Faculty of Mechanical Engineering and Marine Technology

Chair of Modelling and Simulation

Ship dynamics in waves
(Ship Theory II)
Prof. Dr.-Ing. habil. Nikolai Kornev
Rostock
2012
2
Contents
1 Ship motion in regular sea waves 11
1.1 Coupling of dierent ship oscillations . . . . . . . . . . . . . . 11
1.2 Classication of forces . . . . . . . . . . . . . . . . . . . . . . 13
1.3 Radiation force components . . . . . . . . . . . . . . . . . . . 17
1.3.1 Hydrodynamic damping . . . . . . . . . . . . . . . . . 18
1.3.2 Added mass component . . . . . . . . . . . . . . . . . 20
1.4 Hydrostatic component . . . . . . . . . . . . . . . . . . . . . . 24
1.5 Wave exciting force . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6 Motion equations . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.7 Haskinds relation . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2 Free oscillations with small amplitudes 31
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3 Ship oscillations in small transverse waves (beam see) 37
3.1 Hydrostatic forces and moments . . . . . . . . . . . . . . . . . 38
3.2 Hydrodynamic Krylov - Froude force . . . . . . . . . . . . . . 40
3.3 Full Krylov - Froude force and moment . . . . . . . . . . . . . 42
3.4 Force and moment acting on the ship frame in accelerated ow 42
3.5 Full wave induced force and moment . . . . . . . . . . . . . . 43
3.6 Equations of ship heave and roll oscillations . . . . . . . . . . 43
3.7 Analysis of the formula (3.27) . . . . . . . . . . . . . . . . . . 45
3.8 Sway ship oscillations in beam sea . . . . . . . . . . . . . . . . 48
3.9 Ship oscillations at nite beam to wave length ratio and draught
to length ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.10 Eect of ship speed on rolling . . . . . . . . . . . . . . . . . . 52
3
4 Ship oscillations in small head waves 55
4.1 Exciting forces and ship oscillations . . . . . . . . . . . . . . . 55
4.2 Estimations of slamming and deck ooding . . . . . . . . . . . 58
5 Seasickness caused by ship oscillations 61
6 Ship oscillations in irregular waves 65
6.1 Representation of irregular waves . . . . . . . . . . . . . . . . 65
6.1.1 Wave ordinates as stochastic quantities . . . . . . . . . 66
6.1.2 Wave spectra . . . . . . . . . . . . . . . . . . . . . . . 68
6.2 Calculation of ship oscillations in irregular waves . . . . . . . 72
7 Experimental methods in ship seakeeping 75
7.1 Experiments with models . . . . . . . . . . . . . . . . . . . . . 75
7.2 Seakeeping tests with large scale ships . . . . . . . . . . . . . 82
8 Ship oscillation damping (stabilisation) 85
8.1 Damping of roll oscillations . . . . . . . . . . . . . . . . . . . 85
8.1.1 Passive means . . . . . . . . . . . . . . . . . . . . . . . 85
8.1.2 Active stabilizer . . . . . . . . . . . . . . . . . . . . . . 88
8.1.3 Passive Schlingerkiel . . . . . . . . . . . . . . . . . . . 88
8.1.4 Active rudders . . . . . . . . . . . . . . . . . . . . . . 89
8.1.5 Damping of pitch oscillations . . . . . . . . . . . . . . 89
9 Parametric oscillations 95
10 Principles of Rankine source method for calculation of sea-
keeping 101
10.1 Frequency domain simulations . . . . . . . . . . . . . . . . . . 101
10.2 Time domain simulation . . . . . . . . . . . . . . . . . . . . . 106
4
List of Tables
2.1 Frequencies and periods of dierent oscillation types . . . . . . 33
2.2 Referred damping factors for dierent oscillation types . . . . 34
5
6
List of Figures
1.1 Ship motion with 6 degree of freedom (from [1]) . . . . . . . . 12
1.2 Displacement of the center of eort due to change of the ship
draught (from [2]) . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3 Illustration to derivation of damping coecient . . . . . . . . 19
1.4 Added mass and damping coecient of the semi circle frame
at heave oscillations. Here A is the frame area. . . . . . . . . . 20
1.5 Added mass and damping coecient of the box frame at heave
oscillations. Here A is the frame area. . . . . . . . . . . . . . . 21
1.6 Added mass and damping coecient of the semi circle frame
at sway oscillations. Here A is the frame area. . . . . . . . . . 21
1.7 Added mass and damping coecient of the box frame at sway
oscillations. Here A is the frame area. . . . . . . . . . . . . . . 21
1.8 Added mass and damping coecient of the box frame at roll
(heel) oscillations. Here A is the frame area. . . . . . . . . . . 22
1.9 Mirroring for the case 0 . . . . . . . . . . . . . . . . . . . 23
1.10 Mirroring for the case . . . . . . . . . . . . . . . . . . 23
3.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2 Illustration of hydrostatic force . . . . . . . . . . . . . . . . . 39
3.3 Illustration of hydrostatic moment . . . . . . . . . . . . . . . . 39
3.4 Ship as linear system . . . . . . . . . . . . . . . . . . . . . . . 45
3.5 Response function versus referred frequency . . . . . . . . . . 46
3.6 Phase displacement versus referred frequency . . . . . . . . . . 47
3.7 Ship oscillations in resonance case . . . . . . . . . . . . . . . . 47
3.8 Oscillation of a raft with a big metacentric height . . . . . . . 48
3.9 Illustration of the frame in beam waves . . . . . . . . . . . . . 48
3.10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.11 Reduction coecient of the heave oscillations . . . . . . . . . 52
3.12 Sea classication . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.1 Illustration of the ship in head waves . . . . . . . . . . . . . . 55
4.2 Position of ship at dierent time instants in a head wave . . . 58
7
4.3 Curves y = z
max
and y = z(x) . . . . . . . . . . . . . . . . . 59
4.4 Sample for a real ship . . . . . . . . . . . . . . . . . . . . . . . 60
5.1 Inuence of the vertical acceleration on the seasickness de-
pending on the oscillation period . . . . . . . . . . . . . . . . 62
5.2 Inuence of the vertical acceleration on the seasickness de-
pending on the oscillation period . . . . . . . . . . . . . . . . 63
5.3 Number of passengers suering from seasickness on a cruise
liners depending on vertical accelerations . . . . . . . . . . . . 63
5.4 Adaption to seasickness . . . . . . . . . . . . . . . . . . . . . . 64
6.1 Irregular seawaves, 1- two dimensional, 2- three dimensional.
(Fig. from [3]) . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.2 Prole of an irregular wave. (Fig. from [3]) . . . . . . . . . . . 66
6.3 Representation of irregular wave through the superposition of
regular waves. (Fig. from [3]) . . . . . . . . . . . . . . . . . . 67
6.4 p.d.f. of the wave ordinate . . . . . . . . . . . . . . . . . . . . 68
6.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
7.1 Determination oft he inertia moment I
zz
. . . . . . . . . . . . 77
7.2 Determination of I
xx
and z
g
. . . . . . . . . . . . . . . . . . . 78
7.3 Heel test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
7.4 Method of forced rolling . . . . . . . . . . . . . . . . . . . . . 80
7.5 Seakeeping test at MARIN ([4]) . . . . . . . . . . . . . . . . . 81
7.6 Scetch of the MARIN Seakeeping basin ([4]) . . . . . . . . . . 82
7.7 Wave generator of MARIN Seakeeping basin ([4]) . . . . . . . 82
7.8 Method of wave detection . . . . . . . . . . . . . . . . . . . . 82
7.9 Ship motion during large scale tests . . . . . . . . . . . . . . . 83
8.1 U-tube stabilization system of Frahm of the second type . . . 86
8.2 Free surface Type passive Roll stabilization systems of Flume . 87
8.3 Free surface Type passive Roll stabilization systems of Flume . 87
8.4 Active stabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.5 Passive Schlingerkiel . . . . . . . . . . . . . . . . . . . . . . . 89
8.6 Active rudders . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.7 Damping of pitch oscillations . . . . . . . . . . . . . . . . . . 89
8.8 Damping of pitch oscillations . . . . . . . . . . . . . . . . . . 90
8.9 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
8.10 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
8.11 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
8.12 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8.13 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8
8.14 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
8.15 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
8.16 [5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
9.1 Ship oscillations during parametric resonance . . . . . . . . . . 97
9.2 Conditions for parametric resonance appearance . . . . . . . . 97
9
10
Chapter 1
Ship motion in regular sea
waves
1.1 Coupling of dierent ship oscillations
The ship has generally six degrees of freedom which are called as surge ,
sway , heave , heel (or roll) , yaw and pitch (see Fig.1.1) for expla-
nation of each oscillation motion). In this chapter we consider rst the case
of the ship with zero forward speed. Generally, dierent ship oscillations are
strongly coupled. There are three sorts of coupling:
hydrostatic coupling
hydrodynamic coupling
gyroscopics coupling
The hydrostatic coupling is illustrated in Fig.1.2. If the ship draught is
changed, the center of eort of vertical hydrostatic (oating) force is mov-
ing usually towards the ship stern because the frames in the stern are more
full than those in the bow region. The displacement of the center of eort
towards the stern causes the negative pitch angle. Therefore, the heave os-
cillations cause the pitch oscillations and vice versa. With the other words,
the heave and pitch oscillations are coupled.
Hydrodynamic coupling can be illustrated when the ship is moving with ac-
celeration in transverse direction (sway motion). Since the ship is asymmetric
with respect to the midships, such a motion is conducted with appearance
of the yaw moment. Therefore, the sway and yaw oscillations are hydrody-
namically coupled.
11
According to gyroscopic eect, rotation on one axis of the turning around
the second axis wheel produced rotation of the third axis. This rule can be
applied to the ship. For instance, if the ship performs rolling motion and
the transverse force is acting on the ship, it starts to perform the pitch os-
cillations. The gyroscopic eects are present in the equation system (1.13).
They are represented in i-th force equation by products V
j=i

m=j=i
and
by products
j=u

m=j=i
in the i-th moment equation.
In this chapter we consider the ship oscillations with small amplitude. For
such oscillations the coupling mentioned above can be neglected.
Figure 1.1: Ship motion with 6 degree of freedom (from [1])
Figure 1.2: Displacement of the center of eort due to change of the ship
draught (from [2])
12
1.2 Classication of forces
According to the tradition used in ship hydrodynamics since almost a hun-
dred years, the forces acting upon the ship are subdivided into hydrostatic
forces, radiation and diraction forces. This subdivision can be derived for-
mally utilizing the potential theory. The potential theory is still remaining
the theoretical basis for the determination of wave induced forces, since the
most contribution to these forces is caused by processes properly described
by inviscid ow models.
Let us consider the plane progressive waves of amplitude A and direction
w
are incident upon a ship, which moves in response to these waves. The ship
oscillation caused by waves can be written in the form

j
=
0
j
sin t, j = 1, 2, ...6. (1.1)
The corresponding speeds of ship oscillations U
j
, j = 1, 2, ...6 are:
U
j
=
d
j
dt
=
0
j
cos t, j = 1, 2, ...6. (1.2)
and accelerations:
a
j
=
dU
j
dt
=
2

0
j
sin t, j = 1, 2, ...6. (1.3)
Here
0
j
are small ship oscillations amplitudes and is the frequency. Within
the linear theory the ship oscillation frequency is equal to the incident wave
frequency. In what follows we use the linear theory and assume that the both
waves and ship motion are small. The total potential can be written, using
the superposition principle, in the form:
(x, y, z, t) =
6

j=1
U
j

j
(x, y, z) + A
A
(x, y, z) cos t =
=
_
6

j=1

0
j

j
(x, y, z) + A
A
(x, y, z)
_
cos t
(1.4)
where

j
(x, y, z) is the velocity potential of the ship oscillation in j-th motion
with the unit amplitude
0
j
= 1 in the absence of incident waves,
13

A
(x, y, z) is the potential taking the incident waves and their interac-
tion with the ship into account.
The rst potentials
j
(x, y, z) describes the radiation problem, whereas
the second one the wave diraction problem. The potentials
j
(x, y, z)
and
A
are independent only in the framework of the linear theory assuming
the waves and ship motions are small. Within this theory
A
is calculated
for the ship xed in position.
The potentials must satisfy the Laplace equation
j
= 0,
A
= 0 and
appropriate boundary conditions. The boundary conditions to be imposed
on the ship surface are the no penetration conditions (see also formulae 3.18
in the Chapter 3 [6]):
for radiation potentials

1
n
= cos(n, x);

2
n
= cos(n, y);

3
n
= cos(n, z);

4
n
= (y cos(n, z) z cos(n, y));

5
n
= (z cos(n, x) x cos(n, z));

6
n
= (x cos(n, y) y cos(n, x)).
_

_
(1.5)
for wave diraction potentials

A
n
= 0 (1.6)
where n is the normal vector to the ship surface, directed into the body,
(x, y, z) are the coordinates of a point on the ship surface. The r.h.s. of the
conditions (1.5) is the normal components of the ship local velocities caused
by particular oscillating motions.
The diraction potential
A
is decomposed in two parts

A
=

+
p
(1.7)
which

is the potential of incident waves not perturbed by the ship pres-

ence and
p
is the perturbation potential describing the interaction between
14
the incident waves and the ship. The potential of regular waves

is known
(see Chapter 6 in [6]). The boundary condition for
p
on the ship surface is

p
n
=

n
(1.8)
Away from the ship the radiation potentials
j
and the diraction perturba-
tion potential
p
decay, i.e.
p

r
0,
j

r
0.
On the free surface the linearized mixed boundary condition (see formula (6.17)
in [6]) reads

t
2
+ g

z
= 0 on z = 0. (1.9)
Substituting (1.4) in (1.9) yields for
j
(x, y, z) and
A
(x, y, z):

2
g

j
+

j
z
= 0 on z = 0.

