ShipX Vessel Responses (VERES)

Theory Manual

Dariusz Fathi and Jan Roger Hoff

MARINTEK A/S
February 13, 2004
2

CONTENTS

CONTENTS 2

1 HYDRODYNAMIC COEFFICIENTS AND EXCITING FORCES 3

1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Local analysis at each strip . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Low Speed Strip Theory Formulation . . . . . . . . . . . . . . . . . . . . . 7
1.4 High Speed Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2 RESTORING COEFFICIENTS 17

3 VISCOUS ROLL DAMPING COEFFICIENTS 18

3.1 Frictional roll damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Eddy damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Bilge keel damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

BIBLIOGRAPHY 21

INDEX 22

/P
 3

1 HYDRODYNAMIC COEFFICIENTS AND EXCITING FORCES

In this chapter an outline of the main aspects of the strip theory 1 of Salvesen, Tuck & Faltin-
sen [9], and the high speed theory of Faltinsen & Zhao [1] will be given. Further details may
be found in the references.

1.1 General

Consider a ship advancing at constant mean forward speed with arbitrary heading in regular
sinusoidal waves. It is assumed that the resulting oscillatory motions of the ship are linear
and harmonic. A right-handed coordinate system (x, y, z) fixed with respect to the mean
oscillatory position of the ship is used. The z-axis is positive vertically upwards through the
center of gravity of the ship. The origin is located in the plane of the undisturbed free surface.
The ship is assumed to have the x-z plane as a plane of symmetry in its mean oscillatory
position. The x-axis is pointing towards the stern. The translatory displacements in the x, y
and z directions are η1 , η2 and η3 , respectively. η1 is the surge, η2 is the sway and η3 is the
heave displacement. Furthermore, the angular displacements of the rotational motions about
the x, y and z axes are η4 , η5 and η6 , respectively. η4 is the roll, η5 is the pitch and η6 is the
yaw angle.

If viscous effects are neglected, the fluid motion can be assumed to be irrotational, so that
the problem can be formulated in terms of potential flow theory. The total velocity potential
Φ(x, y, z) must then satisfy the Laplace equation and the following exact boundary condi-
tions:
DF
= 0, (1.1)
Dt
on the hull surface where the hull is defined by F (x
, y
, z
) = 0 with (x
, y
, z
) a coordinate
system fixed in the ship, and


Dp D ∂Φ 1 2
= −ρ + |∇Φ| + gz = 0, (1.2)
Dt Dt ∂t 2

on the unknown free surface given by z = Z(x, y; t). In addition, suitable radiation conditions
at infinity must be satisfied. Here g is the acceleration of gravity and ρ is the mass density of
the fluid.

Separating the velocity potential Φ(x, y, z; t) into two parts, one the time-independent steady
contribution due to the forward motion of the ship and the other the time-dependent part
1
Please notice the difference in the definition of the coordinate systems. The theory in this manual has been
re-formulated to fit the coordinate definitions and unit normal definitions used in VERES.

/P
4

associated with the incident wave system and the unsteady body motions, we get

Φ(x, y, z; t) = [Ux + φS (x, y, z)] + φT (x, y, z)eiωt . (1.3)

Here Ux + φS is the steady contribution with U the forward speed of the ship, φ T is the com-
plex amplitude of the unsteady potential and ω is the frequency of encounter in the moving
reference frame. It is understood that real part is to be taken in expressions involving e iωt .

In order to linearize the boundary conditions (1.1) and (1.2), it will be assumed that the
geometry of the hull is such that the steady perturbation potential φ S and its derivatives are
small, and further that by considering only small oscillatory motions, the potential φ T and its
derivatives can also be assumed to be small. Under these assumptions the problem can be
linearized by disregarding higher-order terms in both φ S and φT as well as terms involving
cross products between φS and φT .

Furthermore, in linearizing the problem it will be convenient to linearly decompose the am-
plitude of the time-dependent part of the potential

6
φT = φI + φD + φj ηj , (1.4)
j=1

where φI is the incident wave potential, φ D is the diffraction potential, and φ j is the contri-
bution to the velocity potential from the jth mode of motion.

