M22 ¢ xb + 2 ¢ N22 ¡ V ¢
!e dxb
¯ ¯
¯V V dN22 0 ¯
+¯¯ 2 ¢ N220
+ 2¢ ¢ xb ¯¯
!e ! e dxb
µ 0
¶
0 0 dM22 0
b26 = + N22 ¡ V ¢ ¢ xb ¡ 2V ¢ M22
dxb
¯ 2 ¯
¯V dN22 ¯¯
0
+¯¯ 2 ¢
! e dxb ¯
c026 = 0
12.3. VERTICAL DYNAMIC LOADS 241
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
Then the harmonic lateral shear forces Q2 (x1 ) and the bending moments Q6 (x1 ) in waves in
cross section x1 can be obtained from the horizontal load q2 (xb ) by the following integrals:
Q2 (x1 ) = Q2a cos(! e t + "Q2 ³ )
Zx1
= ¡ q2 (xb ) ¢ dxb
x0
Q6 (x1 ) = Q6a cos(! e t + "Q6 ³ )
Zx1 Zx1
= + q2 (xb ) ¢ xb ¢ dxb ¡ x1 q2 (xb ) ¢ dxb
x0 x0
with:
a011 = +M110
¯ ¯
¯ V dN110 ¯
¯
+¯ 2 ¢ ¯
! e dxb ¯
0
dM11
b011 = 0
+N11 ¡V ¢ + b011V
dxb
c011 = 0
a013 = 0
b013 = 0
c013 = 0
a015 = ¡M11 0
¢ bG
¯ ¯
¯ ¡V dN11 0 ¯
+¯¯ 2 ¢ ¢ bG¯¯
!e dxb
µ 0
¶
dM11
b015 0
= ¡ N11 ¡ V ¢ ¢ bG ¡ b011V ¢ bG
dxb
c015 = 0
with:
a031 = 0
b031 = 0
c031 = 0
a033 = 0
+M33
12.3. VERTICAL DYNAMIC LOADS 243
¯ ¯
¯ V dN330 ¯
+¯¯ 2 ¢ ¯
! e dxb ¯
0
dM33
b033 = +N33
0
¡V ¢
dxb
c033 = +2½g ¢ yw
µ 0
¶
V dM33
a035 = 0
¡M33 0
¢ xb ¡ 2 ¢ N33 ¡ V ¢
!e dxb
¯ ¯
¯ ¡V V dN33 0 ¯
+¯¯ 2 ¢ N33 0
¡ 2¢ ¢ xb ¯¯
!e ! e dxb
µ 0
¶
dM33
b035 0
= ¡ N33 ¡ V ¢ ¢ xb + 2V ¢ M33 0
dxb
¯ 2 ¯
¯V dN33 ¯¯
0
+¯¯ 2 ¢
! e dxb ¯
c035 = ¡2½g ¢ yw ¢ xb
The sectional wave loads for surge and heave are given by:
¤
Xw0 1 (xb ) = +M11 0
¢ ³Äw1
¯ ¯
¯ V dN 0
¤ ¯
+¯¯ ¢ 11
¢ ³ w1 ¯¯
Ä
! ¢ ! e dxb
ï ¯ !
¯!¯ dM 0
¤
+ ¯¯ ¯¯ ¢ N11 0
¡V ¢ 11
¢ ³_ w1
!e dxb
+XF0 K1
¤
Xw0 3 (xb ) = +M330
¢ ³Äw3
¯ ¯
¯ V 0
dN33 ¤ ¯
¯
+¯ ¢ ¢ ³ w3 ¯¯
Ä
! ¢ ! e dxb
ï ¯ !
¯!¯ dM 0
¤
+ ¯¯ ¯¯ ¢ N33 0
¡V ¢ 33
¢ ³_ w3
!e dxb
+XF0 K3
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
Then the harmonic vertical shear forces Q3 (x1 ) and the bending moments Q5 (x1 ) in waves
in cross section x1 can be obtained from the longitudinal and vertical load q1 (xb ) and q3 (xb )
by the following integrals:
Figure 12.10 shows a comparison between measured and calculated distributions of the
vertical wave bending moment amplitudes over the length of the ship.