2
g

A
+

A
z
= 0 on z = 0.
(1.10)
It is obvious from (1.10) that
j
(x, y, z) and
A
(x, y, z) depend on .
Additionally in the wave theory the radiation condition is imposed stating
that the waves on the free surface caused by the potentials are radiated away
from the ship. The potentials introduced above can be found using panel
methods.
The force and the moment on the ship are determined by integrating the
pressure over the wetted ship surface. The pressure can be found from the
Bernoulli equation written in the general form:
p +
u
2
2
+ gz +

t
= C(t) (1.11)
Here the potential is the potential of the perturbed motion. The constant C(t)
which is the same for the whole ow domain is calculated from the condition
that the pressure on the free surface far from the ship is constant and equal
to the atmospheric pressure:
p
a
= C(t) (1.12)
Substituting (1.12) in (1.11) gives:
15
p p
a
=
u
2
2
gz

t
(1.12a)
Remembering that the ship speed is zero and perturbation velocities as well
as the velocities caused by incident waves are small we neglect the rst term
in (1.12a):
p p
a
=
_

t
+ gz
_
(1.13)
Together with (1.4) it gives
p p
a
=
_
6

j=1

0
j

j
(x, y, z) + A
A
(x, y, z)
_
sin t gz (1.14)
The forces and the moment are then calculated by integration of p p
a
over
the wetted ship area

F =
_
S
pndS,

M =
_
S
p(r n)dS. (1.15)
The normal vector direction in (1.15) is into the body. The vertical ordi-
nate z of any point on the wetted area can be represented as the dierence
between the submergence under unperturbed free surface and free surface
elevation
0
. Substituting (1.14) in (1.15) one obtains

F = g
_
S
ndS + g
_
S
n
0
dS
+
6

j=1

0
j

2
sin t
_
S
n
j
dS+
+ (A sin t)
_
S
n(

+
p
)dS
(1.16)
Four integrals in (1.16) represent four dierent contributions to the total
force:
the hydrostatic component (the rst term) acting on the ship oscillating
on the unperturbed free surface (in calm water),
16
the hydrostatic component arising due to waves (the second term),
the damping and the added mass component (the third term) and
the hydrodynamic wave exciting force (the fourth term).
The moment is expressed through similar components.
The third term describes the force acting on the ship oscillating in calm water.
The last term arises due to incident waves acting on the ship. Within the
linear theory keeping only the terms proportional to the amplitude A and
neglecting small terms of higher orders proportional to A
n
, n > 1 one can
show that the integration in the last term can be done over the wetted area
corresponding to the equilibrium state. Thus, the last term describes the
force induced by waves on the ship at rest.
1.3 Radiation force components
Let us consider the second term of the force

F
2
=
6

j=1

0
j

2
sin t
_
S
n
j
dS (1.17)
Each component of this force is expressed as
F
2i
=
6

j=1

0
j

2
sin t
_
S
n
i

j
dS =
6

j=1
c
ji
dU
j
dt
(1.18)
As shown by Haskind [17], the hydrodynamic coecient c
ji
is represented as
the sum of two coecients:
c
ji
=
ji

ji
The term
1

ji
has been introduced to take the fact into account that the
force due to inuence of the free surface depends not only on
2
but also
on . The force is then
F
2i
=
6

j=1
_

ji

ji
_
dU
j
dt
=
6

j=1

ji
a
j

j=1

ji
U
j
_
t
s
= t

2
_
(1.19)
17
As seen from (1.19) the rst component of the force is proportional to the
acceleration a
j
whereas the second one is proportional to the velocity U
j
.
The rst component is called the added mass component, whereas the sec-
ond one- the damping component.
Using the Greens theorem, Haskind derived the following symmetry condi-
tions for the case zero forward speed:
c
ji
= c
ij

_

ji
=
ij

ji
=
ij
1.3.1 Hydrodynamic damping
There are two reasons of the hydrodynamic damping of the ship oscillations
on the free surface. First reason is the viscous damping which is proportional
to the square of the ship velocity C
Dj
U
2
j
2
S. Within the linear theory this
term proportional to the amplitude (
0
j
)
2
is neglected. The main contribution
to the damping is done by the damping caused by radiated waves. When
oscillating on the free surface the ship generates waves which have the me-
chanic potential and kinetic energy. This wave energy is extracted from the
kinetic energy of the ship. Ship transfers its energy to waves which carry it
away from the ship. With the time the whole kinetic energy is radiated away
and the ship oscillations decay.
Similarly to the added mass one can introduce the damping coecients. The
full mechanic energy in the progressive wave with the amplitude A is (see
chapter 6.4 in [6])
E =
1
2
gA
2
(1.20)
per wave length.
The energy transported by waves through sides 1 and 2 (see Fig. 1.3) per
time unit is
E = 2
1
2
gA
2
U (1.21)
where U is the wave group velocity. The damping coecient is dened as
E =
ij
U
2
j
(1.22)
18
Figure 1.3: Illustration to derivation of damping coecient
where
ij
is the coecient of damping in i-th direction when the ship oscil-
lates in j-th motion. U
2
j
is the time averaged square of the ship oscillations
speed. Obviously, U
2
j
=
(

0
j
)
2
2
and
E =
ij
_

0
j
_
2
2
(1.23)
The group velocity (see formulae (6.39) and (6.40) in [6]):
U =
c
2
=
1
2
_
g
k
=
g
2
(1.24)
since kg =
2
(see formula (6.21) in [6]). Equating (1.21) and (1.23) one
obtains with account for (1.24)
gA
2
U =
ij
_

0
j
_
2
2

gA
2
g
2
=
ij
_

0
j
_
2
2

ij
=
g
2

3
_
A

0
j
_
2
(1.25)
The damping coecient
ij
depends on the square of the ratio of the wave
amplitude to the ship oscillation amplitude causing the wave.
The damping coecient of slender body can be found by integration of damp-
ing coecients of ship frames along the ship length
19
B
22
=
L/2
_
L/2

22
dx, B
33
=
L/2
_
L/2

33
dx, B
44
=
L/2
_
L/2

44
dx,
B
55
=
L/2
_
L/2
x
2

33
dx, B
66
=
L/2
_
L/2
x
2

22
dx
(1.26)
The damping coecients of dierent frames are shown in Fig. 1.4, 1.5, 1.6, 1.7
and 1.8 taken from [2]. Solid lines show results obtained from the potential
theory. Generally, the results show the applicability of the potential theory
for calculation of damping coecients. The accuracy of prediction is not
satisfactory for the box B/T=8 in heave and B/T=2 in sway because of the
ow separation at corners which has a sucient impact on hydrodynamics
in these two cases. The agreement for
44
is not satisfactory (see Fig. 1.8)
because of dominating role of the viscosity for this type of damping. For the
semi circle frame the damping coecient in roll is zero
44
= 0 within the
inviscid theory. One hundred per cent of the roll damping is due to viscosity.
Usually
44
are determined using viscous ow models. It is remarkable,
that the damping coecients depend on the frequency and amplitude (see
Fig. 1.8).
Figure 1.4: Added mass and damping coecient of the semi circle frame at
heave oscillations. Here A is the frame area.
1.3.2 Added mass component
When the ship oscillates, the force acting on the ship contains the component
associated with the added mass like in every case of accelerated body motion.
20
Figure 1.5: Added mass and damping coecient of the box frame at heave
oscillations. Here A is the frame area.
Figure 1.6: Added mass and damping coecient of the semi circle frame at
sway oscillations. Here A is the frame area.
Figure 1.7: Added mass and damping coecient of the box frame at sway
oscillations. Here A is the frame area.
The dierence with the case of the motion in unlimited space is the presence
of the free surface. The added mass
ij
have to be calculated with account
for the free surface eect. For their determination the panel methods can be
21
Figure 1.8: Added mass and damping coecient of the box frame at roll
(heel) oscillations. Here A is the frame area.
used. The problem is suciently simplied in two limiting cases 0 and
. The boundary condition (1.10) can be written in the form:

j
z
= 0 for 0,
on z = 0.

j
= 0 for .
(1.27)
The conventional mirroring method can be used for the case 0 (Fig. 1.9).
The mirroring frame is moving in the same direction for surge, sway and yaw.
For the heave, roll and pitch the ctitious frame is moving in the opposite di-
rection. At the free surface, these tricks make the normal components of the
total velocity induced by the actual and the ctitious frames zero, i.e.

j
z
= 0
on z = 0.
In the case the tangential component of the total velocity should
be zero, since
j
= 0
j
=
x
_

j
x
dx = 0

j
x
= 0 on z = 0. The
modied mirroring method is implemented for the case (Fig. 1.10).
The ctitious frame is moving in the opposite direction for surge, sway and
yaw. For the heave, roll and pitch the ctitious frame is moving in the
same direction as that of the original frame. These tricks make the normal
components of the total velocity induced by the actual and the ctitious
frames zero, i.e.

j
z
= 0 at the free surface on z = 0.
Using mirroring method the added mass can be found using the panel with-
out explicit consideration of the free surface since it is taken into account by
ctitious frames.
The added mass of slender body can be found by integration of added mass
22
Figure 1.9: Mirroring for the case 0
Figure 1.10: Mirroring for the case
of ship frames along the ship length
A
22
=
L/2
_
L/2

22
dx, A
33
=
L/2
_
L/2

33
dx, A
44
=
L/2
_
L/2

44
dx,
A
55
=
L/2
_
L/2
x
2

33
dx, A
66
=
L/2
_
L/2
x
2

22
dx
(1.28)
The added mass of dierent frames are shown in Fig. 1.4, 1.5, 1.6, 1.7 and 1.8
taken from [2]. Like in case of damping coecients the results of the potential
theory are not acceptable for roll added mass because of dominating role of
the viscosity. As seen from Fig. 1.4 - Fig. 1.8 the added mass depend on the
23
frequency .
1.4 Hydrostatic component
Let us the ship is in the equilibrium state. The ship weight is counterbalanced
by the hydrostatic lift. Due to small heave motion the equilibrium is violated
and an additional hydrostatic force appears. The vertical component of this
additional hydrostatic force can be calculated analytically from (1.16) for the
case of small heave motion
F

=
_
_
_
g
_
S
cos(nz)zdS
_
_
_
T+

_
_
_
g
_
S
cos(nz)zdS
_
_
_
T
= gA
WP

(1.29)
where is the increment of the ship draught, A
WP
is the waterplane area
and T is the ship draught in the equilibrium state. The roll and pitch hy-
drostatic moments for small change of the roll and pitch angles are
M

= g
0
GM

, (1.30)
M

= g
0
GM
L
, (1.31)
where
and are the roll and pitch angle respectively,
GM

is the transverse metacentric height,

GM
L
is the longitudinal metacentric height and

0
is the ship displacement.
1.5 Wave exciting force
The wave exciting force

F
per
= Asin t
_
S
n(

+
p
)dS contains two com-
ponents. The rst component, determined by the integration of the incident
potential

, Asin t
_
S
n

dS is referred to as the hydrodynamic part of

the Froude-Krylow force. This force called as the Smith eect is calculated by
the integration of wave induced pressure as if the ship is fully transparent for
24
incident waves. The full Froude-Krylow force contains additionally the hydro-
static force arising due to change of the submerged part of the ship caused by
waves (the second term in (1.16)). The second component Asin t
_
S
n
P
dS
takes the diraction eect (the contribution of the scattering potential
p
to
pressure distribution) into account. As shown by Peters and Stokes the
Froude Krylov force is a dominating part of the wave induced forces for os-
cillations of slender ships in directions j=1 (surge), 3 (heave) and 5 (pitch).
1.6 Motion equations
The linearized decoupled motion equations of the ship oscillations are written
in the form
added mass damping hydrostatic wave exciting
force forces forces forces
m

