Professional Documents
Culture Documents
TU DELFT
September 2003
AMARCON
Summary
This report aims to be a guide and help for those people who want to study the theoretical
backgrounds and the algorithms of a ship motions computer code based on the strip theory.
The underlying report describes in detail the theoretical backgrounds and algorithms used by
the first author during the development of his six-degrees-of-freedom ship motions computer
code, called SEAWAY.
The six ship motions of and about the centre of gravity G of the vessel have been defined in
the next figure.
{M
6
j =1
ij
for: i = 1,...6
Because a linear system has been considered here, the forces and moments in the right hand
side of these equations consist of a superposition of:
so-called hydromechanic forces and moments, caused by a harmonic oscillation of the
rigid body in the undisturbed surface of a fluid being previously at rest, and
so-called exciting wave forces and moments on the restrained body, caused by the
incoming harmonic waves.
With this, the system of a with six degrees of freedom moving ship in waves can considered to
be a linear mass-damper-spring system with frequency-dependent coefficients and linear
exciting forces and moments:
{(M
6
j =1
ij
for: i = 1,...6
In here, x i with indices i = 1,2,3 are the displacements of G (surge, sway and heave) and x i
with indices i = 4,5,6 are the rotations about the axes through G (roll, pitch and yaw). The
indices ij present at motion i the coupling with motion j .
The masses in the equations of motion above consist of solid masses or solid mass moments
of inertia of the ship ( M ij ) and added masses or added mass moments of inertia caused
by the disturbed water, the hydrodynamic masses or mass moments of inertia (a ij ). An
oscillating ship generates waves it self too; energy will be radiated from the ship. The
hydrodynamic damping-terms ( bij x& i ) account for this. For the heave, roll and pitch motions,
hydrostatic spring-terms ( c ij x i ) have to be added. The right hand sides of the equations of
motion consist of exciting wave forces and moments ( Fi ).
In the so-called strip theory, the ship will be divided in 20 tot 30 cross sections, of which the
two-dimensional hydromechanic coefficients and exciting wave loads will be calculated. To
obtain the three-dimensional values, these values will be integrated over the ship length
numerically. Finally, the differential equations will be solved to obtain the motions. These
calculations will be performed in the frequency domain.
It was in 1949 that Ursell published his potential theory for determining the hydrodynamic
coefficients of semicircular cross sections, oscillating in deep water in the frequency domain.
Using this, for the first time a rough estimation could be made of the motions of a ship in
regular waves at zero forward speed.
Shortly after that Tasai, Grim, Gerritsma and many other scientists used various already
existing conformal mapping techniques (to transform ship-like cross sections to a semicircle)
together with Ursells theory, in such a way that the motions in regular waves of more realistic
hull forms could be calculated too. Most popular was the 2-parameter Lewis conformal
mapping technique. The exciting wave loads were found from the loads in undisturbed waves
the so-called Froude-Krilov forces or moments completed with diffraction terms
accounting for the presence of the ship in these waves.
Borrowed from the broadcasting technology, Denis en Pierson published in 1953 a
superposition method to describe the irregular waves too. The sea was considered to be the
sum of many simple harmonic waves; each wave with its own frequency, amplitude, direction
and random phase lag. By calculating the responses of the ship on each of these individual
harmonic waves and adding up the responses of the ship, the energy distribution of the ships
behaviour in irregular waves could be found. These irregular motions are characterised by
significant amplitude and average period.
However, these theories provided the motions at zero forward speed only. In 1957, KorvinKroukovski en Jacobs published a method - which was further improved in the sixties - to
account for the effect of forward ship speed.
So at the end of the fifties, all components for an elementary ship motions computer program
for deep water were already available.
Fukuda published in 1962 a calculation technique for the internal sheer forces and bending
moments in a cross section of a ship.
Frank published in 1967 his pulsating source theory to calculate the hydrodynamic
coefficients of a cross section of a ship in deep water directly, without using conformal
mapping. The potential coefficients of a fully submerged cross section (bulbous bow) and
sections with a very low area coefficient (often present in the aft body) could be calculated
now too.
Using Lewis conformal mapping, Keil published in 1974 his theory for obtaining the potential
coefficients in very shallow water.
Useful theories to calculate the added resistance of a ship due to waves were given by Boese
(integrated pressure method) in 1970 and Gerritsma and Beukelman (radiated energy method)
in 1972.
So far, all hydrodynamic coefficients had been determined with the potential theory. However,
roll requires a viscous correction on that. Ikeda, Himeno and Tanaka published in 1978 a very
useful semi-empirical method for determining the viscous roll damping components.
The introduction of personal computers in the early eighties increased the accessibility for
carrying out ship motion calculations considerably; even non-specialists could become users
too. From then on the computer capacity and the computing speed increased very fast, so that
three-dimensional theories could be developed much easier and cheaper now.
Because of the complex problem of forward speed in 3-D theories however, the 2-D approach
(strip theory) is still very favourable for calculating the behaviour of a ship at forward speed.
The many advantages and just a few disadvantages, when comparing 2-D with 3-D, had been
discussed very clearly by Faltinsen and Svensen in 1990.
Among others as a consequence of the work of the researchers mentioned above, a DOS
personal computer strip theory program - called SEAWAY - had been completed by the Delft
University of Technology at the end of the eighties. Recently, a Windows version has been
completed too, see web site www.shipmotions.nl or www.amarcon.com.
Based on the linear strip theory, this program calculates for 6 degrees of freedom in the
frequency domain the hydromechanic loads, wave loads, absolute and relative motions, added
resistance and internal loads of displacement ships, barges and yachts in regular and irregular
waves. When ignoring interaction effects between the two individual hulls, the behaviour of
catamarans and semi-submersibles can be calculated too. The program is suitable for deep
water as well as for very shallow water. Viscous roll damping, bilge keels, free-surface antiroll tanks, external moments and (linear) mooring springs can be added.
The computer code has been verified and validated extensively by the authors, many students
and a large number of industrial users.
Error messages, advises and all type of comments on this technical report are very welcome
by e-mail to J.M.J.Journee@wbmt.tudelft.nl or Leon.Adegeest@amarcon.com.
Table of Contents:
1
Introduction........................................................................................................................11
1.1 About the Authors..........................................................................................................11
1.2 About this Manual........................................................................................................ 12
2
Strip Theory Methods ....................................................................................................... 15
2.1 Definitions .................................................................................................................... 17
2.2 Incident Wave Potential................................................................................................ 20
2.2.1 Continuity Condition............................................................................................ 21
2.2.2 Laplace Equation.................................................................................................. 21
2.2.3 Seabed Boundary Condition................................................................................. 22
2.2.4 Free Surface Dynamic Boundary Condition......................................................... 22
2.2.5 Free Surface Kinematic Boundary Condition....................................................... 24
2.2.6 Dispersion Relationship ........................................................................................ 25
2.2.7 Relationships in Regular Waves........................................................................... 26
2.3 Floating Rigid Body in Waves...................................................................................... 28
2.3.1 Fluid Requirements............................................................................................... 28
2.3.2 Forces and Moments............................................................................................. 30
2.3.3 Hydrodynamic Loads............................................................................................ 31
2.3.4 Wave and Diffraction Loads ................................................................................. 36
2.3.5 Hydrostatic Loads................................................................................................. 38
2.4 Equations of Motion..................................................................................................... 39
2.5 Strip Theory Approaches.............................................................................................. 43
2.5.1 Zero Forward Ship Speed ..................................................................................... 43
2.5.2 Forward Ship Speed.............................................................................................. 44
2.5.3 End-Terms............................................................................................................. 46
2.6 Hydrodynamic Coefficients.......................................................................................... 48
3
2-D Potential Coefficients ................................................................................................ 51
3.1 Conformal Mapping Methods....................................................................................... 53
3.1.1 Lewis Conformal Mapping................................................................................... 54
3.1.2 Extended Lewis Conformal Mapping................................................................... 58
3.1.3 Close-Fit Conformal Mapping.............................................................................. 59
3.1.4 Mapping Comparisons .......................................................................................... 63
3.2 Potential Theory of Tasai.............................................................................................. 65
3.2.1 Heave Motions ...................................................................................................... 66
3.2.2 Sway Motions ....................................................................................................... 76
3.2.3 Roll Motions ......................................................................................................... 88
3.2.4 Low and High Frequencies................................................................................. 100
3.3 Potentia l Theory of Keil.............................................................................................. 103
3.3.1 Notations of Keil................................................................................................. 103
3.3.2 Basic Assumptions.............................................................................................. 104
3.3.3 Vertical Motions.................................................................................................. 106
3.3.4 Horizontal Motions ............................................................................................. 133
3.3.5 Appendices ......................................................................................................... 144
3.4 Potential Theory of Frank ........................................................................................... 153
3.4.1 Notations of Frank .............................................................................................. 153
3.4.2 Formulation of the Problem................................................................................ 155
3.4.3 Solution of the Problem...................................................................................... 157
3.4.4 Low and High Frequencies................................................................................. 161
3.4.5 Irregular Frequencies .......................................................................................... 162
7
10
Introduction
SEAWAY is a frequency-domain ship motions PC program, based on the linear strip theory
for calculating the hydromechanic loads, wave-induced loads, motions, added resistance and
internal loads for six degrees of freedom of displacement ships and yachts, barges, semisubmersibles or catamarans, sailing in regular and irregular waves. The program is suitable for
deep water as well as for shallow water. Viscous roll damping, bilge keels, anti-roll tanks, free
surface effects and (linear springs) can be added.
This computer code has been developed in the late eighties and early nineties under DOS by
the first author. His last SEAWAY for DOS version was released in 2002.
In 2003, the second author took over the software implementation and distribution part of the
job and developed the new SEAWAY for Windows release.
Information can be found at web site www.shipmotions.nl or www.amarcon.com.
13
Chapter 11 describes the determination of the frequency characteristics of the lateral and
vertical shear forces and bending moments and the torsion moments in a way as had been
presented by Fukuda [1962] for the vertical mode. Still water phenomena are described too.
Chapter 12 describes the statistics in irregular waves, by using the superposition principle.
Three examples of normalized wave spectra are given: the somewhat wide wave spectrum 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 resistance, 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 probability
of shipping green water, are defined according to Tasaki [1963]. A theoretically determined
dynamic swell- up had been given too.
Bow slamming phenomena are defined by both the relative bow velocity criterion of Ochi
[1964] and by the peak bottom-impact-pressure criteria of Conolly [1974].
Chapter 13 describes the additions to all algorithms in case of twin- hull ships, such as semisubmersibles and catamarans. However, for interaction effects between the two individual
hulls will not be accounted here.
Chapter 14 shows some typical numerical recipes, as has been used in program SEAWAY.
Finally, Chapter 15 gives a survey of all literature used during the development of this
computer code.
Error messages, advises and all type of comments on this technical report are very welcome
by e-mail: J.M.J.Journee@wbmt.tudelft.nl.
14
The ship is considered to be a rigid body, floating in the surface of an ideal fluid, which is
homogeneous, incompressible, free of surface tension, irrotational and without viscosity. It is
assumed that the problem of the motions of this floating body in waves is linear or can be
linearised. Consequently, only the external loads on the underwater part of the ship are
considered here and the effect of the above water part will be fully neglected.
Faltinsen and Svensen [1990] have discussed the incorporation of seakeeping theories in ship
design clearly. 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 infinitely 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 effective 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 flow theory. This holds that viscous effects are
neglected, which can deliver serious problems when predicting roll motions at resonance
frequencies. In practice, for viscous roll damping effects can be accounted fairly by empirical
methods.
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
however, these ''near zero frequency of encounter problems'' can be solved by forcing the
wave loads going to zero, artificially.
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 become very
important.
The strip theory accounts for the interaction with the forward speed in a very simple way. The
effect of the steady wave system around the ship is neglected and the free surface conditions
are simplified, 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 effect 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 effects of the
hull below the still water level are accounted for. So, when parts of the ship go out of or in to
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.
15
Nevertheless these limitations, seakeeping prediction methods based upon the strip theory
provide a sufficiently good basis for optimisation 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.
16
2.1 Definitions
Figure 2.11 shows a harmonic wave as seen from two different perspectives. Figure 2.11-a
shows what one would observe in a snapshot photo made looking at the side of a (transparent)
wave flume; the wave profile is shown as a function of distance x along the flume at a fixed
instant in time. Figure 2.11-b shows a time record of the water level observed at one location
along the flume; it looks similar in many ways to the other figure, but time t has replaced x
on the horizontal axis.
or
T = 2
2
=
T
Equation 2.11
in which k is the wave number (rad/m) and is the circular wave frequency (rad/s).
17
Obviously, the wave form moves one wave length during one period, so that its speed or
phase velocity, c , is given by:
c= =
T
k
Equation 2.12
Suppose now a sailing ship in waves, with co-ordinate systems as given in Figure 2.12.
18
The right-handed co-ordinate system O( x, y , z ) is moving with the ship's speed V , which
yields:
x0 = V t cos + x cos + y sin
Equation 2.14
From the relation between the frequency of encounter e and the wave frequency :
e = k V cos
Equation 2.15
follows:
= a cos (e t k x cos k y sin )
Equation 2.16
The resulting six ship motions in the O( x, y , z ) system are defined 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 :
sway :
heave :
roll :
pitch :
yaw :
x = xa cos (e t + x )
y = ya cos (e t + y )
z = z a cos(e t + z )
= a cos (e t + )
= a cos(e t + )
= a cos(e t + )
Equation 2.17
The phase shifts of these motions are related to the harmonic wave elevation at the origin of
the O( x, y , z ) system, i.e. the average position of the ship's centre of gravity:
wave : = a cos(e t )
Equation 2.18
The harmonic velocities and accelerations in the O( x, y , z ) system are found now by taking
the derivatives of the displacements, for instance for surge:
surge displaceme nt : x = xa cos (e t + x )
surge velocity :
surge accelerati on :
x& = e xa sin (e t + x )
&x& = e 2 xa cos (e t + x )
Equation 2.19
19
sinh [s ] tanh [s ] s
cosh[s ] 1
Equation 2.22
w
=0
y
and
v 2 w
=
=0
y
y 2
Equation 2.26
Taking this into account, a substitution of Equation 2.21 in Equation 2.25 yields a
homogeneous solution of this equation:
d 2 P( z )
k 2 P( z ) = 0
2
dz
Equation 2.27
with as a homogeneous solution for P( z ) :
P( z ) = C1 e + k z + C2 e k z
Equation 2.28
Using this result from the continuity condition and the Laplace equation, the wave potential
can be written now with two unknown coefficients as:
w (x , z , t ) = (C1 e + k z + C2 e k z ) sin (k x t )
Equation 2.29
in which:
w ( x, z , t )
Equation 2.212
and the wave potential Equation 2.21 becomes:
w ( x, z, t ) = C cosh[k (h + z )] sin (k x t )
Equation 2.213
in which C is an (as yet) unknown constant.
22
w 1
p
+ (u 2 + v 2 + w 2 ) + + g z = 0
t
2
Equation 2.214
In two dimensions, v = 0 , and since the waves have a small steepness (u and w are small),
this equation becomes in a linearised format:
w p
+ + gz = 0
t
Equation 2.215
At the free surface this condition becomes:
w p
+ + g = 0
for: z =
t
Equation 2.216
The constant value p 0 can be included in w t ; this will not influence the velocities
being obtained from the potential w .
With this the equation becomes:
w
+ g = 0
for: z =
t
Equation 2.217
The potential at the free surface can be expanded in a Taylor series, keeping in mind that the
vertical displacement of the wave surface is relatively small:
{ w ( x, z, t )}z = = { w (x , z , t )}z = 0 + w (x , z , t )
z
or:
z = 0
+ .........
w ( x, z, t )
( x, z, t )
2
= w
+O
t
t
z =
z =0
( )
Equation 2.218
which yields for the linearised form of the free surface dynamic boundary condition in
Equation 2.217:
w
+ g = 0
for: z = 0
t
Equation 2.219
With this, the wave surface profile becomes:
1 w
for: z = 0
=
g t
Equation 2.220
A substitution of Equation 2.213 in Equation 2.220 yields the wave surface profile:
C
=
cosh [k h] cos (k x t )
g
or:
23
with: a =
C
cosh[k h ]
g
Equation 2.221
With this, depending on the water depth h , the wave potential in Equation 2.213 will
become:
g cosh [k (h + z )]
w = a
sin (k x t )
cosh [k h]
Equation 2.222
or when is the first of the sine function arguments, as generally will be used in ship motion
equations:
a g cosh [k (h + z )]
w =
sin ( t k x )
cosh [k h]
Equation 2.223
In deep water, the expression for the wave potential reduces to:
a g k z
w =
e sin ( t k x )
(deep water)
Equation 2.224
dt t x dt
for the wave surface: z =
=
+u
t
x
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 Figure 2.22.
24
+g
=0
for: z = 0
2
t
t
or after re-arranging terms:
1 2 w
+
=0
for: z = 0
t g t 2
Equation 2.227
Together with Equation 2.225 this delivers the free surface kinematic boundary condition or
the so-called Cauchy-Poisson condition:
dz 1 2 w
+
=0
for: z = 0
dt g t 2
Equation 2.228
potential (Equation 2.222) in Equation 2.228 gives the dispersion relation for any arbitrary
water depth h :
2 = k g tanh [k h ]
Equation 2.229
In many situations, or T will be known; one must determine k or . This equation will
generally has to be solved iteratively, since k appears in a nonlinear way in Equation 2.229.
In deep water ( tanh [k h ] = 1 ), Equation 2.229 degenerates to a quite simple form which can
be used without difficulty:
2 = k g
(deep water)
Equation 2.230
When calculating the hydromechanical forces and the wave exciting forces on a ship, it is
assumed that x x b , y y b and z z b . In case of forward ship speed, the wave frequency
has to be replaced by the frequency of encounter of the waves e . This leads to the
following expressions for the wave surface in the G( xb , yb , z b ) system:
= a cos(e t k xb cos k yb sin )
Equation 2.231
and the expression for the velocity potential of the regular waves, w , becomes:
a g cosh [k (h + zb )]
w =
cosh [k h]
Equation 2.232
26
27
radiation potential for the oscillatory motion of the body in still water
incident undisturbed wave potential
diffraction potential of the waves about the restrained body
vx =
vy =
vz =
x
y
z
Equation 2.32
The velocity potentials, = r + w + d , have to fulfil a number of requirements and
boundary conditions in the fluid. Of these, the first three are identical to those in the incident
undisturbed waves. Additional boundary conditions are associated with the oscillating floating
body.
1. Continuity Condition or Laplace Equation
As the fluid is homogeneous and incompressible, the continuity condition:
v x v y v z
+
+
=0
x
y
z
Equation 2.33
results into the equation of Laplace:
2 2 2
2 =
+ 2 + 2 =0
x 2
y
z
Equation 2.34
2. Seabed Boundary Condition
The boundary condition on the seabed (no-leak condition), following from the definition of
the velocity potential, is given by:
28
=0
z
for: z = h
Equation 2.35
p
p =
g z
or
+ g z =
t
t
Equation 2.36
At the free surface of the fluid, so for z = ( x, y , z, t ) , the pressure p is constant.
Because of the linearisation, the vertical velocity of a water particle in the free surface
becomes:
dz
=
dt
z
t
Equation 2.37
Combining these two conditions provides the boundary condition at the free surface:
2
+ g
=0
for: z = 0
2
t
z
Equation 2.38
4. Kinematic Boundary Condition on the Oscillating Body Surface
It is obvious that the boundary condition at the surface of the rigid body plays a very
important role. The velocity of a water particle at a point at the surface of the body is equal to
the velocity of this (watertight) body point itself. The outward normal velocity, v n , at a point
P( x, y, z ) at the surface of the body (positive in the direction of the fluid) is given by:
= v n ( x, y , z, t )
n
Equation 2.39
Because the solution is linearised, this can be written as:
6
= v n ( x, y , z, t ) = v j f j
n
j =1
Equation 2.310
in terms of oscillatory velocities, v j , and generalised direction-cosines, f j , on the surface of
the body, S , given by:
29
f 1 = cos (n, x )
f 2 = cos (n, y )
f 3 = cos (n, z )
f 4 = y cos (n, z ) z cos (n, y ) = y f 3 z f 2
f 5 = z cos(n, x ) x cos(n, z ) = z f 1 x f 3
f 6 = x cos (n, y ) y cos (n, x ) = x f 2 y f1
Equation 2.311
The direction cosines are called generalised, because f 1 , f 2 and f 3 have been normalised
(the sum of their squares is equal to 1) and used to obtain f 4 , f 5 and f 6 .
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 , x 2 ,... x 6 , as used here in
the summary.
5. Radiation Condition
The radiation condition states that when the distance R of a water particle to the oscillating
body tends to infinity, the potential value tends to zero:
lim = 0
R
Equation 2.312
6. Symmetric or Anti-symmetric Condition
Since ships and many floating bodies are symmetric with respect to its middle line plane, one
can make use of this to simplify the potential equations:
for sway : ( 2 ) ( x , y ) = ( 2 ) (+ x, y )
for heave :
for roll :
(3 ) ( x, y ) = + (3 ) (+ x, y )
( 4 ) ( x , y ) = ( 4 ) (+ x, y )
Equation 2.313
( )
( )
30
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 ) co-ordinate system.
The pressure p - via the linearised Bernoulli equation - is determined from the velocity
potentials by:
p =
g z
t
r w d
=
+
+
g z
t
t
t
Equation 2.315
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:
r w d
F =
+
+
+ g z n dS
t
t
t
( )
r w d
M =
+
+
+ g z r n dS
t
t
t
S
Equation 2.316
or:
F = F r + F w + Fd + F s
and
M = Mr + Mw +Md + Ms
Equation 2.317
r (x, y, z , t ) = j ( x, y, z , t )
j =1
6
= j ( x, y , z ) v j (t )
j =1
Equation 2.318
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, j ( x, y , z ) in direction j , multiplied
by an oscillatory velocity, v j (t ) in direction j .
This allows the normal velocity on the surface of the body to be written as:
31
r
6
=
j
n
n j =1
6
j
=
vj
j =1 n
Equation 2.319
and the generalised direction cosines are given by:
j
fj =
n
Equation 2.320
With this the radiation terms in the hydrodynamic force and moment becomes:
F = r n dS
t
S
6
= j v j n dS
t j =1
S
( )
M = r r n dS
t
S
( )
= j v j r n dS
t j =1
S
Equation 2.321
The components of these radiation forces and moments are defined by:
F r = ( X r1 , X r 2 , X r3 )
and
M r = ( X r 4 , X r 5 , X r6 )
with:
6
X rk = j v j f k dS
t j =1
S
for: k = 1,...6
6
k
= j v j
dS
t
dn
j
=
1
S
Equation 2.322
Since j and k are not time-dependent in this expression, it reduces to:
6
X rk = X rkj
for: k = 1,...6
j =1
with:
X rkj =
dv j
j k dS
dt
n
S
Equation 2.323
32
This radiation force or moment X rkj 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 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).
Equation 2.323 expresses the force and moment components, X rkj , in terms of still unknown
potentials, j . But not everything is solved yet, a solution for this will be found later in this
Chapter.
2.3.3.1 Oscillatory Motion
Now an oscillatory motion is defined; suppose a motion (in a complex notation) given by:
s j = s aj e i t
Equation 2.324
Then the velocity and acceleration of this oscillation are:
s& j = v j = i saj e i t
&s&j =
dv j
= 2 s aj e i t
dt
Equation 2.325
The hydrodynamic forces and moments can be split into a load in-phase with the acceleration
and a load in-phase with the velocity:
X rkj = M kj &s&j N kj s& j
= s aj 2 M kj + i saj Nkj e i t
= saj 2 j k dS e i t
n
S
Equation 2.326
So in case of an oscillation of the body in the direction j with a velocity potential j , the
hydrodynamic mass and damping (coupling) coefficients are defined by:
M kj = Re j k dS
and
N kj = Im j k dS
n
n
S
Equation 2.327
In case of an oscillation of the body in the direction k with a velocity potential k , the
hydrodynamic mass and damping (coupling) coefficients are defined by:
M jk = Re k j dS
and
N jk = Im k j dS
n
n
S
Equation 2.328
33
2
2
*
*
* j k k j dV = * j nk k n dS
V
S
Equation 2.329
As said before, this theorem is generally valid for all kinds of potentials; it is not necessary
that they fulfil 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 seabed at z = h , the water surface at z = and the wetted surface of the
floating body, S ; see Figure 2.31.
dS
=
S * j n
S * k n dS
Equation 2.330
The boundary condition at the free surface becomes for = e i t :
2 + g
=0
for: z = 0
z
Equation 2.331
or with the dispersion relation, 2 = k g tanh [k h ] :
k tanh [k h ] =
for: z = 0
z
Equation 2.332
34
This implies that at the free surface of the fluid one can write:
1
k tanh [k h ] k = k = k k =
k
z
n
k tanh [k h] n
at the free surface
j j
j
1
k tanh [k h ] j =
=
j =
z
n
k tanh [k h ] n
Equation 2.333
When taking also the boundary condition at the seabed and the radiation condition on the wall
of the cylinder in Figure 2.31:
= 0 for: z = h
and
lim = 0
R
n
Equation 2.334
into account, the integral equation over the surface S * reduces to:
j
k
dS
=
dS
j
k
S
S
n
n
Equation 2.335
in which S is the wetted surface of the oscillating body only.
Notice that j and k still have to be evaluated.
2.3.3.3 Potential Coefficients
The previous subsection provides - for the zero forward ship speed case - symmetry in the
coefficients matrices with respect to their diagonals so that:
M jk = M kj
and
N jk = N kj
Equation 2.336
Because of the symmetry of a ship, some coefficients are zero and
hydrodynamic coefficients for a ship become:
0
M 13
0
M 15
M 11
M 22
0
M 24
0
0
M
0
M 33
0
M 35
Hydrodynam ic mass matrix : 31
M 42
0
M 44
0
0
M
0
M 53
0
M 55
51
0
M 62
0
M 64
0
N11
0
N
Hydrodynam ic damping matrix : 31
0
N
51
0
35
0
N 22
0
N 42
0
N 62
N13
0
N 33
0
N 53
0
0
N 24
0
N 44
0
N 64
N15
0
N 35
0
N 55
0
M 26
0
M 46
0
M 66
0
N 26
0
N 46
0
N 66
For clarity, the symmetry of terms about the diagonal in these matrices (for example that
M 13 = M 31 for zero forward speed) has not been included here. The terms on the diagonals
( M nn ) are the primary coefficients relating properties such as hydrodynamic mass in one
direction to the inertia forces in that same direction. Off-diagonal terms (such as M 13 )
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 effect on the velocity potentials itself, but is not discussed in this
Section. This effect is quite completely explained by Timman and Newman [1962].
d
F w + F d = w +
n dS
t
t
S
( )
d
M w + M d = w +
r n dS
t
t
S
Equation 2.338
The principle of linear superposition allows the determination of these forces on a restrained
body with zero forward speed: n = 0 . This simplifies the boundary condition on the
surface of the body to:
w d
=
+
=0
n
n
n
Equation 2.339
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, w ( x, y, z ) and d (x , y , z ) , multiplied by a
normalised oscillatory velocity, v (t ) = 1 e i t :
w (x , y , z , t ) = w ( x, y, z ) e i t
d (x , y , z , t ) = d ( x, y , z ) e i t
Equation 2.340
This results into:
w
= d
n
n
Equation 2.341
With this and the expressions for the generalised direction-cosines it is found for the wave
forces and moments on the restrained body in waves:
36
= i e i t (w + d )
S
k
dS
n
for: k = 1,...6
Equation 2.342
in which k is the radiation potential.
The potential of the incident waves, w , is known, but the diffraction potential, d , has to be
determined. Green's second theorem provides a relation between this diffraction potential, d ,
and a radiation potential, k :
S d nk dS =S k nd dS
Equation 2.343
and with w n = d n from Equation 2.341 one finds:
S d nk dS = S k nw dS
Equation 2.344
This elimination of the diffraction potential results into the so-called Haskind relations:
X wk = i e i t w k + k w dS
for: k = 1,...6
n
n
S
Equation 2.345
This limits the problem of the diffraction potential because the expression for X wk depends
only on the undisturbed wave potential w and the radiation potential k .
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 floating body with a zero time-averaged speed in all directions.
Newman [1962] however, has generalised the Haskind relations for a body with a constant
forward speed. He derived equations, which differ only slightly from those found by Haskind.
According to Newman's approach the wave potential has to be defined 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 a previous section - now
becomes:
a g k z
w =
e sin ( t k x cos k y sin )
i a g k z i k ( x cos + y sin ) i t
=
e e
e
Equation 2.346
so that the isolated space dependent term is given by:
37
i a g k z i k ( x cos + y sin )
e e
Equation 2.347
= w k + i cos +
sin
n
n
n
n
= w k { f 3 + i ( f 1 cos + f 2 sin )}
Equation 2.348
With this, the wave loads are given by:
X wk = i e i t w f k dS
S
+ i e
i t
k w k { f 3 + i ( f1 cos + f 2 sin )} dS
for: k = 1,...6
Equation 2.349
The first term in this expression for the wave forces and moments is the so-called FroudeKrilov force or moment, which is the wave load caused by the undisturbed incident wave. The
second term is related to the disturbance caused by the presence of the (restrained) body.
( )
Equation 2.350
or more generally:
X sk = g z f k dS
for: k = 1,...6
Equation 2.351
in which the X sk are the components of these hydrostatic forces and moments.
38
0
0
0
0
0
0
0
0
0
0
0
0
0
I zx 0
I zz
Equation 2.41
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 ( m = ) is valid for a free floating structure:
I xx = k xx 2
I yy = k yy
2
I zz = k zz 2
Equation 2.42
When the actual distribution of the solid mass of a ship is unknown, the radii of inertia can be
approximated by:
k xx 0.30 B to 0.40 B
2 KG
k xx 0.289 B 1.0 +
Equation 2.44
in which KG is the height of the centre of gravity, G , above the keel.
For many ships without cargo on board (ballast condition), the mass is concentrated at the
ends (engine room aft and ballast water forward to avoid a large trim), while for ships with
cargo on board (full load condition) the - more or less amidships laden - cargo plays an
39
important role. Thus for normal ships, the radii of inertia, k yy and k zz , are usually smaller in
the full load condition than in the ballast condition.
Notice here that the longitudinal radius of gyration of a long homogeneous rectangular beam
with a length L is equal to about 1 12 L or 0.289 L .
The equations of motions of a rigid body in a space fixed co-ordinate system follow from
Newton's second law. The vector equations for the translations of and the rotations about the
centre of gravity are given respectively by:
d
d
F=
m U
and
M=
H
dt
dt
( )
Equation 2.45
in which:
F
m
U
M
Two important assumptions are made for the loads in the right-hand side of these equations:
a) The so-called hydromechanic forces and moments are induced by the harmonic
oscillations of the rigid body, moving in the undisturbed surface of the fluid.
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 up for obtaining 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 equations of motion for the ship:
d
Surge :
( x& )
= &x&
= X h1 + X w1
dt
d
Sway :
( y& )
= y&&
= X h2 + X w 2
dt
d
( z& )
Heave :
= &z&
= X h3 + X w 3
dt
d
&& = X h 4 + X w 4
Roll :
I xx & I xz & = I xx && I xz
dt
d
Pitch
I yy &
= I xx &&
= X h 5 + X w5
dt
d
&& = X h 6 + X w 6
&& I zx
Yaw
I zz & I zx & = I zz
dt
(
(
(
Equation 2.46
in which:
density of water
40
I ij
X h1 , X h 2 , X h 3
X h4 , X h5 , X h6
X w1 , X w2 , X w3
X w4 , X w5 , X w6
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 antisymmetric motions. Symmetric and anti-symmetric motions of a free-floating structure are not
coupled; they don't have any effect on each other. For instance, a vertical force acting at the
centre 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:
Surge : &x&
X h1 = X w1
Heave : &z&
X h 3 = X w3 symmetric motions
Pitch : I xx &&
X h 5 = X w5
Sway : &y&
Xh 2 = Xw2
( + a11 ) &x&
= X w1
(surge)
(heave)
(pitch)
41
The coupled sway, roll and yaw equations of anti-symmetric motion are:
( + a22 ) &y&
(sway)
(roll)
(yaw)
Equation 2.49
In many applications, I xz = I zx is not known or small; hence their terms are often omitted. In
program SEAWAY they have been introduced in the equations of motion if they can be
calculated from an input of the mass distribution along the ships length, only.
After the determination of the in and out of phase terms of the hydromechanic and the wave
loads, these equations can be solved with a numerical method.
42
43
X h 1 = X h 1' dxb
X w1 = X w1 ' dxb
X h 2 = X h 2 dxb
X w 2 = X w 2 ' dxb
Heave :
X h 3 = X h 3 dxb
X w3 = X w3 ' dxb
Roll :
X h 4 = X h 4 dxb
X w 4 = X w 4 dxb
Pitch :
X h 5 = X h 3 xb dxb
X w5 = X w 3 xb dxb
X h 6 = + X h 2 xb dxb
X w 6 = + X w 2 xb dxb
Surge :
'
Sway :
'
'
'
Yaw :
'
'
'
'
Equation 2.51
in which:
'
X hj
X wj
sectional exciting wave force or moment in direction j per unit ship length
The appearance of two-dimensional surge forces seems strange here. It is strange! A more or
less empirical method has been used in SEAWAY for the surge motion, by defining an
equivalent longitudinal cross section that is swaying. Then, the 2-D hydrodynamic sway
coefficients of this equivalent cross section are translated to 2-D hydrodynamic surge
coefficients by an empirical method based on theoretical results from three-dimensional
calculations and these coefficients 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 hydromechanic coefficients for surge of (slender) ships are of
minor importance, because these coefficients are relatively small.