According to Newton’s second law of dynamics, the harmonic torsional dynamic load per
unit length on the unfastened disk about an longitudinal axis at a distance z1 above the
ship’s center of gravity is given by:
+z1 ¢ q2 (xb )
12.4. TORSIONAL DYNAMIC LOADS 245
with:
a042 = +M42 0
+ OG ¢ a022
¯ ¯
¯ V dN42 0 ¯
¯
+¯ 2 ¢ ¯
! dxb ¯
e
0
dM42
b042 = +N42
0
¡V ¢ + OG ¢ b022
dxb
c042 = 0 + OG ¢ c022
a044 = +M44 0
+ OG ¢ M42 0
+ OG ¢ a024
¯ ¯
¯ V dN44 0
V dN 0 ¯
+¯¯ 2 ¢ + 2 ¢ OG ¢ 42 ¯
! e dxb !e dxb ¯
0
µ 0
¶
0 0 dM44 0 dM42
b44 = +N44 ¡ V ¢ + OG ¢ N42 ¡ V ¢ + b044V + OG ¢ b024
dxb dxb
µ 3 ¶
yw As
c044 = +2½g ¢ ¡ ¢ bG
3 2
µ 0
¶
0 0 V 0 dM42
a46 = +M42 ¢ xb + 2 N42 ¡ V ¢ + OG ¢ a026
!e dxb
¯ ¯
¯V V dN 0 ¯
+¯¯ 2 ¢ N420
+ 2¢ 42
¢ xb ¯¯
!e ! e dxb
µ 0
¶
0 0 dM42 0
b46 = + N42 ¡ V ¢ ¢ xb ¡ 2V ¢ M42 + OG ¢ b026
dxb
¯ 2 ¯
¯V dN42 ¯¯
0
+¯¯ 2 ¢
! e dxb ¯
c046 = 0 + OG ¢ c026
In here, bG is the vertical distance of the centre of gravity of the ship G above the centroid
b of the local submerged sectional area.
The sectional wave load for roll is given by:
¤
Xw0 4 (xb ) = +M44 0
¢ ³Äw4
¯ ¯
¯ V dN 0
¤ ¯
+¯¯ ¢ 44
¢ ³Äw4 ¯¯
! ¢ ! e dxb
ï ¯ !
¯!¯ 0 dM 0
¤
+ ¯¯ ¯¯¢N44 ¡ V ¢ 44
¢ ³_ w4
!e dxb
+XF0 K4
¤
0
+M42 ¢ ³Äw2
246 CHAPTER 12. BENDING AND TORSIONAL MOMENTS
¯ ¯
¯ V dN42 0
¤ ¯
+¯¯ ¢ ³Äw2 ¯¯
! ¢ ! e dxb
ï ¯ !
¯!¯ 0 dM 0
¤
+ ¯¯ ¯¯¢N42 ¡ V ¢ 42
¢ ³_ w2
!e dxb
+OG ¢ Xw0 2
The ”Modi…ed Strip Theory” includes the outlined terms. When ignoring the outlined
terms the ”Ordinary Strip Theory” is presented.
Then the harmonic torsional moments Q4 (x1 ; z1 ) in waves in cross section x1 at a distance
z1 above the ship’s centre of gravity can be obtained from the torsional load q4 (xb ; z1 ) by
the following integral:
² the Neumann wave spectrum, a somewhat wide wave spectrum, which is sometimes
used for open sea areas
² the Bretschneider wave spectrum, an average wave spectrum, frequently used in open
sea areas
² the Mean JONSWAP wave spectrum, a narrow wave spectrum, frequently used in
North Sea areas.
² the average wave period T1 , based on the centroid of the spectral area curve.
To obtain the average zero-crossing period T2 or the spectral peak period Tp , a …xed relation
with T1 can be used not-truncated spectra.
From these wave energy spectra and the transfer functions of the responses, the response
energy spectra can be obtained.
Generally the frequency ranges of the energy spectra of the waves and the responses of
the ship on these waves are not very wide. Then the Rayleigh distribution can be used to
obtain a probability density function of the maximum and minimum values of the waves
and the responses. With this function, the probabilities on exceeding threshold values by
the ship motions can be calculated.
Bow slamming phenomena are de…ned by a relative bow velocity criterium and a peak
bottom impact pressure criterium.
0
J.M.J. Journée, ”Theoretical Manual of SEAWAY, Release 4.19”, Report 1216a, February 2001, Ship
Hydromechanics Laboratory, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands.