The incident wave potential is written as

gζa kz −ik(x cos β+y sin β)
φI = e e , (1.5)
ω0
√
where ζa is the wave amplitude, k is the wave number, β is the heading angle and ω 0 = gk
is the wave frequency which is related to the frequency of encounter by

ω = ω0 + kU cos β. (1.6)

Including only linear terms and applying Taylor expansions about the mean hull position
in the hull boundary condition (1.1) and about the undisturbed free surface, z = 0 in the
free surface condition (1.2), it can be shown that the individual potentials must satisfy the
following linear boundary conditions:

1. The steady perturbation potential, φ S , must satisfy the body condition

∂
[Ux + φS ] = 0, (1.7)
∂n
on the mean position of the hull and the free surface condition
2
2∂φS ∂φS
U 2
+g = 0, (1.8)
∂x ∂z
on z = 0.

/P
 5

2. The incident wave potential, φI , and the diffraction potential, φD , must satisfy

∂φI ∂φD
+ = 0, (1.9)
∂n ∂n
on the mean position of the hull and the free surface condition

2 
∂ ∂
iω + U +g (φI , φD ) = 0, (1.10)
∂x ∂z

on z = 0.

3. The oscillatory potential components φ j , j = 1, 2, . . . , 6), must satisfy

∂φj
= iωnj − Umj , (1.11)
∂n
on the mean position of the hull and the free surface condition

2
∂ ∂
iω + U φj + g φj = 0, (1.12)
∂x ∂z

on z = 0. Here the generalized normal nj , is defined by

(n1 , n2 , n3 ) = n, (1.13)

(n4 , n5 , n6 ) = r × n, (1.14)

with r = (x, y, z) is the position vector with respect to the origin of the coordinate
system and n is the outward unit normal vector pointing into the fluid. Further

1
(m1 , m2 , m3 ) = m
 = (n · ∇)∇(x + φS ) (1.15)
U
1
(m4 , m5 , m6 ) = r × m
 − ∇(x + φS ) (1.16)
U
mj = 0 j = 1, 2, 3, 4 (1.17)
m5 = n3 (1.18)
m6 = −n2 . (1.19)

In addition to these linear boundary conditions the potentials φ S , φI , φD and φj must each
satisfy the Laplace equation in the fluid domain and the appropriate radiation conditions at
infinity.

The pressure in the fluid is obtained from the Bernoulli equation


∂Φ 1 2
p = −ρ + |∇Φ| + gz . (1.20)
∂t 2

/P
6

If the pressure is expanded in a Taylor series about the undisturbed position of the hull and
the pressure expression is then linearized by including only term of first order in φ S and φT ,
it follows that the linearized time-dependent pressure on the hull is


∂
p = −ρ iω + U φT eiωt . (1.21)
∂x

Integration of the pressure over the hull surface yields the hydrodynamic force and moment
amplitudes  

∂
Hj = ρ nj iω + U φT ds, j = 1, 2, . . . , 6. (1.22)
S ∂x
Here the integration is over the mean position of the hull surface S. H 1 , H2 and H3 are the
force components in the x, y and z directions, while H 4 , H5 and H6 are the moments about
the x, y and z axes. The force and moments can be divided into two parts as

Hj = Fj + Gj , (1.23)

where Fj is the exciting force and moment:

 

∂
Fj = ρ nj iω + U (φI + φD ) ds, (1.24)
S ∂x

and Gj is the force and moment due to the body motions η 1 , . . . , η6 :

 

6
∂
Gj = ρ nj iω + U φk ηk ds (1.25)
S ∂x k=1

6
= Tjk ηk . (1.26)
k=1

Here Tjk denotes the hydrodynamic force and moment in the jth direction per unit oscillatory
displacement in the kth mode:
 

∂
Tjk = ρ nj iω + U φk ds. (1.27)
S ∂x

Tjk may be separated into real and imaginary parts as

Tjk = ω 2 Ajk − iωBjk , (1.28)

where Ajk and Bjk are the added mass and damping coefficients, respectively.

1.2 Local analysis at each strip

The diffraction and radiation problems are solved by matching of a near-field solution and
a far-field solution. The near-field solution is obtained by a numerical boundary-element
formulation, while the far-field solution is obtained by an asymptotic analysis.