= A
11

B
11

+F
,per
(t),
m = A
22
B
22
+F
,per
(t),
m

= A
33

B
33

gA
WP
+F
,per
(t),
I
xx
= A
44
B
44
g
0
GM

+M
,per
(t),
I
yy

= A
55

B
55

g
0
GM
L
+M
,per
(t),
I
zz

= A
66

B
66

+M
,per
(t).
(1.32)
The weight is not present in the second equation of the system (1.32) be-
cause it is counterbalanced by the hydrostatic force at rest. The additional
hydrostatic force gA
WP
is the dierence between the weight and the full
hydrostatic force. The system (1.32) is written in the principle axes coordi-
nate system [7].
1.7 Haskinds relation
One of the most outstanding results in the ship oscillations theory is the
relation derived by Max Haskind who developed in 1948 the famous linear
hydrodynamic theory of ship oscillations. Haskind shown how to calculate
the wave induced hydrodynamic force utilizing the radiation potentials
j
and the potential of incident waves

. The determination of the diraction

potential
p
what is quite dicult can be avoided using this relation which
is valid for waves of arbitrary lengths.
25
The Greens formula for two functions and satisfying the Laplace equa-
tion is
_
S
w
_ _

n
_
dS = 0, (1.33)
where S
w
is the ow boundary (wetted ship surface plus the area away from
the ship, see the sample in Chapter/Section 3.2). Particularly, the rela-
tion (1.33) can be applied to radiation potentials
j
. Since the potential
p
satises the Laplace equation and the same boundary conditions as the ra-
diation potentials
j
, the Greens formula (1.33) can also be applied to
j
and
p
_
S
w
_ _

j
n

j

p
n
_
dS = 0 (1.34)
The last term in (1.16) is the wave induced force

F
,per
= (A sin t)
_
S
n(

+
p
) dS =

XAsin t
where

X =
_
S
n(

+
p
) dS. Taking (1.5), (3.34) and (1.8) into account
we get
X
j
=
_
S
w
_
(

+
p
)

j
n
dS (1.35)
_
S
w
_

p

j
n
dS =
_
S
w
_

j

p
n
dS X
j
=
_
S
w
_
_

j
n
+
j

p
n
_
dS

p
n
=

n
= 90

(1.36)
X
j
=
_
S
w
_ _

j
n

j

n
_
dS (1.37)
The formula (1.37) is the Haskinds relation. As seen the wave induced force
can be calculated through the radiation and free wave potentials avoiding
26
the determination of the diraction potential
p
.
The calculation of the integral (1.37) is a complicated problem because the
incident waves dont decay away from the ship and the integral (1.37) should
be calculated over both the surface far from the ship and the ship wetted
surface. Note that the potential

does not decay away from the ship. The

method of the stationary phase [8] allows one to come to the following force
expression using the Haskinds relation (1.37):
F
i,per
= B
ii

i
, where B
ii
=
k
8g(c/2)
2
_
0
|X
i
()|
2
d
Here c is the phase wave velocity (celerity) and is the course angle.
Let us consider the slender ship B(x, z) 0 in a beam wave ( = 90

).
The wetted area is approximately equal to the projection on the symmetry
plane y = 0, S
wetted
= [0, L]

[0, T].
X
j
=
_
S
w
_ _

j
n

j

n
_
dS
_
S
w
_

j
n
dS

3
n
= cos(n, z) =
B
z
X
3
2
_
S
w
_

B
z
dS
The coecient 2 arises due to the integration over two boards y = +B(x, z)
and y = B(x, z). Using the potential of an Airy wave (see formulae (6.18)
in ([6]) estimated at y = 0 one can nd the potential

:
=
Ag

e
kz
sin(ky t)

= ge
kz
/
For the case of a vertical cylinder for which the vertical force does not depend
on the wave course angle the damping coecient B
33
takes a very simple
form [8]:
B
33
=
2g
(c/2)

0
_
T
e
kz
B
z
dz

2
(1.38)
27
1.8 Exercises
1. Schwimmender Balken
(a) Ist die Schwimmlage des Balkens stabil?
(b) Bleibt die Schwimmlage stabil, wenn die H ohe h des Balkens 0, 15 m
betr agt?
(c) Ab welcher Balkenhohe wird die Schwimmlage instabil?
2. ([8]) Berechnen Sie die maximale elektrische Leistung einer Turbine,
die die ganze mechanische Energie einer Welle umwandelt. Die Welle
hat eine H ohe von 1 m, eine Lange von 100 m und eine Breite von 1 km
in Richtung des Wellenkammes.
3. Hinter dem bildet sich ein station ares Wellensystem.
Wie gro ist die Geschwindigkeit der Querwellen, wenn die Schis-
geschwindigkeit 10 m/sek betragt?
4. Welche Lange haben die Querwellen hinter einem Schi, das sich mit
der Geschwindigkeit von 10 m/sek bewegt?
Wie gro ist die Wellenfrequenz?
5. Die Querwelle hinter einem Schi hat die Amplitude 1 m. Sch atzen Sie
den Widerstand des Schies!
Benutzen Sie das Bild 6.12 aus dem Buch von Newman.
28
6. ([8])Das Modell eines Schies wird in einer sehr breiten Schlepprinne
mit der Geschwindigkeit 1 m/s 100 geschleppt. Die Modell ange betr agt 5 m.
Nach 100 Metern wird das Modell gestoppt. Wie viel Querwellen
benden sich in der Schlepprinne, wenn die Reektion von Schlepprin-
nenseiten nicht auftritt?
Hinweis: Das Schi wird als Superposition von zwei Punktstorungen
betrachtet: Bug und Heck. Der Bug erzeugt die Welle. Das Heck
erzeugt die Welle. Gesamtes Wellenbild wird als Summe betrachtet:
Benutzen Sie die Formel (6.21) und die Aufgaben 2 und 3.
7. F ur eine fortschreitende Welle mit der Amplitude 6 m und der L ange 200 m
berechnen Sie die Phasengeschwindigkeit und die maximale Geschwindigkeit
der Wasserteilchen.
In welchen Punkten ist diese Geschwindigkeit maximal?
8. Task:
Develop the theory of vertical oscillations of a very sharp cone with
the draught T=10m and the diameter of 1m in regular and irregular
waves using the Haskinds relation (1.38). The added mass A
33
can be
neglected.
29
30
Chapter 2
Free oscillations with small
amplitudes
2.1 Introduction
Let us consider a ship in the equilibrium position at calm water condition.
The ship has zero forward speed. If a perturbation acts on the ship, it
performs oscillating motions in three directions:
heave,
roll (heel),
pitch.
Yaw, surge and sway motions did not arise at calm water conditions. The
reason is the presence of restoring hydrostatic forces in heave, roll and pitch
directions.
The motion equations of the free oscillation read:
(m + A
33
)

+ B
33

+ gA
WP
= 0,
(I
xx
+ A
44
) + B
44
+ g
0
GM

= 0,
(I
yy
+ A
55
)

+ B
55

+ g
0
GM
L
= 0.
(2.1)
In ship theory the equations (2.1) are written in the normalized form:

+ 2

+
2

= 0,
+ 2

+
2

= 0,

+ 2

+
2

= 0,
(2.2)
31
where

=
B
33
2 (m + A
33
)
,

=
B
44
2 (I
xx
+ A
44
)
,

=
B
55
2 (I
yy
+ A
55
)
(2.3)
are damping coecients and

=
_
gA
WP
m + A
33
,

=
_
g
0
GM

I
xx
+ A
44
,

g
0
GM
L
I
yy
+ A
55
(2.4)
are the eugen frequencies of non damped oscillations.
The equations (2.2) are fully independent of each other. The solutions of the
equations (2.2) written in the general form:

+ 2

+
2
= 0 (2.5)
is given as:
= Ce
pt
(2.6)
Substitution of (2.6) into (2.5) yields the algebraic equation
p
2
+ 2p +
2
= 0 (2.7)
which solution is
p
1,2
=

2
(2.8)
If the system has no damping the solution is
p
1,2
= i = Ce
it
= C(cos t + i sin t) (2.9)
The system oscillates with the constant amplitude and frequency . That is
why the frequency is referred to as the eigenfrequency.
For real ships the damping coecient is smaller than the eigenfrequency <
and the equation (2.8) has two solutions:
p
1
= + i

2
= + i ,
p
2
= i

2
= i ,
(2.10)
In turn, the solution of the dierential equation is
= Ce
t
e
i
= Ce
t
(cos t i sin t) (2.11)
32
It describes damped oscillations with decaying amplitude which decrease is
governed by the factor e
t
, e
t

t
0. The rate of the decay is charac-
terized by the damping coecient . The frequency of damped oscillations
is
=

2
< (2.12)
Due to damping the frequency of the oscillations is shorter whereas the period
is longer:
T =
2

=
2

2
(2.13)
Since
T =
2

=
2

(2.14)
From (2.12) and (2.14) we obtain the frequencies for dierent types of oscil-
lations which are listed in the table below:
Table 2.1: Frequencies and periods of dierent oscillation types
Oscillation Eigenfrequency Frequency of Period of
damped oscillations oscillations
Heave

=
_
gA
WP
m+A
33

=
_

= 2
_
m+A
33
gA
WP
Rolling

=
_
g
0
GM

I
xx
+A
44

=
_

= 2
_
I
xx
+A
44
g
0
GM

Pitch

=
_
g
0
GM
L
I
yy
+A
55

=
_

= 2
_
I
yy
+A
55
g
0
GM
L
The damping is characterized by the logarithmic decrement which is the
logarithm of the ratio of the oscillation amplitude at the time instant t to
that at the time instant t+T, i.e.
(t)
(t + T)
=
e
t
e
(t+T)
= e
T
(2.15)
The logarithm of the ratio (2.15) is
ln
(t)
(t + T)
= ln
e
t
e
(t+T)
= lne
T
=
2

(2.16)
The ratio


is called as the referred damping factor . The decay of the oscil-
lation amplitude is equal to this factor multiplied by 2. Referred damping
factors for dierent types of oscillation can be found from this denition. The
results obtained under assumption are listed in the table 2.2.
33
If the metacentric heights GM

and GM
L
are getting larger, the periods T

=
2
_
I
xx
+A
44
g
0
GM

and T

= 2
_
I
yy
+A
55
g
0
GM
L
as well as the damping factors

=
B
44
2

(I
xx
+A
44
)g
0
GM

and

=
B
55
2

(I
yy
+A
55
)g
0
GM
L
decrease. Therefore, the
smaller are the metacentric heights the larger are the oscillations periods
and the less oscillations are necessary to decay. The time of decay depends
only on damping and doesnt depend on the metacentric height.
Table 2.2: Referred damping factors for dierent oscillation types
Oscillation Referred damping factor
Heave

=
B
33
2

(m+A
33
)gA
WP
Rolling

=
B
44
2

(I
xx
+A
44
)g
0
GM

Pitch

=
B
55
2

(I
yy
+A
55
)g
0
GM
L
2.2 Exercise
1.
0
is the roll angle at t = 0. Find the number of periods N of free roll
oscillations necessary to reduce the amplitude oscillations by factor e
a
.
What is inuence of the metacentric height on N?
2. The period of undamped oscillations is T. The referred damping fac-
tor is 0,2.
Calculate the period of damped oscillations!
Calculate the reduction of the amplitude within the period of damped
and undamped oscillations!
3. Typical periods of roll and pitch oscillations for dierent ships are [9]:
Ship T

, sec T

, sec
tanker 9 ... 15 7 ... 11
ice breaker 8 ... 12 3 ... 5
trawler 6 ... 8 3 ... 4
big cruise liner 20 ... 28 10 ... 12
container ships (20000 30000 t) 16 ... 19 7 ... 9
Explain why T

> T

.
34
4. The periods of oscillations can be estimated from the following simple
empiric formulae [9]
T