Notice 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 defined in the moving axis system with the origin at the timeaveraged position of the centre of gravity, G . All two-dimensional potential coefficients have
been defined here in an axis system with the origin, O , in the water plane; the hydromechanic
and exciting wave moments have to be corrected for the distance OG .
{ }
{ }
Equation 2.52
in which:
44
= V
Dt t
x
Equation 2.53
is a mathematical operator which transforms the potentials ( x0 , y 0 , z 0 , t ) , defined in the
earth bounded (fixed) co-ordinate system, to the potentials ( x, y, z , t ) , defined in the ship's
steadily translating co-ordinate system with speed V .
In waves the motions of the water particles are depending on its local vertical distance to the
mean or still water surface. At each cross section of the ship an average (or equivalent)
constant value has to be found.
Relative to a restrained ship, moving forward with speed V in waves, the equivalent j
*
*
constant components of water particle displacements ( ), velocities ( & ) and
wj
wj
Dt
Dt wj
*
{ }
{ }
Equation 2.54
The effect of the operator in Equation 2.53 can be understood easily when one realises that in
that earth-bound co-ordinate system the sailing ship penetrates through a ''virtual vertical
disk''. For instance, when a ship sails with speed V and constant trim angle through still
water, the relative vertical velocity of a water particle with respect to the bottom of the sailing
ship becomes V sin V .
Two different types of strip theory methods (as has been used in SEAWAY) are discussed
here:
1. Ordinary Strip Theory Method
According to this classic method, the uncoupled two-dimensional
hydromechanic loads and wave loads in an arbitrary direction j are defined by:
D
X hj * =
M jj ' & hj * + N jj ' & hj * + X rsj '
Dt
D
*
'
*
'
*
'
X wj =
M jj &wj + N jj & wj + X fkj
Dt
potential
Equation 2.55
This is the first 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-Kroukovski and Jacobs [1957].
2. Modified Strip Theory Method
According to this modified method, these loads become:
45
X hj * =
D
i
'
N jj ' & hj * + X rsj '
M jj
Dt
e
X wj * =
D
i
'
N jj ' & wj * + X fkj '
M jj
Dt
e
Equation 2.56
This formulation is a more fundamental approach of the forward speed problem, as
published in detail by Tasai [1969] and many others.
'
'
In the equations above, M jj and N jj are the 2-D potential mass and damping coefficients.
'
X rsj is the two-dimensional quasi-static restoring spring term, as generally present for heave,
'
roll and pitch only. X fkj is the two-dimensional Froude-Krilov force or moment which is
calculated by 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 here to calculate the diffraction parts of the total wave
forces and moments.
From a theoretical point of view, one should prefer the use of the modified strip theory
method. However, it appeared from the authors experiences that for ships with moderate
forward speed ( Fn 0.30 ) the ordinary method could provide in some cases a better fit with
experimental data.
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 M jj and damping N jj will appear. These derivatives have to
be determined numerically over the whole ship length in such a manner that the following
relation is fulfilled:
xb ( L )+
x b (0 )
x b (L )
xb ( L )+
df ( xb )
df ( xb )
df ( xb )
df ( xb )
dx
=
dx
+
dx
+
dx b
b
b
b
dxb
dx b
dx b
dx b
xb (0 )
xb (0 )
xb ( 0 )
xb ( L )
= f (0 ) +
df (x b )
dxb f ( L )
dx
x b (0 )
b
xb ( L )
=0
Equation 2.57
with << L , while f (x b ) is equal to the local values of N jj ( xb ) or M jj ( xb ) ; see Figure 2.5
2.
'
46
'
( )
xb 0
df (x b )
dxb = 0
dxb
b
df (x b )
dx
=
dxb b b x (0)f ( x b ) dx b
xb (0 )
b
xb ( L ) +
(L)
xb ( L )
df (x b )
2
dx
=
f (x b ) x b dx b
b
b
dx
b
xb ( 0 )
xb ( 0 )
xb ( L ) +
Equation 2.58
Notice that these expressions are valid for the integration of the potential coefficients over the
full ship length only. They can not be used for calculating local hydromechanic loads. Also for
the wave loads, these expressions can not be used, because there these derivatives are
multiplied with equivalent x b -depending orbital motion amplitudes.
47
48
A more or less empirical procedure has followed by the author for the surge motion. An
equivalent longitudinal cross section has been defined. For each frequency, the twodimensional potential hydrodynamic sway coefficient of this equivalent cross section is
translated to two-dimensional potential hydrodynamic surge coefficients, by an empirical
method based on theoretical results of three-dimensional calculations.
The 3-D coefficients follow from an integration of these 2-D coefficients over the ship's
length. Viscous terms have been be added for surge and roll.
49
50
This Chapter described the various methods, used in the SEAWAY computer code, to obtain
the 2-D potential coefficients:
the theory of Tasai for deep water, based on Ursell's potential theory for circular cylinders
and Lewis and N-parameter conformal mapping
the theory of Keil for very shallow to deep water, based on a variation of Ursell's potential
theory for circular cylinders and Lewis conformal mapping
the theory of Frank for deep water, using pulsating sources on the cross sectional contour.
During the ship motions calculations different co-ordinate systems, as shown before, will be
used. The two-dimensional hydrodynamic potential coefficients have been defined here with
respect to the O( x, y , z ) co-ordinate system for the moving ship in still water.
However, in this section deviating axes systems are used for the determination of the twodimensional hydrodynamic potential coefficients for sway, heave and roll motions. This holds
for the sway and roll coupling coefficients a change of sign. The signs of the uncoupled sway,
heave and roll coefficients do not change.
For each cross section, the following two-dimensional hydrodynamic coefficients have to be
obtained:
M 22 ' and N 22'
2-D potential mass and damping coefficients of sway
M 33 and N33
'
'
2-D potential mass and damping coupling coefficients of roll into sway
2-D potential mass and damping coefficients of heave
M 44 and N
'
44
M42 and N
'
42
2-D potential mass and damping coupling coefficients of sway into roll
'
'
'
'
'
'
The 2-D potential pitch and yaw (moment) coefficients, M 55 , N 55 , M 66 and N 66 , follow
from the previous heave and sway coefficients and the lever of the loads, i.e., the distance of
the cross section to the centre of gravity G .
Finally, an approximation is given for the determination of the 2-D potential surge coefficients
M 11' and N11' .
51
52
z = M s a2 n 1 (2 n 1 )
}
Equation 3.11
with:
z = x + iy
= ie e i
Ms
a 1
a 2 n 1
N
y = + M s ( 1)n a2 n 1 e (2 n 1 )a cos((2n 1) )
n= 0
Equation 3.12
53
The contour of the - by conformal mapping approximated - ship's cross section follows from
putting = 0 in the previous relations in Equation 3.12:
N
y0 = +M s ( 1)n a2 n 1 cos((2n 1) )
n =0
Equation 3.13
The breadth on the waterline of the approximated ship's cross section is defined by:
N
b
b0 = 2 M s a
with: a = {a2 n 1 } and M s = 0
2 a
n =0
Equation 3.14
The draught is defined by:
d 0 = M s b
with: b = ( 1)n a 2 n 1
n =0
}
Equation 3.15
(
(e
Equation 3.17
By putting = 0 is the contour of this so-called Lewis form expressed as:
x0 = M s ((1 + a1 ) sin a3 sin 3 )
y0 = M s ((1 a1 ) cos + a3 cos 3 )
with scale factor:
Bs 2
Ds
Ms =
=
1 + a1 + a3 1 a1 + a3
Equation 3.18
54
Now the coefficients a1 and a 3 and the scale factor M s will be determined in such a manner
that the sectional breadth, draught and area of the approximated cross section and of the actual
cross section are identical.
The half breadth to draught ratio H 0 is given by:
B 2 1 + a1 + a 3
H0 = s =
Ds
1 a1 + a 3
Equation 3.19
An integration of the Lewis form delivers the sectional area coefficient s :
As
1 a1 3 a3
s =
=
Bs Ds 4 (1 + a3 ) 2 a1 2
2
Equation 3.110
in which As is the area of the cross section.
Putting a1 , derived from the expression for H 0 in Equation 3.19, into the expression for s
in Equation 3.110 yields a quadratic equation in a 3 :
c1 a3 + c 2 a3 + c3 = 0
in which:
2
4 s 4 s H 0 1
c1 = 3 +
+ 1
H 0 + 1
c 2 = 2 c1 6
c3 = c1 4
Equation 3.111
The (valid) solutions for a 3 and a1 become:
a3 =
a1 =
c1 + 3 + 9 2 c1
c1
H 0 1
(a + 1)
H 0 +1 3
Equation 3.112
Lewis forms with the other solution of a 3 in the quadratic equation, with a minus sign before
the square root expression:
c1 + 3 9 2 c1
a3 =
c1
55
are looped; they intersect themselves at a point within the fourth quadrant. Since ships are
''better behaved'', these solutions are not considered.
It is obvious that a transformation of a half-immersed circle with radius R will result in
M s = R , a1 = 0 and a 3 = 0 .
Some typical and realistic Lewis forms are presented in Figure 3.12.
for H 0 1.0 :
s <
2
32
H 0
Equation 3.113
for H 0 1.0 :
2
3 +
32
H0
32
4 H 0
Equation 3.114
H 0 1.0 and
3 + 0 < s <
3 +
32
4
32
4 H 0
Equation 3.115
These ranges of the half breadth to draught ratio H 0 and the area coefficient s for the
different typical Lewis forms are shown in Figure 3.13.
for H 0 1.0 :
(2 H 0 ) < s <
10 + H 0 +
32
32
H 0
for H 0 1.0 :
3
1
1
< s <
2
10 + H0 +
32
H0
32
H0
57
If a value of s is outside of this range it has to be set (from a practical point of view) to the
value of the nearest border of this range, to calculate the Lewis coefficients.
Numerical problems, for instance with bulbous or aft cross sections of a ship, are avoided
when the following requirements are fulfilled:
Bs
B
> Ds and Ds > s with for instance = 0.01 .
2
2
(
(e
cos a1 e cos + a3 e
cos 3 a5 e
cos 5
)
Equation 3.121
An integration of the approximated form results into the sectional area coefficient s :
As
1 a1 2 3 a3 2 5 a5 2
s =
=
Bs Ds 4 (1 + a3 )2 (a1 + a5 )2
Equation 3.124
A more complex expression has been obtained by Athanassoulis and Loukakis [1985] for the
relative distance of the centroid to the keel point:
KB
=
=1
Ds
{A
3
i =0
j= 0
k =0
ijk
a2 i 1 a2 j 1 a2 k 1 }
3
H0 s a2 i 13
i =0
in which:
Aijk =
1
1 2 k
1 2k
1 2 k
1 2 k
+
+
4 3 2 (i + j + k ) 1 2 (i j + k ) 1 2 (i + j k ) 1 2 ( i + j + k )
Equation 3.125
The following requirements should be fulfilled when also bulbous cross sections are allowed:
re-entrant forms are avoided when:
1 a1 3 a3 5 a5 > 0
1 + a1 3 a3 + 5 a5 > 0
Equation 3.126
Taking these restrictions into account, the equations above can be solved in an iterative
manner.
z = M s a2 n 1 ( 2 n 1 )
with: a1 = +1 .
n =0
Equation 3.128
Then the contour of the approximated cross section is given by:
N
Ds
{( 1)
N
n= 0
a2 n 1
}
Equation 3.129
The procedure starts with initial values for [M s a 2 n 1 ] . The initial values of M s , a1 and a 3
are obtained with the Lewis method as has been described before, while the initial values of
a 5 through a2 N 1 are set to zero. With these [M s a 2 n 1 ] values, a i -value is determined for
each offset in such a manner that the actual offset ( xi , yi ) lies on the normal of the
approximated contour of the cross section in ( x0 i , y 0 i ) .
Now i has to be determined. Therefore a function F (i ) , will be defined by the distance
from the offset ( xi , yi ) to the normal of the contour to the actual cross section through
(x0i , y 0i ), see Figure 3.14.
60
yi +1 + y i 1
(x i+1 xi 1 )2 + ( y i+1 y i1 )2
Equation 3.130
With this i -value, the numerical value of the square of the deviation of ( xi , yi ) from ( x0 i , y 0 i )
is calculated:
2
2
ei = ( xi x0 i ) + ( yi y0 i )
Equation 3.131
After doing this for all I + 1 offsets, the numerical value of the sum of the squares of
deviations is known:
I
E = {ei }
i= 0
Equation 3.132
The sum of the squares of these deviations can also be expressed as:
2
N
n
x
+
(
1
)
[
M
a
]
sin
(
(
2
n
1
)
i
s
2 n 1
i
I
n= 0
E =
2
N
i =0
n
+ y i ( 1) [M s a 2 n 1 ] cos((2n 1) i )
n =0
Equation 3.133
Then, new values of [M s a 2 n 1 ] have to be determined such that E is minimised. This means
that the derivative of this equation to each coefficient [M s a 2 n 1 ] is zero, so:
E
=0
for: j = 0,... N
{M s a2 j 1 }
Equation 3.134
This provides N + 1 equations:
N
sin
(
(
2
j
1
)
( 1) n [M s a 2n 1 ] sin ((2n 1) i )
i
I
n =0
N
i= 0
n
cos ((2 j 1) i ) ( 1) [M s a2 n 1 ] cos ((2n 1) i )
n= 0
61
for: j = 0,... N
n
(
1
)
[
M
a
]
cos((2 j 2n ) i ) =
s
2 n 1
n= 0
i= 0
{ x
i= 0
for: j = 0,... N
Equation 3.135
To obtain the exact actual breadth and draught, the last two equations ( j = N 1 and j = N )
in Equation 3.135 are replaced by the equations for the breadth at the water line and the
draught:
N
I
n
(
1
)
[
M
a
]
{cos ((2 j 2n ) i )} =
s
2 n 1
n= 0
i= 0
{ x
for : j = 0,...N 2
i= 0
{[M
n =0
{( 1) [M
N
a 2 n 1 ]} = Bs 2
a2 n 1 ] = Ds
n= 0
Equation 3.136
These N + 1 equations can be solved numerically, so that new values for [M s a 2 n 1 ] will be
obtained. These new values are used instead of the initial values to obtain new i -values of
the I + 1 offsets again, etc. This procedure will be repeated several times and stops when the
difference 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:
2
2
E = (I + 1) 0.00005 bmax + d max
Equation 3.137
in which:
bmax
d max
Because a 1 = +1 , the scale factor M s is equal to the final solution of the first coefficient
( n = 0 ). The N other coefficients a 2 n 1 can be found by dividing the final solutions of
[M s a 2n1 ] by this M s -value.
Reference is also given here to a report of de Jong [1973]. In that report several other, suitable
but more complex, methods are described to determine the scale factor M s and the conformal
mapping coefficients a 2 n 1 from the offsets of a cross section.
Attention has been paid in SEAWAY to divergence in the calculation routines and re-entrant
forms. In these cases the number N will be decreased until the divergence or re-entrance
vanish. In the worse case the ''minimum'' value of N will be attained without success. One
62
can then switch to Lewis coefficients with an area coefficient of the cross section, eventually
set to the nearest border of the valid Lewis form area.
Table 3.1-1:
For the least square method in the conformal mapping method, 33 new offsets at equidistant
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.
63
64
Figure 3.21: Tasais axes system for heave, sway and roll oscillations
The figure shows a cross section of an infinite long cylinder in the surface of a fluid. This
cylinder will carry out forced harmonic heave, sway and roll motions, respectively. Using the
approach of Tasai (and de Jong), the determination of the hydrodynamic loads will be showed
in the following Sections.
65
=0
for: = 0
Equation 3.25
B
+
=0
for: x s and y = 0
g
y
2
Equation 3.26
In consequence of the conformal mapping, the free surface condition in Equation 3.26 can be
written as:
N
b
(2n 1) a2 n 1 e ( 2 n 1 )
=0
for: 0 and =
a
2
n= 0
in which:
b 2
2 b0
=
M s
or
b =
(non-dimensional frequency squared)
a
g
2 g
Equation 3.27
From the definition of the velocity potential follows the boundary condition on the surface of
the cylinder for = 0 :
0 ( )
y
= y& 0
n
n
Equation 3.28
in which n is the outward normal of the cylinder surface.
Using the stream function , this boundary condition on the surface of the cylinder ( = 0 )
reduces to:
0 ( )
y
= y& 0
Equation 3.29
Integration results into the following requirement for the stream function on the surface of the
cylinder:
N
Equation 3.210
in which C (t ) is a function of time only.
When defining:
67
2 x0
b0
1 N
n
( 1) a 2 n 1 sin ((2n 1) )
a n =0
Equation 3.211
the stream function on the surface of the cylinder is given by:
b
0 ( ) = y& 0 h( ) + C(t )
2
Equation 3.212
Because of the symmetry of the fluid about the y -axis, it is clear that C (t ) = 0 , so that:
b
0 ( ) = y& 0 h ( )
2
Equation 3.213
For the standing wave system a velocity potential and a stream function satisfying the
equation of Laplace, the symmetrical motion of the fluid 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
de Jong [1973], fulfil these requirements:
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = e 2 m cos (2m )
b N
2n 1
n
( 1)
a 2 n 1 e ( 2 m + 2 n 1) cos((2m + 2 n 1) )
a n = 0
2m + 2 n 1
Equation 3.214
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = e 2 m sin (2m )
b N
2n 1
n
( 1)
a 2 n 1 e (2 m + 2 n 1) sin ((2m + 2 n 1) )
a n = 0
2 m + 2n 1
Equation 3.215
axis, representing such a train of waves at infinity. For this, a function describing a source at
the origin O is chosen.
Tasai [1959], Tasai [1960] and de Jong [1973] gave the velocity potential of the progressive
wave system as:
g a
B =
{Bc (x , y ) cos( t ) + Bs ( x, y ) sin ( t )}
in which:
Bc = + e y cos ( x )
sin (k y ) k cos( k y ) k x
e
dk
k 2 + 2
0
Bs = + e y sin ( x ) +
while:
2
=
g
Bs = e y cos ( x ) +
Equation 3.218
Changing the parameters provides:
g a
B =
{Bc (, ) cos( t ) + Bs (, ) sin ( t )}
Equation 3.219
When calculating the integrals in the expressions for Bs and Bc numerically, the
convergence is very slowly.
Power series expansions, as given by Porter [1960], can be used instead of these last integrals
over k . Summations in these expansions converge much faster than the numeric integration
procedure. This will be shown in the Section 3.2.2 for the sway case.
The total velocity potential and stream function to describe the waves generated by a heaving
cylinder are:
= A + B
= A + B
Equation 3.220
69
So the velocity potential and the conjugate stream function are expressed by:
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
{Q2 m A 2m (, )} sin ( t )
+ Bs (, ) +
m =1
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
)
+
{
Q
)
}
sin
(
t
)
Bs
2m
A2m
m
=
1
Equation 3.221
When putting = 0 , the stream function is equal to the expression in Equation 3.213, found
from the boundary condition on the surface of the cylinder:
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
g a
m =1
0 ( ) =
b
= y& 0 h( )
2
in which:
A 02 m ( ) = sin (2m )
b N
2n 1
n
( 1)
a
sin ((2 m + 2n 1) )
a n= 0
2m + 2n 1 2 n 1
Equation 3.222
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
b0
m =1
= y& 2 g h ( )
a
+ B 0 s ( ) + {Q2 m A 02 m ( )} sin ( t )
m =1
Equation 3.223
The right hand side of this equation can be written as:
b0
y
y&
h( ) = h( ) a b sin ( t + )
2 g a
a
in which:
70
ya
b sin
a
and
B0 =
B 0 c ( ) + {P2 m A 02 m ( )} = h( ) A0
m =1
B 0 s ( ) + {Q2 m A 02 m ( )} = h( ) B0
m =1
Equation 3.225
When putting = 2 , so at the intersection of the surface of the cylinder with the free
surface of the fluid where h ( ) = 1 , we obtain the coefficients A0 and B0 :
A0 = B 0 c ( 2) + {P2 m A 02 m ( 2 )}
m =1
B0 = B 0 s ( 2 ) + {Q2 m A 02 m ( 2)}
m =1
in which:
N
b
m
2n 1
A 02 m ( 2) =
( 1)
a 2 n 1
a
n =0 2m + 2n 1
Equation 3.226
A substitution of A0 and B0 into the set of two equations for each , results for each -value
less than 2 in a set of two equations with the yet unknown parameters P2 m and Q2 m , so:
B 0 c ( ) h( ) B 0 c ( 2 ) = { f 2 m ( ) P2 m }
m =1
B 0 s ( ) h( ) B 0 s ( 2 ) = { f 2 m ( ) Q2 m }
m =1
in which:
f 2 m ( ) = A 02 m ( ) + h( ) A 02 m ( 2 )
Equation 3.227
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 , for instance M = 10 .
Each -value less than 2 will provide an equation for the P2 m and Q2 m series. For a lot of
-values, the best fit values of P2 m and Q2 m are supposed to be those found by means of a
least square method. Notice that at least M values of , less than 2 , are required to solve
these equations.
Another favourable method is to multiply both sides of the equations with . Then the
summation over can be replaced by integration.
Herewith, two sets of M equations have been obtained, one set for P2 m and one set for Q2 m :
71
2
P2 m f 2 m ( ) f 2 n ( ) d = (B 0 c ( ) h( ) B 0 c ( 2 )) f 2 n ( ) d
0
m =1
0
2
M
2
Q2 m f 2 m ( ) f 2 n ( ) d = (B 0 s ( ) h ( ) B 0 s ( 2)) f 2 n ( ) d
0
m =1
0
for: n = 1,...M
M
Equation 3.228
Now the P2 m and Q2 m series can be solved by a numerical method and with these values, the
coefficients A0 and B0 are known too.
From the definition of these coefficients in Equation 3.224 follows the amplitude ratio of the
radiated waves and the forced heave oscillation:
a
b
=
2
2
ya
A +B
0
Equation 3.229
With the solved P2 m and Q2 m values, the velocity potential on the surface of the cylinder
( = 0 ) is known too:
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
g a
m =1
0 ( ) =
in which:
A 02 m ( ) = cos(2 m )
b N
2n 1
n
( 1)
a 2 n 1 cos((2m + 2 n 1) )
a n =0
2m + 2n 1
Equation 3.230
72
0 ( )
t
M
+ B 0 s ( ) + {Q2 m A 02 m ( )} cos ( t )
g a
m =1
B 0 c ( ) + {P2 m A 02 m ( )} sin ( t )
m =1
p ( ) =
Equation 3.231
It is obvious that this pressure is symmetric in .
3.2.1.2 Heave Coefficients
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:
+ 2
dx
'
Fy = p( ) 0 ds
ds
2
2
= 2
dx 0
p () d d
0
Equation 3.232
With this the two-dimensional hydrodynamic vertical force due to heave oscillations can be
written as follows:
g b0 a
'
Fy =
(M 0 cos ( t ) N 0 sin ( t ))
in which:
1
M0 =
n
B 0s () {( 1) (2n 1) a2n 1 cos((2n 1) )} d
n= 0
1 M
(
2n 1)2
m
( 1) Q2 m
a2 n 1
2
2
a m =1
n = 0 (2m ) (2n 1)
N
N m
Q
+
(
1
)
2
m
2
4 a
m =1
n= 0
73
g b0 a
(M 0 cos ( t + ) N 0 sin ( t + ))
a
sin =
A0
y a b
a
cos =
B
y a b 0
Fy =
'
Equation 3.234
the two-dimensional hydrodynamic vertical force can be resolved into components in phase
and out phase with the vertical displacement of the cylinder:
2
g b0 a
Fy ' =
2 b y a
Equation 3.235
This hydrodynamic vertical force can also be written as:
Fy ' = M 33 ' &y& N 33 ' y&
= M 33 ' 2 ya cos ( t + ) + N 33 ' y a sin ( t + )
Equation 3.236
in which:
'
M 33
N 33
'
When using also the amplitude ratio of the radiated waves and the forced heave oscillation,
found before in Equation 3.229, the two-dimensional hydrodynamic mass and damping
coefficients of heave are given by:
b0 2 M 0 B0 + N 0 A0
M 33 ' =
2
A0 2 + B0 2
N 33
'
b0 2 M 0 A0 N 0 B0
=
2
A0 2 + B0 2
Equation 3.237
The signs of these two coefficients are proper in both, the axes system of Tasai and the ship
motions O( x, y, z ) co-ordinate system.
The energy delivered by the exciting forces should be equal to the energy radiated by the
waves, so:
1
Tosc
Tosc
(N
0
'
33
y& ( y& dt ) =
g a c
2
2
Equation 3.238
in which Tosc is the period of oscillation.
74
With the relation for the wave speed c = g at deep water, follows the relation between the
two-dimensional heave damping coefficient and the amplitude ratio of the radiated waves and
the forced heave oscillation:
N33
'
g 2
=
3
a
ya
Equation 3.239
With this amplitude ratio the two-dimensional hydrodynamic damping coefficient of heave is
also given by:
2
2 b0
1
'
N 33 =
2
2
4
A0 + B0
Equation 3.240
'
When comparing this expression for N 33 with the expression found before, the following
energy balance relation is found:
2
M 0 A0 N 0 B0 =
2
Equation 3.241
75
76
+
=0
g
y
for: x
Bs
2
y=0
and
Equation 3.246
In consequence of the conformal mapping, this free surface condition can be written as:
N
b
(2n 1) a2 n 1 e ( 2 n 1 )
=0
for: 0 and =
a
2
n= 0
in which:
b 2
2 b0
=
M s
or
b =
(non-dimensional frequency squared)
a
g
2 g
Equation 3.247
From the definition of the velocity potential follows the boundary condition on the surface of
the cylinder S for = 0 :
0 ( )
x
= x& 0
n
n
Equation 3.248
in which n is the outward normal of the cylinder surface S .
Using the stream function , this boundary condition on the surface of the cylinder ( = 0 )
reduces to:
0 ( )
x
= x& 0
Equation 3.249
Integration results into the following requirement for the stream function on the surface of the
cylinder:
N
Equation 3.250
in which C (t ) is a function of time only.
When defining:
2 y0
g ( ) =
b0
1 N
n
=
( 1) a2 n 1 cos ((2 n 1) )
a n = 0
Equation 3.251
the stream function on the surface of the cylinder is given by:
77
b0
g ( ) + C (t )
2
Equation 3.252
For the standing wave system a velocity potential and a stream function satisfying to the
equation of Laplace, the non-symmetrical motion of the fluid and the free surface condition
has to be found.
The following set of velocity potentials, as they are given by Tasai [1961] and de Jong [1973],
fulfil these requirements:
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = + e ( 2 m +1) sin ((2m + 1) )
b N
2n 1
n
( 1)
a
e (2 m + 2 n ) sin ((2m + 2n ) )
a n = 0
2 m + 2n 2 n 1
Equation 3.253
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = e (2 m +1 ) cos((2m + 1) )
+
b N
2n 1
n
( 1)
a2 n 1 e (2 m + 2 n ) cos ((2m + 2 n) )
a n = 0
2m + 2n
Equation 3.254
78
Bc j = e y sin ( x )
cos(k y ) + k sin (k y ) k x
e
dk
2
2
k
+
Bs j = + e y cos( x )
+
while:
x
x + y2
j = +1 for: x > 0
j = 1 for: x < 0
=
2
g
sin ( x ) +
y
x + y2
sin (k y ) k cos(k y ) k x
e
dk
k 2 + 2
0
)
Equation 3.257
cos(k y ) + k sin (k y ) k x
e dk = {Q sin ( x ) ( S ) cos ( x )} e y
2
2
0
k +
sin (k y ) k cos (k y ) k x
e dk = {Q cos ( x ) + (S ) sin ( x )} e y
2
2
k
+
0
in which:
79
Q = + ln x 2 + y 2 + { pn cos(n )}
n =1
S = + { pn sin (n )}
n =1
x
= arctan
y
(
=
x2 + y2
pn
n n!
= 0.57722..... (Euler constant)
Equation 3.259
The summations in these expansions converge much faster than the numeric integration
procedure. A suitable maximum value for n should be chosen, n = 1,...N .
The total velocity potential and stream function to describe the waves generated by a swaying
cylinder are:
= A + B
= A + B
Equation 3.260
So the velocity potential and the conjugate stream function are expressed by:
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
{Q2 m A 2m (, )} sin ( t )
+ Bs (, ) +
m
=
1
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
Equation 3.261
When putting = 0 , the stream function is equal to the expression found before in Equation
3.252 from the boundary condition on the surface of the cylinder:
+ B 0 c ( ) + {P2 m A 02 m ( )} cos ( t )
g a
m =1
0 ( ) =
+ B 0 s ( ) + {Q2 m A 02 m ( )} sin ( t )
m =1
b
= x& 0 g ( ) + C (t )
2
80
in which:
A 02 m ( ) = cos((2m + 1) )
+
b N
2n 1
n
( 1)
a2 n 1 cos ((2 m + 2n ) )
a n = 0
2 m + 2n
Equation 3.262
+ B 0 c ( ) + {P2 m A 02 m ( )} cos ( t )
m =1
b0
*
= x& 2 g g ( ) + C (t )
a
+ B 0 s ( ) + {Q2 m A 02 m ( )} sin ( t )
m =1
Equation 3.263
When putting = 2 , so at the intersection of the surface of the cylinder with the free
surface of the fluid where g ( ) = 0 , we obtain the constant C * (t ) :
+ B 0 c ( 2 ) + {P2 m A 02 m ( 2 )} cos ( t )
m =1
C * (t ) =
+ B 0 s ( 2 ) + {Q2 m A 02 m ( 2 )} sin ( t )
m =1
in which:
N
m
2n 1
A 02 m ( 2 ) = b ( 1)
a2 n 1
a
n = 0 2m + 2 n
Equation 3.264
A substitution of C * (t ) in the equation for each -value, results for each -value less than
2 into the following equation:
m =1
+ ( ) ( 2) + {Q (
B 0s
2m
A 02 m ( ) A 02 m ( 2))} sin ( t )
B0s
m
=
1
b0
x&
g ( )
2 g a
Equation 3.265
The right hand side of this equation can be written as:
x
b0
x&
g ( ) = g ( ) a b sin ( t + )
2 g a
a
in which:
81
xa
b sin
a
Q0 =
and
This provides for each -value less than 2 a set of two equations with the unknown
parameters P2 m and Q2 m :
B 0 c ( ) B 0 c ( 2) = g ( ) P0 + { f 2 m ( ) P2 m }
m =1
B 0 s ( ) B 0 s ( 2) = g ( ) Q0 + { f 2 m ( ) Q2 m }
m =1
in which:
f 2 m ( ) = A 02 m ( ) + A 02 m ( 2)
Equation 3.267
These equations can also be written as:
B 0 c ( ) B 0 c ( 2) = { f 2 m ( ) P2 m }
m =0
B 0 s ( ) B 0 s ( 2) =
in which:
for m = 0 :
for m > 0 :
{ f ( ) Q }
m =0
2m
2m
f 0 ( ) = g ( )
f 2 m ( ) = A 02 m ( ) + A 02 m ( 2)
Equation 3.268
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 , for instance M = 10 .
Each -value less than 2 will provide an equation for the P2 m and Q2 m series. For a lot of
-values, the best fit values of P2 m and Q2 m are supposed to be those found by means of a
least squares method. Notice that at least M + 1 values of , less than 2 , are required to
solve these equations.
Another favourable method is to multiply both sides of the equations with . Then the
summation over can be replaced by integration.
Herewith, two sets of M + 1 equations have been obtained, one set for P2 m and one set for
Q2m :
2
2
P
f
(
f
(
2m 2 m
= (B 0 c ( ) B 0 c ( 2 )) f 2 n ( ) d
2n
0
m= 0
0
for: n = 0,...M
2
M
2
Q2 m f 2 m ( ) f 2 n ( ) d = (B 0 s ( ) B 0 s ( 2)) f 2 n ( ) d
0
m= 0
0
Equation 3.269
82
Now the P2 m and Q2 m series can be solved by a numerical method and with these values, the
coefficients P0 and Q0 are known now and from the definition of these coefficients in
Equation 3.266 follows the amplitude ratio of the radiated waves and the forced sway
oscillation:
a
b
=
xa
P0 2 + Q0 2
Equation 3.270
With the solved P2 m and Q2 m values, the velocity potential on the surface of the cylinder
( = 0 ) is known too:
M
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
g a
m =1
0 ( ) =
in which:
A 02 m ( ) = + sin ((2m + 1) )
b N
2n 1
n
( 1)
a
sin ((2m + 2 n) )
a n = 0
2m + 2 n 2 n 1
Equation 3.271
+ B 0 s ( ) + {Q2 m A 02 m ( )} cos ( t )
g a
m =1
B 0 c ( ) + {P2 m A 02 m ( )} sin ( t )
m =1
Equation 3.272
It is obvious that this pressure is skew-symmetric in .
3.2.2.2 Sway Coefficients
83
The two-dimensional hydrodynamic horizontal force, acting on the cylinder in the direction of
the x -axis, can be found by integrating the horizontal component of the hydrodynamic
pressure on the surface S of the cylinder:
+ 2
{ p (+ ) p( )}
Fx =
'
=2
dy0
ds
ds
dy0
p( ) d d
0
Equation 3.273
With this the two-dimensional hydrodynamic horizontal force due to sway oscillations can be
written as follows:
g b0 a
Fx ' =
( M 0 cos ( t ) N 0 sin ( t ))
in which:
1
M0 =
n =0
N 1
( 1)m Q2 m (2m + 1) a2 m +1
4 a m =1
N N
(2n 1) (2i 1)
b M
m
(
1
)
a2 n 1 a2 i 1
2 m
2
2
2
a m =1
n =0 i = 0 (2m + 2i ) (2 n 1)
a
sin =
P0
xa b
a
cos =
Q
xa b 0
Equation 3.275
the two-dimensional hydrodynamic horizontal force can be resolved into components in phase
and out phase with the horizontal displacement of the cylinder:
2
g b0 a
'
Fx =
2 b x a
Equation 3.276
84
2
P0 2 + Q0 2
N 22
b0 2 M 0 P0 N 0 Q0
=
2
P0 2 + Q0 2
'
Equation 3.278
The signs of these two coefficients are proper in both, the axes system of Tasai and the ship
motions O( x, y, z ) co-ordinate system.