/P
 7

Sf z

y
So

(III) (I) (II)

2
O - field point φ= 0
S-
S+

Sb

Figure 1.1: The fluid domain and control surfaces

The local analysis in the cross-sectional plane is carried out using Green’s second identity
to represent the velocity potential in terms of a distribution of fundamental two-dimensional
sources and dipoles over a closed surface containing the body surface, the free surface and
a control surface far away from the body. The fluid domain and the control surfaces are
sketched in Figure 1.1.
Straight line segments are used to approximate the control surfaces, and constant values of
the velocity potential and its normal derivatives are assumed at each segment. For each
cross section, the velocity potential at a point (y,z) can be represented using Green’s second
identity: 

∂ log(r) ∂φ
−2πφ = φ − log(r) ds (1.29)
s ∂n ∂n
To obtain our equation system we let the field point approach the center of each element.
At the hull surface, So , the velocity potential φ is the unknown. At the free surface, S f , the
normal velocity ∂φ
∂n
is the unknown, since φ is known through the free surface condition.

This method is applied both in the low speed and high speed theory. The differences are in
the boundary conditions on the free surface, and the integration over the body. The different
theories are described in the following sections.

1.3 Low Speed Strip Theory Formulation

The following section describes the part of the theory which is particular for the low speed
strip theory formulation by Salvesen, Tuck & Faltinsen [9]. Due to different definitions of
coordinate systems, the evaluation of the coefficients is included here, with the definitions
which are applied in VERES.

In the low speed strip theory, the hull condition (1.11) is further simplified by dividing the

/P
8

oscillatory potential into one speed dependent, and one speed independent part
U U
φj = φ0j − φ , (1.30)
iω j
where φ0j is speed independent. This results in the two hull conditions

∂φ0j
= iωnj , (1.31)
∂n
∂φUj
= iωmj . (1.32)
∂n
Now, since both φ0j and φUj must satisfy the Laplace equation, the same free-surface condition
and the same radiation conditions, it follows from the hull conditions above that

φUj = 0, j = 1, 2, 3, 4 (1.33)
φU5 = φ03 , (1.34)
φU6 = −φ02 . (1.35)

Thus, we see that the oscillatory potential components can be expressed in terms of the speed-
independent part of the potential, φ 0j , as

φj = φ0j , j = 1, 2, 3, 4, (1.36)
U
φ5 = φ05 − φ03 , (1.37)
iω
U
φ6 = φ06 + φ02 , (1.38)
iω
where φ0j , (j = 1, 2, . . . , 6) must satisfy the conditions

∂φ0j
= iωnj , (1.39)
∂n
on the mean position of the hull and

2
∂ ∂φ0j
iω + U φ0j + g = 0, (1.40)
∂x ∂z
on z = 0.

Stoke’s theorem may be expressed in the following form [Milne-Thomson [7], 2.51]
  
(n × ∇) × q ds = − dl  × q, (1.41)
S C

where S is a surface situated in the fluid with the closed curve C as boundary. Here q is any
 is the direction element of arc C.
vector function and dl

Letting q = Uφi for the case j = 1, 2, 3 gives

  
  × Uφi.
(n × ∇) × (Uφi) ds = − dl (1.42)
S C

/P
 9

By use of various vector theorems, [Ogilvie and Tuck [8]], the integrand of the surface inte-
gral can be rewritten

(n × ∇) × Uφi = φn × (∇ × Ui) + φ(n · ∇)Ui − n(Ui · ∇φ)

+(n · Ui)∇φ − φn(∇ · Ui)
= φ(n · ∇)Ui − n(Ui · ∇φ). (1.43)

Thus the surface integral is

       
φ(n · ∇)Ui − n(Ui · ∇φ) ds = − φm + n(Ui · ∇φ) ds. (1.44)
S S

The surface integral may now be written as

       
φ(n · ∇)U i − n(U i · ∇φ) ds =
   − n(Ui · ∇φ) ds.
Uφm (1.45)
S S

Neglecting the contribution from the line integral along the free surface one may write

 = dl(i × n),
dl (1.46)

and then use some standard theorems to show that

 
(i × n) × φUi = − n(i · φUi)
= −nφU. (1.47)

This gives the following theorem

    
∂
nj U φ ds = U mj φ ds + U nj φ dl. (1.48)
S ∂x S C

It may be shown [Ogilvie and Tuck (1969)], that this form also applies to the case j = 4, 5, 6.