For updates see web site: http://dutw189.wbmt.tudelft.nl/~johan or http://www.shipmotions.nl.
247
248 CHAPTER 13. STATISTICS IN IRREGULAR WAVES
A comparison of the Neumann, the Bretschneider and the mean JONSWAP wave spectra
is given here for a sea state with a signi…cant wave height of 4 meters and an average wave
period of 8 seconds.
This formulation is accepted by the 2nd International Ship Structures Congress in 1967
and the 12th International Towing Tank Conference in 1969 as a standard for seakeeping
calculations and model experiments. This is reason why this spectrum is also called I.S.S.C.
or I.T.T.C. Wave Spectrum.
The original One-Parameter Pierson-Moskowitz Wave Spectrum for fully developed seas
can be obtained from this de…nition by using a …xed relation between the signi…cantp wave
height and the average wave period in this Bretschneider de…nition: T1 = 3:861 ¢ H1=3 .
A = 0:658
8 Ã !2 9
< !
¡ 1:0 =
!p
B = exp ¡ p
: ¾ 2 ;
° = 3:3 (peakedness factor)
2¼
!p = (circular frequency at spectral peak)
Tp
¾ = a step function of !: if ! < ! p then ¾ = 0:07
if ! > ! p then ¾ = 0:09
The JONSWAP expression is equal to the Bretschneider de…nition multiplied by the fre-
quency function A ¢ ° B .
Sometimes, a third free parameter is introduced in the JONSWAP wave spectrum by
varying the peakedness factor °.
Truncation of wave spectra during numerical calculations, can cause di¤erences between
input and calculated wave periods. Generally, the wave heights will not di¤er much.
In …gure 13.2 and the table below, for ”Open Ocean Areas” and ”North Sea Areas” an
indication is given of a possible average relation between the scale of Beaufort or the wind
velocity at 19.5 meters above the sea level and the signi…cant wave height H1=3 and the
average wave periods T1 or T2 .
13.1. NORMALIZED WAVE ENERGY SPECTRA 251
H1=3 T1 T2 H1=3 T1 T2 °
(kn) (m) (s) (s) (m) (s) (s) (-)
These data are an indication only, a generally applicable …xed relation between wave heights
and wave periods does not exist.
Other open ocean de…nitions for the North Atlantic and the North Paci…c, obtained from
[Bales, 1983] and adopted by the 17th ITTC (1984), are given in the tables below.
The modal or central periods in these tables correspond with the peak period Tp .
For not-truncated spectra, the relations with T1 and T2 are de…ned before.
252 CHAPTER 13. STATISTICS IN IRREGULAR WAVES
North Paci…c
Note:
1) Ambient wind sustained at 19.5 m above surface to generate fully-developed seas.
To convert to another altitude h2 , apply V2 = V1 ¢ (h2 =19:5)1=7 .
2) Minimum is 5 percentile and maximum is 95 percentile for periods given wave
height range.
3) Based on periods associated with central frequencies included in Hindcast Clima-
tology.
13.2. RESPONSE SPECTRA AND STATISTICS 253
This has been visualized for a heave motion in …gures 13.3 and 13.4.
From the spectral density function of a response the signi…cant amplitude can be calculated.
254 CHAPTER 13. STATISTICS IN IRREGULAR WAVES
2
Wave H1/3 = 5.00 m Wave H1/3 = 5.00 m
4 spectrum T2 = 8.00 4 spectrum T2 = 8.00 s
s
3 3
2 2
1 1
0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
Transfer function heave (m/m)
2.0 2.0
Transfer Containership Transfer Containership
function L = 175 metre function L = 175 metre
1.5 1.5
heav Head waves heave Head waves
e
V = 20 knots V = 20 knots
1.0 1.0
0.5 0.5
0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
Spectral density heave (m s)
8 8
2
4 4
2 2
0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
wave frequency (rad/s) frequency of encounter (rad/s)
Figure 13.4: Heave Spectra in the Wave and Encounter Frequency Domain
The signi…cant amplitude is de…ned to be the mean value of the highest one-third part of
the highest wave heights, so:
p
ra1=3 = 2 m0r
A mean period can be found from the centroid of the spectrum by:
m0r
T1r = 2¼ ¢
m1r
An other de…nition, which is equivalent to the average zero-crossing period, is found from
the spectral radius of inertia by:
r
m0r
T2r = 2¼ ¢
m2r
The probability density function of the maximum and minimum values, in case of a spec-
trum with a frequency range which is not too wide, is given by the Rayleigh distribution:
½ ¾
ra ¡ra2
f (ra ) = ¢ exp
m0r 2m0r
This implies that the probability of exceeding a threshold value a by the response amplitude
ra becomes:
½ ¾
Z1
ra ¡ra2
P fra > ag = ¢ exp ¢ dra
m0r 2m0r
a
½ ¾
¡a2
= exp
2m0r
13.2. RESPONSE SPECTRA AND STATISTICS 255
The number of times per hour that this happens follows from:
3600
Nhour = ¢ P fra > ag
T2r
The spectral value of the waves S³ (! e ), based on ! e , is not equal to the spectral value
S³ (!), based on !.