The energy delivered by the exciting forces should be equal to the energy radiated by the
waves, so:
1
Tosc
Tosc
(N
0
'
22
g a c
x& (x& dt ) =
2
2
Equation 3.279
in which Tosc is the period of oscillation.
With the relation for the wave speed c = g at deep water, follows the relation between the
two-dimensional heave damping coefficient and the amplitude ratio of the radiated waves and
the forced sway oscillation:
N 22
'
g 2 a
=
3 xa
Equation 3.280
With this amplitude ratio the two-dimensional hydrodynamic damping coefficient of heave is
also given by:
2 b0 2
1
N 22 ' =
2
4
P0 + Q0 2
Equation 3.281
When comparing this expression for N 22' with the expression found before, the following
energy balance relation is found:
85
M 0 P0 N 0 Q0 =
2
2
Equation 3.282
MR =
'
{ p(+ ) p( )} x
= 2
p () x
+ dx0
dy0
+ y0
ds
ds
ds
dx0
dy
+ y0 0 d
d
d
Equation 3.283
With this the two-dimensional hydrodynamic roll moment due to sway oscillations can be
written as follows:
g b0 2 a
MR' =
(YR cos( t ) X R sin ( t ))
in which:
2
N N
1
YR =
2 a 0
n =0 i =0
M
N
N
1
(2i 1) (2 n 2i) a a
1
)
2
m
2 n 1
2 i 1
2
2
2 a 2 m =1
n = 0 i = 0 (2m + 1) (2 n 2i )
( 1) m Q2 m
N
n m
2
m
+
2
n
2
i
1
)
(
2
i
1
)
b N +
a 2 n 1 a 2 i 1 a 2 m +2 n 2 i 1
+
2n 2i
n = m i =0
8 a 3 m =1 N m N
+ ( 2m 2n + 2i 1) (2i 1) a
2 n 1 a 2 i 1 a 2 m 2 n + 2 i 1
n =0 i = m + n
2n 2i
86
g b0 2 a
(YR cos( t + ) X R sin ( t + ))
a
sin =
P0
xa b
a
cos =
Q
xa b 0
M R' =
Equation 3.285
the two-dimensional hydrodynamic roll moment can be resolved into components in phase
and out phase with the lateral displacement of the cylinder:
g b0 2 a 2
MR' =
2 b xa
((YR Q0 + X R P0 ) cos( t + ) + (YR P0 X R Q0 ) sin ( t + ))
Equation 3.286
This hydrodynamic roll moment can also be written as:
'
'
'
M R = M 42 &x& N 42 x&
= M 42 2 xa cos( t + ) + N42 xa sin ( t + )
'
'
Equation 3.287
in which:
M 42 '
N
'
42
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 coefficients of
sway into roll in Tasai's axes system are given by:
b0 3 YR Q0 + X R P0
'
M 42 =
2
P0 2 + Q0 2
N 42 ' =
b0 3 YR P0 X R Q0
2
P0 2 + Q0 2
Equation 3.288
In the ship motions O( x, y , z ) co-ordinate system these two coupling coefficients will change
sign.
87
Equation 3.290
This forced angular oscillation of the cylinder causes a surface disturbance of the fluid.
Because the cylinder is supposed to be infinitely long, the generated waves will be twodimensional. These waves travel away from the cylinder and a stationary state is rapidly
attained.
Two kinds of waves will be produced:
A standing wave system, denoted here by subscript A .
The amplitudes of these waves decrease strongly with the distance to the cylinder.
A regular progressive wave system, denoted here by subscript B .
These waves dissipate energy. At a distance of a few wavelengths from the cylinder, the
waves on each side can be described by a single regular wave train. The wave amplitude at
infinity a is proportional to the amplitude of oscillation of the cylinder a , provided that
this amplitude is sufficiently 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 fluid has to fulfil the following requirements:
1. The velocity potential must satisfy to the equation of Laplace:
2 2
2 =
+ 2 =0
x 2
y
Equation 3.291
2. Because the sway motion of the fluid is not symmetrical about the y -axis, this velocity
potential has the following anti-symmetric relation:
( x, y ) = (+ x, y )
Equation 3.292
3. The linearised free surface condition in deep water is expressed as follows:
88
+
=0
g
y
for: x
Bs
2
and
y=0
Equation 3.293
In consequence of the conformal mapping, this free surface condition can be written as:
N
b
(2n 1) a2 n 1 e ( 2 n 1 )
=0
for: 0 and =
a
2
n= 0
in which:
b 2
2 b0
=
M s
or
b =
(non-dimensional frequency squared)
a
g
2 g
Equation 3.294
From the definition of the velocity potential follows the boundary condition on the surface of
the cylinder S for = 0 :
0 ( )
r
= r0 & 0
n
s
Equation 3.295
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 , this boundary condition on the surface of the cylinder ( = 0 )
reduces to:
2
2
0 ( ) & x0 + y 0
=
s
s
2
Equation 3.296
Integration results into the following requirement for the stream function on the surface of the
cylinder:
&
0 ( ) = x0 2 + y0 2 + C(t )
2
Equation 3.297
in which C (t ) is a function of time only.
The vertical oscillation at the intersection of the surface of the cylinder and the waterline is
defined by:
b
= 0 = a sin ( t + )
2
Equation 3.298
When defining:
89
x0 + y 0
(b0 2 )2
2
1 N
n
=
( 1) a2 n 1 sin ((2 n 1) )
a n = 0
1 N
n
+ +
( 1) a2 n 1 cos ((2n 1) )
a n = 0
Equation 3.299
the stream function on the surface of the cylinder is given by:
b
0 ( ) = & 0 ( ) + C(t )
4
Equation 3.2100
For the standing wave system a velocity potential and a stream function satisfying to the
equation of Laplace, the non-symmetrical motion of the fluid and the free surface condition
has to be found.
The following set of velocity potentials, as they are given by Tasai [1961] and de Jong [1973],
fulfil these requirements:
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = + e ( 2 m +1) sin ((2m + 1) )
b N
2n 1
n
( 1)
a2 n 1 e (2 m + 2 n ) sin ((2m + 2n ) )
a n = 0
2 m + 2n
Equation 3.2101
g a
A =
{P2 m A 2 m (, ) cos( t )} + {Q2 m A 2 m (, ) sin ( t )}
m =1
m =1
in which:
A 2 m (, ) = e ( 2 m +1) cos ((2 m + 1) )
+
b N
2n 1
( 1)n
a2 n 1 e (2 m + 2 n ) cos((2m + 2n ) )
a n = 0
2m + 2 n
Equation 3.2102
the free surface condition, representing such a train of waves at infinity. For this, a function
describing a two-dimensional horizontal doublet at the origin O is chosen.
Tasai [1961] and de Jong [1973] gave the velocity potential of the progressive wave system
as:
g a
B =
{Bc ( x, y ) cos( t ) + Bs ( x, y ) sin ( t )}
in which:
Bc j = e y sin ( x )
cos(k y ) + k sin (k y ) k x
e
dk
k 2 + 2
0
Bs j = + e y cos( x )
+
while:
x
x + y2
j = +1 for: x > 0
j = 1 for: x < 0
=
2
g
sin ( x ) +
y
x + y2
sin (k y ) k cos(k y ) k x
e
dk
k 2 + 2
0
)
Equation 3.2105
91
Power series expansions, as given by Porter [1960], can be used instead of these last integrals
over k . Summations in these expansions converge much faster than the numeric integration
procedure. This has been showed for the sway case.
The total velocity potential and stream function to describe the waves generated by a swaying
cylinder are:
= A + B
= A + B
Equation 3.2107
So the velocity potential and the conjugate stream function are expressed by:
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
{Q2 m A 2m (, )} sin ( t )
+ Bs (, ) +
m =1
+ Bc (, ) + {P2 m A 2 m (, )} cos ( t )
g a
m =1
(, ) =
Equation 3.2108
When putting = 0 , the stream function is equal to the expression found before in Equation
3.2100 from the boundary condition on the surface of the cylinder:
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
g a
m =1
0 ( ) =
(
)
{
(
)
}
(
)
+
+
Q
sin
B 0s
2m
A 02 m
m =1
b
= & 0 ( ) + C(t )
4
in which:
A 02 m ( ) = cos((2m + 1) )
b N
2n 1
n
+
( 1)
a2 n 1 cos ((2 m + 2n ) )
a n = 0
2 m + 2n
Equation 3.2109
In this expression, B 0 c ( ) and B 0 s ( ) are the values of Bc (, ) and Bs (, ) at the
surface of the cylinder, so for = 0 .
So for each , the following equation has been obtained:
92
+ B 0 c ( ) + {P2 m A 02 m ( )} cos ( t )
m =1
b0
= &
( ) + C* (t )
a
+ B 0 s ( ) + {Q2 m A 02 m ( )} sin ( t )
m =1
Equation 3.2110
When putting = 2 , so at the intersection of the surface of the cylinder with the free
surface of the fluid where ( ) = 1 , we obtain the constant C * (t ) :
C * (t ) = B 0 c ( 2 ) + {P2 m A 02 m ( 2 )} cos( t )
m =1
+ B 0 s ( 2) + {Q2 m A 02 m ( 2 )} sin ( t )
m =1
b0
+ &
4 g a
in which:
N
m
2n 1
A 02 m ( 2 ) = b ( 1)
a2 n 1
a
n = 0 2m + 2 n
Equation 3.2111
A substitution of C * (t ) in the equation for each -value, results for any -value less than
2 into the following equation:
m =1
+ ( ) ( 2 ) + {Q (
( )
( ))} sin ( t )
B 0s
B 0S
2m
A 02 m A 02 m 2
m
=
1
b0
&
{( ) 1}
4 g a
Equation 3.2112
The right hand side of this equation can be written as:
a
b0
&
{( ) 1} = {( ) 1}
b sin ( t + )
4 g a
2 a
in which:
P0 =
a
sin ( )
2 a b
and
Q0 =
a
cos( )
2 a b
Equation 3.2113
This provides for each -value less than 2 a set of two equations with the unknown
parameters P2 m and Q2 m :
93
B 0 c ( ) B 0 c ( 2) = {( ) 1} P0 + { f 2 m ( ) P2 m }
m =1
B 0 s ( ) B 0 s ( 2) = {( ) 1} Q0 + { f 2 m ( ) Q2 m }
m =1
in which:
f 2 m ( ) = A 02 m ( ) + A 02 m ( 2)
Equation 3.2114
These equations can also be written as:
B 0 c ( ) B 0 c ( 2) = { f 2 m ( ) P2 m }
m =0
B 0 s ( ) B 0 s ( 2) =
in which:
for m = 0 :
for m > 0 :
{ f ( ) Q }
m =0
2m
2m
f 0 ( ) = ( ) 1
f 2 m ( ) = A 02 m ( ) + A 02 m ( 2)
Equation 3.2115
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 , for instance M = 10 .
Each -value less than 2 will provide an equation for the P2 m and Q2 m series. For a lot of
-values, the best fit values of P2 m and Q2 m are supposed to be those found by means of a
least squares method. Note that at least M + 1 values of , less than 2 , are required to
solve these equations.
Another favourable method is to multiply both sides of the equations with . Then the
summation over can be replaced by integration. Herewith, two sets of M + 1 equations
have been obtained, one set for P2 m and one set for Q2 m :
2
2
P2 m f 2 m ( ) f 2 n ( ) d = (B 0 c ( ) B 0 c ( 2 )) f 2 n ( ) d
0
m= 0
0
2
M
2
Q2 m f 2 m ( ) f 2 n ( ) d = (B 0 s ( ) B 0 s ( 2)) f 2 n ( ) d
0
m= 0
0
for: n = 0,...M
M
Equation 3.2116
Now the P2 m and Q2 m series can be solved by a numerical method and with these values, the
coefficients P0 and Q0 are known now and from the definition of these coefficients follows
the amplitude ratio of the radiated waves and the forced sway oscillation:
a
b
=
a 2 P0 2 + Q0 2
Equation 3.2117
94
With the solved P2 m and Q2 m values, the velocity potential on the surface of the cylinder
( = 0 ) is known too:
M
+ B 0 c ( ) + {P2 m A 02 m ( )} cos( t )
g a
m =1
0 ( ) =
in which:
A 02 m ( ) = + sin ((2m + 1) )
b N
2n 1
n
( 1)
a2 n 1 sin ((2m + 2 n) )
a n = 0
2m + 2 n
Equation 3.2118
+ B 0 s ( ) + {Q2 m A 02 m ( )} cos ( t )
g a
m =1
B 0 c ( ) + {P2 m A 02 m ( )} sin ( t )
m =1
Equation 3.2119
It is obvious that this pressure is skew-symmetric in .
3.2.3.2 Roll Coefficients
The two-dimensional hydrodynamic moment acting on the cylinder in the clockwise direction
can be found by integrating the roll component of the hydrodynamic pressure on the surface
S of the cylinder:
+ 2
MR =
'
{ p(+ ) p( )} x
= 2
p () x
+ dx0
dy0
+ y0
ds
ds
ds
dx0
dy
+ y0 0 d
d
d
Equation 3.2120
With this the two-dimensional hydrodynamic moment due to roll oscillations can be written as
follows:
95
MR' =
g b0 2 a
(YR cos( t ) X R sin ( t ))
in which:
1
2 a 2
YR =
+
n +i
B 0s ( ) ( 1) (2i 1) a 2n1 a2i 1 sin ((2n 2i ) ) d
0
n =0 i =0
M
N
N
1
(2i 1) (2 n 2i) a a
m
1
)
2 m
2 n 1
2 i 1
2
2
2
2 a m =1
n = 0 i = 0 (2m + 1) (2 n 2i )
( 1)m Q2 m
N n m
(
2m + 2 n 2i 1) (2i 1)
M
b
a 2 n 1 a 2 i 1 a 2 m +2 n 2 i 1
+
2n 2i
3
n = m i =0
8 a m =1 N m N
+ ( 2m 2n + 2i 1) (2i 1) a 2 n 1 a 2 i 1 a 2 m 2 n + 2 i 1
n =0 i = m + n
2n 2i
and X R as obtained from this expression above for YR , by replacing there B0 s ( ) by
B0 c ( ) and Q2 m by P2 m .
Equation 3.2121
These expressions are similar to those found before for the hydrodynamic roll moment due to
sway oscillations.
With Equation 3.2121 in some other format:
g b0 2 a
'
MR =
(YR cos ( t + ) X R sin ( t + ))
2 a
sin =
P0
xa b
2 a
cos =
Q
xa b 0
Equation 3.2122
the two-dimensional hydrodynamic roll moment can be resolved into components in phase
and out phase with the angular displacement of the cylinder:
2
2
2 g b0 a
M R' =
2 b a
in which:
96
M 44 '
'
44
When using also the amplitude ratio of the radiated waves and the forced roll oscillation,
found before, the two-dimensional hydrodynamic mass and damping coefficients of roll in
Tasai's axes system are given by:
+ b0 4 YR Q0 + X R P0
M 44 ' =
8
P0 2 + Q0 2
N 44
+ b0 4 YR P0 X R Q0
=
8
P0 2 + Q0 2
'
Equation 3.2125
The signs of these two coefficients are proper in both, the axes system of Tasai and the ship
motions O( x, y , z ) co-ordinate system.
The energy delivered by the exciting moments should be equal to the energy radiated by the
waves, so:
1
Tosc
Tosc
(N
0
'
44
)(
g a c
& & dt =
2
2
Equation 3.2126
in which Tosc is the period of oscillation.
With the relation for the wave speed c = g in deep water, follows the relation between the
two-dimensional heave damping coefficient and the amplitude ratio of the radiated waves and
the forced sway oscillation:
N 44
'
g2
=
3
a
a
Equation 3.2127
With this amplitude ratio the two-dimensional hydrodynamic damping coefficient of heave is
also given by:
2 b0 4
1
N 44 ' =
2
64
P0 + Q0 2
Equation 3.2128
When comparing this expression for N 44' with the expression found before, the following
energy balance relation is found:
2
YR P0 X R Q0 =
8
Equation 3.2129
97
{ p (+ ) p( )}
Fx =
'
=2
dy0
ds
ds
dy0
p( ) d d
0
Equation 3.2130
With this the two-dimensional hydrodynamic horizontal force due to sway oscillations can be
written as follows:
g b0 a
Fx ' =
( M 0 cos ( t ) N 0 sin ( t ))
in which:
2
N
1
n
M0 =
B 0 s ( ) ( 1) (2n 1) a 2 n 1 sin ((2 n 1) ) d
a 0
n =0
N 1
( 1)m Q2 m (2m + 1) a2 m +1
4 a m =1
N N
(2n 1) (2i 1)
b M
m
(
1
)
2
m
2
n
1
2
i
1
2
2
a 2 m =1
n =0 i = 0 (2m + 2i ) (2 n 1)
2 a
sin =
P0
xa b
2 a
cos =
Q
xa b 0
Equation 3.2132
the two-dimensional hydrodynamic horizontal force can be resolved into components in phase
and out phase with the horizontal displacement of the cylinder:
98
2 g b0 a 2
2 b a
(( M 0 Q0 + N 0 P0 ) cos ( t + ) + (M 0 P0 N 0 Q0 ) sin ( t + ))
Fx ' =
Equation 3.2133
This hydrodynamic vertical force can also be written as:
&& N24 ' &
Fx ' = M 24 '
= M 24' 2 a cos ( t + ) + N 24 ' a sin ( t + )
Equation 3.2134
in which:
M 24 '
N 24'
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 coefficients of
roll into sway are given by:
b0 3 M 0 Q0 + N 0 P0
M 24 ' =
8
P0 2 + Q0 2
N 24 ' =
b0 3 M 0 P0 N 0 Q0
8
P0 2 + Q0 2
Equation 3.2135
In the ship motions O( x, y, z ) co-ordinate system these two coupling coefficients will change
sign.
99
M 22
'
Ds
(1 a1 )2 + 3 a3 2
( 0 ) =
2 1 a1 + a3
}
Equation 3.2136
The 2-D hydrodynamic mass coupling coefficient of sway into roll of a Lewis cross section is
given by Grim [1955] as:
16
Ds
M 42 ' ( 0) = M 22 ' ( 0)
3 1 a1 + a3
4
3
4
12
a1 1 a1 + a3 a1 a3 + a3 2 + a3 a3 2
5
5
7
5
2
2
(1 a1 ) + 3 a3
Equation 3.2137
In Tasai's axes system, M 42 ' will change sign.
The 2-D hydrodynamic mass coefficient of heave of a Lewis cross section goes to infinite, so:
'
M 33 ( 0) =
Equation 3.2138
The 2-D hydrodynamic mass moment of inertia coefficient of roll of a Lewis cross section is
given by Grim [1955] as:
4
16
Bs
M 44 ( 0 ) =
2 (1 + a1 + a3 )
'
8
16 2
2
2
a1 (1 + a3 ) + a1 a3 (1 + a3 ) + a3
9
9
Equation 3.2139
The 2-D hydrodynamic mass coupling coefficient of roll into sway of a Lewis cross section is
given by Grim [1955] as:
100
1 a1 + a3
6
Ds
3
4
12 2
a1 (1 a1 ) (1 + a3 ) + a1 a3 (1 + a3 ) + a3 (1 a1 ) a3
5
5
7
8
16 2
2
2
a1 (1 + a3 ) + a1 a3 (1 + a3 ) + a3
9
9
Equation 3.2140
In Tasai's axes system, M 24 ' will change sign.
All potential damping values for zero frequency will be zero:
'
'
N 22 ( 0 ) = 0
N 44 ( 0 ) = 0
'
N33 ( 0) = 0
N 42 ' ( 0 ) = 0
N 24 ' ( 0 ) = 0
Equation 3.2141
2
Ds
16
(1 a1 + a3 )2 + a3 2
M 22 ( ) =
1 a1 + a3
3
'
Equation 3.2142
The 2-D hydrodynamic mass coefficient of heave of a Lewis cross section is given by Tasai
[1959] as:
2
Bs
(1 + a1 )2 + 3 a3 2
M 33 ( ) =
2 2 (1 + a1 + a3 )
'
)
Equation 3.2143
The 2-D hydrodynamic mass moment of inertia coefficient of roll of a Lewis cross section is
given by Kumai [1959] as:
4
Bs
a12 (1 + a3 )2 + 2 a3 2
M 44 ( ) =
2 (1 + a1 + a3 )
'
)
Equation 3.2144
Information about the 2-D hydrodynamic mass coupling coefficients between sway and roll
of a Lewis cross section could not be found in literature, so:
'
M 42 ( ) = ?
M 24 ' ( ) = ?
Equation 3.2145
101
102
Lewis coefficient
source strength
Aindex
b
B
c
Cindex
Eindex
Findex
g
G
h
H
H indices
HT
I"
kx
ky
amplitude ratio
Lewis coefficient
breadth of body
wave velocity
non-dimensional force or moment
non-dimensional exciting force or moment
hydrodynamic force
acceleration of gravity
function (real part)
water depth
function (imaginary part)
fictive moment arm
water depth - draft ratio
hydrodynamic moment of inertia
wave number in x -direction
wave number in y -direction
m"
M indices
N indices
p
hydrodynamic mass
hydrodynamic moment
hydrodynamic damping coefficient
pressure
r = x 2 + y 2 polar coordinate
103
= 2 g
0 = 2
indices
indices
indices
indices
related to excitation
horizontal or related to horizontal motions
imaginary part
numbering of potential parts
related to transverse motions
real part
related to roll motions
related to vertical or vertical motions
related to waves
104
2 2
2 h
=
tanh
= 0 tanh [0 h]
g
Equation 3.31
The fluid is supposed to be incompressible and inviscid. The flow caused by the oscillating
body in the surface of this fluid can be described by a potential flow. The problem will be
linearised, i.e., contributions of second and higher order in the definition of the boundary
conditions will be ignored. Physically, this yields an assumption of small amplitude motions.
The earth-bounded axes system of the sectional contour is given in Figure 3.32-a.
= vy
= vz
y
z
The value of the stream function increases when - going in the positive direction - the flow
goes in the negative y -direction:
1 < 2
=+
z
y
105
3 > 4
z = y
s
n
s = + n
Equation 3.32
For the imaginary parts, the symbols i and j have been used: i for geometrical variables
(potential and stream function) and j for functions of time.
B
+
= 0 or +
=0
for: y
and z = 0
g
z
z
2
Equation 3.34
3. The seabed is impervious, so the vertical fluid velocity at z = h is zero:
=0
for: z = h
z
Equation 3.35
4. The harmonic oscillating cylinder produces a regular progressive wave system, travelling
away from the cylinder, which fulfils the Sommerfeld radiation condition:
lim y Re{} 0 Im {} = 0
y
y
Equation 3.36
In here, 0 = 2 is the wave number of the radiated wave.
106
5. The oscillating cylinder is impervious too; thus at the surface of the body is the fluid
velocity equal to the body velocity, see Figure 3.32-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
d
ds
body
body
Equation 3.37
Two cases have to be distinguished:
a) The hydromechanical loads, which have to be obtained for the vertically oscillating
cylinder in still water with a vertical velocity equal to:
V = V e j t
The boundary condition on the surface of the body becomes:
d
dy
=V
e j t
ds body
ds body
or:
( y, z , t ) body = V e j t y body + C
Equation 3.38
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 diffraction potential S :
W S
+
n
n
=0
body
W dy W dz S dy S dz
z ds
y ds
z ds
y ds
=
=
or:
dS ( y , z , t )
body
dW dS
+
ds
ds
= dW ( y, z, t )
=
body
body
body
W
W
dy
dz
z
y
body
Equation 3.39
The stream function of an incoming wave - which travels in the negative y -direction, so
= +90 0 - is given in Appendix I of this Section by:
107
W = j
j ( t + 0 y )
e
{sinh [0 z ] tanh [ 0 h ] cosh [0 z ]}
Equation 3.310
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 = +90 0 :
S ( y, z , t ) body = W ( y, z, t ) body
=
j t
e
sinh [0 z ] tanh [0 h ] cosh[0 z ] sinh [0 y ] body
Equation 3.311
In case of another wave direction, this problem becomes three-dimensional and a stream
function can not be written. However, boundary condition (Equation 3.311) provides us a
~
''quasi stream function'' s , i.e. this is the amount of fluid which has to come out of the body
per unit of length, so that - in total - no fluid of the incoming wave enters into the body.
This function can be used as an approximation of the problem:
~
s ( x1 , y1 , z1 , t )
y1
1
W
=
dy W dz
body
z
y
0
z0
body
j t
e
cos(0 x cos )
+ 1 sin 2
y1
Equation 3.312
3.3.3.2 Potentials
3.3.3.2.1 3-D Radiation Potential
Suppose a three-dimensional oscillating cylindrical body in previously still water. To find the
potential of the resulting fluid motions, an oscillating pressure p at the free surface will
replace this body. The unknown amplitude p 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 on both sides of y = 0 ,
so:
108
p ( x, y , z = z0 , t ) = j p (x , y, z = z0 ) e j t
or :
p ( x, y , z = z0 ) = 0
for : y y 2 and x L 2
Equation 3.313
P=
p( x, y , z = z ) dy dx
0
L 2 y 2
+L 2
P (x ) dx <
'
L 2
Equation 3.314
The boundary condition in Equation 3.313 can be fulfilled by pressure amplitude
p ( x, y , z = z 0 ) , which is found by a superposition of an infinite number of harmonic pressures.
From Equation 3.314 follows that the pressure amplitude p (x ) can be integrated, so a
Fourier series expansion follows from:
+
1
p ( x ) = p() cos (kx ( x )) d dk x
0
Because the pressure amplitude p depends on two variables, the Fourier series expansion has
to be two-dimensional too:
1 + +
p ( x, y , z = z0 ) = 2 p(,)
0 0
cos(k y ( y )) cos(k x ( x )) d d dk y dk x
in which k x is the wave number in the x -direction and k y is the wave number in the y direction.
According to Equation 3.313, the pressure amplitude p disappears for y y 2 and
x L 2 , so for this pressure expression remains:
p ( x, y , z = z0 ) =
+ L 2 + y 2
p(,)
2 0 0 L 2 y 2
cos(k y ( y )) cos(k x ( x )) d d dk y dk x
It is assumed that the value of y is small. This means that remains small too. Thus, one
can safely suppose that:
cos (k y ( y )) cos (k y y )
which results in:
+ L 2 + y 2
1
p ( x, y , z = z0 ) = 2 p(,) d cos (k x ( x )) d cos(k y y ) dk y dk x
0 0 L 2 y 2
+L 2
'
1
cos(k y y ) P () cos(k x (x )) d dk y dk x
2
00
L 2
109
This pressure definition leads - as a start - to the following initial definition of the radiation
potential:
cosh k x 2 + k y 2 ( z h )
dk dk d
x
y
2
2
sinh k x + k y h
Equation 3.316
in which the function C (k x , k y ) is still unknown.
This expression in Equation 3.316 for the radiation potential fulfils the Equation of Laplace:
2 2 2
+
+ 2 =0
x 2 y 2
z
Now, the harmonic 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:
p1
j t
=e
C (k x , k y ) cos(k x (x )) cos(k y y )
t
0 0 0
2 g k x 2 + k y 2 tanh k x 2 + k y 2 h
dk dk d
x
y
2
2
tanh k x + k y h
Equation 3.317
The harmonic oscillating pressure is given by:
p1 ( x, y , z = z 0 , t ) = j p1 ( x, y , z = z 0 ) e j t
Equation 3.318
and its amplitude becomes:
p1 ( x, y , z = z0 ) = C (k x , k y ) cos(k x (x )) cos (k y y )
000
2 g k x 2 + k y 2 tanh k x 2 + k y 2 h
dk dk d
x
y
2
2
tanh k x + k y h
Equation 3.319
If this pressure amplitude p 1 is supposed to be equal to the amplitude p , then combining
Equation 3.315 and Equation 3.319 provides the unknown function C (k x , k y ):
110
+L 2
'
1
cos(k y y ) P ( ) cos (k x ( x )) d dk y dk x
2
00
L 2
= p1 ( x, y , z = z 0 )
=
g
cos(k y y ) C (k x , k y ) cos (k x ( x ))
00
0
2
2
2
2
k x + k y tanh kx + k y h
dk dk d
x
y
2
2
tanh k x + k y h
Equation 3.320
Comparing the two integrands provides:
2
2
2
2
g k x + k y tanh k x + k y h
d =
0 C (k x , k y ) cos(k x (x ))
2
2
tanh k x + k y h
+L 2
'
1
P () cos (k x ( x )) d
2
L 2
or:
tanh k x 2 + k y 2 h
0 C (k x , k y ) cos(k x ( x )) d = k 2 + k 2 tanh k 2 + k 2 h
x
y
y
x
+L 2
'
P () cos(k x (x )) d
2
g L 2
Equation 3.321
When defining:
P ( )
A0 () =
g 2
and substituting Equation 3.321 in Equation 3.316 provides the radiation potential:
'
0 r ( x, y , z , t ) = e
j t
+L 2
L 2
A0 ( ) cos (k x ( x ))
cosh k x 2 + k y 2 ( z h ) cos (k y y )
0 cosh k 2 +k 2 h k 2 + k 2 sinh k 2 + k 2 h dk x dk y d
y
x
y
y
x
Equation 3.322
This potential fulfils both, the radiation condition at infinity and the boundary condition at the
free surface.
3.3.3.2.2 2-D Radiation Potential
111
j t
cosh [k ( z h )]
cos(k y ) dk
cosh[k h] k sinh [k h]
0
A0
Equation 3.323
To fulfil also the Sommerfeld radiation condition in Equation 3.36, a term has to be added.
For this, use will be made here of the value of the potential given in Equation 3.323 at a large
distance from the body:
cosh [k ( z h )]
0 r ( y, z , t ) = e j t A0
cos(k y ) dk
cosh
[
k
h
]
sinh
[
k
h
]
0
1
= e j t A0 F (k ) e + i k y dk + F (k ) e i k y dk
2 0
0
with:
F (k ) =
cosh[k ( z h)]
cosh [k h] k sinh [k h ]
112
+i u y
F (u ) e du = ...dk + ...du + ...du + ...du
II
III
IV
= J I + J II + J III + J IV
=0
with:
F (k ) e
+ i k y
dk = lim ( J II + J III + J IV )
R
The location of the singular point follows from the denominator in the expression for
F (k ) :
cosh [0 h ] 0 sinh [0 h] = 0
Because lim ( J III ) = 0 and J IV disappears too for a large y , the singular point itself
R
cosh [0 (z h )]
e+ i 0 y
h sinh [0 h] sinh [0 h ] 0 h cosh [0 h]
cosh [0 h] cosh[0 ( z h )] + i 0 y
= +i
e
0 h + sinh [0 h] cosh [0 h ]
and the searched integral becomes for y :
= i
F (k ) e
+ i k y
dk = i
i u y
F (u ) e du = ...dk + ...du + ...du + ...du
V
VI
VII
= J I + J V + JVI + JVII
=0
with:
F (k ) e
i k y
Because lim ( J VI ) = 0 and J VII disappears too for a large y , the singular point itself
R
F (k ) e
0
i k y
dk = +i
cosh [0 h ] cosh[0 ( z h )] i 0 y
e
0 h + sinh [0 h] cosh[0 h ]
cosh[0 h] cosh [0 (z h )]
sin (0 y )
0 h + sinh [0 h ] cosh [0 h]
Equation 3.324
cosh[0 h ] cosh[0 ( z h )]
cos (0 y )
0 h + sinh [0 h] cosh[0 h ]
Equation 3.325
With Equation 3.324 and Equation 3.325, the radiation potential becomes:
0 r ( y, z , t ) + j 0 j ( y , z , t ) = e j t A0
cosh[k (z h )]
cos(k y ) dk +
0 cosh [k h ] k sinh [k h]
0 h + sinh [0 h] cosh [0 h]
Equation 3.326
From this follows for y :
cosh 2 [0 h ]
= A0
g 0 h + sinh [0 h] cosh[0 h ]
0 ( y , z , t ) = j e j t A0
Equation 3.327
travelling away from both sides of the cylinder.
From the orthogonality condition:
=+
y
z
follows the stream function:
0 r ( y, z , t ) + j 0 j ( y , z , t ) = e j t A0
sinh [k ( z h )]
sin (k y ) dk +
0 cosh [k h ] k sinh [k h]
0 h + sinh [0 h] cosh [0 h]
Equation 3.328
For an infinite water depth, Equation 3.326 and Equation 3.328 reduce to:
114
0 r ( y , z , t ) + j 0 j ( y, z , t ) = e j t A0
e k z
cos(k y ) dk + j e z cos( y )
0 k
j t
0 r ( y , z , t ) + j 0 j ( y, z , t ) = e
A0
e k z
sin (k y ) dk + j e z sin ( y )
0 k
Equation 3.329
Now, the potential and stream functions can be written as:
0 = 0 r + 0 rad + j 0 j
e k z
cos (k y ) dk +
0 k
e k h
sinh [k z ] k cosh[k z ]
= e j t A0
cos (k y )
dk
cosh
[
k
h
]
sinh
[
k
h
]
0
cosh [0 h ] cosh[0 ( z h )]
cos (0 y )
j
0 h + sinh [0 h] cosh[0 h ]
0 = 0 r + 0 rad + j 0 j
e k z
sin (k y ) dk
0 k
e k h
cosh [k z ] k sinh [k z ]
= e j t A0
sin (k y )
dk
cosh[k h] k sinh [k h]
0 k
0 h + sinh [0 h] cosh[0 h ]
Equation 3.330
In here, 0 r is the potential at deep water and 0 rad is the additional potential due to the
finite water depth. j can be written in the same way.