Applying the above theorem in the relationships for the added mass and damping coefficients
will give     
Tjk = ρ nj iωφk ds + Uρ mj φk ds + Uρ nj φk dl. (1.49)
S S CA

Here CA refers to the aftermost cross section of the ship.

The speed-independent part of Tjk may now be defined as

 
0
Tjk = ρiω nj φ0k ds, (1.50)
S

and the speed-independent part of the line integral at any cross section C x as

tjk = ρiω nj φ0k dl. (1.51)
Cx

/P
10

The added mass and damping coefficients can now be expressed in terms of the speed-
independent terms as follows

0 U A
Tjk = Tjk + t , j, k = 1, 2, 3, 4, (1.52)
iω jk
0 U 0 U
T5k = T5k + T3k + tA , k = 1, 2, 3, 4, (1.53)
iω iω 5k
0 U 0 U
T6k = T6k − T2k + tA , k = 1, 2, 3, 4, (1.54)
iω iω 6k
0 U 0 U U2 A
Tj5 = Tj5 − Tj3 + tA + t , j = 1, 2, 3, 4, (1.55)
iω iω j5 ω 2 j3
0 U 0 U U2 A
Tj6 = Tj6 + Tj2 + tA j6 − t , j = 1, 2, 3, 4, (1.56)
iω iω ω 2 j2
0 U2 0 U U2 A
T55 = T55 + 2 T33 + tA + t , (1.57)
ω iω 55 ω 2 53
0 U2 0 U A U2 A
T66 = T66 + 2 T22 + t66 − 2 t62 , (1.58)
ω iω ω

where tAjk refers to the line integral evaluated at the aftermost section. In obtaining the ex-
pressions for T55 and T66 the symmetry relationship for the zero speed coefficients has been
applied
0 0
Tjk = Tkj . (1.59)

Above, the speed-dependent coefficients have been expressed in terms of the speed-
independent surface integral and line integral. The zero-speed terms will now be reduced
to a form suitable for numerical evaluation. This is done by introducing the following strip
theory approximations. If it is assumed that the beam and draft of the ship are much smaller
than the length, then it is consistent with the previous assumptions to set ds = dldξ in the
surface integral   
0
Tjk = ρiω nj φ0k dl dξ = tjk dξ, (1.60)
L Cx L

where L implies that the integration is over the length of the ship and ξ is the variable of
integration in the x-direction. Since the hull is assumed to be long and slender it follows that
in the neighborhood of the hull ∂/∂x  ∂/∂y or ∂/∂z. It also follows that the component of
the hull normal in the x-direction is much smaller than the normal components in the y- and
z-directions
n1  n2 or n3 , (1.61)

so that the components of the three-dimensional generalized normal n j , j = 2, 3, 4, with the

two-dimensional generalized normal in the yz-plane, N j , j = 2, 3, 4 and then set

n5 = −xN3 , (1.62)
n6 = xN2 . (1.63)

/P
 11

In order to reduce the free surface condition it will be necessary to assume that the frequency
of encounter is high, ω  U(∂/∂x), which requires that the wavelength is approximately of
the same order as the ship beam.

Under these assumptions the three-dimensional Laplace equation and the boundary condi-
tions to be satisfied by φ0k for k = 2, 3, 4 reduce to the two-dimensional Laplace equation and
the conditions for the two-dimensional problem of a cylinder with cross section C x oscillating
in the free surface, so that at a given cross section
φ0k = ψk for k = 2, 3, 4, (1.64)
where ψk is the potential for the sectional two-dimensional problem. It also follows that at a
given section
φ05 = −xψ3 , (1.65)
φ06 = xψ2 , (1.66)
while φ01  φ0k , k = 2, 3, . . . , 6. Hence, for j = 2, 3, 4

tjj = ρiω Nj ψj dl = ω 2 ajj − iωbjj , (1.67)
Cx
where ajj and bjj are the sectional two-dimensional added mass and damping coefficients for
sway, heave and roll. Similarly, the sectional sway-roll cross-coupling coefficients is

t24 = ρiω N2 ψ4 dl = ω 2 a24 − iωb24 . (1.68)
Cx

0
The zero-speed added mass and damping coefficients, Tjk can now be expressed in terms of
the sectional two-dimensional added mass and damping coefficients, t 22 , t33 , t44 and t24 . For
ships with lateral symmetry, the only nonzero coefficients are