Because of the requirement of an equal amount of energy in the frequency bands ¢! and
¢! e, it follows:
S³ (! e ) ¢ d! e = S³ (!) ¢ d!
From this the following relation is found:
S³ (!)
S³ (! e ) = d!e
d!
The relation between the frequency of encounter and the wave frequency, of which an
example is illustrated in …gure 13.5, is given by:
! e = ! ¡ kV ¢ cos ¹
So:
kg
d! g ¢ tanh(kh) + h¢cosh2
(kh)
= p
dk 2 kg ¢ tanh(kh)
As can be seen in …gure 13.5, in following waves the derivative d! e =d! can approach from
both sides, a positive or a negative side, to zero. As a result of this, around a wave speed
equal to twice the forward ship speed component in the direction of the wave propagation,
the transformed spectral values will range from plus in…nite to minus in…nite. This implies
that numerical problems will arise in the numerical integration routine.
This is the reason why the spectral moments have to be written in the following format:
Z1 Z1
m0r = Sr (! e) ¢ d! e = Sr (!) ¢ d!
0 0
Z1 Z1
m1r = Sr (! e) ¢ ! e ¢ d! e = Sr (!) ¢ ! e ¢ d!
0 0
Z1 Z1
m2r = Sr (! e) ¢ ! 2e ¢ d! e = Sr (!) ¢ ! 2e ¢ d!
0 0
with: µ ¶2
ra
Sr (!) = ¢ S³ (!)
³a
If Sr (! e ) has to be known, for instance for a comparison of the calculated response spectra
with measured response spectra, these values can be obtained from this Sr (!) and the
derivative d! e =d!. So an integration of Sr (! e ) over ! e has to be avoided.
Because of the linearities, the calculated signi…cant values can be presented by:
ra1=3
versus T1 or T2
H1=3
Z1
RAW
RAW = 2 ¢ S³ (!) ¢ d!
³ 2a
0
Because of the linearities in the motions, the calculated mean added resistance values can
be presented by:
RAW
2
versus T1 or T2
H1=3
13.3. SHIPPING GREEN WATER 257
L
fe = f ¡ 0:75 ¢ B ¢ ¢ Fn2
LE
with:
An oscillating ship will produce waves and these dynamic phenomena will in‡uence the
amplitude of the relative motion. A dynamic swell up should be taken into account.
[Tasaki, 1963] carried out forced oscillation tests with ship models in still water and ob-
tained an empirical formula for the dynamic swell-up at the forward perpendicular in head
waves: s
¢sa CB ¡ 0:45 ! 2e L
= ¢
sa 3 g
with the restrictions:
In this formula sa is the amplitude of the relative motion at the forward perpendicular as
obtained in head waves, calculated from the heave, the pitch and the wave motions.
Then the actual amplitude of the relative motions becomes:
s¤a = sa + ¢sa
The spectral density of the vertical relative motion at the forward perpendicular is given
by:
µ ¤ ¶2
sa
Ss¤ (!) = ¢ S³ (!)
³a
258 CHAPTER 13. STATISTICS IN IRREGULAR WAVES
When using the Rayleigh distribution the probability of shipping green water is given by:
½ ¾
¤ ¡fe2
P fsa > fe g = exp
2m0s¤
The average zero-crossing period of the relative motion is found from the spectral radius
of inertia by: r
m0s¤
T2s¤ = 2¼
m2s¤
The number of times per hour that green water will be shipped follows from:
3600
Nhour = ¢ P fs¤a > fe g
T2s¤
² an emergence of the bow of the ship at 10 percentile of the length aft of the forward
perpendiculars
Ochi de…nes the vertical relative displacement and velocity of the water particles with
respect to the keel point of the ship by:
s = ³ xb ¡ z + xb ¢ µ
s_ = ³_ ¡ z_ + xb ¢ µ_
xb
13.4. BOW SLAMMING 259
with:
Some of these data, as for instance presented by [Lloyd, 1989] as a function of the deadrise
angle ¯, are illustrated in …gure 13.6.