3.3.3.2.3 Alternative Derivation
Assuming that the real part of the potential at an infinite water depth, r , 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 fulfils the free surface
condition and - together with r - also the boundary condition at the seabed.
As a start for the real additional potential will be chosen:
Equation 3.331
115
f (k ) cos (k ) dk = g ()
0
is known:
f (k ) =
1
g ( ) cos (k ) d
0
k cosh [k ( z h )]
The still unknown function C1 (k ) follows from the boundary condition at the seabed:
0 r 0 rad
+
=0
z
z z = h
k e k h
cos (k y ) dk
0 k
j t
= e
A0
C (k ) k cosh[k h] cos (k y ) dk
1
So:
C1 (k ) =
e k h
( k ) cosh[k h]
C 2 (k ) =
k e k h
( k ) { cosh[k h] k sinh [k h]} cosh[k h ]
With this, the real additional potential, as given in Equation 3.330, becomes:
0 rad ( y, z, t ) = e j t A0
e k h sinh [k z ] k cosh[k z ]
0 k cosh [k h ] k sinh [k h] cos(k y ) dk
Equation 3.332
The imaginary part can be obtained as described before.
116
The free surface conditions can not be fulfilled with the potential and the stream function in
Equation 3.330 only.
Additional potentials n are required which fulfil the boundary conditions in Equation 3.33
through Equation 3.36 and together with 0 also fulfil the boundary conditions in Equation
3.37 through Equation 3.39:
( y , z , t ) = A0 0 ' ( y , z , t ) + An n ' ( y , z , t )
n =1
}
( y , z , t )}
A { ( y , z , t ) + ( y , z , t ) +
'
'
nr
nrad
j nj '
n =1
Equation 3.333
Use will be made here of multi-potentials given by Grim [1956, 1957] of which - using the
Sommerfeld radiation condition - the real additional potential nrad and the imaginary
potential part nj will be determined. This results in:
nr ( y , z , t ) = + e j t An
(k + ) k
2 (n 1)
e kz cos (k y ) dk
nrad ( y , z , t ) = + e j t An
(k + ) k
2 (n 1)
e kh
nj ( y , z , t ) = e j t An
sinh [k z ] k cosh [k z ]
cos (k y ) dk
cosh [k h] k sinh [k h ]
0 2 n
cosh [0 (z h )]
cos (0 y )
cosh[0 h] 0 h + sinh [0 h ] cosh[0 h ]
Equation 3.334
The orthogonality condition provides the stream functions:
nr ( y , z , t ) = + e j t An
(k + ) k
2 (n 1)
e kz sin (k y ) dk
nrad ( y , z , t ) = e j t An
(k + ) k
0
2 (n 1)
nj ( y , z , t ) = + e j t An
e kh
cosh[k z ] k sinh [k z ]
sin (k y ) dk
cosh [k h] k sinh [k h ]
0 2 n
sinh [0 ( z h )]
sin (0 y )
cosh[0 h] 0 h + sinh [0 h ] cosh[0 h ]
117
}}
n =0
{
{ {
'
'
'
n =0 + j Anj nr + nrad + Anr nj
A {
nr nr + nrad } Anj nj
= e j t
n =0
+ j {Anj {nr + nrad } + Anr nj }
}}
n =0
{
{ {
'
'
'
n =0 + j Anj nr + nrad + Anr nj
A {
nr + nrad } Anj nj
nr
= e j t
n =0
+ j {Anj {nr + nrad } + Anr nj }
Equation 3.336
Summarised, the complex total potential can now be written as:
( y , z , t ) + i ( y, z, t ) = e j t {A0 r + j A0 j }
cos (k ( y + i ( z h )))
dk
0 cosh[k h] k sinh [k h]
+
cosh
[
h
]
cos
(
(
y
+
i
(
z
h
)
)
)
+ j
0
0
0 h + sinh [0 h ] cosh[0 h ]
{Anr + j Anj }
cos
(
k
(
y
+
i
(
z
h
)
)
)
2
2
2 ( n 1 )
dk
k k
j t
e
0
cosh [k h ] k sinh [k h]
n =1
2n
cos(0 ( y + i ( z h)))
j 0
Equation 3.337
The coefficients Anr and Anj with 0 n have to be determined in such a way that the
instantaneous boundary conditions on the body surface have been fulfilled. These coefficients
are dimensional and it is very practical to determine them for the amplitude of the flow
velocity V ; also if they then have the dimension L2 n +1 :
118
Ar + j Aj = V
'
0 r = e
0 r = e
+
ln
(
r
)
+
Re ( z + i y )m cos ( y ) +
m =1 m m!
m
arctan y +
Im ( z + i y ) sin ( y )
z m =1 m m!
m
m
(
)
+
ln
r
+
Re ( z + i y ) sin ( y )
m =1 m m!
m
arctan y + Im ( z + i y )m cos ( y )
z m =1 m m!
Equation 3.338
with the Euler constant: = 0.57722 .
For r = y 2 + z 2 :
M
(m 1)!
y
0 r = m 2 m Re ( z + i y )m + e z sin ( y )
y
m =1 r
M
y
(m 1)!
0 r = m 2 m Im ( z + i y )m e z cos ( y )
y
m =1 r
Equation 3.339
Mind you that
(m 1)! ( z + i y )m
m r 2m
is semi-convergent.
Equation 3.340
For the expansion of the remaining potential parts use has been made of the following
relations as derived in Appendix II:
119
k 2t
cosh [k z ] cos(k y ) =
Re ( z + i y )2 t
t = 0 (2t )!
k 2t
2t
Im (z + i y )
t = 0 (2t )!
sinh [k z ] sin (k y ) =
k 2 t +1
Re ( z + i y )2 t +1
t = 0 (2t + 1)!
sinh [k z ] cos(k y ) =
k 2 t +1
cosh [k z ] sin (k y ) =
Im ( z + i y )2 t +1
t = 0 (2t + 1)!
e k h
0 rad =
dk
0 ( k ) ( cosh[k h] k sinh [k h])
k 2t+ 1
k 2t
2 t +1
Re
(
z
+
i
y
)
Re ( z + i y )2 t
t = 0 (2t + 1)!
t = 0 (2t )!
k 2 t +1 e k h
dk
t =0
2 t +1
2t
(2t + 1) Re (z + i y )
Re ( z + i y )
}
{
})
G(2t + 1)
=
( Re {( z + i y ) } (2t + 1) Re{(z + i y ) })
(2t + 1)!
t =0
2 t +1
2t
Equation 3.341
It is obvious that:
G(2t + 1)
0 rad =
(2t + 1) Im ( z + i y )2 t Im ( z + i y )2 t +1
(2t + 1)!
t =0
})
Equation 3.342
The function:
k t e k h
dk
(k ) ( cosh [k h ] k sinh [k h])
0
will be treated in the next Section.
G (t ) =
120
k 2 2 k 2 ( n 1 ) e k h
nrad =
dk
0 (k ) ( cosh [k h] k sinh [k h ])
2 t +1
2t
k
2 t +1
k
Re ( z + i y )2 t
k
t = 0 (2t + 1)! Re ( z + i y )
t = 0 (2t )!
k 2 t + 2 n +1 e k h
dk
1 0 (k ) ( cosh[k h ] k sinh [k h])
(2t + 1)!
=
k 2 t + 2 n 1 e k h
2
dk
t =0
Re ( z + i y )2 t +1 (2t + 1) Re ( z + i y )2 t
G(2t + 2n + 1) 2 G(2t + 2n 1)
(2t + 1)!
=
t =0
2 t +1
(2t + 1) Re ( z + i y )2 t
Re ( z + i y )
})
})
Equation 3.343
It is clear that:
nrad
G(2t + 2n + 1) 2 G(2t + 2n 1)
(2t + 1)!
=
t =0
2 t +1
(2t + 1) Im ( z + i y )2 t
Im ( z + i y )
})
Equation 3.344
For the imaginary parts can be written:
+ cosh 2 [ 0 h ]
0 j =
0 h + sinh [ 0 h] cosh [ 0 h ]
2 t +1
0 2 t
2t
0
2 t +1
Re ( z + i y ) tanh [0 h]
Re (z + i y )
t = 0 (2t )!
t =0
(2t + 1)!
cosh 2 [0 h]
=
0 h + sinh [ 0 h] cosh [ 0 h ]
0 j
0 2 t
2t
Im (z + i y )
t = 0 (2t )!
2 t +1
0
2 t +1
Im ( z + i y )
tanh [0 h]
t =0
(2t + 1)!
Equation 3.345
121
nj =
0 2 n
0 h + sinh [ 0 h] cosh [ 0 h ]
0 2 t
Re ( z + i y )2 t
t = 0 (2t )!
2 t +1
0
h
]
Re ( z + i y )2 t +1
t =0
(2t + 1)!
2 t +1
0 2 t
Im (z + i y )2 t tanh [0 h] 0
Im ( z + i y ) 2 t +1
t = 0 (2t )!
t =0
(2t + 1)!
= 0 2 ( n 1) 0 2 2 0 j
= 0
nj =
2 ( n 1)
} tanh [
0 0 j
2
+ 0 2 n
0 h + sinh [ 0 h] cosh [ 0 h ]
Equation 3.346
3.3.3.4 Function G(t)
The function:
k t e k h
dk
(k ) ( cosh [k h ] k sinh [k h])
0
G (t ) =
[ ]
with unit L1 t has two singular points: k = and k = 0 , see Figure 3.34.
ut eu
du
(u h ) ( h cosh [u ] u sinh [u ])
0
Equation 3.347
A substitution of:
w = u +i v =
2 ei
2
122
provides:
J =
wt e w
dw
(w h) { h cosh [w] w sinh [w]}
II
III
IV
= ...du + J I + J II + J III + J IV
0
=0
From this follows:
u t eu
0 (u h ) { h cosh [u] u sinh [u ]} du = Re{J I + J II + J III + J IV }
J I and J II are imaginary because they are residues and J III = 0 for R .
So, it remains:
ut eu
0 (u h ) { h cosh [u ] u sinh [u ]} du
= Re{J IV }
wt e w
= Re
dw
IV (w h) { h cosh [w] w sinh [w]}
~ = 2 ei
with real: w
2
the nominator of this integral will be made real by removing.
So:
123
wt e w
~] w
~ sinh [w
~]}
(w h ) { h cosh[w
'
G (t ) = Re
dw
IV ( w h ) { h cosh [w] w sinh [w]}
~ h) { h cosh [w
~] w
~ sinh [w
~]}
( w
t
= cos
2(h )2 1 tan t cos + e + 1 + tan t sin
4
4
4
4
d
0 2
2
h
)
2
2
2
cosh [ ] 2 h + 2h tanh []
+ 2 2 h 2 2 cos + 2h sin
Equation 3.348
Because t is always odd:
d
4 0
2 2h( h )
cosh[] 2 2 h 2 + 2 2h tanh [ ]
2 2
2
+ 2 h cos + 2h sin
for t = 1,5,9,......
'
(
{
d
4 0
2 2h ( h )
cosh [ ] 2 2 h 2 + 2 2h tanh []
2 2
2
+ 2 h cos + 2h sin
'
for t = 3,7,11,......
u e u
'
G (1) = G(1) =
du
(u h ) ( h cosh [u ] u sinh [u ])
0
Equation 3.349
124
This integral converges fast for small h -values. This will be approximated by:
u e u
lim G1 (1) =
du
h 0
(u h ) h u 2
0
Equation 3.350
This can be written as:
e u
du +
1 h 0 u h
1
e
lim G1 (1) = lim
du +
h 0
h 0 2 h
h 1 0 u h
1
e
du
2 h h + 1 0 u + h
From:
eu
am
a
du
=
+
ln
a
+
0 u a
m =1 m m!
follows:
e h
(
h )m
1 h + ln ( h ) + m m! +
m =1
+ e h
ln ( h ) ( h ) 2
h 0
h 0
2
m =1 m m!
2 h 1 h
+ h
e
ln ( h )
( h) 2
+
+ ( 1)m
2
m m!
2 h 1+ h
m =1
or:
e h
(
h )m
1 h + ln ( h) + m m! +
m =1
ln ( h)
(
h ) 2
1 + h e h
lim G1 (1) = lim
+
+ h +
h 0
h 0
2
m =2 m m!
2 h (1 h)
m
1 h e + h
ln
(
h
)
m ( h ) 2
+
h + ( 1)
2
m m!
2 h (1 h )
m= 2
or:
125
e h
(
h)m
+ ln ( h ) +
+
m =1 m m!
1 h
1 cosh h sinh h + ln ( h ) +
1 h
2
h
cosh h sinh h
1
m 1
m
+ e h ( h ) 2 + e h ( h) 2
m = 2 m m!
m = 2 m m!
m 1
m
+ h
m ( h ) 2
m ( h ) 2
+ h
e
( 1)
+e
( 1)
m m!
m m!
m =2
m =2
or:
]
[
lim G1 (1) = 1 ln ( h )
h 0
Equation 3.351
The imaginary part of integral in Equation 3.348 has been treated in Appendix III.
3.3.3.5 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
t
= j
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 90 0 is required, which means a
multiplication with j . Then the pressure is:
=
p dyn =
The hydrodynamic force on the body is equal to the integrated pressure on the body. In the
two-dimensional case, this is a force per unit length.
The vertical force becomes:
FV = FVr + j FVj
= p dy
S
)
{A
n= 0 S
nr
'
)}
'
'
Equation 3.352
126
The real part of this force is equal to the hydrodynamic mass coefficient times the oscillatory
acceleration, from which the hydrodynamic mass coefficient follows:
F
F Vr
m" = Vr =
b
V
or non-dimensional:
m"
F Vr
CV =
=
2
2
B
B V
8
8
Equation 3.353
The imaginary part of the force must be equal to the hydrodynamic damping coefficient times
the oscillatory velocity, from which the damping coefficient follows:
F Vj
NV =
V
Instead of this coefficient, generally the ratio between amplitude of the radiated wave and
the oscillatory motion z will be used. The energy balance provides:
2
AV
=
z
2
NV
=
g 2 cgroup
=
=
0
sinh [2 0 h]
N
g 2 0 h + sinh [2 0 h] V
2
cosh 2 [0 h]
F Vj
V 0 h + sinh [0 h ] cosh[0 h ]
Equation 3.354
In deep water, the hydrodynamic mass for 0 becomes infinite, because the potential in
Equation 3.338 becomes:
lim 0 r = + ln ( r )
0
Equation 3.355
and the non-dimensional mass of a circle becomes:
m"
lim CV = lim
0
0
2
B
8
1
= 2 ln ( r ) +
2
n =1 n 4 n 1
and the amplitude ratio in this deep water case becomes:
d AV
= B
lim
0
d
The hydrodynamic mass for 0 in shallow water remains finite. Because the multipotentials - just as in deep water provide finite contributions, the radiation potential has to
127
be discussed only, which is decisive (infinite mass) in deep water. The change-over borderline
''deep to shallow'' water provides for this radiation potential:
lim (0 r + 0 rad ) = + ln ( r ) ln ( h )
0
r
= + ln
h
Equation 3.356
It is obvious that Equation 3.356 - just as Equation 3.355 - provides an infinite value.
When the contributions of the multi-potentials (which disappear here for the borderline case
0 ) are ignored, it follows from Equation 3.327 for the amplitude ratio in shallow water:
AV =
z
cosh 2 [0 h]
= A0
g z 0 h + sinh [0 h] cosh [0 h]
cosh 2 [0 h]
= A0
z 0 h + sinh [0 h ] cosh[0 h]
A0
cosh 2 [0 h ]
0 h + sinh [0 h] cosh [0 h ]
V
sinh [0 h ] cosh[0 h]
'
= A0 0
0 h + sinh [0 h] cosh [0 h ]
=
Because:
A '
=1
lim 0
0
B
follows:
sinh [0 h] cosh[0 h ]
lim AV = B 0
0
0 h + sinh [0 h] cosh [0 h ]
and:
d AV
d A d
= lim V 0
lim
0
d 0 0 d0 d
=
B
1
lim
2 0 0 0 h + sinh [0 h ] cosh [0 h]
=
Thus, AV ( B 2) has at B 2 = 0 a vertical tangent.
The fact that the hydrodynamic mass goes to infinity for zero frequency can be explained
physically as follows. The smaller the frequency becomes, the longer becomes the radiated
wave and the faster travels it away from the cylinder. In the borderline case 0 has the
wave an infinite length and it travels away - just as the pressure (incompressible fluid) - with
an infinite velocity. This means that all fluid particles are in phase with the motions of the
body. This means that the hydrodynamic force is in phase with the motion of the body, which
holds too that:
128
'
F Vi
= arctan F Vj ' = 0
HT = = arctan
F Vr
F Vr
'
This condition is fulfilled only when F Vj = 0 or F Vj ' = m" = .
'
F Vj
NV
g2
AV 2
2
=
=
=
AV = 2
V 4
Because lim AV = B follows lim F Vj ' = B 2 . The term F Vj ' = m" has to be infinite.
0 0
0 0
The finite 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 in an incompressible fluid) not travel with an
infinite velocity; its maximum velocity is g h . 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:
F Vi '
0
HT = arctan
'
F Vr
So, F Vr ' = m" has to be finite.
3.3.3.6 Wave Loads
The wave forces FE on the restrained body in waves consist of:
forces F1 in the undisturbed incoming waves (Froude-Krylov hypothesis) and
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.
Thus:
FE = F1 + F2 + j F3
These forces will be determined from the undisturbed wave potential W and the disturbance
potential S . As mentioned before, for 900 only an approximation will be found.
FE = F1 + F2 + j F3
= ( W + S ) dy
S
= e j t
j xcos
e 0
n =0 S
)}
Equation 3.357
129
In here:
2 j ( t 0 xcos )
F1 =
e
y sin )} dy
E1 =
E2 =
F2
g B
=
E3 =
Anr ' nr Anj ' nj dy
B n= 0 S
F3
g B
Anr ' nj + Anj ' nr dy
B n = 0 S
Equation 3.358
In case of = 900 , so beam waves, the theory of Haskind-Newman see Haskind [1957] or
Newman [1962] - can be used too to determine the amplitudes E1 , E 2 and E 3 .
When W = W e j t is the potential of the incoming wave and S = S e j t is the potential
of the disturbance by the body at a large distance from the body with velocity amplitude
V = 1 , then:
h
FE = e j t W
W dz
y
y
0
Equation 3.359
According to Equation 3.3105 in Appendix I is:
W =
{cosh [0 z ] tanh [0 h] sinh [0 z ]} cos(0 y )
From the previous subsections follows the asymptotic expression for the disturbance potential
in still water with V = 1 :
130
cosh 2 [0 h ]
0 h + sinh [0 h ] cosh [0 h]
{cosh [0 z ] tanh [0 h] sinh [0 z ]} sin (0 y )
( y ) =
) {(
'
'
2
'
'
2 ( n 1 )
2
A0 r + j A0 j 0 Anr + j Anj 0
n =1
0 h + sinh [0 h ] cosh [0 h]
h
) {(
) {(
'
'
2
2
A
+
j
n =1
j t
= g e
'
'
2
'
'
2 (n 1)
2
A0 r + j A0 j 0 Anr + j Anj 0
n =1
Non-dimensional:
E1 + E 2 =
{ }
Re F E
g B
) {
) {
'
2
'
2 ( n 1 )
2
= A0 r 0 Anr 0
B
n =1
E3 =
{ }
Im F E
g B
'
A0 j 0 2 2 Anj ' 0 2 (n 1 )
B
n =1
Equation 3.360
3.3.3.7 Solution
The Lewis transformation of a cross section is given by:
y + i z = e + i + a e i + b e i 3
Equation 3.361
Then, the co-ordinates of the cross section are:
y = (1 + a ) cos( ) + b cos(3 )
z = (1 a ) sin ( ) b sin (3 )
Equation 3.362
Then:
131
z + i y = i ( y i z ) = i e i + a e + i + b e +i 3
)
Equation 3.363
All calculations will be carried out in the Lewis domain. Scale factors are given in the table
below.
Breadth
Draught
Water depth
Wave number
Acceleration
Forces
Ship
Lewis form
Ship
Lewis form
BR
TI
HT TI
0
g
FG
2 (1 + a + b)
1 a + b
WT = HT (1 a + b )
WF = 0 TI / (1 a + b )
g
F
BR {2 (1 + a + b )}
TI (1 a + b )
TI (1 a + b )
(1 a + b ) TI
1
{TI (1 a + b )}2
Equation 3.364
To determine the unknowns An , an equal number of equations has to be formulated. Because
only Lewis forms are used here, a simple approach is possible.
All stream function parts and boundary conditions can be given as a Fourier series:
n = {c mn sin (2 m ) + d mn cos[(2m + 1) ]}
m= 0
or with:
2
2 2
(
2m + 1)
cos ((2 m + 1) ) = 1
+
sin
(
2
k
k =1 k 4k 2 (2m + 1)
in:
n = an 0 + {a nm sin (2m )}
m =1
'
The solution of the by equating coefficients generated equations provide the unknowns An .
132
= vn body
n body
also in still water two boundary conditions:
1. For sway:
= vn
n body
=
body
dy dz
z ds y ds
d
ds
body
body
U e j t
dz
ds
body
or:
d
body
= U e j t dz
body
body
+C
Equation 3.365
2. For roll:
= vn
body
=
=
body
dy dz
z ds y ds
d
ds
body
body
= e j t
dr
r
ds
body
or:
d
body
= e j t r dr
body
body
( y, z , t ) = e j t
y2 + z2
body
+C
Equation 3.366
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 I), respectively:
j t
S body ( y , z , t ) =
e
body
( x, y1 , z1 , t ) = e j t sin (0 x cos )
0 1 sin 2
y1
sin (
Equation 3.368
3.3.4.2 Potentials
3.3.4.2.1 2-D Radiation Potential
In a similar way as Equation 3.322 for heave, the three-dimensional radiation potential for
sway and roll can be derived as:
0 r ( x, y , z , t ) = e j t
+L 2
L 2
A0 ( ) cos (k x ( x ))
k x 2 + k y 2 cosh k x 2 + k y 2 ( z h)
0 cosh k 2 + k 2 h k 2 + k 2 sinh k 2 + k 2 h
y
x
y
y
x
sin (k y y ) dk x dk y d
Equation 3.369
This expression reduces for the two-dimensional case ( k x = 0 and k y = k ) into:
134
0 r ( y, z , t ) = e j t A0
k cosh [k ( z h )]
Equation 3.370
With the Sommerfeld radiation condition in Equation 3.36 and Appendix III, the total
radiation potential becomes:
0 r ( y, z , t ) + j 0 j ( y , z , t ) = e j t A0
k cosh[k (z h )]
sin (k y ) dk +
0 cosh [k h] k sinh [k h ]
0 h + sinh [0 h] cosh [0 h ]
Equation 3.371
and the stream function becomes:
0 r ( y, z , t ) + j 0 j ( y , z , t ) = e j t A0
k sinh [k ( z h )]
cos(k y ) dk +
0 cosh [k h] k sinh [k h ]
0 h + sinh [0 h] cosh [0 h ]
Equation 3.372
Potential and stream function are divided in:
0 = 0 r + 0 rad + j 0 j = e j t A0
k e k z
sin (k y ) dk +
0 k
k e k h sinh [k z ] k cosh[k z ]
sin (k y ) dk +
0 k cosh [k h] k sinh [k h ]
cosh[0 ( z h )] cosh [0 h]
j 0
sin (0 y )
0 h + sinh [ 0 h] cosh[ 0 h ]
Equation 3.373
135
0 = 0 r + 0 rad + j 0 j = e j t A0
k e k z
cos (k y ) dk +
0 k
k e k h cosh[k z ] k sinh [k z ]
cos
(
k
y
)
dk
+
0 k cosh [k h ] k sinh [k h]
sinh [0 (z h )] cosh [ 0 h]
j 0
cos(0 y )
0 h + sinh [0 h] cosh [ 0 h]
Equation 3.374
nr ( y , z , t ) = e j t An (k + ) k 2 n 1 e k z sin (k y ) dk
0
nrad ( y , z , t ) = e j t An (k + ) k 2 n 1 e k h
0
sinh [k z ] k cosh[k z ]
sin (k y ) dk
cosh [k h ] k sinh [k h]
nj ( y , z , t ) = + e j t An
0 2 n +1
cosh [0 ( z h )]
sin (0 y )
cosh[0 h] 0 h + sinh [0 h] cosh [0 h ]
Equation 3.375
The related stream function is:
nr ( y , z , t ) = + e
j t
An (k + ) k 2 n 1 e k z cos( k y ) dk
0
nrad ( y , z , t ) = e j t An (k + ) k 2 n 1 e k h
0
cosh[k z ] k sinh [k z ]
cos (k y ) dk
cosh [k h ] k sinh [k h ]
nj ( y , z , t ) = + e j t An
0 2 n +1
sinh [0 ( z h )]
cos(0 y )
cosh [0 h] 0 h + sinh [0 h ] cosh[0 h]
Equation 3.376
136
k sin {k ( y + i ( z h))}
dk +
0 cosh [k h ] k sinh [k h]
0 h + sinh [0 h] cosh[0 h ]
e j t
( Anr + j Anj )
sin
{
k
(
y
+
i
(
z
h
)
)
}
2
2
2
n
+
1
dk
k k
cosh[k h] k sinh [k h]
n =1
2 n +1
sin {0 ( y + i ( z h))}
j 0
Equation 3.377
Writing this in a similar way as for heave provides:
A (
nr nr + nrad ) Anj nj
( y , z , t ) = +e j t
n =0
j {Anj (nr + nrad ) + Anr nj }
Equation 3.378
and
A (
)
nr
nr + nrad Anj nj
( y , z , t ) = +e j t
n= 0
+ j {Anj (nr + nrad ) + Anr nj }
Equation 3.379
with:
Ar + j Aj = U Ar ' + j Aj '
Ar + j Aj = Ar ' + j Aj '
for sway
for roll
An has dimension L2 n + 2 .
'
[ ]
An n and An n have for sway dimension [L] and for roll dimension L2 .
'
'
'
The determination of the coefficients An follow from the boundary conditions at the body
contour.
3.3.4.3 Expansion of Potential Parts
For r = y 2 + z 2 0 :
137
0 r
0 r
+
ln
(
r
)
+
Re ( z + i y )m sin ( y )
m =1 m m!
y
= 2
+ e z
y +z
m
m
arctan y +
Im ( z + i y ) cos( y )
z
m
m
!
m =1
z
=+ 2
e z
y + z2
m
m
(
)
+
ln
r
+
Re ( z + i y ) cos ( y ) +
m =1 m m!
m
arctan y + Im ( z + i y ) m sin ( y )
z m =1 m m!
Equation 3.380
with the Euler constant: = 0.57722 .
For r = y 2 + z 2 :
0 r =
0 r = +
M
y
y
(m 1)!
m
z
+
m 2 m Im (z + i y ) e cos ( y )
2
2
y +z
y
m =1 r
M
z
y
(m 1)!
m
z
m 2 m Re ( z + i y ) e sin ( y )
2
2
y +z
y
m =1 r
Equation 3.381
nr = ( 1)n +1 (2n )! Re ( y + i z ) (2 n +1 ) +
Im ( y + i z ) 2 n
2n
nr = ( 1)n +1 (2n )! Im ( y + i z ) (2 n +1 ) Re ( y + i z ) 2 n
2n
Equation 3.382
G(2t + 3)
G (2t + 1)
0 rad =
Im ( z + i y )2 t +1
Im (z + i y )2 t
(2t )!
t = 0 (2t + 1)!
G(2t + 3)
G(2t + 1)
2 t +1
2t
0 rad =
Re ( z + i y )
Re ( z + i y )
(2t )!
t = 0 (2t + 1)!
Equation 3.383
nrad
nrad
G( 2t + 2n + 3) 2 G(2t + 2n + 1)
+
Im ( z + i y )2 t +1
(
2
t
+
1
)
!
=
G (2t + 2n + 1) 2 G(2t + 2n 1)
t =0
Im (z + i y )2 t
(2t )!
G( 2t + 2n + 3) 2 G(2t + 2n + 1)
+
Re ( z + i y )2 t +1
(2t + 1)!
=
G (2t + 2n + 1) 2 G(2t + 2n 1)
t =0
Re ( z + i y )2 t
(2t )!
Equation 3.384
138
oj = + 0
cosh 2 [0 h]
0 h + sinh [0 h] cosh [0 h]
0 2 t +1
Im ( z + i y )2 t +1
t = 0 (2t + 1)!
2
t
tanh [ h] 0 Im (z + i y )2 t
0
t = 0 (2t )!
cosh 2 [0 h]
oj = + 0
0 h + sinh [0 h] cosh [0 h]
0 2 t +1
Re ( z + i y )2 t +1
t = 0 (2t + 1)!
2t
tanh [ h] 0 Re ( z + i y )2 t
0
t = 0 (2t )!
(
)
(
Equation 3.385
)
)
nj = 0 2 (n 1) 0 2 2 0 j
nj = 0 2 (n 1
0j
Equation 3.386
+ i = An
(2 n +1 )
(2 n +1 )
+ ( y + i z i 2mh)
n= 0
( y + i z + i 2mh)
m =1
Equation 3.387
For Lewis forms this becomes:
139
i ( 2 n +1 )
p
p 2n + p
a e i 2 + b e i 4 +
( 1)
e
p =0
p
p 2 p + 2n + 1
+ i = An
( 1)
n =0
2 p +1
(2 n +1)
(i 2m H )
2
p
+
1
i 2
+i
m =1
p = 0 e
+ a e i + b e i 3
2
m
2n + p p p l l i (2 n + 2 p +2 l +1)
p
a b e
( 1)
+
p l
l = 0
p = 0
p +n
2 ( p + n +1 )
= An
( 1) (2m H )
n =0
2 p +1 l
2n + 2 p + 1 2 p + 1 l l k k i (2 p 2 l 2 k +1)
2
a b e
m =1 p = 0
2
p
+
1
l
l
=
0
k
=
0
Equation 3.388
These sums converge as long as:
e + i + a e i + b e i 3 e + i + a e i + b e i 3
=
<1
2m H
2 m HT (1 a + b )
Equation 3.389
Because m 1 , it follows the condition:
!
1 + a e i 2 + b e i 4
2 HT
1 a + b
Equation 3.390
The potential converges too when:
1+ a + b
B
2 HT
=
or
1 a + b 2T
B 4 h
Equation 3.391
{A
n =0 S
nr
'
)}
140
FR = FRr + j FRj
= 2 e j t
{A
n =0 S
)}
'
nr
Equation 3.393
The hydrodynamic moment at sway oscillations in still water becomes:
M Q = M Qr + j M Qj
= U e j t
{A
n =0 S
'
nr
)}
{A
n= 0 S
'
nr
)}
'
Of course, the coefficients Anr and Anj of sway and roll will differ.
Fictive moment levers are defined by:
H Qr =
H Qj =
M Qj
U NQ
H Rr
H Rj =
N R B
2 F Rj
M Qr
U m"
I " 2
=
F Rr
T
U T 2
2
2
AQ =
2
2
cosh 2 [0 h ]
F Qj
U 0 h + sinh [0 h] cosh[0 h ]
H Qr
M Qr
=
T
T F Qr
H Qj
M Qj
=
T
T F Qj
Equation 3.396
and for roll motions:
141
CR =
I"
M Rr
=
T 4 2 T 4
8
8
AR =
2
B2
4
H Rr
M Rr
=
T
T F Rr
H Rj
M Rj
=
T
T F Rj
2
cosh 2 [ 0 h ]
4 2
M Rj
2 B 2 0 h + sinh [ 0 h ] cosh[0 h]
Equation 3.397
= e j t
j 0 x cos
(
)
(
)
cosh
[
z
]
tanh
[
h
]
sinh
[
z
]
sin
sin
dz
0
0
0
0
S
n =0
)}
Equation 3.398
M E = M1 + M 2 + j M 3
= e j t
j 0 xcos
(
cosh
[
z
]
tanh
[
h
]
sinh
[
z
]
)
sin
(
sin
)
( y dy + z dz )
0
0
0
0
S
n =0
)}
Equation 3.399
The separate parts are:
142
2 j ( t 0 xcos )
e
F1 = +
S
2 j ( t 0 xcos )
e
M1 = +
S
Dimensionless:
F1
g 0 Ax
tanh [ 0 h ]
=+
Ax
E1 =
y sin ) dz
E2 =
F2
g 0 Ax
E3 =
tanh [ 0 h]
'
'
Anr nr Anj nj dz
Ax
n= 0 S
F3
g 0 Ax
tanh [ 0 h]
Anr ' nj + Anj ' nr dz
Ax
n= 0 S
H Wr
M1 + M2
=
T
T F1 + F 2
H Wj
M3
=
T
T F3
Equation 3.3100
The Haskind-Newman relations see Newman [1962] - are valid too here:
143
E1 + E 2 =
{ }
Re F E
g 0 Ax
(
{ }
) {
) {
'
2
'
2 (n 1)
A0 r + 0 2 Anr 0
Ax
n =1
Im F E
E3 =
g 0 Ax
'
2
'
2 ( n 1 )
A0 j + 0 2 Anj 0
Ax
n =1
Equation 3.3101
3.3.4.7 Solution
To determine the unknowns An , an equal number of equations have 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:
or with
cos (2m ) = 1 +
16
m2
sin ((2k + 1) )
k = 0 (2k + 2m + 1) (2 k 2m + 1) (2 k + 1)
in:
n = {a nm sin ((2 m + 1) )}
m= 0
'
The solution of the by equating coefficients generated equations provide the unknowns An .