2.5

T, T

2.4

T, T

cB/
_
GM

,
where T and B are draught and beam. The empiric coecient c is
equal approximately 0, 8...0, 85 for big cruise liners.
Calculate the change of the period of roll oscillations of a cruise liner if
a load with mass 1 ton is elevated in vertical direction from 10 m from
the keel line to the 1 meters from the keel line! The ship displacement
is 20 000 t and the beam is 30 m.
35
36
Chapter 3
Ship oscillations in small
transverse waves (beam see)
The formalism developed in this chapter is based on the following assump-
tions:
waves are regular,
waves amplitudes related to the wave lengths are small. Wave slope is
small.
wave length is much larger than the ship width,
The ship has zero forward speed.
From the rst two assumptions it follows, that the collective action of waves
on ship can be considered through the superposition principle. Therefore,
the theory can be developed for the interaction of the ship with a single wave
with given length and amplitude. The eects of dierent waves are then
summed. For the case of small waves the oscillations are decoupled. The
hydrodynamic, hydrostatic and gyroscopic coupling eects are neglected.
The perturbation forces (see the last column in the equation system (1.32))
arise due to wave induced change of the hydrostatic forces and due to hy-
drodynamic eects caused by orbital motion in waves. The orbital motion
causes the hydrodynamic pressure change which results in the wave induced
hydrodynamic forces.
In each frame, the pressure gradient induced by waves is assumed to be
constant along the frame contour and equal to the pressure gradient at the
centre A on the free surface. When considering the roll and pitch oscillations
37
in transverse waves it is additionally assumed that the ship draught change
and the ship slope relatively to the free surface are constant along the ship.
The wave ordinate is given by the formula derived for the progressive wave
(see Chapter 6 in [6])

0
= Asin (t + k) (3.1)
where A is the amplitude, is the wave propagation direction and is the
frequency.
In this section the incident waves are perpendicular to the ship (see Fig.3.1).
The wave propagation direction is in direction, i.e. = . The waves induce
roll and heave oscillations. The curvature of the free surface is neglected, the
free surface is considered as the plane performing angular oscillations and
translational oscillations in vertical direction.
Figure 3.1:
3.1 Hydrostatic forces and moments
The hydrostatic forces during the heave oscillations are calculated neglecting
the wave surface slope. The hydrostatic force acting on the ship with draft
increment in the wave with the ordinate
0
(see Fig.3.2) is
F
hydr

= gA
WP
(
0
) = gA
WP
( Asin t) (3.2)
where = 0 at the point A. Since the wave slope is neglected, the de-
pendence of
0
on is not considered. The rst part gA
WP
is the
38
Figure 3.2: Illustration of hydrostatic force
restoring hydrostatic force which is already present in (1.32). The second
part gA
WP
Asin t is the wave induced hydrostatic force.
The additional hydrostatic moment is determined from the analysis of the
Figure 3.3.
Figure 3.3: Illustration of hydrostatic moment
The relative slope of the ship to the free surface is , where is the wave
39
surface slope
=
d
0
d

0
= Ak cos t =

2
g
Acos t =
A
cos t (3.3)

A
is the amplitude of the angular water plane oscillations. The hydrostatic
pressure increases linearly in direction perpendicular to the free surface plane.
Therefore, the restoring moment is the same as in the case if the free surface is
horizontal and the ship is inclined at the angle . The restoring moment
is known from the ship hydrostatics
M
hydr

= g
0
GM

( ) = g
0
GM

(
A
cos t) (3.4)
The rst part g
0
GM

is the restoring hydrostatic moment, whereas the

second part g
0
GM

A
cos t is the wave induced hydrostatic moment.
3.2 Hydrodynamic Krylov - Froude force
Hydrodynamic forces arise due to wave induced hydrodynamic pressures.
From the Bernoulli equation the pressure is (see (1.13)).
p =
u
2
2
gz

t
+ p
a
(3.5)
The constant pressure p
a
does not need to be considered since being inte-
grated over the ship wetted surface results in zero force and moment. The
rst term in (3.5) is neglected within the linear theory under consideration.
The second term results in force and moment considered above in the sec-
tion 3.1. The remaining term p
unst
=

t
is responsible for hydrodynamic
eects caused by waves. If the interaction between the ship and incident
waves is neglected (Krylov - Froude formalism) the potential can be written
as the potential of uniform unsteady parallel ow:
= u

(t) (3.6)
where u

(t) is unsteady velocity of the ow in the wave

u

(t) =
d
0
dt
(3.7)
Since the unsteady pressure p
unst
=

t
=
u

t
is zero at = 0 the
total pressure is equal to the atmospheric pressure p = p
a
(see (3.5)). The
gradient of the hydrodynamic pressure in vertical direction reads:
p
unst

t
_
=

t
_

_
=
u

(t)
t
=

0
=
2
Asin t
(3.8)
40
The unsteady pressure at the point < 0 is then:
p
unst
() = p
unst
( = 0)

_
0
p
unst
z
dz =
0
_

p
unst
z
dz
The force caused by p
unst
on each frame is calculated by the integration of
the pressure over the frame wetted area
dF
dyn

=
_
p
unst
cos(n)dC =
_
_
_
0
_

p
unst
z
dz
_
_
cos(n)dC =
=
_
_
_
0
_

2
Asin tdz
_
_
cos(n)dC =
2
Asin t
_
cos(n)dC
(3.9)
Here the normal vector is the inward normal vector.
Since the integral
_
cos(n)dC is equal to the frame area taken with the
opposite sign, i.e. A
f
, the hydrodynamic force caused by waves takes the
form:
dF
dyn

=
2
Asin tA
f
(3.10)
Being integrated along the ship length this force gives the force acting on the
whole ship length
F
dyn

=
L
_
0
dF
dyn

d =
L
_
0

2
Asin tA
f
d =
=
2
Asin t
L
_
0
A
f
d = m
2
Asin t = m

0
(3.11)
The hydrodynamic moment acting on the ship frame
dM
dyn

=
_
p
unst
( cos(n) cos(n))dC =
=
2
Asin t
_
( cos(n) cos(n))dC
(3.12)
Within the linear theory considering small ship slopes the last integral in (3.12)
is zero, i.e.
dM
dyn

= 0 M
dyn

=
L
_
0
dM
dyn

= 0 (3.13)
41
3.3 Full Krylov - Froude force and moment
The full Froude Krylov force takes the form:
F
l

= gA
WP
Asin t m
2
Asin t = gA
WP
Asin t + m

0
(3.14)
The rst term is caused by hydrostatic eect, whereas the second one by
hydrodynamic eects. The second term is referred in the literature to as the
Smith eect.
The full Froude Krylov moment contains only the wave induced hydrostatic
component:
M
l

= g
0
GM

A
cos t (3.15)
3.4 Force and moment acting on the ship frame
in accelerated ow
These forces are determined using the concept of the relative motion. Let
us
j
is a ship displacement in j-th direction. As it has been explained in pre-
vious chapters, the force acting on the body moving with the acceleration

j
in a liquid at rest is equal to the product of added mass with the acceleration
taken with opposite sign, i.e. A
jj

j
. If the liquid moves with the accelera-
tion

jL
relative to motionless body, the force acting on the body is towards
the acceleration direction, i.e. A
jj

jL
. If both body and liquid move with ac-
celerations the total force is A
jj
(

jL
). Similarly, the damping force can
introduced being proportional to the relative velocity B
jj
(

jL
). The
rst components of both forces A
jj

j
and B
jj

j
are already represented
by the rst and the second columns in the motion equations (1.32). The sec-
ond components A
jj

jL
and B
jj

jL
represent the hydrodynamic forces due
to interaction between the incident waves and oating body. Remembering
that

jL
= Acos t and

jL
=
2
Asin t we obtain the lift force caused
by the interaction between the ship and incident wave:
F
2

= A
33

2
Asin t + B
33
Acos t (3.16)
In roll oscillations the ship moves with the angular velocity and angular
acceleration . The free surface oscillates with the angular velocity and
acceleration . Taking from (3.3) we obtain the roll moment caused by
the interaction between the ship and incident wave:
M
2

= A
44

A
cos t B
44

A
sin t = A
44

4
g
Acos t B
44

3
g
Asin t
(3.17)
42
3.5 Full wave induced force and moment
In the section 1.5 we divided the wave induced forces into the Froude Krylov
part and the interaction force. Commonly, the Froude Krylov force is the
dominating part of the wave induced forces.
To calculate the full wave induced force we have to note that the Smith eect
is already represented in the force A
jj

jL
+B
jj

jL
. All hydrodynamic eects
are taken into account. Only the hydrostatic part of the Froude Krylov force
should be added to A
jj

jL + B
jj

j
L to get the full wave induced force:
F
,per
=
_
gA
WP

2
A
33

Asin t + B
33
Acos t (3.18)
The full moment is the sum of (3.15) and (3.17):
M
,per
=
_
g
0
GM

A
44

2
_

2
g
Acos t B
44

3
g
Asin t (3.19)
3.6 Equations of ship heave and roll oscilla-
tions
Substitution of all forces derive above into the original dierential equations
results in two following decoupled ordinary dierential equations
(m + A
33
)
_

0
_
+ B
33
_

0
_
+ gA
WP
(
0
) = m

0
(3.20)
(I
xx
+ A
44
) ( ) + B
44
( ) + g
0
GM

( ) = I
xx
. (3.21)
The solution of both equations can be represented as the sum =
inh
+
free
=
inh
+
free
, where
free
and
free
are free heave oscillations:

free
= Ce

t
(cos

t i sin

t),
free
= Ce

t
(cos

t i sin

t) satis-
fying the homogeneous equations:
(m + A
33
)

+ B
33

+ gA
WP
= 0
(I
xx
+ A
44
) + B
44
+ g
0
GM

= 0.
When the free oscillations decay
free
,
free

t
0, the solutions of the
equation (3.20) and (3.21) tend to the solutions of inhomogeneous equations:
(m + A
33
)

+ B
33

+ gA
WP
= A
33

0
+ B
33

0
+ gA
WP

0
(I
xx
+ A
44
) + B
44
+ g
0
GM

= A
44
+ B
44
+ g
0
GM

.
(3.22)
43
The inhomogeneous equation (3.21) is written in terms of relative roll an-
gle
(r)
= in the normalized form:

(r)
+ 2

(r)
+
2

(r)
=

2
1 + k

A
cos t, (3.23)
where k

= A
44
/I
xx
. The solution of (3.23) is seeking in the form

(r)
=
(r)
A
cos (t

) (3.24)
Substituting (3.24) into (3.23) and separating terms proportional to cos t
and sin t gives two equations:

(r)
A
_
(
2

2
) cos

+ 2

sin

=

2
1 + k

A
(3.25)

(r)
A
_
2

cos

+ (
2

2
) sin

= 0 (3.26)
It follows from (3.25) and (3.26)
_

(r)
A
_
2
_
_

2
_
2
cos
2

+ 4
2

2
sin
2

+ 4

sin

2
_
cos

=

4
(1 + k

)
2

2
A
_

(r)
A
_
2
_
_

2
_
2
sin
2

+ 4
2

2
cos
2

cos

2
_
sin

= 0.
The sum of two last equations
_

(r)
A
_
2
_
_

2
_
2
+ 4
2

2
cos
2

_
=

4
(1 + k

)
2

2
A
allows one to nd the ratio
(r)
A
/
A

(r)
A

A
=

2

/ (1 + k

)
_
_
1
2

_
2
+ 4
2

, (3.27)
where

and

. Eigenfrequency

and damping coecient

are given by formulae (2.3) and (2.4). The phase of the response relative to
that of the input (phase displacement) is found from (3.26):

= arctg
_
2

1
2

_
(3.28)
44
Similar solutions are obtained for the heave oscillations:

(r)
A
A
=

2

/(1 + k

)
_
(1
2

)
2
+ 4
2

(3.29)

= arctg
_
2

1
2

_
(3.30)
with k

= A
33
/m,

and

.
3.7 Analysis of the formula (3.27)
The formula (3.27) can be rewritten as follows:

(r)
A

A
=

A

A
=

2

/ (1 + k

)
_
_
1
2

_
2
+ 4
2

or

A
= 1 +

2

/ (1 + k

)
_
_
1
2

_
2
+ 4
2

(3.31)
The physical meaning of terms in (3.31) is obvious from the following expres-
sion
amplitude of ship roll oscillations
amplitude of wave angle oscillations
= 1 + enhancement (3.32)
Since the function

2

/(1+k

(
1
2

)
2
+4
2

is positive

2

/(1+k

(
1
2

)
2
+4
2

> 0 the ship

roll amplitude is larger than the the amplitude of the angular water plane
oscillations
A
, i.e.