0
T22 = t22 dξ, (1.69)

0 0
T26 = T62 = ξt22 dξ, (1.70)

0
T66 = ξ 2 t22 dξ, (1.71)

0
T33 = t33 dξ, (1.72)

0 0
T35 = T53 = − ξt33 dξ, (1.73)

0
T55 = ξ 2 t33 dξ, (1.74)

0
T44 = t44 dξ, (1.75)

0 0
T24 = T42 = t24 dξ, (1.76)

0 0
T46 = T64 = ξt24 dξ, (1.77)

/P
12

where the integration is over the length of the ship.

This implies that the following expressions are obtained for the added mass and damping
coefficients


U A
A22 = a22 dξ − b , (1.78)
L ω 2 22

B22 = b22 dξ + UaA 22 , (1.79)
L 
U
A24 = A42 = a24 dξ − 2 bA , (1.80)
ω 24
L
B24 = B42 = b24 dξ + UaA 24 , (1.81)
 L
U
A33 = a33 dξ − 2 bA , (1.82)
L ω 33

B33 = b33 dξ + UaA 33 , (1.83)
L
U
A44 = a44 dξ − 2 bA , (1.84)
ω 44
L
B44 = b44 dξ + UaA 44 , (1.85)
L 
U U
A53 = − xa33 dξ − 2 b33 dξ + 2 xA bA
33 , (1.86)
L ω L ω

/P
 13
 
B53 = − xb33 dξ + U a33 dξ − UxA aA 33 , (1.87)
 L  L
U U
A62 = xa22 dξ + 2 b22 dξ − 2 xA bA 22 , (1.88)
ω L ω
L 
B62 = xb22 dξ − U a22 dξ + UxA aA 22 , (1.89)
L L
 
U U
A64 = xa24 dξ + 2 b24 dξ − 2 xA bA 24 , (1.90)
L ω L ω
 
B64 = xa24 dξ − U a24 dξ + UxA aA 24 , (1.91)
L L
 
U U U2 A
A35 = − xa33 dξ + 2 b33 dξ + 2 xA bA 33 + a , (1.92)
L ω L ω ω 2 33
 
U2 A
B35 = − xb33 dξ − U a33 dξ − UxA aA 33 + b , (1.93)
L L ω 2 33
 
U U U2 A
A26 = xa22 dξ − 2 b22 dξ − 2 xA bA 22 − a , (1.94)
L ω L ω ω 2 22
 
U2 A
B26 = xb22 d + U a22 d + UxA aA 22 − b , (1.95)
L L ω 2 22
 
U U U2 A
A46 = xa24 dξ − 2 b24 dξ − 2 xA bA 24 − a , (1.96)
L ω L ω ω 2 24
 
U2 A
B46 = xb24 dξ + U a24 dξ + UxA aA 24 − b , (1.97)
L L ω 2 24
 
2 U2 U U2
A55 = x a33 dξ + 2 a33 dξ − 2 x2A bA 33 − 2
xA aA
33 , (1.98)
L ω L ω ω
 
2 U2 U2
B55 = x b33 dξ + 2 b33 dξ + UxA a33 − 2 xA bA
2 A
33 , (1.99)
L ω L ω
 
2 U2 U U2
A66 = x a22 dξ + 2 a22 dξ − 2 x2A bA 22 − 2
xA aA
22 , (1.100)
L ω L ω ω
 
2 U2 U2
B66 = x b22 dξ + 2 b22 dξ + Ux2A aA 22 − xA bA
22 . (1.101)
L ω L ω2

The exiting force and moment is expressed as

 