3.3.5 Appendices
3.3.5.1 Appendix I: Undisturbed Wave Potential
The general expression of the complex potential of a shallow water wave, travelling in the
negative y -direction, is:
cos{0 ( y + i ( z h )) + t }
W + i W = c
with : c =
sinh [ 0 h]
0
=
=
cosh[0 (z h )] cos(0 y + t )
cosh[0 (z h )] cos(0 y + t )
Equation 3.3102
144
W = +
cosh [0 ( z h)] cos(0 y + t )
cosh[0 h]
e j t cosh[0 ( z h )] e j 0 y
cosh [0 h]
j t
e
{cosh[0 z ] tanh [0 h ] sinh [0 z ]} e j 0 y
W =
sinh [0 ( z h)] sin (0 y + t )
cosh[0 h ]
= j
e j t sinh [0 ( z h )] e j 0 y
cosh[0 h]
= j
j t
e
{sinh [0 z ] tanh [0 h] cosh [0 z ]} e j 0 y
For the vertical motions is the in y -symmetrical part of the potential significant. For the
horizontal motions is the in y -point-symmetrical part (multiplied with j , so a phase shift of
900 ) important.
j t
e
{cosh[0 z ] tanh [0 h ] sinh [0 z ]} cos(0 y )
j t
=
e
{sinh [0 z ] tanh [0 h ] cosh [0 z ]} sin (0 y )
W V = +
W V
j t
e
{cosh [0 z ] tanh [0 h] sinh [0 z ]} sin (0 y )
j t
=
e
{sinh [0 z ] tanh [0 h ] cosh[0 z ]} cos(0 y )
W H =
W H
Equation 3.3103
When the wave travels in the x w -direction, the potential becomes:
W =
j ( t 0 xw )
e
{cosh [ 0 z ] tanh [ 0 h] sinh [ 0 z ]}
With:
x w = x cos y sin
y w = x sin + y cos
Equation 3.3104
the potential becomes:
j t j 0 ( y sin x cos )
W V =
e
e
{cosh[0 z ] tanh [0 h ] sinh [0 z ]}
Equation 3.3105
This results for the vertical motions in:
145
W V = +
j ( t 0 x cos )
e
{cosh[0 z ] tanh [0 h ] sinh [0 z ]} cos(0 y sin )
Equation 3.3106
Equation 3.3107
y + z e
2
y
i arctan
z
= r e i
the following series expansions can be found.
cosh [k z ] cos(k y ) = cosh [k z ] cosh[i k y ]
1
= {cosh[k (z + i y )] + cosh[k ( z i y )]}
2
1
= cosh k r e+ i + cosh k r e i
2
2t
1 (k r )
(k r )2t e i 2t
=
e + i 2t +
2 t = 0 (2t )!
t = 0 (2t )!
(k r )2 t
(2t )!
t =0
]}
cos(2t )
k 2t
2t
Re ( z + i y )
t =0 (2t )!
1 (k r )
(k r )2t +1 e i (2t +1)
=
e + i ( 2 t +1 ) +
2 t = 0 (2t + 1)!
t = 0 (2t + 1)!
{ [
]}
2 t +1
(
k r)
=
cos((2t + 1) )
t = 0 (2t + 1)!
k 2 t +1
2 t +1
Re ( z + i y )
t = 0 (2t + 1)!
146
i (k r )
(
k r)
+ i 2t
=
e
e i 2 t
2 t = 0 (2t )!
t = 0 (2t )!
]}
2t
(
k r)
=
sin (2t )
(2t )!
t =0
k 2t
2t
=
Im ( z + i y )
t = 0 (2t )!
i (k r )
(
k r)
+ i (2 t +1 )
=
e
e i ( 2 t +1)
2 t = 0 (2t + 1)!
t = 0 (2t + 1)!
{ [
]}
2 t +1
(
k r)
=
sin ((2t + 1) )
t = 0 (2t + 1)!
k 2 t +1
2 t +1
=
Im (z + i y )
t = 0 (2t + 1)!
147
0 r ( y, z , t ) + j 0 j ( y , z , t ) = e j t A0
cosh[k (z h )]
lim
cos(k y ) dk
0
0 j cosh [k h ] k sinh [k h ]
g
= e j t A0
lim
0
j
sinh [k z ] k cosh[k z ]
k h
g
e
cos(k y ) dk
j
k j cosh [k h ] k sinh [k h]
g
g
ek z
cos(k y ) dk +
j
k
g
e k h
= lim
0
0 j k
j
sinh [k z ] k cosh [k z ]
cos (k y ) dk
j
cosh[k h] k sinh [k h ]
Re ( z + i y )2 t +1 Re (z + i y )2 t
g
(2t + 1)!
(2t )!
h
2
t
+
1
= lim
k
0
dk
t = 0
0 j k j cosh[k h] k sinh [k h]
g
g
Re ( z + i y )2 t +1 Re ( z + i y )2 t
j
= lim
g
(
2t + 1)!
(
2t )!
0
t =0
148
e k h
k i
G(t ) + i H (t ) = lim
0
kt
0
dk
cosh [k h] k sinh [k h ]
i
will be normalised as done before:
G ' (t ) + i H ' (t ) = h t 1 (G(t ) + i H (t )) =
t
u
u e
= lim
du
0
h
0
h
cosh [u ] u sinh [u ]
h i
This is a complex integral and must be solved in the complex domain with w = u + i v .
The integrand has a singularity for:
h
w1 = h i
g
and w2 is the solution of the equation:
h
h i
cosh[w] w sinh [w] = 0
g
149
e w wt
dw
=
lim
...du + ...dw + ...dw
0 (w w ) (w cosh[w ] w sinh [w ])
0
1
1
0
I
II
lim
= lim ...du +J I + J II
0
0
=0
Also:
t
u
u e
G ' (t ) + i H ' (t ) = lim
du
0
h
0
u h i
h
h i
cosh[u ] u sinh [u ]
lim
0 and R
{J I + J II }
Because:
lim J I = 0
R
follows:
wt e w
G' (t ) + i H ' (t ) = lim
dw
0
wt e w
G' (t ) = Re lim
dw
0 II (w w1 ) (w1 cosh [w] w sinh [w])
wt e w
'
H (t ) = Im lim
dw
0 II (w w1 ) (w1 cosh [w] w sinh [w])
150
wt e w
dw
0 (w w ) (w cosh [w ] w sinh [w ])
1
1
= lim {J II + J III + J IV }
lim
J IV disappears for R .
It can be found that:
Re {J II } = Re{J III }
Then it follows:
H ' (t ) = Im lim J II
0
Im {J II } = + Im{J III }
and
= ( 0 h )
t 1
cosh 2 [ 0 h ]
tanh t 1 [0 h]
0 h + sinh [0 h] cosh [0 h]
Re (z + i y )
0 jad = H (2t + 1)
(2t )!
(2t + 1)!
t =0
2t
0 cosh 2 [0 h]
2 t
[0 h] + sinh [0 h] cosh [0 h]
=
2t
2 t +1
t = 0 Re ( z + i y )
Re
(
z
+
i
y
)
(2t )!
(2t + 1)!
cosh 2 [0 h ]
[0 h] + sinh [0 h] cosh [0 h]
2 t +1
0 2 t
0
2t
2 t +1
(
)
[
]
Re ( z + i y )
t =0 (2t )! Re z + i y tanh 0 h
t = 0 (2t + 1)!
e z cos( y )
The same will be found as a difference between Equation 3.345 and Equation 3.329.
151
152
(m )
( m)
J ij
153
( m)
N
N ( )
ni
(m )
(m )
PV
(m)
pa
pv
Qj
(m )
(m )
Q j +N
(m )
sj
T
t
(m )
vi
= 2 g
k
j
wave number
k -th irregular wave number for adjoint interior problem
abscissa of the j -th input point
density of fluid
(m )
154
( m ) ( x, y, t ) = Re (m ) (x , y ) e i t
Equation 3.41
satisfying the following conditions:
1. The Laplace equation:
2 ( m ) 2 (m )
2 (m) =
+
=0
x 2
y 2
Equation 3.42
in the fluid domain, i.e., for y < 0 and outside the cylinder.
2. The free surface condition:
2 (m )
( m )
+
g
=0
t 2
y
Equation 3.43
on the free surface y = 0 outside the cylinder, while g is the acceleration of gravity.
3. The seabed boundary condition for deep water:
155
(m )
=0
y
Equation 3.44
4. The condition of the normal velocity component of the fluid at the surface of the
oscillating cylinder being equal to the normal component of the forced velocity of the
cylinder. i.e., if v n is the component of the forced velocity of the cylinder in the direction
of the outgoing unit normal vector n , then
n ( m ) = v n
Equation 3.45
This is the kinematic boundary condition on the oscillating body surface, being satisfied at
the mean (rest) position of the cylindrical surface.
5. The radiation condition that the disturbed surface of the fluid takes the form of regular
progressive outgoing gravity waves at large distances from the cylinder.
According to Wehausen and Laitone [1960], the complex potential at z of a pulsating point
source of unit strength at the point in the lower half plane is:
i k (z )
1
e
G* ( z, , t ) =
ln ( z ) ln z + 2 PV
dk cos( t )
2
k
0
e i ( z ) sin ( t )
Equation 3.46
so that the real point-source potential is:
H (x , y, ,, t ) = Re G* (z , , t )
Equation 3.47
where:
z = x +i y
Letting:
G( z , ) =
= + i
= i
= 2 g
i k ( z )
1
e
Re ln (z ) ln z + 2 PV
dk
2
0 k
i Re e i (z )
Equation 3.48
then:
H (x , y, ,, t ) = Re G (z , , t ) e i t
}
Equation 3.49
Equation 3.49 satisfies the radiation condition and also Equation 3.41 through Equation
3.44.
Another expression satisfying all these conditions is:
156
i t
H x, y, ,, t
= Re i G( z , ) e
2
}
Equation 3.410
Since the problem is linear, a superposition of Equation 3.49 and Equation 3.410 results in
the velocity potential:
( m ) ( x, y , t ) = Re Q( s ) G( z , ) e i t ds
C0
Equation 3.411
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:
Re n Q(s ) G( z, ) ds = 0
C0
Im n Q(s ) G( z, ) ds = A( m ) n (m )
C0
( )
( )
Equation 3.412
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 Equation 3.412 represent a set of two coupled
integral equations for the real functions Re {Q(s )} and Im{Q( s )} . The solution of these integral
equations and the evaluation of the kernel and potential integrals are described in the
following section and in Appendices II and III, respectively.
157
(i +1 i )2 + (i +1 i )2
Equation 3.414
while the angle made by the i -th segment with the positive x -axis is given by:
i
i = arctan i +1
i +1 i
Equation 3.415
The outgoing unit vector normal to the cross section at the i -th midpoint ( xi , yi ) is:
n i = i sin i j cos i
Equation 3.416
where i and j are unit vectors in the directions of increasing x and y , respectively.
158
The cylinder is forced into simple harmonic motion with radian frequency , according to the
displacement equation:
S ( m ) = A(m ) cos ( t )
Equation 3.417
for m = 2,3, 4 corresponding to sway, heave or roll, respectively.
The rolling oscillations are about an axis through a point (0, y0 ) in the symmetry plane of the
cylinder.
In the translation modes, any point on the cylinder moves with the velocity:
(2)
sway :
= i A( 2 ) sin ( t )
Equation 3.418
heave :
(3 )
= j A(3 ) sin ( t )
Equation 3.419
Ri = xi + ( yi y0 )
2
y y0
i = arctan i
x
y y0
= arcsin i
Ri
and
x
= arccos i
Ri
Therefore, by elementary two-dimensional kinematics, the unit vector in the direction of
increasing is:
i = i sin i + j cosi
y y0
x
= i
i + i j
Ri
Ri
so that:
roll :
(4 )
= Ri S (4 ) i
= A( 4 ) ( yi y0 ) i xi j sin ( t )
Equation 3.420
The normal components of the velocity vi
( xi , yi ) are:
sway :
heave :
roll :
( m)
= ni v
(m )
vi (3 ) = + A( 3 ) cos i sin ( t )
vi ( 4 ) = + A(4 ) {( yi y0 ) sin i + xi cosi } sin ( t )
Equation 3.421
Defining:
159
vi ( m )
A( m ) sin ( t )
then - consistent with the previously mentioned notation - the direction cosines for the three
modes of motion are:
sway : ni ( 2 ) = sin i
ni ( m ) =
heave :
roll :
ni (3 ) = + cosi
ni ( 4 ) = +( yi y0 ) sin i + xi cos i
Equation 3.422
(2 )
(3)
(3 )
= n+ i . Swaying and
( 4)
(4 )
+ Qj
j =1
N
(m )
I ij
(m)
}+ {Q
N
N+ j
j =1
(m )
J ij
(m )
}= 0
Q j ( m ) J ij ( m ) + QN + j (m ) I ij (m ) = A( m ) ni ( m )
j =1
j =1
Equation 3.423
where the superscript
(m )
( m)
The hydrodynamic pressure at ( xi , yi ) along the cylinder is obtained from the velocity
potential by means of the linearized Bernoulli equation:
(m )
p ( m ) ( xi , yi ,, t ) =
( xi , yi , , t )
t
Equation 3.424
as:
p ( m ) ( xi , yi , , t ) = p a
( m)
( xi , yi , ) cos ( t ) + pv (m ) ( xi , yi ,) sin ( t )
Equation 3.425
160
(m )
where p a and pv are the hydrodynamic pressures in-phase with the displacement and inphase with the velocity, respectively and denotes the density of the fluid.
As indicated by the notation of Equation 3.424 and Equation 3.425, 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
M (m ) () = 2 p a ( m ) ( xi , yi , ) ni ( m ) s i
i =1
N
N (m ) () = 2 p v ( m ) ( xi , yi , ) ni ( m ) s i
i =1
Equation 3.426
for the added mass and damping forces or moments, respectively.
The velocity potentials for very small and very large frequencies are derived and discussed in
the next section.
=0
y
Equation 3.427
on the surface of the fluid outside the cylinder, whereas for extremely large frequencies, i.e.,
when , the free-surface condition becomes the ''impulsive'' surface condition:
=0
Equation 3.428
on y = 0 and outside the cylinder.
Equation 3.42, Equation 3.44 and Equation 3.45 remain valid for both asymptotic cases.
The radiation condition is replaced by a condition of boundedness at infinity.
Therefore, there is a Neumann problem for the case 0 and a mixed problem when
. The appropriate complex potentials for a source of unit strength at a point in the
lower half plane are:
1
G0 ( z , ) =
ln ( z ) ln z + K0
2
)}
Equation 3.429
and:
G (z , ) =
)}
1
ln ( z ) ln z + K
2
Equation 3.430
161
for the Neuman and mixed problems, respectively, where K0 and K are not yet specified
constants.
Let:
a ( x, y ,,) = Re{Ga ( z , )}
so that the velocity potentials for the m -th mode of motion are:
(m )
( m)
a ( x, y ) = Qa (s ) a (x , y, ,) ds
C0
Equation 3.431
(m )
for a = 0 , and a = , where Qa is the expression for the source strength as a function of
position along the submerged contour of the cross section C0 .
An analysis similar to the one in the section on formulating the problem leads to the integral
equation:
( m)
n Qa (s ) a (x , y, ,) ds = A( m ) n (m )
( )
C0
Equation 3.432
which - after application at the N segmental midpoint - yields a set of N linear algebraic
equations in the N unknown source strengths Q j .
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 difficulty, since as z , G (z , ) 0 . In fact K = 0 , which
can be inferred from the pulsating source-potential Equation 3.48 by letting .
Considering the Neumann problem, note that the constant K0 in the Green's function equation
(Equation 3.429) 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 closed.
But since n i
( m)
K n
= n+ i
(m )
(m )
ds = 0
C0
K n
(3 )
ds 0
C0
162
solutions were subject to the provisions that no point of the immersed surface of the body
would be outside a cylinder drawn vertically downward from the intersection of the body with
the free surface and that the free surface would be intersected orthogonally by the body in its
mean or rest position.
John [1950] also showed that for a set of discrete ''irregular'' frequencies the Green's functionintegral equation method failed to give a solution. He demonstrated that the irregular
frequencies occurred when the following adjoint interior-potential problem had eigenfrequencies.
Let (x, y ) be such that:
1.
2.
2 2
+
= 0 inside the cylinder in the region bounded by the immersed surface of the
x 2 y 2
body and the extension of the free surface inside the cylinder.
k = 0 on the extension of the free surface inside the cylinder, k being the wave
y
number corresponding to the irregular frequency k , k = 1,2,3,etc.
T
1 = coth
B
B
Equation 3.434
Keeping T fixed in Equation 3.434 but letting B vary and setting b = B , then from the
Taylor expansion:
3
1
b T (b T )
b coth[b T ] = b
+
+ ......
3
45
b T
1 1 T
Equation 3.435
is 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 : 1 = 1.48 , while for B T = 2.0 : 1 = 1.71 .
At an irregular frequency the matrix of influence coefficients of Equation 3.423 becomes
singular as the number of defining points per cross section increases without limit, i,e., as
N . In practice, with finite 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 defining 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 first irregular frequency 1 is less for the midsection than for
any other cross section.
As an example, for a ship with a 7:1 length-to-beam and a 5:2 beam-to-draft, the first irregular
wave encounter frequency - in non-dimensional form with L denoting the ship length occurs at 1 L g 5.09 , which is beyond the range of practical interest for ship-motion
analysis.
Therefore, for slender surface vessels, the phenomenon of the first irregular frequency of
wave encounter is not too important.
Increasing the number of contour line elements (or panels in 3-D) does not remove the
irregular frequency, but tends to restrict the effects to a narrower band around it; see for
instance Huijsmans [1996]. It should be mentioned too that irregular frequencies appear for
free surface piercing bodies only; fully submerged bodies do not display these characteristics.
An effective method to reduce the effects of irregular frequencies is ''closing'' the body by
means of a discretisation of the free surface inside the body, i.e. putting a ''lid'' on the free
surface inside the body.
See here the computed added mass and damping of a hemisphere in Figure 3.42. The solid
line in this figure results from including the ''lid''.
20 00
1 25 0
su rge /sw a y
1 00 0
Damping (ton/s)
Mass (ton)
15 00
h ea ve
10 00
75 0
h ea ve
50 0
su rge /sw a y
5 00
25 0
0
0
F re qu e ncy (ra d/ s)
F re qu en cy (ra d /s)
3.4.6 Appendices
3.4.6.1 Appendix I: Evaluation of Principle Value Integrals
The real and imaginary parts of the principle value integral:
i k (z )
e
PV
dk
k
0
are used in evaluating some of the kernel and potential integrals.
The residue of the integrand at k = is e i ( z ) , so that:
i k (z )
i k ( z )
e
e
PV
dk =
dk i e i (z )
0 k
0 k
Equation 3.436
where the path of integration is the positive real axis indented into the upper half plane about
k = .
Notice hereby that:
= 2 g > 0
Im {z} < 0
Im { } 0
The transformation = i (k ) z converts the contour integral on the right hand side
of Equation 3.436 to:
i k ( z )
e
ew
i
(z )
dk
=
dw
0 k
w
i (z )
x > 0
i 2
= e i (z ) E1 i z +
for :
x < 0
0
+ ln i z +
i 2
x > 0
= e i ( z ) ( 1)n i z n +
for :
0
x
<
0
n n!
n =1
where = 0.57722 is the well-known Euler-Mascheroni constant.
The definition of E1 has been given by Abramovitz and Stegun [1964].
Setting:
Im i z
+
r = i z
and
= arctan
Re i z
the following expression is obtained for Equation 3.436:
))
)}
( )}
{
{
165
(
(
)}
)}
PV
0
e i k (z )
dk = e ( y + ) {cos ( ( x )) i sin ( ( x ))}
k
r n cos (n )
+
ln
(
r
)
+
n n!
n =1
x > 0 r n sin (n )
i
2 for : x < 0 +
n n!
n =1
Equation 3.437
Separating Equation 3.437 into its real and imaginary parts yields:
k( y+ )
e
cos(k ( x ))
C (r , ) cos( ( x ))
PV
dk = e ( y + )
k
+ S (r , ) sin ( ( x ))
0
C (r , ) sin ( ( x ))
e k ( y + ) sin (k ( x ))
dk = e ( y + )
k
S (r , ) cos( ( x ))
0
PV
Equation 3.438
provided that:
r n cos(n )
C (r , ) = + ln (r ) +
n n!
n =1
x > 0
r sin (n )
S (r , ) = +
for :
+
n n!
x < 0
n =1
2
1
ln ( z ) ln z +
2
ds ( 1)m
i k (z )
1
e
s + j PV
dk
0
(m )
I ij = Re n i
1 ln z + ln ( z + ) +
2
ds
i k (z + )
s j 1 PV e
dk
k
0
z = zi
{(
)}
Equation 3.439
and:
J ij
(m )
= Re ni
i (z )
m
i ( z + )
e
ds ( 1) e
ds
s
s j
+j
z = zi
Equation 3.440
166
dF ( z )
Re n i F (z ) z = z i = Re i ei i
dz z = z i
{(
Considering the term containing ln ( z ) , it is evident that the kernel integral is singular
when i = j , so that the indicated differentiation cannot be performed under the integral sign.
However, in that case one may proceed as follows.
Since:
= + i
and
d = d + i d
= ds cos j + i ds sin j
i
= e j ds
for along the j -th segment.
Therefore:
ds = e i j d
and:
Re n i
i e i j
ln ( z ) ds = Re d j+1 i j
s
dz e d ln ( z )
j
z= z j
z = z j
j
i
j+ 1
= Re d
ln (z ) d
dz
z = z j
j
z j
d
'
'
Re i
ln d
= arg (z j j ) arg (z j j + 1 )
dz z j +1
z =z j
( )
Equation 3.441
If i j , differentiation under the integral sign may be performed, so that:
( L1 ) = Re ni
ln ( z ) ds
s
j
z= zi
(x ) + ( y
(
)
= sin ln
(x ) + ( y
j +1
j )
j +1 )
y j
y j + 1
arctan i
+ cos(i + j ) arctan i
xi j +1
xi j
167
( L2 ) = Re ni
ln z ds
s
j
z = zi
(x )2 + ( y + )2
i
j
i
j
= sin (i + j ) ln
2
2
(xi j +1 ) + ( yi + j +1 )
y + j
y + j +1
arctan i
+ cos (i + j ) arctan i
j
j +1
i
i
Equation 3.443
The kernel integral, containing the principal value integrals, is:
( L5 ) = Re n i
i k (z )
ds PV
dk
s
k
0
j
z= zi
j
i k ( z i )
d
e
i ( i + j )
= Re i e
d
PV
dk
k
d
j +1
k ( yi + j )
e
cos (k (xi j ))
+ PV
dk
k
0
= sin (i + j )
(
y
+
dk
PV
k
0
k ( yi + j )
e
sin (k (x i j ))
dk
+ PV
k
0
cos(i + j )
(
y
+
e i j +1 sin (k (xi j +1 ))
dk
PV
k
0
Equation 3.444
The first integral on the right hand side of Equation 3.440 becomes:
168
( L7 ) = Re n i
i (z )
e
ds
s
j
z = zi
+ e ( yi + j ) cos ( (xi j ))
= sin (i + j ) ( y + )
e i j +1 cos ( (xi j +1 ))
+ e ( y i + j ) sin ( (xi j ))
+ cos(i + j ) ( y + )
e i j +1 sin ( (xi j + 1 ))
Equation 3.445
The kernel integrals over the image segments are obtained from Equation 3.443 through
Equation 3.445 by replacing j , j +1 and j with j = j , ( j +1) = j +1 and j = j ,
respectively.
3.4.6.3 Appendix III: Potential Integrals
The velocity potential of the m -th mode of oscillation at the i -th midpoint ( xi , yi ) is:
( m ) ( xi , yi , t ) =
i k (z i )
e
dk ds
ln ( zi ) ln zi + 2 PV
k
s j
0
1
m
Q j Re ( 1)
2 j =1
i k ( z i + )
ln z + ln (z + ) + 2 PV e
i
i
0 k dk ds
s
j
cos( t )
m
m QN + j Re e i (z i ) ds ( 1) e i ( z i + ) ds
j =1
s j
sin ( t )
sj
Equation 3.446
The integration of the ln ( zi ) term is straight forward, yielding:
169
Re ln ( z i ) ds = + cos j
s j
+ (x ) ln (x )2 + ( y )2 + ( )
i
j
i
j
i
j
j
j +1
2
2
(
)
(
)
(
)
ln
+
y
i
j +1
i
j +1
i
j +1
yi j
y i j +1
+ ( y i j +1 ) arctan
( y i j ) arctan
xi j +1
x
i
j
+ sin j
+ (y ) ln (x )2 + ( y )2 + ( )
i
j
i
j
i
j
j
j +1
2
2
(
y
ln
(
x
)
+
(
y
i
j +1
i
j +1
i
j +1
y i j
y j +1
(xi j +1 ) arctan i
+ (xi j ) arctan
x i j +1
x
i
j
Equation 3.447
In the integration of the ln z term, note that j and j +1 are replaced by j and j +1 ,
respectively.
To evaluate the potential integral containing the principal value integral, proceed in the
following manner.
For an arbitrary z in the fluid domain:
j +1
j +1
i k ( z )
e
dk
i j
i k (z )
d
ds PV 0 k dk = e PV 0 k e
=e
i j
e i k z
PV
dk
k
0
j +1
i k
i k
e i k z e j +1 ei k j
dk
k
k
0
where the change of integration is permissible since only one integral requires a principle
value interpretation.
i j
= e
PV
After dividing by and multiplying by k + k under the integral sign, the last expression
becomes:
i k
i
i k (z j+1 )
i k (z j )
i e j i k z e j+1 ei k j
e
e
dk + PV
dk PV
dk
k
k
k
0
0
0
Equation 3.448
Regarding the first integral in Equation 3.448 as a function of z :
170
F(z ) = e
i k z
i k
j+1
e i k j
dk
k
Equation 3.449
i k ( z j +1 )
i k ( z j )
F ' ( z ) = i e
dk e
dk
0
0
1
1
=
z 1 z j +1
So:
F ( z ) = ln z j ln z j +1 + K
) (
Equation 3.450
where K is a constant of integration to be determined presently. Since F (z ) is defined and
analytic for all z in the lower half plane and since by Equation 3.449, lim {F ( z )} = 0 , it
z
( K5 ) = Re ds PV e
dk
k
s j
ln z i j ln z i j +1
i e i j
(
z
(
z
i
i
= Re
j +1
j
e
e
+ PV
dk PV
dk
0
0
+ sin j
(xi j )2 + ( yi + j )2
ln
2
2
(xi j +1 ) + ( yi + j +1 )
k ( yi + j )
e k ( yi + j+1 ) cos (k (x i j +1 ))
e
cos(k (xi j ))
PV
dk PV
dk
1
k
k
0
0
=
+ cos j
y i + j
y i + j +1
arctan
+ arctan
xi j +1
x
i
j
k ( yi + j )
k ( yi + j +1 )
e
sin
(
k
(
x
)
)
e
sin
(
k
(
x
)
)
i
j
i
j
+
1
PV
dk PV
dk
k
k
0
0
) (
Equation 3.451
The integration of the potential component in-phase with the velocity over s j gives:
171
( K7 ) = Re e i (z ) ds
i
s j
( y i + j )
sin ( (xi j ) j )
1 + e
= ( y + )
e i j +1 sin ( (xi j + 1 ) j )
Equation 3.452
172
500
H eave
S wa y
Mass Coefficient
22
R oll
M 33
400
4250
'
M 44
4 000
4000
200
300
3 000
3750
200
2 000
3500
1 000
3250
M id sh i p s e cti o n
o f a c o n ta i n e r sh ip
100
100
0
0
0
20 0
0 .5 1 .0 1 .5 2 .0 2.5
S way
0
0
150
'
Damping Coefficient
5 000
'
'
0 .5 1 .0 1 .5 2 .0 2 .5
He ave
3000
0 .5 1 .0 1 .5 2 .0 2 .5
0
3 00
20 0
N 33
'
Fra nk
15 0
100
0 .5 1 .0 1 .5 2 .0 2 .5
Ro ll
'
N 22
S way - Ro ll
Roll - Sway
'
'
M 2 4 = M 42
N 44
C lo se-fit
2 00
15 0
10 0
10 0
50
1 00
50
L ew is
50
Sw ay - Ro ll
Roll - S way
'
'
N 24 = N 42
0
0
0 0 . 5 1 .0 1 .5 2 .0 2 .5
0
0 0 .5 1 .0 1.5 2.0 2. 5
f re qu ency (rad/ s)
0
0
0 .5 1.0 1 .5 2.0 2 .5
f req ue ncy (ra d/ s)
0 0 .5 1.0 1.5 2. 0 2 .5
f requ en cy (rad /s)
173
174
dx
(1 a1 )2 + 3 a3 2
M 22 ( 0) =
2 1 a1 + a3
'
}
Equation 3.62
When using this formula for surge, the total potential mass of surge is defined by:
M 11 ( 0) = B M 11* ( 0)
Equation 3.63
A frequency-independent total hydrodynamic mass coefficient is estimated empirically by
Sargent and Kaplan [1974] as a proportion of the total mass of the ship :
a
M 11 (S & K ) =
2a
with:
a=
1 b2
b3
1 + b
ln
2 b
1 b
B
b = 1
L
where
Equation 3.64
175
176
Viscous Damping
The strip theory is based on the potential flow theory. This holds that viscous effects are
neglected, which can deliver serious problems when predicting roll motions at resonance
frequencies. In practice, viscous roll damping effects can be accounted for by empirical
formulas. For surge and roll, additional damping coefficients 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.
177
L
V
This total resistance coefficient Ct is given in Figure 4.11 as a function of the ship length.
178
{ln ( Rn ) 2}
Equation 4.13
in which:
S
Rn
From this empirical formula follows the pure viscous part of the additional damping
coefficient at forward ship speed V :
d {R f (V )}
B11v =
dV
Equation 4.14
which can be obtained numerically.
179
=
0
=
B44 + B44 v
2 ( I xx + A44 ) C44
Equation 4.23
This damping coefficient is written as a fraction between the actual damping coefficient,
B44 + B44 v , and the critical damping coefficient, B44 cr = 2 (I xx + A44 ) C44 ; so for critical
damping: cr = 1 .
Herewith, the equation of motion can be re-written as:
2
&& + 2 0 & + 0 = 0
Equation 4.24
Suppose the vessel is deflected to an initial heel angle, a , in still water and then released. The
solution of the equation of motion of this decay becomes:
= a e t cos(0 t ) +
sin (0 t )
0
Equation 4.25
180
(t + T )
Equation 4.26
Because = 0 2 for the natural frequency oscillation and the damping is small so that
2
2 << 0 , one can neglect 2 here and use 0 ; this leads to:
2
0 T T = 2
Equation 4.27
The non-dimensional total roll damping is given now by:
(t )
1
=
ln
2 (t + T )
0
= ( B44 + B44 v )
2 C44
Equation 4.28
The non-potential part of the total roll damping coefficient follows from the average value of
by:
2 C44
B44 v =
B44
0
Equation 4.29
1
ln a i
2 a i+2
versus: a =
a i + a i+2
2
Equation 4.210
2. To avoid spreading in the successively determined -values, caused by a possible zeroshift of the measuring signal, double amplitudes can be used instead:
a i+1
a i a i+1 + a i+ 2 a i+3
1
=
ln a i
versus: a =
2 a i+2 a i +3
4
) (
Equation 4.211
3. Sometimes the results of free rolling tests are presented by:
a
versus: a
a
with the absolute value of the average of two successive positive or negative
maximum roll angles, given by:
a + a i+1
a = i
2
and the absolute value of the difference of the average of two successive positive or
negative maximum roll angles, given by:
a = a i a i+1
Then the total non-dimensional roll damping coefficient becomes:
2+ a
1
a
=
ln
2
a
2
a
Equation 4.212
The decay coefficient can therefore be estimated from the decaying oscillation by
determining 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 a .
Note that these decay tests provide no information about the relation between the potential
coefficients 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 is much greater than about 0.2 and is in any case
strictly valid for small values of only. Luckily, this is generally the case.
Be aware that this damping coefficient is determined by assuming an uncoupled roll motion
(no other motions involved). Strictly, this damping coefficient is not valid for the actual
coupled motions of a ship that will be moving in all directions simultaneously.
182
The successively found values for , plotted on base of the average roll amplitude, will often
have a non-linear behaviour as illustrated in Figure 4.22.
0.04
Produc t c arrier, V = 0 knots
mean linear and cubic damping
mean linear and quadratic damping
second ex periment, negative angles
second ex periment, positiv e angles
first experiment, negative angles
first experiment, pos itive angles
0.03
0.02
0.01
0
0
According to Miller [1974], the non-dimensional total roll damping coefficient, , can be
obtained by:
= 1 + 2 a
with:
2
3
Fn
L
L Fn Fn
1 = CV 0.00085
+
+ 2
B GM Cb Cb
Cb
l bk
rb 3
2 = 19.25 Abk
+ 0.0024 L B
3
rb
L B d Cb
Equation 4.215
where:
Abk = lbk hbk
lbk
hbk
rb
L
B
d
Cb
GM
Fn
a
CV
Generally CV = 1.0 , but (according to an experienced user of computer code SEAWAY) for
slender ships, like frigates, a suitable value for CV seems to be:
CV = 4.85 3.00 GM
Equation 4.216
184
an equivalent linear roll damping dissipates an identical amount of energy as the non-linear
(2 )
roll damping. This results for a linearised quadratic roll damping coefficient, B44v , into:
B44 v
(1 )
Equation 4.217
For the estimation of the non-potential parts of the roll damping, use has been made of work
published by Ikeda, Himeno and Tanaka [1978]. A few sub-ordinate parts have been modified
and this empirical method is called here the ''Ikeda method''.