An equivalent deadrise angle ¯ is de…ned here by the determination of an equivalent wedge.
The contour of the cross section inside 10 percentile of the half breadth B=2 of the ship
has been used to de…ne an equivalent wedge with a half breadth: b = 0:10 ¢ B=2.
The accessory draught t of the wedge follows from the section contour.
In the forebody of the ship, this draught can be larger than 10 percentile of the amidships
draught T . If so, the section contour below 0:10 ¢ T has been used to de…ne an equivalent
wedge: t = 0:10 ¢ T .
If this draught is larger than the local draught, the local draught has been used.
The accessory half breadth b of the wedge follows from the section contour.
with:
Then: ½ ¾
¡Ds2 ¡s_ 2cr
P fslamg = exp +
2m0s 2m0s_
Note that, because of including the forward speed e¤ect, the spectral moment of the ve-
locities does not follow from the spectral density of the relative displacement as showed in
the de…nition of Ochi.
The average period of the relative displacement is found by:
r r
m0s m0s
T2s = 2¼ = 2¼
m2s m0s_
Then the number of times per hour that a slam will occur follows from:
3600
Nhour = ¢ P fslamg
T2s
Chapter 14
Twin-Hull Ships
When not taking into account the interaction e¤ects between the two individual hulls, the
wave loads and motions of twin-hull ships can be calculated easily. Each individual hull
has to be symmetric with respect to its centre plane. The distance between the two centre
planes of the single hulls should be constant. The coordinate system for the equations of
motion of a twin-hull ship is given in …gure 14.1.
263
264 CHAPTER 14. TWIN-HULL SHIPS
L
266 CHAPTER 14. TWIN-HULL SHIPS
Z
¡ ¢
+ XF0 K1 (ps) + XF0 K1 (sb) ¢ dxb
L
+z1 ¢ qT 2 (xb )
In here:
The calculation procedure of the forces and moments is similar to the procedure given
before for mono-hull ships.
270 CHAPTER 14. TWIN-HULL SHIPS
.
Chapter 15
Numerical Recipes
Some typical numerical recipes, as used in the strip theory program SEAWAY, are described
in more detail here.
15.1 Polynomials
Discrete points can be connected by a …rst degree or a second degree polynomial, see …gure
15.1-a,b.
Figure 15.1: First and Second Order Polynomials Through Discrete Points
f (x) = ax + b
271
272 CHAPTER 15. NUMERICAL RECIPES
f (x0 ) ¡ f (xm )
a =
x0 ¡ xm
b = f (x0 ) ¡ ax0
f (xp ) ¡ f (x0 )
a =
xp ¡ x0
b = f (x0 ) ¡ ax0
Notify that only one interval is required for obtaining the coe¢cients in that interval.
f (x) = ax2 + bx + c
Notify that two intervals are required for obtaining these coe¢cients, valid in both intervals.
15.2 Integrations
Numerical integrations can be carried out by either the trapezoid rule or Simpson’s general
rule.
SEAWAY uses Simpson’s general rule as a standard. Then, the integrations have to be
carried out over a number of sets of two intervals, see …gure 15.1-b. Numerical inaccuracies
can be expected when x0 ¡xm ¿ xp ¡x0 or xp ¡x0 ¿ x0 ¡xm . In those cases the trapezoid
rule has to be preferred, see …gure 15.1-a.