∂
Fj = ρ nj iω + U (φI + φD )ds, j = 1, 2, . . . , 6. (1.102)
S ∂x
It is convenient to separate the exciting force into two parts, the incident wave part F jI , and
the diffraction part FjD , so that
Fj = FjI + FjD , (1.103)
with  

∂
FjI =ρ nj iω + U φI ds, (1.104)
S ∂x
and  

∂
FjD =ρ nj iω + U φD ds. (1.105)
S ∂x

/P
14

The incident wave part of the exciting force, denoted the Froude-Krylov force, may now be
written as  
I
Fj = ρi nj (ω − kU cos β) φI ds, (1.106)
S

which may be reduced to  

FjI = ρiω0 nj φI ds. (1.107)
S

Applying the Stokes theorem to the diffraction part of the exciting force gives
  
D
Fj = ρ (iωnj + Umj ) φD ds + ρU nj φD dl. (1.108)
S CA

The sectional Froude-Krylov force is defined by


−ikx cos β
fj (x) = ige Nj ekz e−iky sin β dl, j = 2, 3, 4. (1.109)
Cx

The sectional diffraction force is obtained by writing the diffraction potential as

φD = ψD ζa e−ikx cos β , (1.110)


−ikx cos β
hj (x) = iωe Nj ψD dl, j = 2, 3, 4. (1.111)
Cx

This enables the exciting force and moment to be written in the final form

U
Fj = ρζa (fj + hj )dξ + ρζa hA , j = 2, 3, 4, (1.112)
L iω j


U U
F5 = −ρζa x(f3 + h3 ) − h3 dξ − ρζa xA hA 3, (1.113)
L iω iω


U U
F6 = ρζa x(f2 + h2 ) − h2 dξ + ρζa xA hA 2. (1.114)
L iω iω

F1  Fk , k = 2, 3, . . . , 6, and is neglected. hA
j refers to hj (x) evaluated at the aftermost
section.

1.4 High Speed Formulation

The high-speed formulation [1] is based on a strip theory approach, where the free-surface
condition is used to step the solution in the downstream direction. The solution is started
assuming that both the velocity potential and its x-derivative are zero at the first strip, counted
from the bow.

In the solution procedure, the radiation and diffraction potentials are re-written as

φ = e−i(ω/U )x Ψ(y, z) (1.115)

/P
 15

where Ψ(y, z) is independent of x and the time dependency is in the first term.

The following boundary conditions have to be satisfied at each cross-section for Ψ, where the
conditions holds for both radiation and diffraction:

• Laplace’s equation in the fluid:

∂2Ψ ∂2Ψ
+ = 0. (1.116)
∂y 2 ∂z 2
• Free-surface condition:
2
2∂Ψ ∂Ψ
U 2
+g = 0 on z = 0. (1.117)
∂x ∂z
This free-surface condition differs from the ordinary strip theory formulation in that
the terms proportional to U 2 are retained, while they are neglected in the ordinary strip
theory.

• Body boundary condition:

∂Ψj
= iωnj − Umj , j = 2, . . . , 6 (1.118)
∂n
∂ΨD
= (inx − nz )ω0 ζa ekz+i(ω0 /U )x−iky sin β (1.119)
∂n
• Furthermore, the potentials and their x-derivatives are set to zero at the foremost part
of the vessel, i.e:

Ψ = 0 (1.120)
on x = xB
∂Ψ
= 0
∂x
where xB denotes the first strip. This solution procedure requires that there are no
upstream waves. This condition is satisfied when the waveform parameter τ = ωU/g >
1/4.

When the radiation potentials Ψj and the diffraction potential ΨD has been solved for each
strip, the total solution can be found by applying Eqn. 1.115 and integrating over the hull
surface:

• Hydrodynamic coefficients
The added mass and damping coefficients are calculated by the equations
Re(Tjk ) −Im(Tjk )
Ajk = , Bjk = (1.121)
ω2 ω
where
 

∂
Tjk = ρ nj iω + U φk ds (1.122)
S ∂x

/P
16

• Exciting forces
The wave excitation forces are calculated by integrating the Froude-Krylov and diffrac-
tion forces over the hull surface, i.e:
 