This Ikeda method estimates the following linear components of the roll damping coefficient
of a ship:
B44 S a correction on the potential roll damping coefficient due to forward speed,
B44 F the frictional roll damping coefficient,
B44 E the eddy making roll damping coefficient,
B44 L the lift roll damp ing coefficient and
B44 K the bilge keel roll damping coefficient.
So, the additional - mainly viscous - roll damping coefficient B44V is given by:
B44V = B44 S + B44 F + B44 E + B44 L + B44 K
Equation 4.218
Ikeda, Himeno and Tanaka [1978] 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 confirmed 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 sufficiently.
For numerical reasons three restrictions have to be made:
if, locally, s > 0.999 then: s = 0.999 .
density of water
a
L
B
D
CM
CB
SL L D
Sf
roll amplitude
length of the ship
breadth
amidships draught
amidships section coefficient
block coefficient
lateral area
wetted hull surface area
OG
Bs
Ds
As
s
H0
a1
a3
Ms
rf
LO
hk
Lk
rk
fk
Cp
pressure coefficient
lm
rb
LR
186
+ (2 A1 A2 1) e
Equation 4.219
with:
B44
= V g
D = 2 D g
A1 = 1.0 + D 1 .2 e 2 D
A2 = 0.5 + D
1 .0
2 D
r
0.512 f
a
Rn =
Equation 4.220
In here, for ship forms the average distance between the roll axis and the hull surface can be
approximated by:
S
(0.887 + 0.145 CB ) f + 2 OG
L
rf =
Equation 4.221
with a wetted hull surface area S f , approximated by:
S f = L (1.7 D + CB B)
Equation 4.222
The relation between the density, kinematic viscosity and temperture of fresh water and sea
water are given in Figure 4.23.
187
1030
2.0
Sea Water
D ensity (kg/m )
Kinematic Viscosity (m s)
1020
1010
1.5
Sea Water
Fresh Water
1.0
1000
Fresh Water
990
0.5
0
10
20
30
10
20
30
Temperature ( C)
T emperature ( C)
Figure 4.23: Relation between density, kinematic viscosity and temperature of water
When eliminating the temperature of water, the following relation can express the kinematic
viscosity into the density of water in the kg-m-s system:
fresh wate r : 10 6 = 1.442 + 0.3924 ( 1000 ) + 0.07424 ( 1000)2
sea water : 106 = 1.063 + 0.1039 ( 1025) + 0.02602 ( 1025)2
Equation 4.223
as given in Figure 4.24.
Salt Water
20
20
25
Fresh Water
25
15
10
15
10
5
Viscosity Actual
Viscosity Polynomial
Temperature
0
997
998
Viscosity Actual
Viscosity Polynomial
T emperature
999
1000
3
0
1023
1024
1025
1026
1027
1028
1
rf 3 S f C f
2
Equation 4.225
This frictional roll damping component increases slightly with forward speed.
Semi-theoretically, Tamiya deduced a modification coefficient for the effect 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 coefficient:
1
V
B44 F (2 ) = r f 3 S f C f 1.0 + 4.1
2
L
Equation 4.226
Then, the equivalent linear roll damping coefficient due to skin friction is expressed as:
8
1
V
B44 F = a rf 3 S f C f 1.0 + 4.1
L
3
2
Equation 4.227
Ikeda confirmed the use of his formula for the three-dimensional turbulent boundary layer
over the hull of an oscillating ellipsoid in roll motion.
4.2.4.4 Eddy Making Damping, B44 E
At zero forward speed the eddy making roll damping for the naked hull is mainly caused by
vortices, generated by a two-dimensional separation. From a number of experiments with twodimensional cylinders it was found that for a naked hull this component of the roll moment is
proportional to the roll frequency squared and the roll amplitude squared. This means that the
corresponding quadratic roll damping coefficient does not depend on the
period parameter but on the hull form only.
When using a simple form for the pressure distribution on the hull surface it appears that the
pressure coefficient C p is a function of the ratio of the maximum relative velocity U max to
the mean velocity U mean on the hull surface:
U
= max
U mean
Equation 4.228
The C p relation was obtained from experimental roll damping data of two-dimensional
models. These experimental results are fitted by:
C p = 0.35 e 2.0 e0 .187 + 1.50
Equation 4.229
The value of around a cross section is approximated by the potential flow theory for a
rotating Lewis form cylinder in an infinite fluid.
An estimation of the local maximum distance between the roll axis and the hull surface, rmax ,
has to be made.
Values of rmax () have to be calculated for:
189
= 2 =
and
0.5
a (1 + a3 )
cos 1
4 a3
Equation 4.230
f3
rmax +
OG
2 Ds H 0 s +
D
s
2 Ms
a 2 + b 2
H
with:
Bs
2 Ds
As
s =
Bs D s
Bs
Ds
Ms =
=
2 (1 + a1 + a3 ) 1 a1 + a3
H0 =
H = 1 + a1 + 9 a3 + 2 a1 (1 3 a3 ) cos(2 ) 6 a3 cos(4 )
a = 2 a3 cos(5 ) + a1 (1 a3 ) cos(3 ) +
2
{(6 3 a ) a + (a
{(6 + 3 a ) a + (a
3 a1 a3 + a1 cos()
b = 2 a3 sin (5 ) + a1 (1 a3 ) sin (3 ) +
2
+ 3 a1 a3 + a1 sin ()
f 3 = 1 + 4 e 1 .6510 (1 s )
5
Equation 4.233
With this a quadratic sectional eddy making damping coefficient for zero forward speed
follows from:
190
B44 E 0
(2 )'
1
4 r
= Ds max
2
Ds
C p
f 1 rb OG f1 rb
f1 rb
1 +
+ f H
1
Ds
Ds
Ds 2 0
Ds
with:
rb = 2 Ds
H 0 (s 1)
4
rb = Ds
B
rb = s
2
Equation 4.235
For three-dimensional ship forms the zero forward speed eddy making quadratic roll damping
coefficient is found by an integration over the ship length:
(2 )
(2 )'
B44 E 0 = B44 E 0 dxb
L
Equation 4.236
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 analysed this
forward speed effect by experiments and the result has been given in an empirical formula.
With this the equivalent linear eddy making damping coefficient at forward speed is given by:
V
8
1
B44 E =
a B44 E 0 ( 2 )
with: K =
2
0.04 L
1+ K
3
Equation 4.237
1
OG
OG
2
LR
LO LR
Equation 4.238
The slope of the lift curve CL is defined by:
191
CL
2 D
B
=
+ 4.1 0.045
L
L
kN =
Equation 4.239
in which the coefficient is given by Ikeda in relation to the amidships section coefficient
CM :
CM < 0.92 :
= 0.00
0.92 < CM < 0.97 :
= 0.10
0.97 < CM < 0.99 :
= 0.30
Equation 4.240
These data are fitted here by:
= 106 (CM 0.91)2 700 (CM 0.91)3
with the restrictions:
if CM < 0.91 then = 0.00
if CM > 1.00 then = 0.35
Equation 4.241
a component B44 K S
(2)
due to the pressure on the hull surface, created by the bilge keels.
(2)
deduced from experimental results of oscillating flat plates. The drag coefficient CD depends
on the period parameter or the Keulegan-Carpenter number. Ikeda measured this non-linear
drag also by carrying out free rolling experiments with an ellipsoid with and without bilge
keels. This resulted in a quadratic sectional damping coefficient:
hk
CD = 22.5
+ 2.40
( 2 )'
3
2
rk a f k
B44 K N = rk hk f k CD
with:
f k = 1.0 + 0.3 e 160 (1 .0 s )
Equation 4.242
The approximation of the local distance between the roll axis and the bilge keel rk is given as:
2
r
OG
r
rk = Ds H 0 0.293 b + 1.0 +
0.293 b
Ds
Ds
Ds
Equation 4.243
192
The approximation of the local radius of the bilge circle rb in here is given before.
Assuming a pressure distribution on the hull caused by the bilge keels, a quadratic sectional
roll damping coefficient can be defined:
B44 K S
(2 )'
k
1
2
2
= rk f k C p lm dh
2
0
Equation 4.244
Ikeda carried out experiments to measure the pressure on the hull surface created by bilge
+
keels. He found that the coefficient C p the pressure on the front face of the bilge keel does
not depend on the period parameter, while the coefficient C p 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 defines an equivalent length of a constant negative pressure region S0 over the height of
the bilge keels, which is fitted to the following empirical formula:
S0 = 0.3 f k rk a + 1.95 hk
Equation 4.245
+
The pressure coefficients on the front face of the bilge keel, C p , and on the back face of the
C p = 1.20
C p = 22.5
and
hk
1.20
f k rk a
Equation 4.246
l m dh = Ds A C p + B C p
2
with:
193
A = (m3 + m4 ) m8 m7 2
B=
m1 =
m43
(1 m1 )2 (2 m3 m2 ) + m (m m + m m )
+
1
3
5
4
6
3 ( H0 0.215 m1 )
6 (1 0.215 m1 )
rb
Ds
OG
Ds
m3 = 1.0 m1 m2
m2 =
m4 = H 0 m1
S0
0.25 m1
Ds
= 0.0
m7 =
m8 = m7 + 0.414 m1
S
= m7 + 1.414 m1 1 cos 0
rb
The equivalent linear total bilge keel damping coefficient can be obtained now by integrating
the two sectional roll damping coefficients over the length of the bilge keels and linearizing
the result:
8
B44 K =
a B44 K N ' + B44 K S ' dxb
3
Lk
Equation 4.248
Experiments of Ikeda have shown that the effect of forward ship speed on this roll damping
coefficient can be ignored.
4.2.4.7 Calculated Roll Damping Components
In Figure 4.25 an example is given of the several roll damping components, as derived with
Ikeda's method, for the S-175 container ship design.
194
195
196
Hydromechanical Loads
With the 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
defined by:
D
X hj ' =
M jj ' &hj * + N jj ' &hj * + X RSj '
(Ordinary Strip Theory, OST)
Dt
In the ''Modified Strip Theory'', as has been published later by for instance Tasai [1969] and
others, these loads become:
D
i
'
X hj ' =
N jj ' &hj * + X RSj '
(Modified Strip Theory, MST)
M jj
Dt
e
*
In these definitions of the two-dimensional hydromechanical load, &hj is the harmonic
'
'
oscillatory motion, M jj and N jj are the two-dimensional potential mass and damping and
'
the non-diffraction part X RSj is the two-dimensional quasi-static restoring spring term.
At all following pages, the hydromechanical load has been calculated in the G( xb , yb , zb ) axes
system with the centre of gravity G in the still water level, so OG = 0 .
Some of the terms in the hydromechanical loads have been outlined there:
the ''Modified Strip Theory'' (OST) includes these outlined terms, but
when ignoring these outlined terms, the ''Ordinary Strip Theory'' (MST) has been
presented.
197
Equation 5.11
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 of
the water particles, relative to the cross section of an oscillating ship in still water, are defined
by:
h 1* = x + bG
bG
&h 1* = x& + bG & V
xb
x& + bG &
bG &
2 bG
&&h 1* = x&& + bG && 2 V
+V 2
xb
xb 2
&x& + bG &&
Equation 5.12
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 defined by:
D
X h 1' =
M 11' &h 1* + N11' &h1*
Dt
dM 11' & *
h1
= M 11' &&h1* + N11' V
dxb
Equation 5.13
According to the ''Modified Strip Theory'' this hydromechanical force becomes:
D
i
'
X h 1' =
N11 ' &h1*
M 11
Dt
e
'
'
'
V dN
dM 11 & *
'
*
= M 11 + 2 11 &&h 1 + N11 V
h1
dx
dx
b
b
Equation 5.14
198
X w1
Equation 5.15
with:
'
V
dN
a11 = + M 11 dxb + 2 11 dxb
e L dxb
L
'
' dM 11 '
dxb + b11V
b11 = + N11
dx
b
L
c11 = 0
a13 = 0
b13 = 0
c13 = 0
a15 = M
L
V dN
bG dxb + 2 11 bG dxb
e L dxb
'
'
11
'
dM 11
b15 = N11 ' V
bG dxb b11V BG
dxb
L
c15 = 0
Equation 5.16
The ''Modified Strip Theory'' includes the outlined terms. When ignoring these outlined terms
the ''Ordinary Strip Theory'' is presented.
A small viscous surge damping coefficient b11V , derived from the still water resistance
approximation of Troost [1955], has been added here.
199
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled surge equation become:
a11 = + M 11 ' dxb
L
c11 = 0
a13 = 0
b13 = 0
c13 = 0
V dN11
bG dxb
2
e L dxb
'
'
'
dM 11
b15 = N11 V
bG dxb b11V BG
dxb
L
c15 = 0
200
Equation 5.21
The lateral and roll motions of the water particles, relative to the cross section of an oscillating
ship in still water, are defined by:
h 2 * = y xb OG
h 4* =
&h 2 * = y& xb & + V OG &
&h 4* = &
&& + 2 V & OG &&
&&h 2 * = y&& xb
&&h 4* = &&
Equation 5.22
dM 22 ' & *
dM 24 ' & *
h 2 + M 24 ' &&h 4 * + N 24 ' V
h 4
= M 22 ' &&h 2 * + N 22 ' V
dx
dx
b
b
Equation 5.23
According to the ''Modified Strip Theory'' this hydromechanical force becomes:
D
i
D
i
'
'
'
*
'
'
*
Xh2 =
N 22 &h 2 +
N 22 &h 2
M 22
M 22
Dt
e
e
Dt
'
'
V dN
dM 22 & *
= M 22 ' + 2 22 &&h 2 * + N 22 ' V
h 2
dxb
dxb
e
V dN '
dM 24 ' & *
+ M 24 ' + 2 24 &&h 4 * + N24 ' V
dxb
dxb h 4
e
Equation 5.24
201
X w2
with:
a 22 = + M 22
'
V
dN 22 '
dxb + 2
dxb
e L dxb
dM 22 '
b22 = + N 22 ' V
dxb
dxb
L
c 22 = 0
a 24 = + M 24 ' dxb + OG M 22 ' dxb
L
dN
dN
V
V
+ 2 24 dxb + 2 OG 22 dxb
dxb
e L dxb
e
L
'
'
dM 24 '
dM 22 '
'
'
b24 = + N 24 V
dxb + OG N 22 V dx dxb
dx
b
b
L
L
c 24 = 0
a 26 = + M 22
L
'
'
V
dM 22
'
xb dxb + 2 N 22 V
dx
dx b b
e L
'
V
V
dN
'
N 22 dxb + 2 22 xb dx b
2
e L
e L dxb
dM 22 '
'
xb dxb 2 V M 22 ' dxb
b26 = + N 22 V
dxb
L
L
V 2 dN 22
+ 2
dxb
e L dxb
'
c 26 = 0
Equation 5.25
The ''Modified Strip Theory'' includes the outlined terms. When ignoring these terms the
''Ordinary Strip Theory'' is presented.
202
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled sway equation become:
a 22 = + M 22 dxb
'
c 22 = 0
a 24 = + M 24 dxb + OG M 22 dxb
'
'
c 24 = 0
a 26 = + M 22 ' x b dxb +
L
V
N 22 ' dxb
2
e L
c 26 = 0
Equation 5.26
So no terms have been added for the ''Modified Strip Theory''.
203
Equation 5.31
The vertical motions of the water particles, relative to the cross section of an oscillating ship
in still water, are defined by:
h 3 * = z + xb
&h 3 * = z& + xb & V
&&h 3 * = &z& + xb && 2 V &
Equation 5.32
According to the ''Ordinary Strip Theory'' the two-dimensional potential hydromechanical
force on a heaving cross section in still water is defined by:
D
X h3' =
M 33' &h 3 * + N33' &h 3 * + 2 g yw h 3*
Dt
= M 33
'
'
dM 33 ' & *
*
&
&
h 3 + 2 g yw h 3*
h 3 + N 33 V
dxb
Equation 5.33
D
i
'
X h3' =
N33' &h 3* + 2 g y w h 3 *
M 33
Dt
e
'
'
= M 33 + 2
h 3 + N 33 V
h 3 + 2 g yw h 3
dxb
dxb
e
Equation 5.34
204
X w3
Equation 5.35
with:
a31 = 0
b31 = 0
c31 = 0
V
dN 33 '
dxb
e 2 L dxb
dM 33 '
b33 = + N 33 ' V
dxb
dxb
L
c33 = +2 g y w dxb
L
a 35 = M 33
L
'
'
V
dM 33
'
x b dxb 2 N33
dx
dxb b
e L
V
V dN
N 33 ' dxb + 2 33 xb dxb
2
e L
e L dxb
'
'
dM
'
33
xb dxb + 2 V M 33 ' dxb
b35 = N 33 V
dxb
L
L
'
V2
dN
+ 2 33 dxb
e L dx b
c 35 = 2 g y w xb dxb
L
Equation 5.36
The ''Modified Strip Theory Method'' includes the outlined terms. When ignoring these terms
the ''Ordinary Strip Theory'' is presented.
205
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled heave equation become:
a31 = 0
b31 = 0
c31 = 0
a33 = + M 33 ' dxb
L
c33 = +2 g y w dxb
L
a 35 = M 33 ' xb dxb
L
V
N 33 ' dxb
2
e L
c 35 = 2 g y w xb dxb
L
Equation 5.37
So no terms have been added for the ''Modified Strip Theory''.
206
Equation 5.41
The roll and lateral motions of the water particles, relative to the cross section of an oscillating
ship in still water, are defined by:
h 4* =
h 2 * = y xb OG
&h 4* = &
&h 2 * = y& xb & + V OG &
&&h 4* = &&
D
'
' & *
' & *
X h2 =
M 44 h 4 + N 44 h 4 + 2 g
bG h 4 *
Dt
2
3
D
M 42 ' & h 2 * + N 42 ' &h 2 *
Dt
y w 3 As
dM 44 ' & *
h 4 *
= M 44 ' &&h 4 * + N 44 ' V
+
2
bG
h
4
dx
2
b
dM 42 ' & *
+ M 42 ' &&h 2 * + N42 ' V
h 2
dx
b
Equation 5.43
According to the ''Modified Strip Theory'' this hydromechanical moment becomes:
y w 3 As
D
i
'
'
' & *
X h2 =
N 44 h 4 + 2 g
bG h 4 *
M 44
Dt
e
2
D
i
'
'
*
+
N 42 & h 2
M 42
Dt
e
'
V dN
dM 44 ' & *
= M 44 ' + 2 44 &&h 4 * + N 44 ' V
h4
dx
dx
b
b
3
y
A
+ 2 g w s bG h 4 * +
2
3
'
'
V dN 42 && *
dM 42 & *
'
'
+ M 42 + 2
+ N V
dxb h 2 42
dxb h 2
e
Equation 5.44
207
+ c42 y
+ c44
&& + b46 & + c46 =
+ ( I xz + a46 )
X w4
Equation 5.45
with:
a 42 = + M 42
'
V
dN 24 '
dxb + OG a 22 + 2
dxb
e L dxb
dM 42 '
b42 = + N 42 ' V
dxb + OG b22
dx
b
L
c 42 = 0 + OG c22
a 44 = + M 44 ' dxb + OG M 42 ' dx b + OG a24
L
'
'
V
dN
V
dN
+ 2 44 dxb + 2 OG 42 dxb
dxb
e L dxb
e
L
'
'
dM 44
dM 42
'
'
b44 = + N 44 V
dx
+
OG
dx + b + OG b24
L 42
dxb b
dxb b 44V
L
y 3 A
c 44 = +2 g w + s bG dxb + OG c24
3
2
L
= + g GM
a 46 = + M 42 xb dx b +
'
'
V
dM 42
'
dx + OG a 26
42
dxb b
e 2 L
'
V
V
dN
'
+ 2 N 42 dxb + 2 42 xb dxb
e L
e L dxb
'
dM 42
'
'
b46 = + N 42 V
xb dxb 2 V M 42 dx b + OG b26
dxb
L
L
'
V 2 dN 42
dxb
e 2 L dxb
c 46 = 0 + OG c26
Equation 5.46
The ''Modified Strip Theory'' includes the outlined terms. When ignoring these terms the
''Ordinary Strip Theory Method'' is presented.
208
A viscous roll damping coefficient b44V , derived for instance with the empirical method of
Ikeda [1978], has been added here.
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled roll equation become:
a42 = + M 42 ' dxb + OG a22
L
c24 = 0
a44 = + M 44 dxb + M 42 dxb + OG a24
'
'
c44 = + g GM
a46 = + M 42 ' xb dxb +
L
V
N42 ' dxb + OG a26
2
e L
'
c46 = 0
Equation 5.47
So no terms have been added for the ''Modified Strip Theory''.
209
Equation 5.51
According to the ''Ordinary Strip Theory'' the two-dimensional potential hydromechanical
moment on a pitching cross section in still water is defined by surge and heave contributions:
'
'
dM
'
'
*
11
&
&
bG &h1*
X h 5 = M 11 bG h1 N11 V
dxb
'
dM 33 '
*
&
&
xb &h 3 * 2 g y w xb h 3*
M 33 xb h 3 N 33 V
dxb
'
Equation 5.52
According to the ''Modified Strip Theory'' this hydromechanical moment becomes:
'
'
'
V dN
dM 11
'
'
*
*
X h 5 = M 11 + 2 11 bG &&h1 N11 V
bG &h1
dxb
dxb
e
'
V dN 33'
&& * N ' V dM 33 x & *
M 33 ' + 2
dxb b h 3 33
dxb b h 3
e
2 g yw xb h 3 *
Equation 5.53
210
X w5
Equation 5.54
with:
a51 = M 11' bG dxb +
L
V dN11'
bG dxb
e 2 L dxb
'
'
dM
11
bG dxb b11V BG
b51 = N11 V
dxb
L
c51 = 0
a53 = M 33
'
V dN 33'
xb dxb + 2
xb dxb
e L dxb
'
dM 33'
b53 = N33 V
xb dxb
dxb
L
c53 = 2 g yw xb dxb
L
'
2
V
dN
a55 = + M 11 bG dxb + 2 11 bG dxb
e L dxb
L
2
'
+ M 33' xb 2 dxb +
L
' dM 33 '
V
N33
xb dxb
2
dxb
e L
V
V
dN 33'
'
2
dx
+
xb dxb
b
b
2 33
2
dx
e L
e L b
'
'
2
2
dM
11
bG dxb +b11V BG
b55 = + N11 V
dxb
L
dM 33 ' 2
'
+ N33 ' V
xb dxb 2 V M 33 xb dxb
dx
L
b
L
V 2 dN 33'
+
xb dxb
e 2 L dxb
c55 = +2 g yw xb 2 dxb
L
211
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled pitch equation become:
V dN11'
a51 = M bG dxb + 2
bG dxb
e L dxb
L
'
11
'
dM 11'
bG dxb b11V bG
b51 = N11 V
dx
b
L
c51 = 0
V
N33 ' dxb
2
e L
c53 = 2 g yw xb dxb
L
+ M 33' xb 2 dxb +
L
2
V
dN11'
bG
dxb
e 2 L dxb
V
V
N 33' xb dxb + 2 M 33 ' dxb
2
e L
e L
V
N33 ' xb dxb
2
e L
'
2
2
dM 11'
b55 = + N11 V
bG
dx
+
b
BG
b
11
V
dxb
L
+ N 33 xb dxb +
'
V2
'
N33 dxb
2
e L
c55 = +2 g yw xb 2 dxb
L
Equation 5.56
212
Equation 5.61
According to the ''Ordinary Strip Theory'' the two-dimensional potential hydromechanical
force on a yawing cross section in still water is defined by sway contributions:
'
'
dM
'
'
*
22
&
&
xb &h 2 *
X h 2 = M 22 xb h 2 + N22 V
dxb
+ M 24
'
'
dM 24 '
*
&
&
xb &h 4 *
xb h 4 + N 24 V
dxb
Equation 5.62
V dN 22
'
'
&& * + N ' V dM 22 x & *
X h 2 = M 22 + 2
dxb b h 2 22
dxb b h 2
e
V dN '
dM 24 '
xb &h 4 *
+ M 24 ' + 2 24 xb &&h 4 * + N 24' V
e
dxb
dxb
Equation 5.63
213
X w6
Equation 5.64
with:
a 62 = + M 22 ' x b dxb +
L
'
V
dN 22
xb dxb
2
e L dxb
'
dM
'
22
x dxb
b62 = + N22 V
dxb b
L
c 62 = 0
dN
dN
V
V
+ 2 24 xb dxb + 2 OG 22 xb dxb
dxb
e L dxb
e
L
'
'
dM 24 '
dM 22 '
'
xb dxb + OG N 22 ' V
x dxb
b64 = + N24 V
dxb
dxb b
L
L
c 64 = 0
a 66 = + M 22 ' x b 2 dxb +
L
'
V
dM 22
'
x dxb
22
2
dxb b
e L
'
V
V
dN
'
2
N 22 xb dx b + 2 22 xb dxb
2
e L
e L dxb
'
dM 22 2
'
'
b66 = + N22 V
xb dx b 2 V M 22 xb dx b
dxb
L
L
'
V 2 dN 22
xb dxb
e 2 L dxb
c 66 = 0
Equation 5.65
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
214
After simplification (see the end-terms in Section 2-5-3), the expressions for the total
hydromechanical coefficients in the coupled yaw equation become:
a 62 = + M 22 x b dxb +
'
V
'
N 22 dx b
2
e L
b62 = + N 22 x b dxb + V M 22 dx b
'
'
c 62 = 0
a 64 = + M 24 ' x b dxb + OG M 22 ' xb dxb
L
V
V
'
'
N24 dxb + 2 OG N 22 dxb
2
e L
e
L
b64 = + N 24 x b dxb + V M 24 dx b
'
'
+ OG N 22 x b dxb + V OG M 22 ' dx b
'
c 64 = 0
a 66 = + M 22 ' x b 2 dxb +
L
V
V2
'
dx
+
M 22 ' dxb
22
b
b
2
2
e L
e L
V
N22 ' xb dxb
2
e L
b66 = + N 22
L
'
V 2
x b dxb +
N 22 ' dxb
2
e L
2
c 66 = 0
Equation 5.66
215
216
cosh [k h]
Equation 6.11
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 defined.
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:
D & '
D
&&wj ' =
wj
with:
= V
for: j = 1,2,3, 4
Dt
Dt t
xb
{ }
Equation 6.12
With this, the relative velocities and accelerations in the different directions can be found:
Surge direction:
w
&w1 ' =
xb
+ k g cos cosh[k (h + zb )]
=
cosh[k h]
cosh[k (h + zb )]
'
&&w1 = k g cos
a sin (e t k xb cos k yb sin )
cosh [k h ]
Equation 6.13
Sway direction:
w
'
&w 2 =
y b
+ k g sin cosh [k (h + z b )]
cosh [k h ]
cosh [k (h + z b )]
= k g sin
a sin (e t k xb cos k y b sin )
cosh [k h ]
=
&&w 2 '
Equation 6.14
Heave direction:
217
w
&w 3 ' =
z b
k g sinh [k (h + z b )]
=
cosh[k h]
sinh [k (h + zb )]
'
&&w 3 = k g
a cos (e t k xb cos k yb sin )
cosh [k h]
Equation 6.15
Roll direction:
& ' & '
&w 4 ' = w 2 w 3 = 0
zb
yb
&& ' = 0
w4
Equation 6.16
These zero solutions are obvious, because the potential fluid is free of rotation.
The pressure in the fluid follows from the linearised equation of Bernoulli:
cosh [k (h + z b )]
p = g zb + g
a cos (e t k xb cos k yb sin )
cosh [k h ]
p
p
p
= p0 +
dxb +
dyb +
dzb
xb
yb
zb
Equation 6.17
with the following expressions for the pressure gradients:
p
cosh [k (h + z b )]
= + g k cos
a sin (e t k xb cos k yb sin )
xb
cosh [k h ]
p
cosh[k (h + zb )]
= + g k sin
a sin (e t k xb cos k yb sin )
yb
cosh[k h ]
p
cosh [k (h + zb )]
= g + g k
a cos(e t k xb cos k yb sin )
zb
cosh [k h]
Equation 6.18
These pressure gradients can be expressed in the orbital accelerations too:
p
= &&w1 '
xb
p
= + &&w 2 '
yb
p
= g + &&w3 '
zb
Equation 6.19
218
Equation 6.21
According to the ''Ordinary Strip Theory'' the exciting wave forces for surge on a restrained
cross section of a ship in waves are defined by:
D
X w1 ' =
M ' & * + N11' &w1* + X FK 1'
Dt 11 w1
dM 11' & *
w1 + X FK1 '
= M 11 ' &&w1* + N11' V
dxb
Equation 6.22
According to the ''Modified Strip Theory'' these forces become:
D
i
'
'
'
*
'
X w1 =
N11 &w1 + X FK 1
M 11 +
Dt
e
'
'
&& * '
& *
V
dN
dM
'
11
11
w1 + N11 V
w1 + X FK 1'
= M 11 + 2
dxb
dxb
e
Equation 6.23
+ yb
p
dyb dzb
xb
T yb
X FK1 =
+ yb
&&
'
w1
dyb dz b
T y b
Equation 6.24
After neglecting the second order terms, the Froude-Krilov force can be written as:
X FK1 = Ach ( k g cos ) a sin (e t k xb cos )
with:
0
sin ( k yb sin ) cosh [k (h + z b )]
Ach = 2
yb dz b
k yb sin
cosh [k h ]
T
Equation 6.25
When expanding the Froude-Krilov force in deep water with >> 2 yw and >> 2 T
in series, it is found:
1
X FK1 = A + k S y + k 2 I y + ... ( k g cos ) a sin (e t k xb cos )
2
with:
0
A = 2 yb dzb
T
S y = 2 yb z b dzb
T
I y = 2 yb z b 2 dzb
T
Equation 6.26
The acceleration term k g cos a in here is the amplitude of the longitudinal component
of the relative orbital acceleration in deep water at z b = 0 .
The dominating first term in this series consists of a mass and this acceleration. The mass
term A is used to obtain from the total Froude-Krilov force an equivalent longitudinal
component of the orbital acceleration of the water particles:
*
X FK1 = A &&w1
Equation 6.27
This holds that the equivalent longitudinal components of the orbital acceleration and velocity
*
are equal to the values at z b = 0 in a wave with reduced amplitude a 1 :
&& * = k g cos * sin ( t k x cos )
w1
a1
+ k g cos
&w1* =
a 1* cos (e t k xb cos )
with:
a 1* =
Ach
A a
Equation 6.28
This equivalent acceleration and velocity will be used in the diffraction part of the wave force
for surge.
From the foregoing follows the total wave loads for surge:
220
X w1 = + M &&w1* dxb +
'
11
V
dN11' && *
w1 dxb
e L dxb
dM 11 ' & *
+
N11' V
w1 dxb
e
dxb
L
+ X FK1 ' dxb
L
Equation 6.29
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
Equation 6.210
According to the ''Ordinary Strip Theory'' the exciting wave forces for sway on a restrained
cross section of a ship in waves are defined by:
D
X w 2' =
M 22 ' & w2 * + N22 ' &w 2 * + X FK 2 '
Dt
'
dM 22 ' & *
' && *
w2 + X FK 2 '
= M 22 w 2 + N22 V
dxb
Equation 6.211
According to the ''Modified Strip Theory'' these forces become:
D
i
'
'
'
*
'
Xw2 =
N 22 &w 2 + X FK 2
M 22 +
Dt
e
'
'
V dN 22 && * '
dM 22 & *
'
'
= M 22 + 2
w2 + N22 V
w2 + X FK 2
dxb
dxb
e
Equation 6.212
221
y
b
T yb
=
+ yb
&&
w2
'
dyb dzb
T y b
Equation 6.213
After neglecting the second order terms, the Froude-Krilov force can be written as:
X FK 2 = Ach ( k g sin ) a sin (e t k xb cos )
with:
0
sin ( k yb sin ) cosh [k (h + z b )]
Ach = 2
yb dz b
k yb sin
cosh [k h ]
T
Equation 6.214
When expanding the Froude-Krilov force in deep water with >> 2 yw and >> 2 T
in series, it is found:
1
X FK 2 = A + k S y + k 2 I y + ... ( k g sin ) a sin (e t k xb cos )
2
with:
0
A = 2 yb dzb
T
S y = 2 yb z b dzb
T
I y = 2 yb z b 2 dzb
T
Equation 6.215
The acceleration term k g sin a in here is the amplitude of the lateral component of the
relative orbital acceleration in deep water at z b = 0 .
The dominating first term in this series consists of a mass and this acceleration.
This mass term A is used to obtain from the total Froude-Krilov force an equivalent lateral
component of the orbital acceleration of the water particles:
*
X FK 2 = A &&w 2
222
This holds that the equivalent lateral components of the orbital acceleration and velocity are
*
equal to the values at z b = 0 in a wave with reduced amplitude a 2 :
&& * = k g sin * sin ( t k x cos )
w2
&w1*
a2
+ k g sin
=
a 2 * cos (e t k xb cos )
with:
a 2 * =
Ach
A a
Equation 6.216
This equivalent acceleration and velocity will be used in the diffraction part of the wave force
for sway.