A

A
> 1.
Figure 3.4: Ship as linear system
45
The ship can be considered as a system with the waves as input and the
resulting motion as the output (Fig.3.4). As seen from (3.31) this system is
linear for small amplitude oscillations. In terms of linear system theory the
formula (3.31) reads
output
input
= 1 + enhancement (3.33)
The linear system is time invariant. The output produced by a given in-
put is independent of the time at which the input is applied. The function
1 + enhancement which characterizes the system response in the frequency
domain is called the frequency response function.
The enhancement function

2

/(1+k

(
1
2

)
2
+4
2

goes to zero if referred frequency

becomes zero. At very large frequencies ,

2

/(1+k

(
1
2

)
2
+4
2

1/ (1 + k

).
The enhancement is maximum in the resonance case

=
1

12
2

12

. Strictly speaking the resonance frequency =

12
2

is not
equal to the eigenfrequency

, i.e. >

. Since

is small, this discrep-

ancy can be neglected

. Typical dependence of the ratio

(r)
A

A
on the
referred frequency is presented in Fig. 3.5.
Figure 3.5: Response function versus referred frequency
Typical dependence of the phase displacement on the referred frequency is
presented in Fig. 3.6.
46
Figure 3.6: Phase displacement versus referred frequency
The phase displacement is equal to /2 in the resonance case

= 1 for
every damping. For

0 the phase displacement disappears. For

the phase displacement tends to . The largest relative

roll angle occurs in the resonance case either at wave crests or at wave
troughs (Fig. 3.7). Indeed, the magnitude of the relative roll angle in the
resonance case
(r)
=
(r)
A
cos
_
t

2
_
=
(r)
A
sin t attains the maximum

(r)
A
at sin t = 1. It corresponds to wave crest
0
= Asin t = A and wave
trough
0
= A.
Figure 3.7: Ship oscillations in resonance case
At very large metacentric height GM

the eigenfrequency is also get-

ting large

=
_
g
0
GM

I
xx
+A
44
. The referred frequency for a limited wave
frequency tends to zero

0. The relative roll angle am-

plitude and phase displacement are zero. The oating body moves together
with the free surface as shown in Fig. 3.8 like a raft.
Similar results are obtained from analysis of the heave oscillations formu-
47
Figure 3.8: Oscillation of a raft with a big metacentric height
lae (3.29) and (3.30).
3.8 Sway ship oscillations in beam sea
The equation describing the sway oscillations is (see the second equation in
the system (1.32)):
m = A
22
B
22
+ F
,per
(t) (3.34)
The wave exciting force F
,per
(t) consists of two components of hydrostatic
and hydrodynamics nature. As seen from Fig. 3.9 the hydrostatic force is
F
hyd

= g
0
=
0

2
Acos t = m
2
Acos t (3.35)
Figure 3.9: Illustration of the frame in beam waves
The horizontal oscillations of the wave surface can be presented in harmonic
48
form (see formula (6.23) in [6]):

0
(y, z) = Ae
kz
cos(ky t),

0
(0, 0) = A cos(t)
0
(0, 0) = Asin(t)

0
= A
2
cos t
(3.36)
The hydrodynamic component of the wave induced force is written in the
similar form as (3.16):
F
2

= A
22

0
+ B
22

0
(3.37)
Substitution of (3.37) and (3.35) into (3.34) gives:
m = A
22
B
22
+ m
0
+ A
22

0
+ B
22

0
(3.38)
or
(m + A
22
)(
0
) + B
22
(
0
) = 0 (3.39)
The solution of the equation is written in the form:
(
0
) = Ce
t
(3.40)
which substitution into (3.39) allows one to nd
=
B
22
m + A
22
(3.41)
The parameter is positive. Therefore, (
0
) = Ce
t

t
0
t

0
.
As soon the transitional process is nished, the ship oscillates together with
the wave
=
0
= Asin t (3.42)
3.9 Ship oscillations at nite beam to wave
length ratio and draught to length ratio
The analysis presented above was carried out for the case of a very long wave,
i.e. both the beam to length ratio B/L and the draught to length ratio T/L
are small. The results for roll oscillation obtained for the case B/L 0,
T/L 0 are extended to the case B/L 0(1), T/L 0(1) using reduction
coecients. According to this traditional in shipbuilding approach the wave
amplitude is multiplied with the reduction coecient , i.e.
A
red
= A (3.43)
The ship oscillations at B/L 0(1), T/L 0(1) are smaller than these at
B/L 0, T/L 0 due to two reasons
49
Hydrostatic force is smaller because the submerged volume is smaller
due to wave surface curvature,
Hydrodynamic force is smaller because the velocities caused by the
orbital motion are not constant as assumed above. They decay with
the increasing submergence as exp(kz).
The rst reduction factor is mainly due to the nite beam to length ratio
B/L 0(1).
Let us consider rst the reduction coecient for the heave oscillations. The
factor
B
considers the reduction of the hydrostatic force due to the nite
beam to length ratio. To estimate
B
the xed ship is considered at the time
instant t = /2 when the wave crest is in the symmetry plane (Fig. 3.10).
Figure 3.10:
The free surface ordinate

0
= Asin
_

2
+ k
_
= Acos k
The hydrostatic force obtained in the previous analysis is
R
0
= gAA
wp
(3.44)
50
whereas the actual one is calculated by the integral:
R
true
= gA
_
A
wp
cos kdd = 2gA
L/2
_
L/2
B()/2
_
0
cos kdd =
=
2gA
k
L/2
_
L/2
sin
kB()
2
d
(3.45)
Using the Taylor expansion for sin kB()
sin
kB()
2
=
kB
2

(kB)
3
48
+ ...
the nal formula for R
true
takes the form:
R
true
=
2gA
k
L/2
_
L/2
sin
kB()
2
d
2gA
k
L/2
_
L/2
_
kB/2
(kB)
3
48
_
d =
= gAA
wp

gAk
2
24
L/2
_
L/2
B
3
d = gAA
wp
_
1
k
2
2
I
A
wp
_
,
(3.46)
where
A
wp
=
L/2
_
L/2
Bd, I =
1
12
L/2
_
L/2
B
3
d,
The reduction of the hydrostatic force can be taken by the following coe-
cient into account:

B
=
gAA
wp
_
1
k
2
2
I
A
wp
_
gAA
wp
= 1
k
2
2
I
A
wp
(3.47)
The second reduction factor is mainly due to the nite draught to length ratio
T/L 0(1). The factor
T
considers the reduction of the hydrodynamic
force due to the nite draught to length ratio. The reduction coecient is
given here without derivation:

T
= 1
_
2
T
L
_
+

2(2 )
_
2
T
L
_
2

6(3 2)
_
2
T
L
_
3
(3.48)
51
where is the coecient of the lateral area = A
LA
/(LT).
The total reduction coecient

is calculated as the product of

B
and
T
neglecting their mutual inuence:

=
B

T
(3.49)
The formula (3.49) is valid at
L
B
> 4,
L
T
> 8. For heave calculations one can
use the formula (3.29) with A

instead of A.
Reduction coecient of the roll oscillations can be calculated from the ex-
pression gained from regression of experimental data:

= exp
_
4.2 (R

)
2
_
, R =

_
_
BT
r

/GM

2g
_
1/2
(3.50)
Here r

is the metacentric radius. Amplitude of roll oscillations is found

from (3.31) with
A

instead of
A
. A sample of the reduction coecient
for a real ship is presented in Fig. 3.11.
Figure 3.11: Reduction coecient of the heave oscillations
3.10 Eect of ship speed on rolling
In the previous chapters the ship speed was assumed to be zero. If the ship
moves in waves with speed v in x-direction the following modications should
be made in theory
52
Figure 3.12: Sea classication
the added mass A
ij
and damping coecients B
ij
depend on the en-
counter frequency

e
=
2
v cos
wave
g
where is encounter angle (Fig. 3.12)
reduction coecient x (3.43) depends on
In the hydrodynamic theory one uses time derivative in ship coordinate
system

t
which can be expressed through the time derivative in the
inertial reference system

t
0
as

t
0
=

t
v
The formulae (3.27 - 3.30) can be used also in the case v = 0 with the
substitution
e
instead of .
53
54
Chapter 4
Ship oscillations in small head
waves
4.1 Exciting forces and ship oscillations
Let us consider the ship oscillations in small head waves coming from the
stern (
wave
= 0

), where
wave
is the wave course angle. The wave ordinate,
wave orbital motion velocity and acceleration are:

0
= Asin(t k) (4.1)

0
= Acos(t k) (4.2)

0
=
2
Asin(t k) (4.3)
Figure 4.1: Illustration of the ship in head waves
Within the linear theory of ship oscillation the ship is considered at rest. The
perturbation force acting on the section AB (Fig. 4.1) can be represented as
the sum of
the hydrostatic Froude Krylov force gB()
0
(),
the hydrodynamic force A
33

0
() + B
33

0
()
55
dF
,per
d
= A
33

0
() + B
33

0
() + gB()
0
() (4.4)
Here we used the principle of relative motion (see section 3.4) for a ship
frame. Integrating
dF
,per
d
over the ship length we obtain the whole wave
induced force F
,per
. If the ship is symmetric with respect to the midship
B() = B(), A
f
() = A
f
(), A
33
() = A
33
(), B
33
() = B
33
(), the
terms with sin(k) are neglected and the formula for F
,per
is simplied to:
F
,per
=
2
A
_
L
A
33
cos kd sin t + A
_
L
B
33
cos kd cos t+
+ A
_
L
gB() cos kd sin t = A
_
L
_
gB()
2
A
33
_
cos kd sin t+
+ A
_
L
B
33
cos kd cos t = F
,per
sin (t
,per
)
(4.5)
where
F
,per
= A

_
_
_
_
L
(gB()
2
A
33
) cos kd
_
_
2
+
2
_
_
_
L
B
33
cos kd
_
_
2
,

,per
= arc tan
_
_

_
L
B
33
cos kd
_
L
(gB()
2
A
33
) cos kd
_
_
;
(4.6)
The wave exciting moment is calculated by multiplication of
dF
,per
d
with the
arm :
M
,per
=
_
L

dF
E,
d
d
= A
_
L
_
gB()
2
A
33
_
sin k cos t
A
_
L
B
33
sin k sin t = M
,per
sin (t
,per
)
(4.7)
56
where
M
,per
= A

_
_
_
_
L
(gB()
2
A
33
) sin kd
_
_
2
+
2
_
_
_
L
B
33
sin kd
_
_
2
,

,per
= + arc tan
_
_
_
L
(gB()
2
A
33
) sin kd

_
L
B
33
sin kd
_
_
;
(4.8)
Substitution (4.5) and (4.7) in the third and sixth equations of the sys-
tem (1.3.2) gives:
m

= A
33

B
33

gA
WP
+ F
,per
sin (t
,per
) (4.9)
I
yy

= A
55

B
55

g
0
GM
L
+ M
,per
sin (t
,per
) (4.10)

(m + A
33
)

+ B
33

+ gA
WP
= F
,per
sin (t
,per
) (4.11)
(I
yy
+ A
55
)

+ B
55

+ g
0
GM
L
= M
,per
sin (t
,per
) (4.12)
Dividing both equations by the coecient of the rst term one obtains:

+ 2

+
2

= f

sin (t
,per
)

+ 2

+
2

= f

sin (t
,per
)
(4.13)
where
f

=
F
,per
m + A
33
, f

=
M
,per
I
yy
+ A
55
,

=
B
33
2(m + A
33
)
,

=
B
55
2(I
yy
+ A
55
)
,

=
_
gA
WP
m + A
33
,

g
0
GM
L
I
yy
+ A
55
.
Solution of (4.13) is seeking in the form
=
A
sin
_
t
,per

per

_
, =
A
sin
_
t
,per

per

_
. (4.14)
After some simple manipulations the amplitudes of the heave and pitch oscil-
lations as well as the phase displacements are obtained from (4.13) and (4.14):

A
=
f

_
(
2

2
)
2
+ 4
2

2
,
per

= arc tan
_
2

2
_
(4.15)
57

A
=
f

_
(
2

2
)
2
+ 4
2

2
,
per

= arc tan
_
2

2
_
(4.16)
In the resonance case the phase displacement is equal to /2, i.e.

per

= /2 in case

= and
per

= /2 in case

= .
4.2 Estimations of slamming and deck ood-
ing
Results of ship oscillations obtained in the previous section can be used for
practically useful estimations. For instance, we can estimate the slamming
and deck ooding. Using the relations derived above

0
= Asin(t k), =
A
sin(t
,per

per

), =
A
sin(t
,per

per

)
(4.17)
one can display the ship positions in head waves as shown in Fig. 4.2
Figure 4.2: Position of ship at dierent time instants in a head wave
Let us represent the formulae (4.17) in the form:

0
= Asin t cos k Acos t sin k
=
A
sin t cos
_

,per
+
per

A
cos t sin
_

,per
+
per

_
=
A
sin t cos
_

,per
+
per

A
cos t sin
_

,per
+
per

_
.
(4.18)
The local change of the draft is:
z(x) =
0
+ x = f
1
(x) cos t + f
2
(x) sin t (4.19)
where
f
1
(x) = Acos k +
A
sin(
,per
+
per

) x
A
sin(
,per
+
per

)
f
2
(x) = Asin k
A
cos(
,per
+
per

) + x
A
cos(
,per
+
per

)
(4.20)
58
Figure 4.3: Curves y = z
max
and y = z(x)
A sample of the curve y = z(x) is shown in Fig. 4.3. The maximum draft is
then
z
max
(x) =
_
f
2
1
(x) + f
2
2
(x) (4.21)
The curve y = +z
max
(x) shows the contour of maximum wave elevations
along the ship board in the symmetry plane whereas the curve y = z
max
(x)
the minimum wave elevations. Both curves are symmetric with respect to
the equilibrium water plane.
Deck ooding takes place if z
max
(x) > H, where H is the board height.
Slamming takes place if z
max
(x) > T
There three zones limited by curves y = z
max
(x) can be distinguished along
the ship board (see Fig. 4.3):
Allways dry area (white),
Allways wetted area (red),
Intermediate area (orange).
A sample of ooding curves for a real ship is given in Fig. 4.4
59
Figure 4.4: Sample for a real ship
60
Chapter 5
Seasickness caused by ship
oscillations
Symptoms of the seasickness are giddiness (Schwindelgef uhl), headache (Kopf-
schmerz), sickness (

Ubelkeit) and vomiting (Erbrechen). The seasickness

is the reason of work capacity reduction, memory decline (R uckgang der
Ged achtnisleistung), motion coordination (Bewegungskoordinierung), reduc-
tion of muscular strength, etc. Diagram of Sain Denice (Fig. 5.1) shows the
inuence of the vertical acceleration on the seasickness depending on the
oscillation period.
For the irregular sea state the similar diagram was proposed by Krappinger
(Fig. 5.2) who estimated the percentage of people suering from the seasick-
ness depending on the root mean square deviation and frequency.
According to standards developed in US Navy the oscillations have no sig-
nicant eect on the work capacity if the amplitude of the roll oscillations
is under eight degrees, the amplitude of pitch oscillation is below three de-
grees, the vertical accelerations does not exceed 0.4 g whereas the transversal
accelerations 0.2 g. The upper limit of the roll angle for the deck works is
20 degrees which corresponds to the reduction of the work capacity of about
50 percent.
At present the seasickness has insuciently been studied in medical science.
As shown in the study by Vosser, the seasickness is developed at a certain
level of overloads and then can remain even the ship oscillations decay. It
is shown in diagram 5.3 presenting the number of passenger on a cruise
liner n

/n suering from the seasickness during eighty hours of the journey.

At the journey beginning the vertical acceleration

/g was less than 0.1 and
only 16 percent of passenger were sick. As soon as the vertical acceleration
61
Figure 5.1: Inuence of the vertical acceleration on the seasickness depending
on the oscillation period
attained 0.4g more than 80 percent of passengers were sick. In spite of the
ship oscillation decay after 30 hours of the way the number of sick passen-
gers is not reduced. On the contrary this number is slightly increased during
the next 24 hours. Only after 36 hours the seasickness retreated. The next
diagram 5.4 illustrates the fact that the adaption to seasickness is relatively
weak.
Diathesis to seasickness depends on the individual properties of organisms.
There are many people who had never had problems with seasickness. How-
ever there are experienced seamen who suers from this sickness the whole
professional life.
62
Figure 5.2: Inuence of the vertical acceleration on the seasickness depending
on the oscillation period
Figure 5.3: Number of passengers suering from seasickness on a cruise liners
depending on vertical accelerations
63
Figure 5.4: Adaption to seasickness
64
Chapter 6
Ship oscillations in irregular
waves
6.1 Representation of irregular waves
The irregular waves can be both two dimensional and three dimensional (Fig. 6.1).
Figure 6.1: Irregular seawaves, 1- two dimensional, 2- three dimensional.
(Fig. from [3])
A feature of the irregular waves distinguishing them from regular ones is
the non-recurrence of their form in time (Fig. 6.2). The following relations
between wave lengths L and wave heights h are recommended in practical
calculations for swell:
Within the linear theory the irregular waves can be represented as the super-
position of regular waves with dierent amplitudes, frequencies and course
angles, as shown in Fig. 6.3.
65
Figure 6.2: Prole of an irregular wave. (Fig. from [3])
h = 0.17L
3/4
Zimmermann,
h = 0.607L
1/2
British Lloyd,
h = 0.45L
0.6
det Norske Veritas.
6.1.1 Wave ordinates as stochastic quantities
The wave ordinate is the stochastic function with a certain probability density
function (see Fig. 6.1). The p.d.f. distribution of the real irregular wave
ordinates is Gaussian. i.e,
p.d.f. =
1
_
2D

e
(
0
)
2
/(2
2
)
, (6.1)
where
0
is the mathematical expectation (in our case
0
= 0), is the
standard deviation:

2
= (
0
)
2
= D

(6.2)
D

is the dispersion. Probability P(

1
< <
2
) of the event, that the
ordinate lies in the range between
1
and
2
is then
P(
1
< <
2
) =

2
_

1
p.d.f.()d =
1
_
2D

2
_

1
e

2
/(2
2
)
d =
1

2
/

2D

1
/

2D

e
t
2
dt
(6.3)
The last integral is known as the probability integral
(x) =
2

x
_
0
e
t
2
dt (6.4)
66
Figure 6.3: Representation of irregular wave through the superposition of
regular waves. (Fig. from [3])
satisfying the following properties
(x) = (x), () = 1, () = 1 (6.5)
Using the probability integral, the probability P(
1
< <
2
) takes the form
P(
1
< <
2
) =
1
2
_

_

2
_
2D

_

1
_
2D

__
(6.6)
The probability P(< <
2
) = P( <
2
) is the probability of the event
that does not exceed
2
:
P( <
2
) =
1
2
_
1 +
_

2
_
2D

__
(6.7)
The probability P(
1
< < ) = P(
1
< is the probability of the event
that larger than
1
:
P(
1
< ) =
1
2
_
1
_

1
_
2D

__
(6.8)
67
Figure 6.4: p.d.f. of the wave ordinate
In the probability theory is shown that the p.d.f. of the amplitude of a
stochastic quantity having the Gaussian p.d.f. distribution satises the Raleigh
law:
p.d.f.(
a
) =

a
D

2
a
/(2D

)
(6.9)
The probability that the amplitude is larger than

is
P(
a
>

) =

a
D

2
a
/(2D

)
d
a
= e

2
/(2D

)
(6.10)
When evaluating the wave height an observer determines the middle height
of one third of the highest waves. This height is referred to as the signicant
wave height and designated as h
1/3
.
Dependence between the dispersion and the signicant wave is
D

= 0.063h
2
1/3
(6.11)
6.1.2 Wave spectra
Irregular waves are considered as the superposition of innite number of
regular waves of dierent frequencies, amplitudes and course angles (Fig. 6.3).
68
According to this concept the wave elevation (x, y, t) is represented in form
of Fourier - Stieltjes integral:
(x, y, t) = Real
__
dA(, )exp[ik(x cos + y sin ) + it + (, )]
(6.12)
Here is the wave frequency, k is the wave number k =
2
/g, is the
wave course angle and (, ) is the phase angle. The quantity dA(, , t) is
the function of the amplitude corresponding to the wave propagating at the
course angle < < + with the frequency < < + . The
mean square elevation is obtained from time averaging the quadrat of the
elevation:

2
(x, y) = lim
T
1
T
T
_
0

2
(x, y, t)dt =
=
__
dA(, )exp[ik(x cos + y sin ) + it + (, )]
__
dA

(
1
,
1
)exp [
ik
1
(x cos
1
+ y sin
1
) i
1
t (, )] =
1
2
__
dA(, )dA

(, )
(6.13)
Here the superscript stands for the complex conjugate amplitude function.
Rigorous derivation of the formula (6.13) can be found in [10]. Multiply-
ing
2
(x, y) with g
g
2
(x, y) =
g
2
__
dA(, )dA

(, ) (6.14)
and comparing the result with the expression for the energy (6.34) derived
in [6]
E = T
Fl
+ E
p
=
gA
2
2
L 1m
One can conclude that g
2
(x, y) is the time averaged energy per surface
unit. Using the representation of the integral
__
dA(, )dA

(, ) = 2
2
_
0

_
0
S

(, )dd (6.15)
we introduce the spectral density of the irregular waves S

(, ) which is
the contribution of the wave with the frequency < < + and the
69
course angle < < + to the irregular wave energy. Commonly the
function S

(, ) is called shortly the wave spectrum.

At present there is no much information on the energy distribution both on
the frequency and the course angle. The typical measurements with buoy
do not provide information about the dependency of wave elevations on the
course angle. In the ship theory is assumed that the irregular waves have a
preferential propagation direction and the wave have long wave crest. The
waves are approximately two dimensional. Such rough sea can fully be char-
acterized by the frequency spectrum S

() dening as
S

() =
2
_
0
S

(, )d (6.16)
The spectrum of the wave state S

() shows the distribution of the wave en-

ergy on frequencies. The two dimensional spectrum S

(, ) can be restored
from the one dimensional one S

() using the following simple approxima-

tion:
S

(, ) =
4
3
S

() cos
4

To determine the spectrum, the wave ordinates are measured and represented
in Fourier series. The energy E( < < + ) is calculated as the
squared wave ordinate for each interval of the frequencies . The spectral
density of waves is calculated as
S

() = lim

E( < < + )

(6.17)
From the probability theory:
D

_
0
S

()d (6.18)
One of the most popular wave spectral densities is the spectrum of Pierson
and Moskowitz (PM):
S

() =
g
2

5
exp
_

_
g
U
_
4
_
, (6.19)
where = 0.0081, = 0.74, U is the wind velocity at the height of 19.4
m over the free surface. The spectrum (6.19) has been obtained by approx-
imation of data measured in 1964 in the North Atlantic region. Fig. 6.5
70
Figure 6.5:
illustrates the PM spectra depending on the wind velocity U.
The mean wave height is

h = 2

_
0

a
p.d.f.(
a
)d
a
= 2

_
0

a
D

2
a
/(2D

)
d
a
= (2D

)
1/2
(6.20)
The signicant wave height is:
h
1/3
=
2

a
p.d.f.(
a
)d
a

1
p.d.f.(
a
)d
a
(6.21)
where the amplitude
1
is chosen from the condition

1
p.d.f.(
a
)d
a
= 1/3 e

2
1
/(2D

)
= 1/3
1
=
_
2D

ln3 (6.22)
It follows from (6.9) and (6.21):
h
1/3
= 4
_
D

(6.23)
71
The middle frequency is dened as
=
_

_
0

2
S

()d

_
0
S

()d
_

_
1/2
=
_

_
0

2
S

()d
D

()
_

_
1/2
(6.24)
Substitution of (6.19) into (6.18), (6.23) and (6.24) results in
h
1/3
= 2
U
2
g
_

_
1/2
, = ()
1/4
(g/U) (6.25)
6.2 Calculation of ship oscillations in irregu-
lar waves
Using the assumption of small waves we can substitute the superposition of
regular waves into equations (3.20) and (3.21) describing the heave and roll
oscillations. Since the equations are linear the responses of the ship to each
regular wave can be calculated separately. In this case one can obtain the
history of oscillations in time.
However, from point of view of practical applications only the statistical pa-
rameters of oscillations are of importance. To determine them, the ship is
considered as the dynamic system. The seaway is the input which is trans-
formed by the ship into oscillations considered as the output. In the statisti-
cal theory shown, that if the input signal has the Gaussian p.d.f. distribution
the output signal has also the Gaussian p.d.f. distribution. With the other
words, the ship oscillation parameters (roll angle, etc) obey the normal Gaus-
sian law whereas the amplitudes of oscillation parameters satisfy the Raleigh
law. The only unknown value in these distributions laws is the dispersion D.
Let us consider the roll oscillations of a ship with the zero forward speed.
As shown in the previous lectures the ratio of the roll oscillations amplitude
(output signal) to the wave slope amplitude (input signal) is given by the
formula

(r)
A

A
=

A

A
=

2
/(1 + k

)
_
(1
2
)
2
+ 4
2

2

A
=
=

2
/(1 + k

)
_
(1
2
)
2
+ 4
2

2
+ 1 = ()
72
where
() =

2
/(1 + k

)
_
(1
2
)
2
+ 4
2

2
+ 1 (6.26)
is the so called response function. Since the wave spectral density is propor-
tional to the wave ordinates squared and taking the superposition principle
into account, we obtain the following relation between the spectral density
of the seaway and the spectral density of oscillations
S