SEAWAY makes the choice bewteen the use of the trapezoid rule andr Simpson’s rule
automatically, based on the following requirements:
x0 ¡ xm x0 ¡ xm
Trapezoid rule if: < 0:2 or: > 5:0
xp ¡ xm xp ¡ xm
x0 ¡ xm
Simpson’s rule if: 0:2 < < 5:0
xp ¡ xm
15.2. INTEGRATIONS 273
Zxp Zxp
f (x) dx = (ax + b) dx
x0 x0
· ¸xp
1 2
= ax + bx
2 x0
with:
f (xp ) ¡ f (x0 )
a =
xp ¡ x0
b = f (x0 ) ¡ ax0
Integration over two intervals results into:
Zxp
(x0 ¡ xm ) f (xm ) + (xp ¡ xm ) f (x0 ) + (xp ¡ x0 ) f (xm )
f (x) dx =
2
xm
Zxp Zxp
¡ ¢
f (x) dx = ax2 + bx + c dx
xm xm
· ¸xp
1 3 1 2
= ax + bx + cx
3 2 xm
with:
Some algebra leads for the integration over these two intervals to:
Zxp (
x0 ¡ xm ¡ xp ¡x
2
0
f (x) dx = f (xm )
x0 ¡ xm
xm
(xp ¡ xm )2
+ f (x0 )
2 (x0 ¡ xm ) (xp ¡ x0 )
¾
xp ¡ x0 ¡ x0 ¡x
2
m
xp ¡ xm
+ f (xp ) ¢
xp ¡ x0 3
The in-phase and out-phase parts of the wave loads have to be obtained from longitudinal
integrations:
Z Z
0 0
Fw1 = Fw1 (xb ) dx = fw1 (xb ) cos xb dxb
L L
Z Z
0 0
Fw2 = Fw2 (xb ) dx = fw2 (xb ) sin xb dxb
L L
0 0
Direct numerical integrations of Fw1 and Fw2 over the ship length, L, require integration
intervals, ¢xb , which are much smaller than the smallest wave length, ¢xb 6 ¸min =10.
This means that a large number of cross sections are required.
0 0 0
This can be avoided by writing Fw1;2 in terms of fw1;2 (xb ) cos xb and fw1;2 (xb ) sin xb , in
0
which the integrands fw1;2 (xb ) vary very slow over short wave lengths. These functions
0
fw1;2 (xb ) can be approximated by second degree polynomials:
f (x) = ax2 + bx + c
Z
cos x dx = + sin x
Z
x cos x dx = + cos x + x sin x
Z
¡ ¢
x2 cos x dx = +2x cos x + x2 ¡ 2 sin x
and:
15.3. DERIVATIVES 275
Z
sin x dx = ¡ cos x
Z
x sin x dx = + sin x ¡ x cos x
Z
¡ ¢
x2 sin x dx = +2x sin x ¡ x2 ¡ 2 cos x
the following expressions can be obtained for the in-phase and out-phase parts of the wave
loads, integrated from xm through xp , so over the two intervals x0 ¡ xm and xp ¡ x0 :
Zxp Zxp
F (x) dx = f (x) cos x dx
xm xm
Zxp
¡ ¢
= ax2 + bx + c cos x dx
xm
Zxp Zxp Zxp
= a x2 cos x dx + b x cos x dx + c cos x dx
xm xm xm
= [+ (f (x) ¡ 2a) sin x + (2ax + b) cos x]xxpm
Zxp Zxp
F (x) dx = f (x) sin x dx
xm xm
Zxp
¡ ¢
= ax2 + bx + c sin x dx
xm
Zxp Zxp Zxp
2
= a x sin x dx + b x sin x dx + c sin x dx
xm xm xm
= [¡ (f (x) ¡ 2a) cos x + (2ax + b) sin x]xxpm
with coe¢cients obtained by:
15.3 Derivatives
First and second degree functions, of which the derivatives have to be determined, have
been given in …gure 15.2-a,b.
276 CHAPTER 15. NUMERICAL RECIPES
df (x)
for x < x0 : = am
dx
df (x)
for x > x0 : = ap
dx
It is obvious that, generally, the derivative at the left hand side of x0 - with index m
(minus) - and the derivative at the right hand side of x0 - with index p (plus) - will di¤er:
"µ ¶ # "µ ¶ #
df df
6=
dx x=x0 dx x=x0
m (minus or left of x0 ) p (plus or right of x0 )
df
A mean derivative at x = x0 can be obtained by:
dx
h¡ ¢ i h¡ ¢ i
µ ¶ df
(x0 ¡ xm ) ¢ dx + (xp ¡ x0 ) ¢ df
df x=x0 m dx x=x0
p
=
dx x=x0 xp ¡ xm
Zxp
smp = ds
xm
Zxp p
= dx2 + dy 2
xm
with:
xp ¡ xm
cos ® = q
(xp ¡ xm )2 + (yp ¡ ym )2
yp ¡ ym
sin ® = q
(xp ¡ xm )2 + (yp ¡ ym )2
15.4. CURVE LENGTHS 279
.