From the foregoing follows the total wave loads for sway:
X w 2 = + M 22
'
&&w 2 * dxb +
V
dN 22 ' && *
w 2 dxb
e L dxb
dM 22 ' & *
'
+
N 22 V
w 2 dxb
e
dxb
L
+ X FK 2 ' dxb
L
Equation 6.217
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
Equation 6.218
According to the ''Ordinary Strip Theory'' the exciting wave forces for heave on a restrained
cross section of a ship in waves are defined by:
D
X w 3' =
M 33 ' &w 3* + N33 ' &w3 * + X FK 3 '
Dt
'
dM 33' & *
' && *
w3 + X FK 3 '
= M 33 w3 + N33 V
dxb
Equation 6.219
According to the ''Modified Strip Theory'' these forces become:
223
Xw3 =
'
D
i
'
'
*
'
N33 &w 3 + X FK 3
M 33 +
Dt
e
'
'
V dN 33 && *
dM 33 & *
'
'
'
= M 33 + 2
w3 + N 22 V
w3 + X FK 3
dxb
dxb
e
Equation 6.220
z
b
T y b
=
(g + && ) dy
+ yb
'
w3
dzb
T yb
Equation 6.221
After neglecting the second order terms, the Froude-Krilov force can be written as:
2 yw sin ( k yb sin )
X FK 3 =
with:
0
sin ( k yb sin ) sinh [k (h + zb )]
Ash = 2
yb dzb
k yb sin
cosh [k h]
T
Equation 6.222
When expanding the Froude-Krilov force in deep water with >> 2 yw and >> 2 T
in series, it is found:
2 yw
1
X FK 3 =
+ A + k S y + k 2 I y + ... ( k g ) a cos (e t k xb cos )
2
k
with:
224
A = 2 yb dzb
T
S y = 2 yb z b dzb
T
I y = 2 yb z b 2 dzb
T
Equation 6.223
??????
*
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 fluid and can be obtained by.
*
T3 = ???
This holds that the equivalent vertical components of the orbital acceleration and velocity are
*
equal to the values at zb = T3 :
???
When expanding the Froude-Krilov force in shallow water with k h 0 and in long waves
with ??? in series, it is found:
C3 = ???
with:
???
*
So in shallow water, T3 can be obtained by.
T3 = ???
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 definition for T3 is valid in deep water too, because:
???
These equivalent accelerations and velocities will be used to determine the diffraction part of
the wave forces for heave.
*
From the foregoing follows the total wave loads for heave:
X w 3 = + M 33' &&w 3* dxb +
L
V
dN '
33 &&w 3* dxb
e L dxb
dM 33 ' & *
+
N33 ' V
w 3 dxb
e
dxb
L
+ X FK 3 ' dxb
L
Equation 6.224
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
X w 4 = X w 4 dxb
'
Equation 6.225
According to the ''Ordinary Strip Theory'' the exciting wave moments for roll on a restrained
cross section of a ship in waves are defined by:
D
X w 4' =
M ' & * + N 42' &w 2 * + X FK 4 ' + OG X w 2 '
Dt 42 w2
dM 42 ' & *
w 2 + X FK 4 ' + OG X w2 '
= M 42 ' &&w 2* + N42 ' V
dxb
Equation 6.226
According to the ''Modified Strip Theory'' these moments become:
D
i
'
'
'
*
'
'
Xw4 =
N 42 &w 2 + X FK 4 + OG X w 2
M 42 +
Dt
e
'
'
'
V dN
dM 42 & *
'
*
'
'
= M 42 + 2 42 &&w2 + N24 V
w2 + X FK 4 + OG X w2
dxb
dxb
e
Equation 6.227
p
X FK 4 =
zb +
yb dyb dzb
yb
z b
T yb
{ &&
+ yb
w2
'
) }
'
z b + g + &&w 3 yb dyb dz b
T y b
Equation 6.228
After neglecting the second order terms, the Froude-Krylov moment can be written as:
S
C
226
sin ( k yw sin )
cos ( k y w sin )
k y w sin
= 2
yw 3
2
( k yb sin )
C yw
yb zb dzb
k yb sin
cosh[k h]
T
0
S ych = 2
sin ( k yb sin )
cos( k yb sin )
k yb sin
cosh [k (h + zb )] 3
I zsh = 2
yb dzb
2
cosh[k h]
( k yb sin )
T
0
Equation 6.229
For deep water, the cosine-hyperbolic expressions in here reduce to exponential expressions.
From the foregoing follows the total wave loads for roll:
X w 4 = + M 42
'
*
&&w 2 dxb +
V
dN 42 ' && *
w 2 dxb
e L dxb
'
dM 42 & *
+
N 42 ' V
w 2 dxb
dxb
L e
+ X FK 4 dxb + OG X w 2
'
Equation 6.230
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
with:
X w 5 = X w1 bG X w 3 xb
'
'
'
Equation 6.231
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:
227
V
dN11'
*
&
&
= M bG w1 dxb +
bG &&w1* dxb
e L dxb
L
Xw5
'
11
'
dM 11
'
*
N11 V
bG &w1 dxb
e
dxb
L
'
33
V
dN33 '
*
&
&
xb w3 dxb +
xb &&w 3* dxb
e L dxb
dM 33 '
N 33' V
xb &w3 * dxb
e
dxb
L
X FK 3 xb dxb
'
Equation 6.232
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
with:
X w 6 = + X w 2 xb
'
'
Equation 6.233
From this follows the total wave loads for yaw:
X w 6 = + M 22
'
xb &&w 2* dxb +
V
dN 22 '
xb &&w2 * dxb
e L dxb
dM 22 '
+
N 22' V
xb & w2 * dxb
e
dxb
L
+ X FK 2 ' xb dxb
L
Equation 6.234
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
228
for: j = 2,3,4
Equation 6.31
'
The 2-D hydrodynamic loads X hi in the sway, heave and roll directions i , exercised by the
fluid on a cross section of the cylinder, can be obtained from the 2-D velocity potentials and
the linearised equations of Bernoulli. The velocity potentials have been obtained by using the
work of Ursell [1949] and N -parameter conformal mapping. These hydrodynamic loads are:
g ja
'
X hi = 2 g ywl
Aij cos t + x j + Bij sin t + x j
)}
Equation 6.32
in where j is the mode of oscillation and i is the direction of the load. The phase lag x j is
defined as the phase lag between the velocity potential of the fluid and the forced motion
x j . The radiated damping waves have an amplitude ja and ywl is half the breadth of the
cross section at the waterline. The potential coefficients Aij and Bij and the phase lags x j ,
expressed in terms of conformal mapping coefficients, are given in a foregoing chapter.
'
These loads X hi can be expressed in terms of in-phase and out-phase components with the
harmonic oscillations:
2
a g ja
X hi ' = 2 ij
x ja
{( A
ij
with:
229
a22 = 2
a24 = 4 y wl
a33 = 2
a42 = 4 ywl
a44 = 2
x ja
2
P0 j =
y wl sin x j
ja
g
Q0 j = +
x ja
2
y wl cos x j
ja
g
Equation 6.33
The phase lag x j between he velocity potentials and the forced motion is incorporated in the
coefficients P0 j and Q0 j and can be obtained by using:
P
x j = arctan 0 j
+ Q0 j
Equation 6.34
This equation will be used further on for obtaining wave load phases.
Generally, these hydrodynamic loads are expressed in terms of potential mass and damping
coefficients:
'
X hi = M ij &x& j N ij x& j
= M ij 2 x ja cos ( t ) + N ij x ja sin ( t )
with:
M ij = bij
Nij = bij
b22 = 2 y wl
b24 = ywl 3
b33 = 2 y wl 2
b42 = 2 y wl 3
b44 = 2 y wl 4
Equation 6.35
Note that the phase lag information x j is vanished here.
Tasai [1965] has used the following potential damping coupling coefficients in his formulation
of the hydrodynamic loads for roll:
N 44 '
'
'
'
'
N 42 = '
and
N 24 = N22 l w
lw
230
'
(N ) = (N )
' 2
' 2
24
42
N 22 '
N 22 '
Equation 6.37
This relation - which has been confirmed by numerical calculations with SEAWAY - will be
used further on for obtaining the wave loads for roll from those for sway.
c
2 k h
cwave =
and
c group = wave =
k
2
sinh [2 k h]
Equation 6.38
Consider a cross section which is harmonic oscillating with a frequency = 2 T and an
'
for: j = 2,3,4
= X hja ' cos hj ' cos( t ) X hja ' sin hj ' sin ( t )
Equation 6.39
The energy required for this oscillation should be equal to the energy radiated by the damping
waves:
1 T
1 T '
'
X hj x& j dt = N jj x& j x& j dt
T 0
T 0
1
2
= 2 g a cgroup
2
or:
1
1
'
'
'
2
X hja x ja sin hj = N jj 2 x ja
2
2
= g a 2 cgroup
Equation 6.310
From the first part of Equation 6.310 follows:
X hja' sin hj '
x
'
= N jj ja
a
a
Equation 6.311
231
From the second part of Equation 6.310 follows the amplitude ratio of the oscillatory
motions and the radiated waves:
x ja 1
2 g c group
=
a
N jj '
Equation 6.312
Combining these last two equations provides for the out-phase part - so the damping part - of
the oscillatory force:
X hja' sin hj '
'
= 2 g c group N jj
for: j = 2,3,4
a
In here, X hja sin hj is the in-phase with the velocity part of the exciting force or moment.
'
'
Equation 6.313
in which wj is the phase lag with respect to the wave surface elevation at the center of the
cross section.
This leads for the amplitude of the exciting wave force to:
X wja'
'
= 2 g c group N jj for: j = 2,3,4
a
'
Equation 6.314
which is in principle the same equation as the previous one for the out-phase part of the
oscillatory force in still water.
'
However, for the phase lag of the wave force, wj , an approximation has to be found.
232
Figure 6.31 Vector diagrams of wave components for sway and heave
6.3.3.1 Heave Mode
The vertical wave force on a restrained cross section in waves is:
X w 3 ' = X w 3 a ' cos t + w3 '
The phase lag of a radiated wave, wR 3 , at the intersection of the ship's hull with the waterline,
'
yb = y wl , is wR 3 = k ywl . The phase lag of the wave force, w3 , has been approximated by
this phase:
'
'
w 3 = wR 3 = k y wl
'
'
Equation 6.317
Then, the in-phase and out-phase parts of the wave loads are:
233
Equation 6.318
'
'
'
in which a 3 and v 3 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:
X '
a 3 ' = w31'
M 33
v 3' =
X w32 '
N33 '
Equation 6.320
The phase lag of an incoming undisturbed wave, wI 2 , at the intersection of the ship's hull
with the waterline, yb = y wl , is:
'
wI 2 ' = k y wl sin
if sin < 0 then : wI 2 ' = wI 2 ' +
Equation 6.323
234
In very short waves - so at high wave frequencies - the ship's hull behaves like a
'
vertical wall and all waves will be diffracted. Then, the phase lag of the wave force, w 2 , is
equal to:
'
'
w 2 ( ) = wI 2
Equation 6.324
The acceleration and velocity amplitudes of the water particles in the undisturbed surface of
the incoming waves are:
a2 ' stillwater surface = k g sin
( )
(v )
a2 k g sin
=
=
'
'
Equation 6.325
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 very
'
much over the draft of the section. Apparently, the phase lag of the wave force, w 2 , can be
approximated by:
'
'
X
+ M 22 ( k g sin )
'
w 2 ( 0) = arctan FK 2
' k g sin
N 22
Equation 6.326
When plotted against , the two curves w 2 ( 0 ) and w 2 ( ) will intersect each
'
'
other. The phase lag of the wave force, w 2 , can now be approximated by the lowest of these
two values:
w2 ' = w 2 ' ( )
'
'
Equation 6.328
Now the in-phase and out-phase terms of the wave force in beam waves are:
X FK 2 ' + X w 21' = X w 2 a ' sin w 2 '
= a 2 g cgroup N 22 ' sin w 2 '
Equation 6.329
235
'
'
'
in which a 2 and v 2 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 influence of the wave direction - the equivalent
acceleration and velocity amplitudes of the water particles are:
X '
a 2 ' = w 21' sin
M 22
v 2' =
X w 22 '
sin
N 22 '
Equation 6.331
(N ) = (N )
=
' 2
N 44
'
' 2
24
42
N 22 '
N 22 '
The horizontal wave moment on a restrained cross section in beam waves is:
'
'
'
X w 4 = X w 4 a cos t + w4
Equation 6.332
of which the amplitude is equal to:
X w 4a = a
2 g c group N 44
'
'
(N )
' 2
= a
2 g c group
= a
2 g c group N 22
24
N 22
'
= X w 2a
'
'
N 24 '
N 22 '
N 24 '
N 22 '
Equation 6.333
The in-phase and out-phase parts of the wave moment in beam waves are:
236
X FK 4 + X
'
'
w 41
= X FK 2 + X
'
X w 42 = X w 22
'
'
N24
'
w 21
N 24 '
) N
'
22
'
N22 '
Equation 6.334
'
'
'
in which a 24 and v 24 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 influence of the wave direction - the equivalent
acceleration and velocity amplitudes of the water particles are:
X w 41'
'
a 24 =
sin
M 24 '
v 24 ' =
X w42 '
sin
N 24 '
Equation 6.336
a1'
Equation 6.337
237
Figure 6.42: Comparison of simple diffraction wave loads with DELFRAC data
238
After dividing the left and right hand terms by the wave amplitude a , two sets of six coupled
equations of motion are available.
The 6 variables in the coupled equations for the vertical plane motions are:
xa
xa
Surge :
cos x and
sin x
a
a
z
za
Heave : a cos z and
sin z
a
a
a
a
Pitch :
cos and
sin
a
a
The 6 variables in the coupled equations for the horizontal plane motions are:
ya
ya
Sway :
cos y and
sin y
a
a
a
a
Roll :
cos and
sin
a
a
a
a
Yaw :
cos and
sin
a
a
These sets of motions have to be solved by a numerical method. A method that provides
continuous good results, given by de Zwaan [1977], has been used in the strip theory program
SEAWAY.
239
The associated phase shifts of these motions relative to the wave elevation are:
x
y
z
The transfer functions of the translations are non-dimensional. The transfer functions of the
rotations can be made non-dimensional by dividing the amplitude of the rotations by the
amplitude of the wave slope k a in lieu of the wave amplitude a :
xa
ya
za
a
a
a
a
a
a
k a k a k a
Some examples of calculated transfer functions of a crude oil carrier and a containership are
given in Figure 7.11, Figure 7.12 and Figure 7.13.
1 .5
1 .5
C ru de O il C a rrie r
V = 0 kn
P it ch
C ru de O il C arrier
V = 0 kn
H e a ve
1 .0
= 90
= 180
0 .5
1 .0
= 18 0
0 .5
= 90
0
0
0 .25
0 .5 0
0 .7 5
1 .0 0
0 .2 5
0. 50
0 .7 5
1. 00
W a ve F req u en cy (rad /s )
-90
-90
= 90 0
-180
-270
-360
= 180
-450
= 1800
-180
Phase
Phase z (deg)
-270
-450
-540
-540
-630
-630
-720
= 90
-360
-720
0
0.25
0.50
0.75
1.00
0.25
0.50
0.75
1.00
240
1 .5
1 .5
C ru de O il C a rrie r
V = 1 6 kn
P it ch
C ru de O il C arrier
V = 1 6 kn
H e a ve
1 .0
= 90
= 18 0
0 .5
1 .0
= 1 80
0 .5
= 90
0
0
0
0 .25
0 .5 0
0 .7 5
1 .0 0
0 .2 5
0. 50
0 .7 5
1. 00
W a ve F req u en cy (rad /s )
-90
= 90
-180
-270
= 180
-360
-90
= 180 0
-180
Phase
Phase z (deg)
-450
-270
-450
-540
-540
-630
-630
-720
= 90
-360
-720
0
0.25
0.50
0.75
1.00
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 = 10 k nots
10
V = 20 k nots
V = 20 k nots
V = 10 knots
1.00
V = 0 knots
0.75
0.50
0.25
Containership
L pp = 175 metre
0
Containership
L pp = 175 metre
0
0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
242
= x Pa cos e t + xP
= yPa cos e t + yP
= zPa cos e t + z P
With the six motions of the centre of gravity, the harmonic motions of any point P( xb , yb , zb )
on the ship in the ship-bound xb , yb and zb directions - or in the earth bound system in x , y
and z directions - can be calculated by using the previous equations.
243
= e xPa sin e t + x P
= e x Pa cos e t + x P 2
= e y Pa sin e t + y P
= e yPa cos e t + y P 2
= e z Pa sin e t + z P
= e z Pa cos e t + z P 2
244
Longitudinal acceleration:
&x&P = &x& yb
&& + z b &&
= e xPa cos e t + x P
2
= e xPa cos e t + xP
2
Lateral acceleration:
&& zb &&
&y&P = &y& + xb
= e yPa cos e t + y P
2
= e y Pa cos e t + y P
Vertical acceleration:
&z&P = &z& xb && yb &&
= e z Pa cos e t + z P
2
= e z Pa cos e t + z P
2
Longitudinal acceleration:
&x&P = &x& yb
&& + zb && g
Lateral acceleration:
&y&P = &y& + xb
&& zb && + g
Vertical acceleration:
&z&P = &z& xb && yb &&
= e 2 z Pa cos e t + z P
246
= s Pa cos e t + s P
with:
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.
Notice the different behaviours of the absolute and relative vertical motions, as given in
Figure 7.51.
5
5
C on tai ne rs h ip
H ea d w av e s
C o nta ine rs h ip
H e ad w av e s
4
V = 20 k no ts
3
V = 10 k no ts
V = 0 k n ots
1
R A O te nd s
to 1 .0
V = 2 0 k n o ts
V = 1 0 k n ots
3
V = 0 k n ots
2
R A O te nd s
to 1 .0
R A O te nd s
to 0 .0
R AO ten d s
to 0.0
0
0
0.5
1 .0
1 .5
0 .5
1.0
1 .5
w a v e fre q ue nc y (r ad /s )
247
248
Anti-Rolling Devices
Since the disappearance of sails on ocean-going ships, with their stabilising wind effect on the
rolling motions, naval architects have been concerned in reducing the rolling of ships among
waves. With bilge keels they performed a first successful attack on the problem of rolling, but
in several cases these bilge keels did not prove to be sufficient. Since 1880, numerous other
more or less successful ideas have been tested and used.
Four types of anti-rolling devices and its contribution to the equations of motion are described
here:
bilge keels
passive free-surface tanks
active fin stabilisers
active rudder stabilisers.
The active fin and rudder stabilisers are not built into the program SEAWAY yet.
249
coefficient C p 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 defines an equivalent length
of a constant negative pressure region S 0 over the height of the bilge keels and a twodimensional roll-damping component can be found.
The total bilge keel damping has been obtained by integrating these two two-dimensional rolldamping components over the length of the bilge keels.
Experiments of Ikeda showed that the effect of forward speed on the roll damping due to the
bilge keels could be ignored.
The equivalent linear total bilge keel damping has been obtained by linearising the result, as
has been shown in a separate Chapter.
250
0 =
g
h
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 modified to be useful for any
arbitrary vertical position of the centre of rotation by Journe [1997]. For low and high
frequencies and the frequencies close to the natural frequency, different approaches have been
used. A calculation routine has been made to connect these regions.
8.2.1.1 Low and High Frequencies
The harmonic roll motion of the tank is defined by:
= a sin ( t )
In the axis-system of Figure 8.21 and after linearisation, the vertical displacement of the tank
bottom is described by:
z = s + y
and the surface elevation of the fluid is described by:
z = s + h +
Relative to the bottom of the tank, the linearised surface elevation of the fluid is described by:
= h + y
Using the shallow water theory, the continuity and momentum equations are:
v
+v
+
=0
t
y
y
v
v
+v + g
+ g = 0
t
y
y
In these formulations, $v$ denotes the velocity of the fluid 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:
a 2 b << g
The boundary conditions for v have been determined by the velocity produced in the
horizontal direction by the excitation. Between the surface of the fluid and the bottom of the
tank, the velocity of the fluid v varies between vs and vs cosh [k h ] with a mean velocity:
vs (k h ) . 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 ) &
at: y =
2
252
For small values of a , the continuity equation and the momentum equation can be given in a
linearised form:
+ h
=0
t
y
v
+g
+ g = 0
t
y
The solution of the surface elevation in these equations, satisfying the boundary values for
v , is:
(s + h ) 2
b 0 1 +
g
sin y
=h
b
cos
0
Now, the roll moment follows from the quasi-static moment of the mass of the frozen liquid
l b h and an integration of over the breadth of the tank:
M = g l b h s +
+b 2
h
+ g l y dy
2
b 2
This delivers the roll moment amplitude for low and high frequencies at small water depths:
h
M a = g l b h s + a +
2
3
2
0
(
s + h ) 2 0
3
g l b 1 +
tanh
a
2
2 0
For very low frequencies, so for the limit value 0 , this will result into the static moment:
h b3
M = g l b h s + + a
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 for high frequencies:
M = 0
<< 0
for:
M =
>> 0
8.2.1.2 Natural Frequency Region
For frequencies near to the natural frequency 0 , the expression for the surface elevation of
the fluid goes to infinity. Experiments showed the appearance of a hydraulic jump or a bore
at these frequencies. Obviously, then the linearised equations are not valid anymore.
Verhagen and van Wijngaarden [1965] solved the problem by using the approach in gas
dynamics, when a column of gas has been oscillated at small amplitude, e.g. by a piston. At
253
frequencies near to the natural frequency at small water depths, they found a roll moment
amplitude, defined by:
4
2
l b3 4
2 a h 2 b ( 0 )
M a = g
1
12
3b
32 g a
The phase shifs between the roll moment and the roll motion at small water depths are given
by:
M = 2 +
<< 0
for:
M = 2
>> 0
with:
2 b ( )2
0
= 2 arcsin
24 g a
2 b ( 0 )2
arcsin
2
2
96 g a 3 b ( 0 )
Because that the arguments of the square roots in the expression for M have to be positive,
the limits for the frequency are at least:
0
24 g a
24 g a
< < 0 +
2
b
b 2
254
The roll moments have been calculated here for low and high frequencies and for frequencies
near to the natural frequency of the tank. A calculation routine connects these three regions.
255
a4 = 0
b4
c4
Kta
sin t
a
=
K
= ta cos t
a
It is obvious that for an anti-rolling free-surface tank, built into a ship, it holds:
a = a
and
e =
So it can be written:
= a cos ( t + )
Kt = Kta cos ( t + + t )
Then, an additional moment has to be added to the right-hand side of the equations of motion
for roll:
X 4 tank = a 44 tank && + b44 tank & + c44 tank
with:
256
a44 tank = 0
Kta
sin t
a
b44 tank =
Kta
c44 tank =
cos t
a
This holds that the anti-rolling coefficients a44 tank , b44 tank and c44 tank have to be subtracted
from the coefficients a 44 , b44 and c44 in the left-hand side of the equations of motion for roll.
30
T ra w ler
T r aw l er
L = 2 3.9 0 m e tre
L = 2 3.9 0 m e tre
W itho ut ta nk
25
W i th ou t ta nk
30
20
W ith tan k
20
W ith tan k
15
10
10
5
0
0
0.5
1.0
1 .5
2 .0
2.5
S ig nific a nt w a v e he ig ht (m )
257
&& + b6 & + c6
X 6 fin = a6
with:
a2 = 2 sin a
b2 = 2 sin b
c2 = 2 sin c
c
1
a = s fin fin
2
2
C
+ L C (k )
fin
1
C
c = V 2 Afin L C (k )
2
fin
In here:
c
1
V Afin fin
2
2
C L
fin
C (k )
c
k= e r
2 V
Afin
s fin
span of fin
c fin
The nominal lift curve slope of a fin profile in a uniform flow is approximated by:
CL
1.80 ( ARE )
=
2
(
ARE )
1.80 + cos
+ 4.0
cos 4
with:
V ( ) = V
with: < BL
BL
Vx
0 .2
BL = 0.377 x fin Rx
with: Rx =
in which:
V ( ) flow velocity inside boundary layer
V
forward ship speed
The kinematic viscosity of seawater can be found from the water temperature T in degrees
centigrade by:
1.78
106 =
m2 /s
2
1.0 + 0.0336 T + 0.000221 T
It is assumed here that the total lift of the fin can be found from:
s fin
1
1
CL V 2 ( ) c ( ) d = CLfin V 2 Afin
2
2
0
where c ( ) is the chord at span-wise location .
For rectangular fins, this is simply an assumption of a uniform loading.
Because:
crfin + ctfin
mean chord of fin
2
the correction to the lift curve slope is:
2
2 crfin ctfin
BL
EBL = rfin 1
1 BL
c fin 9 s fin
2
8 s fin
c fin =
= EBL
fin
1.80 + ( ARE ) fin 2 + 4.0
Generally a fin is mounted close to the hull, so the effective aspect ratio is about twice the
geometric aspect ratio:
s
( ARE ) fin = 2 ( AR ) fin = 2 fin
c fin
260
a2 = + a
b2 = +b
c2 = +c
a4 = zb rud a
b4 = zb rud b
b4 = zb rud c
a6 = + xb rud a
b6 = + xb rud b
c6 = + xb rud c
and:
3
1
c
a = s rud rud
2
2
C
+ L C (k )
rud
1
C
c = V 2 Arud L C (k )
2
rud
b =
1
c
V Arud rud
2
2
In here:
Vrud 1.125 V
C L
rud
c
k = e rud
2 V
Arud
srud
crud
=
rud 1.80 + ( ARE ) 2 + 4.0
rud
Generally a rudder is not mounted close to the hull, so the effective aspect ratio is equal to the
geometric aspect ratio:
( ARE )rud = ( AR)rud = srud
crud
262
Suppose a linear spring connected to point P on the ship, see Figure 8.41.
The linear spring coefficients in the three directions in a certain point P are defined by
(C px , C py , C pz ) . The units of these coefficients are N/m or kN/m.
263
Surge:
c12 = 0
c13 = 0
c14 = 0
c15 = + C px z p
c16 = C px y p
c21 = 0
c22 = +C py
Sway:
c23 = 0
c24 = C py z p
c25 = 0
c26 = +C py x p
c31 = 0
c32 = 0
Heave:
c33 = +C pz
c34 = +C pz y p
c35 = C pz x p
c36 = 0
c41 = 0
c42 = C py z p
Roll:
c43 = +C pz y p
c44 = +C py z p 2 + C pz y p 2
c45 = C pz x p y p
c46 = C py x p z p
264
c51 = +C px z p
Pitch:
c52 = 0
c53 = C pz x p
c54 = C pz x p y p
c55 = +C px z p + Cpz x p
2
c56 = C px y p z p
c61 = C px y p
c62 = +C py x p
Yaw:
c63 = 0
c64 = C py x p z p
c65 = C px y p z p
c66 = +C px y p 2 + C py x p 2
It is obvious that in case of several springs, a linear superposition of the coefficients can be
used.
When using linear springs, generally 12 sets of coupled equations with the in and out of phase
terms of the motions have to be solved. Because of these springs, the surge, heave and pitch
motions will be coupled then with the sway, roll and yaw motions.
265
Displaceme nt
266
Resistance (kN)
2 00 0
Re sist an ce
1 50 0
S till wat er re sista n ce R SW
1 00 0
50 0
0
0
10
20
30
T im e (s)
an integrated pressure method, as introduced by Boese [1970], suitable for all wave
directions.
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:
267
Raw
a 2
In a non-dimensional way the transfer function of the mean added resistance is presented as:
Raw
Raw" =
g a 2 B 2 L
in which:
L
lengt h between perpendiculars
B
maximum breadth of the waterline
Both methods will be described here.
268
P=
'
33
*2
Vz dxb dt
0 L
in which:
'
b33
Vz
Te
The speed dependent hydrodynamic damping coefficient for the vertical motion of a cross
section is defined here as showed before:
dM 33 '
b33' = N33' V
dxb
The harmonic vertical relative velocity of a point on the ship with respect to the water
particles is defined by:
D
'
(z xb + yb )
Vz = &w3
Dt
= &w3 ' z& xb & + V + yb &
For a cross section of the ship, an equivalent harmonic vertical relative velocity has to be
found, defined here by:
Vz * = &w 3* z& xb & + V
= Vza cos e t + V *
*
With this the radiated energy during one period of oscillation is given by:
'
'
dM 33
*2
P=
N33 V
V
dxb
za
e L
dxb
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:
c
P = Raw V
T
cos e
= Raw
2
k cos
From this the transfer function of the mean added resistance according to Gerritsma and
Beukelman can be found:
269
Equation 10.11
This method gives good results in head to beam waves. However, in following waves this
method could fail.
When the wave speed in following waves approaches the ship speed the frequency of
encounter in the denominator tends to zero, e 0 . At these low frequencies, the potential
sectional mass is very high and the potential sectional damping is almost zero. The damping
multiplied 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 an extreme
positive or negative added resistance.
270
cosh[k h ]
cosh[k (h + zb )]
= g z + g
a cos (e t k xb cos k yb sin )
cosh[k h ]
The horizontal force on an oscillating cross section is given by:
f (xb , t ) =
p dz
D s +z x
2 + ( Ds + z x )2
= g
+
( + Ds z x )
2
tanh [k h]
with: z x = z xb .
As the mean added resistance during one period will be calculated, the constant term and the
first harmonic term can be ignored. So:
2 + z x 2 ( z x )
*
f ( xb , t ) = g
+
2
tanh [k h ]
Te
f ( x , t ) dt
*
g a 2
z 2 2 sa cos ( k xb cos s )
1 + xa2 +
2
a
a tanh [k h]
271
dy w
Raw1 = 2 f * ( xb )
dxb
dxb
L
g a 2
z 2 2 sa cos( k xb cos s ) dy w
=
1 xa2
dx dxb
2
a tanh [k h]
a
b
L
where yw is the still water line.
For deep water, this part of the mean added resistance reduces to:
g
dy
Raw1 =
sa 2 w dxb
(as given by Boese for deep water)
2
dxb
L
The integrated vertical hydromechanical and wave forces in the ship-bounded system vary not
only in time but also in direction with the pitch angle.
From this follows a second contribution to the mean added resistance:
T
Raw 2
1 e
=
{Z (t ) + Z w (t )} (t ) dt
Te 0 h
T
1 e
=
&z&(t ) (t ) dt
Te 0
For this second contribution can be written:
1
Raw 2 = e 2 za a cos(z )
2
So the transfer function of the total mean added resistance according to Boese is given by:
Raw 1
z xa 2 2 sa cos ( k xb cos s ) dy w
dxb
=
L a 2
a 2 2
a tanh [k h ]
dxb
+
1
e 2 za a cos(z )
2
Equation 10.21
272
273
274
The input values for the calculation of shear forces and bending and torsion moments are
often more or less inaccurate. Mostly small adaptations are necessary, for instance to avoid a
remaining calculated bending moment at the forward end of the ship.
The total mass of the ship is found by an integration of the mass per unit length:
m = m' ( xb ) dxb
Lm
It is obvious that this integrated mass should be equal to the mass of displacement, calculated
from the underwater hull form:
m =
Both terms will be calculated from independently derived data, so small deviations are
possible. A proportional correction of the masses per unit length m' (xb ) can be used, see
Figure 10.33.
Then m' (xb ) will be replaced by:
m ' (xb )
m
32 ( xB xG )
Lm 3
In here:
xA
Lm
In here:
xA
Lm
277
'
The transverse radius of inertia k xx is found from the distribution of the radii of inertia of the
masses per unit length:
2
1
k yy 2 = m' ( xb ) k xx ' dxb
m Lm
If this value of k xx differs from a required value k xx (new) of the radius of inertia, a
proportional correction of the longitudinal distribution of the radii of inertia can be used:
k xx ' (new)
'
'
k xx ( xb , new) = k xx ( xb , old ) '
k xx (old )
Consider a section of the ship with a length dxb to calculate the shear forces and the bending
and the torsion moments.
278
Q(x1 ) =
x1
M (x1 ) = Q( xb ) dxb
x0
xb dQ( x )
b
= +
dxb dxb
dxb
x0 x0
x1
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:
279
x1
Q(x1 ) = q( xb ) dxb
x0
x1
M (x1 ) = + q( xb ) ( x1 xb ) dxb
x0
x1
x1
x0
x0
= + q( xb ) xb dxb x1 q( xb ) dxb
For the torsion 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 differ generally, so numerical integrations of these two terms have to be carried out
separately.
280
Q3 sw ( x1 ) = q3 sw (xb ) dxb
x0
x1
x1
x0
x0
Q5 sw ( x1 ) = + q3 sw ( xb ) xb dxb x1 q3 sw ( xb ) dxb
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 showed in the following Sections.