() =
2
()S

() (6.27)
or
S

() = S

()
_

2
/(1 + k

)
_
(1
2
)
2
+ 4
2

2
+ 1
_
2
(6.28)
Dispersion of the roll oscillations and the standard deviation are found from
the denitions (6.2) and (6.18)
D

_
0
S

()d =

_
0

2
()S

()d,

=
_
D

(6.29)
Similar formulae can be obtained for angular roll velocity and acceleration
D

=

_
0

2
S

()d,

=
_
D

(6.30)
D

=

_
0

4
S

()d,

=
_
D

(6.31)
The dispersions obtained from (6.29), (6.30) and (6.31) determine fully the
irregular ship oscillations in heavy seaway. Using them the following further
parameters can be calculated
Most probable amplitude of oscillations corresponding to the maximum
of the p.d.f.(
a
) distribution

m
=

(6.32)
The probability that the roll amplitude exceeds
m
is 60.6 %.
Averaged amplitude of roll oscillations (mathematical expectation)
=
_

1.25

(6.33)
The probability that the roll amplitude exceeds is 45.6 %.
73
The probability that the roll amplitude exceeds the value

:
p(
A
>

) = e
0.5(

)
2
(6.34)
Averaged frequency and averaged period of oscillations:

,

T

=
2

= 2

(6.35)
Number of ship inclinations (semi periods) within the time interval t:
N
t
=
2t

(6.36)
Number of ship inclinations within the time interval t provided the roll
angle amplitude is larger than

:
N

= N
t
P(
A
>

) =
2t

e
0.5(

)
2
(6.37)
The formulae (6.32) - 6.37) are derived under assumption that the oscillations
obey the Gaussian distribution law.
74
Chapter 7
Experimental methods in ship
seakeeping
7.1 Experiments with models
The seakeeping experiments are performed under condition that the Froude
number of the model Fn
m
and the large scale ship Fn
s
are equal:
F
n
m
= F
n
s

V
m
_
gD
m
1/3
=
V
s
_
gD
s
1/3
Since the periodic motion are considered, the similarity of Strouhal numbers
should also be satised:
Sh
m
= Sh
s


m
L
m
V
m
=

s
L
s
V
s
,
Where is the frequency. Unfortunately the similarity with respect to
Reynolds (viscosity eects) and Weber (spray eects) numbers are not ful-
lled:
Re
m
= Re
s
, We
m
= We
s
.
It is recommended to choose the model length from the condition Re 10
6
.
If is the scale factor:
L
s
= L
m
The following relations derived from the similarity conditions are valid
75
angles
s
=
0

m
linear accelerations

s
=
0

m
linear velocities

s
=
1/2

m
periods
s
=
1/2

m
angular velocities
s
=
1/2

m
frequencies
s
=
1/2

m
angular acceleration
s
=
1

m
displacement, mass, forces F
s
=
3
F
m
inertia moments I
s
=
5
I
m
(7.1)
The following parameters of the model are to be determined before experi-
ment:
weight G
positions of the center of gravity x
g
, z
g
metacentric heights GM

, GM
l
inertia moments I
xx
, I
yy
, I
zz
They should satisfy the similarity conditions:
G
s
=
3
G
M
(x
gs
, z
gs
) = (x
gm
, z
gm
)
GM
,ls
= GM
,lm
I
xxs
=
5
I
xxm
, I
yys
=
5
I
yym
I
zzs
=
5
I
zzm
.
Since the draught and beam are approximately equal B T one assumes
that I
yy
I
zz
.
For the determination of I
zz
the model is hanged out as shown in Fig. 7.1.
The model is oscillating in horizontal plane about the vertical axis as shown
in Fig. 7.1. The period of oscillation is measured. The inertia moment is
calculated then from formula:
I
zz
= I
yy
=
G
m
a
2

2
4
2
l
The additional loads are placed on the ship and they are shifted along the
x-axis as long as the condition
76
Figure 7.1: Determination oft he inertia moment I
zz
I
zzs
=
5
I
zzm
Is fullled.
The inertia moment I
xx
and the center of gravity z
g
is determined using the
setup shown in Fig. 7.2.
To determine z
g
only one load is used, which causes the model heeling. If is
the heel angle, the distance a calculated from moment equilibrium equation:
a G
m
= Pl a =
Pl
G
m

The gravity center ordinate is then:

z
g
= z
n
a
The inertia moment I
xx
is determined when the model is forced to roll with
the period of free roll oscillations. The moment I
xx
is then calculated as
I
xx
= G
m
((

2
)
2

a
g
)
The similarity conditions
I
xxs
=
5
I
xxm
and
zg
s
= z
gm
77
Figure 7.2: Determination of I
xx
and z
g
are fullled by vertical and horizontal shifts of loads P.
The metacentric height GM

is determined from heel tests (see Fig. 7.3).

GM

m
=
Pl
G
m

The forces arising in roll oscillations are found from free roll oscillations with-
out waves. The model is brought from the equilibrium state and experiences
free decaying oscillations
=
0
e

cos
1
t,
where
1
=
_

. Since
2

the frequencies
1
and

are ap-
proximately equal, i.e.
1

.
Once the period of free oscillations T

is measured, the frequency

=
2
T

and the sum I

xx
+ A
44
are calculated (see table 2.1):
I
xx
+ A
44
=
g
0m
GM
m

(7.2)
Since the inertia moment is known, the added mass A
44
is found from (7.2).
78
Figure 7.3: Heel test
A
44
=
g
0m
GM
m

I
xx
The damping factor

is determined from the denition formula (2.16).

ln
(t)
(t + T

)
=
2

where the (t) and (t + T

) are measured.
The damping coecient B
44
is also calculated from its denition
B
44
= 2

(I
xx
+ A
44
).
The results are represented depending on the magnitude of . Since the
measurements are performed only at

, the results at =

are not
obtained in these measurements.
For ships with large damping the free oscillations decay very quickly. It causes
big error in data analysis. For this case the method of forced oscillations is
applied. The model is forced to roll using horizontally oscillating load as
shown in Fig. 7.4. The load produce the perturbation moment
M
per
= M
0
sin t
The roll oscillations without incident waves are described by the ordinary
dierential equation of the second order:
79
Figure 7.4: Method of forced rolling
+ 2

+
2

=
2

M
0
GGM

sin t
which has the solution
=
0
sin(t

)
where

0
=
M
0
GGM

1
_
(1
2
)
2
+ 4
2

2
(7.3)

, =

= a tan
2

1
2
(7.4)
The measured quantities are M
per
and roll angle .
The damping factor is calculated from (7.3) at dierent frequencies . The
method of forced roll oscillations is very accurate only at

, i. e. in
the resonance case.
For reliable determination of inertial and damping forces depending on fre-
quency it is necessary to apply more complicated setups than these de-
scribed above. A description of these setups can be found in [11].
80
The main purpose of seakeeping measurements in regular waves is the ex-
perimental determination of dependencies

A
(r)

A
(),

(),

A
(r)
A
() and

()
(see formulae (3.27 - 3.30) and Fig. 3.5, 3.6).
Knowledge of response functions

A
(r)

A
() and

A
(r)
A
allows to calculate the
spectra of ship oscillations.
The seakeeping tests in regular and irregular waves Fig. (7.5) are performed
in seakeeping and manoeuvring basin. The seakeeping basins can be open or
closed. In the closed basin the irregular waves are generated using segmented
wave generators consisting of hinged aps. Each ap is controlled separately
by a driving motor.
The seakeeping basin of MARIN (Fig. 7.6) has dimensions 170 x 40 x 5m.
The wave generator (Fig. 7.7) produces waves with signicant wave height
of 0,45 m and a peak period of 2 sec. The irregular waves have a prescribed
spectrum. The model is either self propelled (free running test) or carried
by the carriage with the speed up to 6m/sec. Model length range is from 2m
to 8m. Additionally to waves the wind is generated by an adjustable 10m
wide platform with electrical fees. Free running tests are performed such
that the model follows an arbitrary pre-dened track through the basin. The
seakeeping tests in open basin are performed under condition that the free
waves, generated naturally, have desirable heights and periods.
Figure 7.5: Seakeeping test at MARIN ([4])
One of the most important aims of seakeeping tests in irregular waves is the
evaluation of slamming (Fig. 7.5) and ooding.
81
Figure 7.6: Scetch of the MARIN Seakeeping basin ([4])
Figure 7.7: Wave generator of MARIN Seakeeping basin ([4])
7.2 Seakeeping tests with large scale ships
The most reliable evaluation of ship seakeeping performances can be gained
on the base of tests with large scale ships.
The rst task within framework of such measurements is the determination
of sea state. The wave heights are measured using bues, hydrostatic pressure
sensors, or by stereo photography from airplanes (see Fig. 7.8).
Figure 7.8: Method of wave detection
The aim of these measurements is the wave spectra S

(, ).
The ship performs tack motions as long as the one hundred full oscillations
82
occur (Fig. 7.9).
Figure 7.9: Ship motion during large scale tests
The time history of all kinematic parameters is documented by dierent
sensors. Fig. 7.9 illustrates the dependence of the roll angle on the time
(t). The time signals of kinematic parameters are evaluated using Fourier
analysis. After that the spectra of ship kinematic parameters are calculated.
Knowing the spectra, the repones function can be determined from (6.27) as
the nal aim of the large scale tests:
() =

()
S

()
83
84
Chapter 8
Ship oscillation damping
(stabilisation)
8.1 Damping of roll oscillations
8.1.1 Passive means
U-tube Passive Roll stabilization system

eigenfrequency of the ship

z
eigenfrequency of uid oscillations in tank
The Frahm systems ([12]) belong to the resonance type. The duct and air
channel ore selected to fulll the conditions:

/
z
1 (8.1)

z
=

2
85
Disadvantage: The draught of ship is changed in operations.

is changed.
But
z
remains constant The system becomes non ecient.
U-tube stabilization system of Frahm of the second type
Figure 8.1: U-tube stabilization system of Frahm of the second type

z
1
Free surface Type passive Roll stabilization systems of Flume
Simplied mathematical model.
Equation without stabilizer:
A
44
+ B
44
+ g
0
GM
j
= F

A
44
- eective mass moment of inertia
86
Figure 8.2: Free surface Type passive Roll stabilization systems of Flume

z
is changed in a wide
range due to change of
water level in tank.
Figure 8.3: Free surface Type passive Roll stabilization systems of Flume
B
44
- damping coecient
F

- wave induced forces

Equation with stabilizer:
(A
44
+ m
T
) + (B
44
+ N
T
) + (B
T
+ g
0
GM
j
) = F

m
T
- eective mass moment of inertia of water in stabilizer unit
N
T
- damping coecient of stabilizer unit
B
T
- restoring moment coecient of stabilizer unit
The values m
T
, N
T
and B
T
are related only to the dierence between an
active stabilizer and a condition where the uid is replaced by a solid mass
corresponding to the frozen liquid.
A
44
+ B
44
+ (tg
0
GM
j
) = F

F
T
m
T
+ N
T
+ B
T
= F
T
87
Solution:
=
A
cos t
F
T
= F
TA
cos(t
FT
)
F
TA
A
=
_
(B
T
m
T

2
)
2
+ (N
T
)
2

Ft
= tan
N
T

B
T
m
t

A
destabilized

A
=
N
44
_
_
(g
0
GM
j
A
44

2
)
2
+ B
2
44

2
+ B
2
44

A
stabilized

A
=
N
44
_
(g
0
GM
j
A
44

2
+ B
T
m
T

2
)
2
+ (N
T
+ B
44
)
2

2
8.1.2 Active stabilizer
Figure 8.4: Active stabilizer
8.1.3 Passive Schlingerkiel
p =
S
k
LB
; r
0
=
r
k
B
b
k
l
k
= 0.01...0.02
88
Figure 8.5: Passive Schlingerkiel
Figure 8.6: Active rudders
8.1.4 Active rudders
8.1.5 Damping of pitch oscillations
Figure 8.7: Damping of pitch oscillations
= k
1