Bibliography
[Bales, 1983] Bales, S. L. (1983). Wind and Wave Data for Seakeeping Performance As-
sessment. Technical report, Prepared for Seakeeping Committee ITTC, Athens, Greece.
[Boese, 1970] Boese, P. (1970). Eine Einfache Methode zur Berechnung der Wiederstand-
serhöhung eines Schi¤es in Seegang. Technical Report 258, Institüt für Schi¤bau der
Universität Hamburg, BRD.
[Bosch and Vugts, 1966] Bosch, J. J. v. d. and Vugts, J. H. (1966). Roll Damping by
Free Surface Tanks. Technical Report 83-S, Netherlands Ship Research Centre TNO,
Shipbuilding Department, Delft, The Netherlands.
[Conolly, 1974] Conolly, J. E. (1974). Standards of Good Seakeeping for Destroyers and
Frigates in Head Seas. In International Symposium on the Dynamics of Marine Vehicles
and Structures in Waves, London, UK, number 8.
[Dimitrieva, 1994] Dimitrieva, I. (1994). DELFRAC, 3-D Potential Theory Including Wave
Di¤raction and Drift Forces Acting on the Structures. Technical Report 1017, Delft
University of Technology, Ship Hydromechanics Laboratory, The Netherlands.
[Frank, 1967] Frank, W. (1967). Oscillation of Cylinders in or below the Free Surface of
Deep Fluids. Technical Report 2375, Naval Ship Research and Development Centre,
Washington DC, USA.
[Fukuda, 1962] Fukuda, J. (1962). Coupled Motions and Midship Bending Moments of a
Ship in Regular Waves. Journal of the Society of Naval Architects of Japan.
[Grim, 1953] Grim, O. (1953). Berechnung der durch Schwingungen eines Schi¤skörpers
Erzeugten Hydrodynamischen Kräfte. Jahrbuch der Schi¤sbautechnischen Gesellschaft,
47:277–299.
281
282 BIBLIOGRAPHY
[Grim, 1955] Grim, O. (1955). Die Hydrodynamischen Kräfte beim Rollversuch. Schi¤-
stechnik, 3.
[Haskind, 1957] Haskind, M. (1957). The Exciting Forces and Wetting of Ships in Waves
(in Russian). Izvestia Akademii Nauk SSSR, Otdelenie Tskhnicheshikh Nauk, (7):65–79.
[Huijsmans, 1996] Huijsmans, R. H. M. (1996). Motions and Drift Forces on Moored Ves-
sels in Current. PhD thesis, Delft University of Technology, The Netherlands.
[Ikeda et al., 1978] Ikeda, Y., Himeno, Y., and Tanaka, N. (1978). A Prediction Method
for Ship Rolling. Technical Report 00405, Department of Naval Architecture, University
of Osaka Prefecture, Japan.
[John, 1950] John, F. (1950). On the Motions of Floating Bodies, Part II. Comm. on Pure
and Applied Mathematics, pages 45–100.
[Journée, 1992] Journée, J. M. J. (1992). Strip Theory Algorithms, Revised Report 1992.
Technical Report 912, Delft University of Technology, Ship Hydromechanics Laboratory,
The Netherlands.
[Journée and van ’t Veer, 1995] Journée, J. M. J. and van ’t Veer, A. P. (1995). First
Order Wave Loads in Beam Waves. In Proceedings 5th International O¤shore and Polar
Engineering Conference, The Hague. (Internet: http://dutw189.wbmt.tudelft.nl/~johan
or http://www.shipmotions.nl).
BIBLIOGRAPHY 283
[Keil, 1974] Keil, H. (1974). Die Hydrodynamische Kräfte bei der periodischen Bewegung
zweidimensionaler Körper an der Ober‡äche ‡acher Gewasser. Technical Report 305,
University of Hamburg, BRD.
[Kerczek and Tuck, 1969] Kerczek, C. v. and Tuck, E. O. (1969). The Representation of
Ship Hulls by Conformal Mapping Functions. Journal of Ship Research, 13(4).
[Korkut and Hebert, 1970] Korkut, M. and Hebert, E. (1970). Some Notes on Static An-
chor Chain Curve. In O¤shore Technology Conference, 1970, Dallas, Texas, USA.,
number OTC 1160, Dallas, Texas, USA.