281
'
22
V dN 22 '
+ 2
e
dxb
dM 22 '
dxb
c22 ' = 0
a24 ' = +M 24' + OG M 22 ' +
b24 = + N
'
'
24
V dN24 ' V
dN 22 '
OG
e 2 dxb
e 2
dxb
dM 24 '
dM 22 '
'
V
+ OG N 22 V
dxb
dxb
c24 = 0
'
'
V
e 2
dM 22 ' V
V dN 22 '
'
N22 ' V
+
N
+
xb
22
2
2
dx
dx
b
b
e
e
'
'
'
dM 22
V 2 dN22
'
= + N 22 V
x 2 V M 22 + 2
dxb b
dxb
e
c26 ' = 0
282
'
&&w 2 * +
V
dN 22 ' && *
w2
e dxb
dM 22 ' & *
'
+
N 22 V
e
dxb w 2
'
+ X FK 2
The ''Modified 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 q 2 ( xb ) by the following integrals:
Q2 ( x1 ) = Q2 a cos e t + Q 2
x1
q2 (xb ) dxb
x0
Q6 ( x1 ) = Q6 a cos e t + Q 6
x1
x1
x0
x0
+ q2 ( xb ) xb dxb x1 q2 ( xb ) dxb
283
h3
w3
(
(
)
)
V dN11'
e 2 dxb
b11 = + N11
dM 11
+ b11V
dxb
'
'
'
c11' = 0
a13' = 0
b13' = 0
c13' = 0
a15 ' = M 11' bG +
'
V dN11
bG
2
e dxb
'
'
dM 11
b15 = N11 V
bG b11V bG
dxb
'
c15 = 0
'
284
a33 = + M 33
'
'
V dN 33'
+ 2
dxb
e
b33 = + N33 V
'
'
dM 33 '
dxb
c33 = +2 g yw
'
V
e 2
dM 33 ' V
V dN '
+ 2 N33 ' + 2 33 xb
N33 '
dxb e
e
dxb
dM 33'
V 2 dN33 '
'
b35 ' = N33 ' V
dx
+
2
dx
+
b
33
b
2
b
dx
dxb
b
L
L
e
c35 ' = 2 g yw xb
The sectional wave loads for surge and heave are given by:
X w1 ' = + M 11' &&w1* +
V
dN '
11 &&w1*
e dxb
dM 11 ' & *
+
N11' V
e
dxb w1
'
+ X FK1
'
'
*
X w 3 = + M 33 &&w 3 +
V
dN 33' && *
w 3
e dxb
dM 33 ' & *
'
+
N33 V
e
w3
dx
b
+ X FK 3 '
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
285
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 11.32 shows a comparison between measured and calculated distributions of the
vertical wave bending moment amplitudes over the length of the ship.
286
&& + g )
m' ( xb ) k xx ' && z m ' ( &y& + xb
+ z1 q2 ( xb )
2
287
V dN 24 '
e 2 dxb
'
b42 = + N 42
'
dM 42
V
+ OG b22 '
dxb
'
c 24 ' = 0
a 44 ' = + M 44 ' + OG M 42 ' + OG a 24 ' +
b44 = + N 44
'
c 44
'
b46
'
dM 44 '
dM 42 '
'
V
+ OG N 42 V
dxb
dxb
'
y w 3 As
= +2 g
+
bG
2
3
a 46 = + M 42
'
V dN 44 '
V
dN 42 '
OG
dx b
e 2 dxb
e 2
'
V
xb + 2
e
dM 42 '
V
V dN 42 '
'
'
'
N 42 V
+ OG a 26 + 2 N 42 + 2
xb
dxb
dx b
e
e
dM 42 '
'
= + N 42 V
dxb
'
2
dxb
e
V
dN 42 ' && *
w2
e dxb
dM 42 ' & *
'
+
N 42 V
e
w2
dx
b
+ X FK 4 ' + OG X w2 '
The ''Modified Strip Theory'' includes the outlined terms. When ignoring the outlined terms
the ''Ordinary Strip Theory'' is presented.
Then the harmonic torsion moments Q4 ( x1 , z1 ) in waves in cross section x1 at a distance z1
above the centre of gravity can be obtained from the torsion load q4 (xb , z1 ) by the following
integral:
Q4 ( x1 , z1 ) = Q4 a cos e t + Q 4
x1
= q4 (xb , z1 ) dxb
x0
288
289
3832 H1 / 3
S () =
6 exp 2 2
5
T1
T1
290
with:
A = 0.658
2
1.0
p
B = exp
2
= 3.3
(peakedness factor)
2
p =
(circular frequency at spectral peak)
Tp
= a step function of :
12.1.4Definition of Parameters
The nth order spectral moments of the wave spectrum, defined as a function of the circular
wave frequency , are:
291
mn = S () n d
0
= 1
m2
m0 m4
= 1.073 T2
=
T2
= 1.287 T2
= 0.772 Tp
Truncation of wave spectra during numerical calculations can cause differences between input
and calculated wave periods. Generally, the wave heights will not differ much.
In Figure 12.12 and Table 12.11 - 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 significant wave height H1 / 3 and the
average wave periods T1 or T2 .
Notice that these data are an indication only. A generally applicable fixed relation between
wave heights and wave periods does not exist.
292
293
Other open ocean definitions for the North Atlantic and the North Pacific, obtained from
Bales [1983] and adopted by the 17th ITTC (1984), are given in Table 12.12. The modal or
central periods in these tables correspond with the peak period T p . For not-truncated spectra,
the relations with T1 and T2 are defined before.
294
r
S r () = a S ()
a
or
r
S r (e ) = a
a
S (e )
This has been visualized for a heave motion in Figure 12.21and Figure 12.22.
295
W av e
s pe c tru m
H 1/3 = 5 .0 0 m
W av e
s p ec trum
T 2 = 8.00 s
H 1/3 = 5.00 m
T = 8 .0 0 s
2
0
0
2.0
0.5
1.0
2 .0
0
2 .0
T ra ns fer
fun c ti on
h eav e
1.5
1 .5
C onta ine rs hi p
L = 17 5 m e tre
H ead w av e s
V = 20 k no ts
1.0
0.5
1.0
T ran s fer
fu nc tio n
he av e
1 .5
1.5
2.0
C on tain ers h ip
L = 1 75 m etre
H ea d w av e s
V = 2 0 k n ots
1 .0
0.5
0 .5
0
0
0.5
1.0
2 .0
0.5
1.0
1.5
2.0
H e av e
s p ec tru m
1 .5
a1 /3
= 1.92 m
z
6
T z = 7.74 s
a1/3
= 1 .92 m
T z = 7 .74 s
0
0
0.5
1.0
1 .5
2 .0
0.5
1 .0
1.5
2 .0
fr eq ue nc y of en c ou nter ( rad /s )
Figure 12.22: Heave Spectra in the Wave and Encounter Frequency Domain
The moments of the response spectrum are given by:
mnr = Sr (e ) e de
n
From the spectral density function of a response the significant amplitude can be calculated.
The significant amplitude is defined to be the mean value of the highest one-third part of the
highest wave heights, so:
ra 1 / 3 = 2 m0 r
A mean period can be found from the centroid of the spectrum by:
m
T1 r = 2 0 r
m1 r
Another definition, which is equivalent to the average zero-crossing period, is found from the
spectral radius of inertia by:
m0 r
T2 r = 2
m2 r
The probability density function of the maximum and minimum values, in case of a spectrum
with a frequency range that is not too wide, is given by the Rayleigh distribution:
296
r2
ra
exp a
m0 r
2 m0 r
This implies that the probability of exceeding a threshold value a by the response amplitude
ra becomes:
ra 2
ra
P{ra > a} =
exp
dra
m0 r
a
2 m0 r
a2
= exp
2 m0 r
The number of times per hour that this happens follows from:
3600
N hour =
P{ra > a}
T2 r
The spectral value of the waves S (e ) , based on e , is not equal to the spectral value
S () , based on . Because of the requirement of an equal amount of energy in the
297
So:
d
=
dk
kg
h cosh 2 [k h ]
2 k g tanh [k h]
g tanh [k h] +
As can be seen in Figure 12.23, in following waves the derivative de 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 infinite to minus infinite. 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:
m0r
m1r
=
=
S ( ) d
r
S ( ) d
S ( )
m2r
de
S () d
r
S () d
r
S ()
r
with:
2
r
S r () = a S ()
a
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 de d . So an integration of Sr (e ) over e has to be avoided.
Because of the linearities, the calculated significant values can be presented by:
ra 1 / 3
versus T1 or T2
H1 / 3
with:
H1 / 3 significant wave height
T1 ,T2 average wave periods
298
R
RAW = 2 aw2 S () d
0 a
Because of the linearities of the motions, the calculated mean added resistance values can be
presented by:
RAW
versus T1 or T2
H1/ 32
299
sa
3
g
with the restrictions:
block coefficient:
Froude number:
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:
*
sa = sa + s a
Then, shipping green water is defined by:
*
sa > f e at the forward perpendicular
The spectral density of the vertical relative motion at the forward perpendicular is given by:
s*
Ss * () = a S ()
a
2
mns * = S s * () e d
n
300
When using the Rayleigh distribution the probability of shipping green water is given by:
f e 2
P sa * > f e = exp
2 m0 s *
The average zero-crossing period of the relative motion is found from the spectral radius of
inertia by:
m0 s *
T2 s * = 2
m2 s *
The number of times per hour that green water will be shipped follows from:
3600
*
N hour =
P s a > fe
T2 s *
301
12.4.1Criterium of Ochi
Ochi [1964] translated slamming phenomena into requirements for the vertical relative
motions of the ship.
He defined slamming by:
an emergence of the bow of the ship at 10 percentile of the length aft of the forward
perpendiculars
an exceeding of a certain critical value at the instance of impact by the vertical relative
velocity, without forward speed effect, between the wave surface and the bow of the ship
Ochi defines the vertical relative displacement and velocity of the water particles with respect
to the keel point of the ship by:
s = x b z + xb
s& = & z& + x &
xb
with:
xb = a cos(e t k xb cos )
& xb = e a sin (e t k xb cos )
So a forward speed effect (V -term) is not included in this vertical relative velocity. The
spectral moments of the vertical relative displacements and velocities are defined by m0 s and
m0 &s .
Emergence of the bow of the ship happens when the vertical relative displacement amplitude
sa at 0.90 L is larger than the ship's draft Ds at this location.
The probability of emergence of the bow follows from:
Ds 2
P{sa > Ds } = exp
2 m0 s
The second requirement states that the vertical relative velocity exceeds a threshold value.
According to Ochi, 12 feet per second can be taken as a threshold value for a ship with a
length of 520 feet.
Scaling results into:
s&cr = 0.0928 g L
The probability of exceeding this threshold value is:
302
s& 2
P{s&a > s&cr } = exp cr
2 m0 s&
Both occurrences, emergence of the bow and exceeding the threshold velocity, are statistically
independent. In case of slamming both occurrences have to appear at the same time.
So the probability on a slam is the product of the both independent probabilities:
P{slam} = P{s a > Ds } P{s&a > s&cr }
2
Ds2
s&cr
= exp
+
2 m0 s 2 m0 s&
12.4.2Criterium of Conolly
Conolly [1974] translated slamming phenomena into requirements for the peak impact
pressure of the ship.
He defined slamming by:
an emergence of the bow of the ship
an exceeding of a certain critical value by the peak impact pressure at this location.
The peak impact pressure is defined by:
1
p = C p s&cr 2
2
The coefficient C p has been taken from experimental data of slamming drop tests with
wedges and cones, as given in literature.
Some of these data, as for instance presented by Lloyd [1989] as a function of the deadrise
angle , are illustrated in Figure 12.41.
The accessory draught t of the wedge follows from the section contour. In the fore body 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 define 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.
0 2
Critical peak impact pressures pcr have been taken from Conolly [1974]. He gives measured
impact pressures at a ship with a length of 112 meter over 30 per cent of the ship length from
forward. From this, a lower limit of pcr has been assumed. This lower limit is presented in
Figure 12.43.
This approach can be translated into local hull-shape-depending threshold values of the
vertical relative velocity too:
2 pcr
s&cr =
Cp
The vertical relative velocity, including a forward speed effect, of the water particles with
respect to the keel point of the ship is defined by:
D
s& =
z + xb
Dt xb
= & xb z& + xb & V
with:
xb = a cos(e t k xb cos )
& = sin ( t k x cos )
xb
Then:
2
Ds 2
s&cr
P{slam} = exp
+
2 m0 s 2 m0 s&
Notice that, because of including the forward speed effect, the spectral moment of the
velocities does not follow from the spectral density of the relative displacement as showed in
the definition of Ochi.
The average period of the relative displacement is found by:
m0s
m0 s
T2 s = 2
= 2
m2s
m0 s&
Then the number of times per hour that a slam will occur follows from:
3600
N hour =
P{slam}
T2 s
305
306
13 Twin-Hull Ships
When not taking into account the interaction effects 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 (2 yT ) should be constant. The co-ordinate system for the equations of
motion of a twin-hull ship is given in Figure 12.41.
307
T &x&
T &y&
T &z&
&&
ITxx && I Txz
ITyy &&
&&
&& I Tzx
ITzz
X Th 1
X Th 2
X Th 3
X Th 4
X Th 5
X Th 6
=
=
=
=
=
=
X Tw1
X Tw 2
X Tw 3
X Tw 4
X Tw 5
X Tw 6
in which:
T
I Tij
X Th1 , X Th 2 , X Th 3
X Th 4 , X Th 5 , X Th 6
X Tw1 , X Tw 2 , X Tw3
X Tw 4 , X Tw5 , X Tw 6
308
X Th 2
X Th 3
X Th 4
= + 2 a42 &y&
+ 2 a44 &&
+ 2 yT 2 a33 &&
&&
+ 2 a46
X Th 5
X Th 6
= + 2 a62 &y&
+ 2 a64 &&
&&
+ 2 a66
+ 2 b42 y&
+ 2 b44 &
+ 2 yT 2 b33 &
+ 2 b46 &
+ 2 c42 y
+ 2 c44
+ 2 yT 2 c33
+ 2 c46
+ 2 b62 y&
+ 2 b64 &
+ 2 b66 &
+ 2 c62 y
+ 2 c64
+ 2 c66
2
2
&& + 2 yT b11 & + 2 yT 2 c11
+ 2 yT a11
309
cosh [k h]
This holds that for the port side ( ps ) and starboard (sb ) hulls the equivalent components of
the orbital accelerations and velocities in the surge, sway, heave and roll directions are equal
to:
*
*
&&w1 ( ps ) = k g cos a 1 sin (e t k xb cos k yT sin )
&&w1* (sb ) = k g cos a 1* sin (e t k xb cos + k yT sin )
+ k g cos
*
*
&w1 ( ps ) =
a 1 cos(e t k xb cos k yT sin )
+ k g cos
&w1* (sb ) =
a 1* cos(e t k xb cos + k yT sin )
*
*
&&w 2 ( ps ) = k g sin a 2 sin (e t k xb cos k yT sin )
&&w 2 * (sb ) = k g sin a 2 * sin (e t k xb cos + k yT sin )
+ k g sin
*
*
&w 2 ( ps ) =
a 2 cos (e t k xb cos k yT sin )
+ k g sin
&w 2 * (sb ) =
a 2 * cos (e t k xb cos + k yT sin )
*
*
&&w 3 ( ps ) = k g a 3 cos (e t k xb cos k yT sin )
&&w 3* (sb ) = k g a 3 * cos (e t k xb cos + k yT sin )
+kg
*
*
&w 3 ( ps ) =
a 3 sin (e t k xb cos k yT sin )
+
k
g
&w 3* (sb ) =
a 3* sin (e t k xb cos + k yT sin )
From this follows the total wave loads for the degrees of freedom. In these loads on the
following pages, the ''Modified Strip Theory'' includes the outlined terms. When ignoring
these outlined terms the ''Ordinary Strip Theory'' is presented.
310
'
*
*
X Tw1 = + M 11 &&w1 ( ps ) + &&w1 ( sb ) dxb
L
V
dN '
11 &&w1* ( ps ) + &&w1* (sb ) dxb
e L dxb
dM 11' & *
'
+
N11 V
w1 ( ps ) + &w1* (sb ) dxb
e
dxb
L
'
*
*
X Tw 2 = + M 22 &&w 2 ( ps ) + &&w 2 ( sb ) dxb
L
V
dN 22 ' && *
+
dM 22 ' & *
+
N22 ' V
w 2 ( ps ) + &w 2 * ( sb ) dxb
e
dxb
L
'
*
*
X Tw3 = + M 33 &&w3 ( ps ) + &&w 3 (sb ) dxb
L
V
dN '
33 &&w 3* ( ps ) + &&w3 * (sb ) dxb
e L dxb
dM 33' & *
+
N 33' V
w3 ( ps ) + &w 3* (sb ) dxb
e
dxb
L
311
V
dN '
42 &&w2 * ( ps ) + &&w 2* (sb ) dxb
e L dxb
'
dM 42 & *
'
*
+
N42 V
w 2 ( ps ) + &w 2 (sb ) dxb
e
dxb
L
+ OG X Tw 2 + yT X Tw 3
V
dN11'
*
*
+
N11 V
bG &w1* ( ps ) + &w1* (sb ) dxb
e
dxb
L
'
V
dN
*
*
33 xb &&w3 ( ps ) + &&w 3 (sb ) dxb
e L dxb
'
dM 33 '
N33 ' V
xb &w 3* ( ps ) + &w3 * (sb ) dxb
e
dxb
L
X FK 3 ( ps ) + X FK 3 ( sb ) xb dxb
'
'
312
V
dN '
22 xb &&w 2* ( ps ) + &&w 2* ( sb ) dxb
e L dxb
dM 22 '
'
+
N22 V
xb &w 2 * ( ps ) + &w 2 * (sb ) dxb
e
dxb
L
+ yT X Tw1
313
with:
*
*
Vz ( ps ) = &w3 ( ps ) z& xb & + V yT &
Vz * (sb ) = &w3 * (sb ) z& xb & + V + yT &
(
(
)
)
+kg
*
*
&w 3 ( ps ) =
a 3 sin (e t k xb cos k yT sin )
+kg
&w 3* (sb ) =
a 3* sin (e t k xb cos + k yT sin )
Raw 1
z xa 2 2 sa cos ( k xb cos s ) dyw
dxb
2
L a 2
2
a tanh [k h ]
a
ps dxb
2
2 s a cos( k xb cos s ) dy w
1
z xa
dxb
g 1 2
tanh
[
k
h
]
a
L
a
sb dxb
1
e 2 za a cos(z ) ps
2
1
+ e 2 za a cos(z )sb
2
+
with:
314
qT 2 (xb ) = + X Th 2 ( xb ) + X Tw 2 ( xb ) + 2 g As
&& + g
&& z m '
mT ' (xb ) &y& + xb
'
'
qT 3 (xb ) = + X Th 3 ( xb ) + X Tw 3 (xb )
mT ' (xb ) &z& xb &&
'
'
qT 4 ( xb , z1 ) = + X Th 4 ' ( xb ) + X Tw 4 ' ( xb )
&& z (&y& + x
&& + g )
mT (xb ) kTxx
m
b
+ z1 qT 2 (xb )
'
In here:
mT '
kTxx
As
'2
'
The calculation procedure of the forces and moments is similar to the procedure given before
for mono-hull ships.
315
316
14 Numerical Recipes
Some typical numerical recipes, as used in the strip theory program SEAWAY, are described
in more detail here.
14.1 Polynomials
Discrete points can be connected by a first degree or a second degree polynomial, see Figure
14.11-a,b.
Figure 14.11: First and Second Order Polynomials Through Discrete Points
Notice that only one interval is required for obtaining the coefficients in that interval.
317
x p x0
x0 xm
a=
x p xm
b=
f (x p ) f ( x0 )
a (x p + x0 )
x p x0
c = f ( x0 ) a x0 2 b x0
Notice that two intervals are required for obtaining these coefficients, valid in both intervals.
318
14.2 Integrations
Either the trapezoid rule or Simpsons general rule can carry out numerical integration.
SEAWAY uses Simpson's general rule as a standard. Then, the integration has to be carried
out over a number of sets of two intervals, see Figure 14.11-b. Numerical inaccuracies can
be expected when x0 xm << x p x0 or x p x0 << x0 xm . In those cases the trapezoid rule
has to be preferred, see Figure 14.11-a.
SEAWAY makes the choice between the use of the trapezoid rule and Simpson's rule
automatically, based on the following requirements:
x0 xm
x0 xm
Trapezoid rile, if:
< 0.2
or
> 5.0
x p xm
x p xm
x xm
Simpsons rule, if:
0.2 < 0
< 5.0
x p xm
xp
x0
x0
f (x ) dx = (a x + b ) dx
x
p
1
= a x2 + b x
2
x0
with:
a=
f (x p ) f ( x0 )
x p x0
b = f ( x0 ) a x0
xp
f ( x ) dx = (a x
xm
+ b x + c dx
xm
p
1
1
3
2
= a x + b x + c x
2
3
xm
with:
f (x p ) f ( x0 ) f (x0 ) f ( xm )
x p x0
x0 xm
a=
x p xm
f (x p ) f ( x0 )
a (x p + x0 )
x p x0
b=
c = f ( x0 ) a x0 2 b x0
Some algebra leads for the integration over these two intervals to:
x p x0
x0 xm 2
f ( xm ) +
x0 x m
xp
2
x x
(
x p xm )
p
m
(
)
(
)
x f x dx = 2 ( x0 xm ) (x p x0 ) f x0 + 3
x xm
x p x0 0
f (xp )
x p x0
The in-phase and out-phase parts of the total wave loads have to be obtained from
longitudinal integration of sectional values. Direct numerical integration of
'
'
Fwa cos F ' dxb and Fwa sin F ' dxb over the ship length, L , require integration
intervals, xb , which are much smaller than the smallest wave length, xb min 10 . This
means that a large number of cross sections are required.
This can be avoided by writing the integrands in terms of f 1 ( x ) cos x dx and f 2 ( x ) sin x dx ,
in which the integrands f 1, 2 ( x ) vary much slower over short wave lengths as the harmonics
itself.
320
2 sin x
and:
sin x dx = cos x
x sin x dx = sin x x cos x
x sin x dx = +2 x sin x (x
2
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 x p , so over the two intervals x0 xm and x p x0 :
xp
xp
xm
xm
F (x ) dx = f (x ) cos x dx
xp
(a x
+ b x + c cos x dx
xm
xp
xp
xp
xm
xm
xm
xp
xp
xm
xm
F (x ) dx = f (x ) sin x dx
xp
(a x
+ b x + c sin x dx
xm
xp
xp
xp
xm
xm
xm
321
f (x p ) f ( x0 ) f (x0 ) f ( xm )
x p x0
x0 xm
a=
x p xm
b=
f (x p ) f ( x0 )
a (x p + x0 )
x p x0
c = f ( x0 ) a x0 2 b x0
With this approach, the wave loads on a barge, for instance, can be calculated by using two
section intervals only for any length of the barge.
322
14.3 Derivatives
First and second degree functions, of which the derivatives have to be determined, have been
given in Figure 14.31-a,b.
dx x = x 0 m (minus or left of zero dx x = x0 p (plusor right of zero
A mean derivative df dx at x = x0 can be obtained by:
df
=
dx x = x0
+ (x x ) df
p
0
dx x = x0 m
dx x = x 0 p
( x0 xm ) df
x p xm
The two polynomials - each valid over two intervals below and above x = x0 - are given by:
f (x ) = am x 2 + bm x + cm
for x < x0 :
for x > x0 :
f (x ) = a p x 2 + bp x + c p
df
dx x = xm 2
2 (xm 1 xm 2 ) ( x0 xm1 ) { f ( xm 1 ) f ( xm 2 )}
2
( xm1 xm 2 ) { f (x0 ) f ( xm 1 )}
2
+ ( x0 xm1 ) { f ( xm1 ) f ( xm 2 )}
=
( xm1 xm2 ) ( x0 xm1 ) (x0 xm 2 )
df
dx x = x m1
+ ( xm1 xm 2 )2 { f ( x0 ) f (xm 1 )}
+ ( x0 xm1 )2 { f ( xm1 ) f ( xm 2 )}
=
( xm1 xm2 ) (x0 xm1 ) (x0 xm 2 )
df
dx x = x0
+ 2 ( xm1 xm 2 ) ( x0 xm 1 ) { f (x0 ) f ( xm 1 )}
2
+ ( xm1 xm 2 ) { f ( x0 ) f (xm 1 )}
2
( x0 xm1 ) { f ( xm1 ) f (xm 2 )}
=
(xm1 xm 2 ) ( x0 xm1 ) ( x0 xm2 )
324
df
dx x = xm 2
2 (x p 1 xm 2 ) (x p 2 x p 1 ) { f (x p 1 ) f (x0 )}
2
(
)
{
(
)
(
)
}
f
x
f
x
p1
0
p2
p1
2
+ (x p 2 x p 1 ) { f (x p1 ) f ( x0 )}
=
(x p1 x0 ) (x p 2 x p1 ) (x p2 x0 )
df
dx x = x m1
+ (x p 1 x0 )2 { f (x p 2 ) f (x p 1 )}
2
+ (x p 2 x p 1 ) { f (x p1 ) f ( x0 )}
=
(x p1 x0 ) ( x p2 x p1 ) (x p2 x0 )
df
dx x = x0
+ 2 (x p 1 x0 ) (x p 2 x p1 ) { f (x p 2 ) f (x p 1 )}
2
+ (x p 1 x0 ) { f (x p 2 ) f (x p 1 )}
2
(x p 2 x p 1 ) { f (x p1 ) f ( x0 )}
=
(x p1 x0 ) (x p 2 x p1 ) (x p 2 x0 )
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 differ:
df
df
dx x = x 0 m (minus or left of zero dx x = x0 p (plusor right of zero
A mean derivative df dx at x = x0 can be obtained by:
df
dx x = x0
df
df
d m + d p
dx x = x0 m
dx x = x 0 p
=
dm + d p
with:
x xm 2
x0 xm1 m1
( x0 x m 2 )
2
dm =
3 ( x0 xm1 )
x x p1
x p1 x0 p 2
(x p 2 x0 )
2
dp =
3 (x p 1 x0 )
325
smp = ds
xm
xp
(dx ) + (dy )
2
xm
( x0 xm )2 + ( y0 ym )2 + (x p x0 )2 + ( y p y0 )2
+ p 1 + p 2 + ln p + 1 + p 2
0
0
0
0
= p2
p1 1 + p1 2 ln p1 + 1 + p12
with:
326
cos =
sin =
(x
(x
x p xm
xm )2 + ( y p ym )2
yp ym
xm )2 + ( y p ym )2
p0 = + 2
(x
(x
xm ) cos + ( y p y m ) sin
p x0 ) cos + ( y p y0 ) sin
p1 =
(x
p=
(x
x0 ) cos + ( y p y0 ) sin
x0 ) cos + ( y p y0 ) sin
4
327
328
15 References
Abramowitz and Stegun [1964]
M. Abramowitz and I.A. Stegun,
Handbook of Mathematical Functions with Formulas and Mathematical Tables,
National Bureau of Standards, U.S. Department of Commerce, Applied Mathematics
Series 55, June 1964.
Adegeest [1994]
L.J.M. Adegeest,
Non-linear Hull Girder Loads in Ships,
PhD Thesis, 1994, Delft University of Technology, The Netherlands.
Athanassoulis and Loukakis [1985]
G.A. Athanassoulis and T.A. Loukakis,
An Extended-Lewis Form Family of Ship Sections and its Applications to Seakeeping
Calculations,
International Shipbuilding Progress, 1985, Vol. 32, No. 366.
Bales [1983]
S.L. Bales,
Wind and Wave Data for Seakeeping Performance Assessment,
Technical Report, 1983, Prepared for the 17th ITTC Seakeeping Committee, Athens,
Greece.
Boese [1970]
P. Boese,
Eine Einfache Methode zur Berechnung der Wiederstandserhhung eines Schiffes in
Seegang,
Technical Report 258, 1970, Institt fr Schiffbau der Universitt Hamburg, Germany.
van den Bosch and Vugts [1966]
J.J. van den Bosch and J.H. Vugts,
Roll Damping by Free Surface Tanks,
Technical Report 83-S, 1966, Netherlands Ship Research Centre TNO, Shipbuilding
Department, Delft, The Netherlands.
Conolly [1974]
J.E. Conolly,
Standards of Good Seakeeping for Destroyers and Frigates in Head Seas,
International Symposium on the Dynamics of Marine Vehicles and Structures in
Waves, 1974, No. 8, London, U.K.
Dimitrieva [1994]
I. Dimitrieva,
DELFRAC, 3-D Potential Theory Including Wave Diffraction and Drift Forces Acting
on the Structures,
(Description of the 3-D Computer Code DELFRAC, written by J.A. Pinkster),
Technical Report 1017, 1994, Delft University of Technology, Ship Hydromechanics
Laboratory, The Netherlands.
329
330
Huijsmans [1996]
R.H.M. Huijsmans,
Motions and Drift Forces on Moored Vessels in Current,
PhD Thesis, 1996, Delft University of Technology, The Netherlands.
Ikeda, Himeno and Tanaka [1978]
Y. Ikeda, Y. Himeno and N. Tanaka,
A Prediction Method for Ship Rolling,
Technical Report 00405, 1978, Department of Naval Architecture, University of Osaka
Prefecture, Japan.
John [1950]
F. John,
On the Motions of Floating Bodies, Part II,
Comm. on Pure and Applied Mathematics, 1950, Vol. 3, Interscience Publishers, Inc,
New York, pages 45-100.
de Jong [1973]
B. de Jong,
Computation of Hydrodynamic Coefficients of Oscillating Cylinders,
Technical Report 145-S, 1973, Netherlands Ship Research Centre TNO, Shipbuilding
Department, Delft, The Netherlands.
Journe [1991]
J.M.J. Journe,
Motions of Rectangular Barges,
Proceedings 10th International Conference on Offshore Mechanics and Arctic
Engineering}, 1991, Stavanger, Norway.
Journe [1992]
J.M.J. Journe,
Strip Theory Algorithms, Revised Report 1992,
Technical Report 912, 1992, Delft University of Technology, Ship Hydromechanics
Laboratory, The Netherlands.
Journe [1997]
J.M.J. Journe,
Liquid Cargo and Its Effect on Ship Motions,
STAB'97 Conference, 1997, Varna, Bulgaria, Internet: www.shipmotions.nl.
Journe [2001]
J.M.J. Journe,
Verification and Validation of Ship Motions Program SEAWAY,
Technical Report 1213a, 2001, Delft University of Technology, Ship Hydromechanics
Laboratory, The Netherlands, Internet: www.shipmotions.nl.
Journe and van t Veer [1995]
J.M.J. Journe and A.P. van t Veer,
First Order Wave Loads in Beam Waves,
331
Proceedings 5th International Offshore and Polar Engineering Conference, 1995, The
Hague, the Netherlands, Internet: www.shipmotions.nl.
Journe and Massie [2001]
J.M.J. Journe and W.W. Massie,
Offshore Hydromechanics, Lecture Notes,
First Edition, January 2001, Delft University of Technology, Ship Hydromechanics
Laboratory, The Netherlands, Internet: www.shipmotions.nl.
Keil [1974]
H. Keil,
Die Hydrodynamische Krfte bei der periodischen Bewegung zweidimensionaler
Krper an der Oberflche flacher Gewasser,
Bericht Nr. 305, 1974, Institut fr Schiffbau der Universitt Hamburg, Deutschland.
von Kerczek and Tuck [1969]
C. von Kerczek and E.O. Tuck,
The Representation of Ship Hulls by Conformal Mapping Functions,
Journal of Ship Research, 1969, Vol. 13, No. 4.
Kirsch [1969]
M. Kirsch,
Zur Tauch- und Stampfbewegung eines Schiffes,
Schiffstechnik, Band 16, 1969.
Kirsch [1969]
M. Kirsch,
Die Berechnung der Bewegungsgren der gekoppelten Tauch- und Stampfschwingungen nach der erweiterten Streifenmethode von Grim und die Berechnung
der Wahrscheinlichkeit fr das berschreiten bestimmter Schranken durch diese
Gren,
Bericht Nr. 241, 1969, Institut fr Schiffbau der Universitt Hamburg, Deutschland.
Korkut and Hebert [1970]
M. Korkut and E. Hebert,
Some Notes on Static Anchor Chain Curve,
Offshore Technology Conference, 1970, No. OTC 1160, Dallas, Texas, U.S.A.
Korvin-Kroukovsky [1955]
B.V. Korvin-Kroukovsky,
Investigations of Ship Motions in Regular in Regular Waves,
Transactions SNAME, 1955, No. 63, pages 386-435.
Korvin-Kroukovsky and Jacobs [1957]
B.V. Korvin-Kroukovsky and W.R. Jacobs,
Pitching and Heaving Motions of a Ship in Regular Waves,
Transactions SNAME, 1957, No. 65, pages 590-632.
Kumai [1959]
T. Kumai,
332
W.R. Porter,
Pressure Distribution, Added Mass and Damping Coefficients for Cylinders
Oscillating in a Free Surface,
Technical Report 82, 1960, University of California, Institute of Engineering
Research, Berkeley, California, U.S.A.
Reed and Nowacki [1974]
M.A. Reed and H. Nowacki,
Interactive Creation of Fair Ship Lines,
Journal of Ship Research, 1974, Vol. 18, No. 2.
Salvesen, Tuck and Faltinsen [1970]
N. Salvesen, E.O. Tuck and O.M. Faltinsen,
Ship Motions and Sea Loads,
SNAME, 1970, Vol. 78, pp 250-287
Sargent and Kaplan [1974]
T.P. Sargent and P. Kaplan,
Modifications to Lloyds Register of Shipping Strip Theory Computer Program (LR
2570,
Technical Report 74-103, 1974, Oceanics Inc.
Schmitke [1978]
R.T. Schmitke,
ROLLRFT, a Fortran Program to Predict Ship Roll, Sway and Yaw Motions in Oblique
Waves, Including the Effect of Rudder, Fin and Tank Roll Stabilizers,
Technical Report 78/G, 1978, Defence Research Establishment, Atlantic, Dartmouth
N.S., Canada.
Stoker [1957]
J.J. Stoker,
Water Waves,
Interscience publishers, Inc, New York, 1957.
Tasai [1959]
F. Tasai,
On the Damping Force and Added Mass of Ships Heaving and Pitching,
Technical Report, Research Institute for Applied Mechanics, Kyushu University,
Japan, 1959, Vol. VII, No 26.
Tasai [1960]
F. Tasai,
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, 1960, Vol. VIII, No 31.
Tasai [1961]
F. Tasai,
334
335
336