∂φD
Fj = ρ nj iω0 φI + iωφD + U ds. (1.123)
S ∂x

/P
 17

2 RESTORING COEFFICIENTS

The hydrostatic restoring coefficients which are independent of frequency and forward speed,
follow directly from hydrostatic considerations.
For a vessel in the free surface symmetric about the x−z plane, the only coefficients different
from zero are

C33 = ρg bdx = ρgAwp , (2.1)
L 
C35 = C53 = −ρg bxdx, (2.2)
L 
C44 = ρg∇(zB − zG ) + ρg y 2 bdx = ρg∇GM T , (2.3)
 L

C55 = ρg∇(zB − zG ) + ρg x2 bdx = ρg∇GM L , (2.4)

L

Here b is the sectional breadth, Awp is the waterplane area, ∇ is the displaced volume of water
and zB and zG are the z-coordinates of the center of buoyancy and center of gravity, respec-
tively. GM L and GM T is the longitudinal and transverse metacentric heights, respectively.

/P
18

3 VISCOUS ROLL DAMPING COEFFICIENTS

3.1 Frictional roll damping

For turbulent flow, Kato [6] applies Hughes formula for the frictional coefficient, giving the
following contribution to the roll damping coefficients:
V1 8
B44 = · 0.9275 · ρ · S · rs2 · ω 0.5 · ν 0.5 (3.1)
3π
V2
B44 = 0.00755 · ρ · S · rs2.772 · ω −0.114 · ν 0.114 · η4a
−0.228
. (3.2)
V2
Note that the nonlinear damping coefficient B44 is dependent on the roll amplitude. This
requires that the wave amplitude must be given as input to VERES .

Correction for forward speed effects

The coefficients are corrected for forward speed effects, as explained in Himeno [2], giving
the corrected damping coefficient as


U
B = B0 1 + 4.1 · , (3.3)
ωL
where the coefficient B0 represents the friction damping coefficient at zero forward speed.

3.2 Eddy damping

This damping component is caused by flow separation at the bilge of the cross section. Based
on results from forced roll tests for a number of two dimensional cylinders without bilge
keels, Ikeda et.al. [3] has proposed a prediction method giving the following contribution to
the roll damping coefficients:
V1
B44 = 0.0 (3.4)

V2 2
B44 = 0.5 · ρ · rmax cp (s)(s)ds (3.5)
S

where rmax is the maximum distance from the roll axis to the hull surface, c p (s) is the pressure
coefficient and (s) is the roll moment lever. The integration is taken over the wetted surface.

Correction for forward speed effects

The coefficients are corrected for forward speed effects as explained in Himeno [2] giving the
corrected eddy damping coefficient as
0.04 · ω 2 · L2
B = B0 · 2 (3.6)
U + 0.04 · ω 2 · L2

/P
 19

3.3 Bilge keel damping

Damping due to normal forces on bilge keels

The contribution to the roll damping coefficients is based on Ikeda et.al. [4] giving:
V1 8
B44 = · 22.5 · b2bk · rbk
2
·f ·ω (3.7)
3π 2
V2
B44 = 2.4 · bbk · f 2 · rbk
3
, (3.8)

where bbk is the breadth of the bilge keel, rbk is the distance from the roll axis to the bilge
keel and f is a correction factor for velocity increment at the bilge given as

f = 1.0 + 0.3 · exp {−160 · (1.0 − σ)} . (3.9)

Here σ is the area coefficient of the cross-section.

Damping due to hull pressure created by bilge keels

The contribution to the roll damping coefficients is based on Ikeda et.al. [5] giving:

• Damping due to pressure increase on the hull:

+
B1H = 0.0 (3.10)
+
B2H = 0.5 · f 2 · rbk
2
· 1.2 · (I2 + I4 ), (3.11)

where  s3
I2 = (s)ds (3.12)
s2
and  s6
I4 = (s)ds. (3.13)
s5

• Damping due to pressure decrease on the hull:

− 4
B1H = · f · rbk · bbk · ω · 22.5(I1 + I3 ) (3.14)
3π 2
−
B2H = 0.5 · f 2 · rbk
2
· 1.2 · (I1 + I3 ), (3.15)

where  s2
I1 = (s)ds (3.16)
s1
and  s5
I3 = (s)ds. (3.17)
s4

/P
20

• Total contribution:

The total contributions from the hull pressure damping due to bilge keels can then be
expressed as:

V1 + −
B44 = B1H + B1H (3.18)
V2 + −
B44 = B2H + B2H (3.19)

/P
 21

BIBLIOGRAPHY

[1] FALTINSEN , O.M. AND Z HAO , R. Numerical predictions of ship motions at high for-
ward speed. In Phil. Trans. R. Soc. Lond. A, vol. 334, pp. 241–252, 1991.