[Kumai, 1959] Kumai, T. (1959). Added Mass Moment of Inertia Induced by Torsional
Vibration of Ships. Reports of the Research Institute for Applied Mechanics, VII(28).
[Lamb, 1932] Lamb, H. (1932). Hydrodynamics. Dover Publications, Inc., New York, 6
edition.
[Lewis, 1929] Lewis, F. M. (1929). The Inertia of Water Surrounding a Vibrating Ship. In
Transactions SNAME.
[Miller, 1974] Miller, E. R. (1974). Unknown Title of a Report on Roll Damping. Technical
Report 6136-74-280, NAVSPEC.
[Newman, 1962] Newman, J. N. (1962). The Exciting Forces on Fixed Bodies in Waves.
Journal of Ship Research, 6(4):10–17.
[Ochi, 1964] Ochi, M. K. (1964). Prediction of Occurence and Severity of Ship Slamming
at Sea. In Proceedings of 5th O.N.R. Symposium, Bergen, Norway.
[Ogilvie, 1963] Ogilvie, T. F. (1963). First and Second Order Forces on a Cylinder Sub-
merged under a Free Surface. Journal of Fluid Mechanics, pages 451–472.
[Porter, 1960] Porter, W. R. (1960). Pressure Distribution, Added Mass and Damping
Coe¢cients for Cylinders Oscillating in a Free Surface. Technical Report 82, University
of California, Institute of Engineering Research, Berkeley, California, USA.
[Reed and Nowacki, 1974] Reed, M. A. and Nowacki, H. (1974). Interactive Creation of
Fair Ship Lines. Journal of Ship Research, 18(2).
[Sargent and Kaplan, 1974] Sargent, T. P. and Kaplan, P. (1974). Modi…cations to Lloyds
Register of Shipping Strip Theory Computerprogram (LR 2570). Technical Report 74-
103, Oceanics Inc.
284 BIBLIOGRAPHY
[Tasai, 1959] Tasai, F. (1959). On the Damping Force and Added Mass of Ships Heav-
ing and Pitching. Technical report, Research Institute for Applied Mechanics, Kyushu
University, Japan. Vol. VII, No 26.
[Tasai, 1960] Tasai, F. (1960). Formula for Calculating Hydrodynamic Force on a Cylinder
Heaving in the Free Surface, (N-Parameter Family). Technical report, Research Institute
for Applied Mechanics, Kyushu University, Japan. Vol. VIII, No 31.
[Tasai, 1961] Tasai, F. (1961). Hydrodynamic Force and Moment Produced by Swaying
and Rolling Oscillation of Cylinders on the Free Surface. Technical report, Research
Institute for Applied Mechanics, Kyushu University, Japan. Vol. IX, No 35.
[Tasai, 1965] Tasai, F. (1965). Ship Motions in Beam Waves. Technical report, Research
Institute for Applied Mechanics, Kyushu University, Japan. Vol. XIII, No 45.
[Tasai, 1969] Tasai, F. (1969). Improvements in the Theory of Ship Motions in Longitudi-
nal Waves. In Proceedings 12th I.T.T.C.
[Timman and Newman, 1962] Timman, R. and Newman, J. N. (1962). The Coupled
Damping Coe¢cients of a Symmetric Ship. Journal of Ship Research, 5(4):1–7.
[Troost, 1955] Troost, L. (1955). A Simpli…ed Method for Preliminary Powering of Single
Screw Merchant Ships. New England Section of the Society of Naval Arcitects and Marine
Engineers, October Meeting.
[Ursell, 1949] Ursell, F. (1949). On the Heaving Motion of a Circular Cylinder on the
Surface of a Fluid. Quarterly Journal of Mechanics and Applied Mathematics, II.
[Verhagen and van Wijngaarden, 1965] Verhagen, J. and van Wijngaarden, L. (1965).
Non-linear oscillations of ‡uid in a container. Journal of Fluid Mechanics, 22(4):737–751.
[Zwaan, 1977] Zwaan, A. P. d. (1977). Method for Solving Equations with Constant or
Variable Right Hand Terms (in Dutch). Technical Report 5, Delft University of Tech-
nology, Department for Shipbuilding and Shiphandling, Centrale Werkgroep Wiskunde,
The Netherlands.