Q
+ k
3
_

Q

Q
_
3
Utilizes Hoppe/Flume data with data from > 300 vessels seakeeping
test and > 2000 tank model test
89
Figure 8.8: Damping of pitch oscillations
Figure 8.9: [5]
Database allows unique analysis on vessel-tank combination of existing
vessels
high accuracy for initial tank design consulting
90
Figure 8.10: [5]
Figure 8.11: [5]
91
Figure 8.12: [5]
Figure 8.13: [5]
92
Figure 8.14: [5]
Figure 8.15: [5]
93
Figure 8.16: [5]
94
Chapter 9
Parametric oscillations
Parametric oscillations arise when one of parameters characterizing the os-
cillating system depends periodically on time. Parametric ship oscillations
arise due to periodic change of the metacentric height:
GM

= GM
0

+ A
GM
cos t (9.1)
which is in previous chapters assumed to be constant, i.e.
GM

= GM
0

.
The motion equation of free roll oscillations with variable metacentric height
is:
I
xx
+ A
44
B
44
+ g
0
(GM
0

+ A
GM
cos t) = 0
or:
+ 2

+
2

cos t,
where

=
B
44
2(I
xx
+ A
44
)
,
2

=
g
0
GM
0

I
xx
+ A
44

=
A
GM
g
0
I
xx
+ A
44
(9.2)
Analyzing (9.2) one can state that change of the metacentric height can cause
the perturbation moment resulting in parametric oscillations.
Physical reason for the appearance of the additional perturbation moment is
the eect of the ship submergence change during roll and vertical oscillations.
The change of the roll angle causes the moment.
95

0
gGM

(9.3)
The change of the ship draught results in an additional moment
g
0
A
GM
cos t
which is proportional to and .
The equation (9.2) has no analytic solution in elementary functions. How-
ever, properties of its solution are well known. It can be shown that if the
frequency of perturbation moment wave frequency is twice as large as the
free roll oscillation frequency, i.e. = 2

, the parametric resonance takes

places. Since the perturbation moment depends on heave , the parametric
resonance takes place when the frequency of vertical oscillations is approx-
imately twice as large as the frequency of roll oscillations. With the other
words, during semi period of roll oscillations the heave change performs the
full period oscillation as shown schematically on Fig. 9.1. The parametric
resonance is typical for ships with big distance between the center of gravity
and water plane surface (Fig. 9.2).
For parametric oscillation it is necessary that ship is brought from the equi-
librium state by a certain perturbation. The natural reason of such per-
turbation for roll oscillations is beam seaway. However, the parametric roll
oscillation can arise also in head waves. If the ship has a certain roll angle in
head waves, the hydrostatic lift force becomes larger at the wave crest and
smaller at wave valley. This results in a perturbation moment depending on
roll angle.
This kind of parametric oscillations depends on the wave lengths and ship
altitude.
Periodic change of the metacentric height results in asymmetry of roll oscil-
lations. Due to phase displacement between heave oscillations and waves

the averaged additional moment is dierent for positive and negative . In

the resonance case of vertical oscillations

, the middle roll angle is [9]

0
=

2
g
A
z
g
T
GM

T
(9.4)
where
z
g
- the position of the center of gravity with respect to keel,
96
T - ship draught,
=
v
A
wp
T
- are reduction factors (see 3.9)
v - ship volume,
A
wp
- water plane surface.
The larger z
g
T is, the bigger is the middle roll angle
0
.
For beam sea
0
is in the direction of waves.
Figure 9.1: Ship oscillations during parametric resonance
Figure 9.2: Conditions for parametric resonance appearance
97
Vertikale Schwingungen und Rollschwingungen
I
xx
= A
uu
B
uu
g
0

GM
j

u
GM
j
= GM
0
j
A
GM
cos t
. .
Ursache der
parametrischen
Rollschwingungen
I
xx
= +A
uu
+ B
uu
+ g
0
(GM
j
0
+ A
GM
cos t) = 0
(I
xx
+ A
uu
) + B
uu
+ g
0
(GM
j
0
+ A
GM
cos t) = 0
+ 2

2
=

cos t

= B
uu
/2(I
xx
+ A
uu
)
2
= g
0
GM
0
j
/(I
xx
+ A
uu
)
keine L osung in Elementenfunktionen.
Gleichung wurde gr undlich untersucht.
Wichtige Bedingungen f ur parametrische Resonanz:
Parametrische Resonanz
Frequenz der vertikalen Schwingungen > 2 Frequenz der Rollschwingungen
98
Q
0
=
e
Zg T
h

Ax
2

XT

e
= x
0
X =
V
ST
99
100
Chapter 10
Principles of Rankine source
method for calculation of
seakeeping
10.1 Frequency domain simulations
Generally there are two basic methods of simulation of time dependent pro-
cesses. In the rst method called the simulation in frequency domain the
unsteady process is considered as the sum of the mean part which is time
independent and the periodical part. The mean part can be calculated using
nonlinear strategy, i.e. the seeking parameters depends on ship kinematic
parameters in non linear manner. The periodic part is considered as small
and found from linear theory. This formalism is based on Fourier analysis of
unsteady processes. Each unsteady quantity q is represented in form:
q = q
0
+ qe
it
(10.000)
where
q
0
is the mean value,
q is amplitude and
is frequency.
The potential around the ship is represented as the sum of four terms [13]
= (V x +
s
) + (
w
+
I
) (10.1)
where
101
V x potential of incident uniform ow,
V ship speed,

w
potential of incident wave,

I
remaining unsteady potential,

s
potential of the steady ow disturbance.
The terms in the rst parenthesis describes the steady ow around the ship
with account of free surface eects. The second parenthesis represent the
periodic ow due to waves. determination of the potential (V x +
s
) is
discussed in the wave resistance potential theory. This problem can be solved
utilizing either full nonlinear or linear formalism. Numerical linear method
for two dimensional case is described in Chapter 6.7 (Kornev N., Ship Theory
I, Manuscript, 2008). The boundary conditions for
I
are linearized. The
following boundary conditions should be satised [13]:
no penetration on the hull,
kinematic boundary condition on the free surface,
dynamic boundary condition on the free surface,
decay of disturbances far away from the ship,
radiation condition.
The Laplace equation and decay condition are automatically satised within
the Rankine source method. The unsteady potential is decomposed into
radiation and diraction components:

I
=
6

i=1

i
u
i
+
d
(10.2)
where u
i
- ship velocities.
For the sake of simplicity we consider fully linear formalism (
s
and
I
are
small). The kinematic and dynamic boundary conditions are used to derive
mixed boundary condition on the free surface.
The dynamic boundary condition written in the inertial reference system ist
(see page 110 in [6]).

t
0
+
1
2
_
_

x
_
2
+
_

y
_
2
+
_

z
_
2
_
+ gz = 0 (10.3)
Its linear version is
102

t
0
+ gz = 0 (10.4)
The relation between the derivative on time in inertia /t
0
and ship xed
coordinate system /t is

t
0
=

t

V (10.5)
where

V is the speed of the reference system substituting (10.5) into (10.4)
we obtain:

V + gz = 0, z = (10.6)
The linearized version of the kinematic boundary conditions (see page 111
in [6]) reads:

t
0
=

z
(10.7)
Dierentiating (10.5) on time

t
2
0
+ g

t
0
= 0 (10.8)
and substituting (10.7) into (10.8) one obtains

t
2
0
+ g

z
= 0 (10.9)
using relation

t
0
=

t

V
we get

t
2
0
=

t
0
_

t
0
_
=

t
0
_

V
_
=
=

t
_

V
_
v

V
_
=
=

2

t
2

t
+

V
2

2
=
=

2

t
2
2

t
103
The mixed boundary condition in ship xed system reads

t
2
2

t
+ g

z
= 0 (10.10)
Substituting =
0
+
1
, where
0
= V x+
s
and
1
=

+
I
into 10.10
one obtains

1
t
2
2

1
t
+ g

1
z
+ g

0
z
= 0 (10.11)
Since
1
=
1
e
it
10.11 takes the form:

2

1
2i

V
1
+ g

1
z
+ g

0
z
e
it
= 0
within the fully linear formalism developed here

0
z
= 0 at the free surface.
Therefore the problem is reduced to the problem with respect to ampli-
tudes
I
since

is given

= Re(ic

Ae
ik(xcos y sin )
e
it
)

2

I
2i

V
I
+ g

I
z
= +
2

+
+ 2i

z
;
(10.12)
The no penetration condition yields on the ship hull
n
(1)
+

u ( mi
e
n) +

[x ( mi
e
n) +
+n
(0)

= 0
(10.13)
where m = (n)
(0)

= Re
_
ic

Ae
kz
e
ik(xcos y sin )
e
i
e
t
_
u = (u
1
, u
2
, u
3
)
T
describes the translations,
= (u
4
, u
5
, u
6
)
T
= (
1
,
2
,
3
)
T
the rotations.
The panels are distributed on the hull and on the free surface around the
ship.
To get the boundary conditions for diraction potential, the potentials
i
are
set to zero in equations (10.12) and (10.13). If the radiation potential, say

(3)
, is to be found, the diraction potentials
d
and
i(i=3)
are set to zero
in boundary conditions (10.12) and (10.6). The boundary conditions (10.12)
104
and (10.13) are fullled at centres of each panel.
This results is a system of linear algebraic equations with respect to the
source strengths, which can be solved by Gauss method.
Once the source strengths are known, all potentials and derivatives (veloci-
ties) can be calculated.
The unknown motion amplitudes u
i
are computed from momentum equa-
tions:
m
_

u +

x
g
_
= a

G +
_
S
_
p
(1)

Ua
g
+a (x a
g
)
__
ndS
(10.14)
m
_
x
g

u
_
+ I

= x
g

_


G
_
+
+
_
_
p
(1)
[ua
g
+ (x
g
)]
_
( x n) dS
(10.15)
x
y
center of gravity
G = mg ship weight
I matrix of the moments of inertia of the ship
M =
_
_
_
_
_
_
_
_
m 0 0 0 mz
g
0
0 m 0 mz
g
0 mx
g
0 0 m 0 mx
g
0
0 m
z
g 0 Q
xx
0 Q
xz
mz
g
0 mx
y
0 Q
yy
0
0 mx
g
0 Q
xz
0 Q
zz
_
_
_
_
_
_
_
_
Q
xx
=
_
(y
2
+ z
2
)dm, Q
xz
=
_
xzdm
I =
_
_
Q
xx
0 Qxz
0 Q
yy
0
Q
xz
0 Q
zz
_
_
The harmonic pressure p
(1)
is decomposed into parts due to incident wave,
due to diraction, and due to radiation.
p
(1)
= p

+ p
d
+
6

i=1
p
i
u
i
105
These components can be calculated from the linearized Bernoulli equation:
p
i
=
_

i
t
+
(0)

i
_
Two momentum vector equations (10.14) and (10.15) form a linear system
of six equations for the six motions u
i
which is easily solved.
10.2 Time domain simulation
Consideration of nonlinear eects. Seaway is computed as superposition of
elementary waves.
The wave frequencies
j
are chosen such that the area under the sea
spectrum between
j
and
j+1
is the same for all j. This results in
constant amplitudes for all elementary waves regardless of frequency.
The frequency interval for simulation is divided into subintervals. These
subintervalls are larger where S

or the important RAOs are small and

vice versa. In each subinterval a frequency
j
is chosen randomly (based
on constant probability distribution). Encounter angles are chosen ran-
domly.
The frequencies, encounter angles, and phase angles chosen before the simu-
lation must be kept during the whole simulations.
106
Bibliography
[1] Brix J. Manoeuvring technical manual. Seehafen Verlag, 1993.
[2] Kleinau D. Theorie des Schies, Manuskript. University of Rostock,
2001. (in German).
[3] Makov J. Ship oscillations. Kaliningrad, 2007.
[4] http://www.marin.nl/web/Facilities-Tools/Basins/Seakeeping-
Manoeuvring-Basin/Seakeeping-and-manoeuvring-basin-pdf.htm.
[5] Winkler S. Flume passive anti-roll tanks. application on merchant and
naval ships. present at the University of Rostock.
[6] Kornev N. Schistheorie I, page 162 S.
[7] Kornev N. Ship Theory I (ship manoeuvrability). 2011. Manuscript.
[8] Newman J. Marine hydrodynamics. MIT Press, 1984.
[9] Sisov V. Ship Theory. Odessa, 2004.
[10] Price W.G. and Bishop R.E. Probabilistic theory of ship dynamics.
Halsted, London, 1974.
[11] Vugts J.H. The hydrodynamic coecients for swaying, heaving and
rolling cylinders in a free surface. Int. Shipbuilding Progress, 15:pp. 224
276, 1968.
[12] Frahm H. Results of trials of the anti-rolling tanks at sea. Trans. of the
Institution of Naval Architects, Vol.53, 1911.
[13] Bertram V. Practical ship hydromechanics, page 270 p. Butterworth-
Heinemann, 2000.
107