[2] H IMENO , Y. Prediction of ship roll damping – state of the art. Technical Report 239,
Dept. of Naval Architecture and Marine Engineering, University of Michigan, 1981.

[3] I KEDA , Y. ET. AL . On eddy making component of roll damping force on naked hull.
Technical Report 00403, Dep. of Naval Arch., University of Osaka Prefecture, 1978.

[4] I KEDA , Y. ET. AL . On roll damping force of ship - effect of friction of hull and normal
force on bilge keels. Technical Report 00401, Dep. of Naval Arch., University of Osaka
Prefecture, 1978.

[5] I KEDA , Y. ET. AL . On roll damping force of ship - effect of hull surface pressure cre-
ated by bilge keels. Technical Report 00402, Dep. of Naval Arch., University of Osaka
Prefecture, 1978.

[6] K ATO , H. On the frictional resistance to the rolling of ships. Journal of Zosen Kiokai,
102:115, 1958.

[7] M ILNE -T HOMSON , L.M. Theoretical Hydrodynamics. MacMilan Co., New York, fifth
edition, 1968.

[8] O GILVIE , T.F. AND T UCK , E.O. A Rational Strip-Theory of Ship Motion: Part I. Tech-
nical Report 013, Department of Naval Architecture, The University of Michigan, 1969.

[9] S ALVESEN , N., T UCK , E.O. AND FALTINSEN , O. Ship motions and sea loads. In
Transactions of the Society of Naval Architects and Marine Engineers, vol. 78, pp. 250–
287, 1970.

/P
22

INDEX

The italic numbers denote the pages where the corresponding entry is described, numbers
underlined point to the definition, the suffix n refers to footnotes and all others indicate the
places where it is used.

A L
added mass . . . . . . . . . . . . . . . . . . . . 6, 15 Laplace’s equation . . . . . . . . . . 3, 5, 8, 11, 15

B M
Bernoulli’s equation . . . . . . . . . . . . . . . . . 5 metacentric heights . . . . . . . . . . . . . . . . 17
boundary conditions ................ 3
N
C normal vector, generalized . . . . . . . . . . . . 5
coordinate systems . . . . . . . . . . . . . . . . . 3 P
pressure . . . . . . . . . . . . . . . . . . . . . . . . . 5
D
damping . . . . . . . . . . . . . . . . . . . . . . 6, 15 R
diffraction forces . . . . . . . . . . . . . . . . 14, 16 restoring coefficients . . . . . . . . . . . . . . . 17

E S
exciting forces . . . . . . . . . . . . . . 3, 6, 13, 16 Salvesen, Tuck & Faltinsen . . . . . . . . . . . 3, 7
strip theory formulation
F
high speed . . . . . . . . . . . . . . . . . . 14–16
Faltinsen & Zhao . . . . . . . . . . . . . . . . 3, 14
low speed . . . . . . . . . . . . . . . . . . . 7–14
Froude-Krylov forces . . . . . . . . . . . . . 14, 16
V
G velocity potential . . . . . . . . . . . . . . . . . . 3
Green’s second identity . . . . . . . . . . . . . . 7 viscous roll damping
bilge keel damping . . . . . . . . . . . . 19–20
H
coefficients . . . . . . . . . . . . . . . . . . . 18
high speed formulation . . . . . . . . . . . . . . 14
eddy damping . . . . . . . . . . . . . . . . . 18
hydrodynamic coefficients . . . . . . . . . . . . 3
frictional damping . . . . . . . . . . . . . . 18
hydrodynamic forces and moments . . . . . . . 6
hydrostatic coefficients . . . . . . . . . . . . . . W
. . . . . . . . . . . see restoring coefficients wave potential .................... 4

/P