OSE522, KAIST

Hydroelastic analysis of two dimensional floating structure

x3
𝑛�⃗

x1 SB Incident
𝑛�⃗�
h

Green function

○ For finite depth case :

) ln ( x1 − ξ1 ) 2 + ( x3 − ξ3 ) 2 + ln ( x1 − ξ1 ) 2 + (2h + x3 + ξ3 ) 2 − 2ln h
G ( x1, x3 ; ξ1, ξ3 =
∞  (K + τ )e −τ h cosh τ (ξ3 + h)cosh τ ( x3 + h)cosτ x1 − ξ1 e −τ h 
∫
− 2 P.V .
0

 τ (τ sinh τ h − K cosh τ h)
+
τ 
 dτ

(1)

( K + k0 )e − k0 h cosh k0 (ξ3 + h)cosh k0 ( x3 + h)cos k0 x1 − ξ1

+ 2π i
k0 (sinh k0 h + k0 h cosh k0h − Kh sinh k0h)

where h = water depth

xi , ξi = spatial point, source point

P.V . = Cauchy principle value

K = ω 2 / g , k0 sinh k0 h − K cosh k0 h =
0 : dispersion relation

g = acceleration of gravity

ω = angular frequency

○ In series representation ( x3 = 0) :

−ik x1 −ξ1 ∞ − k n x1 −ξ1

iLke Lk ne
=G ( x1; ξ1 ) 2π
 2
− 2π ∑  2
(2)
h  2 2  Lω 2   + Lω n =1 h  2 2  Lω 2   − Lω
L k −  L kn +  
L   g



 g L   g



 g
   

1
 OSE522, KAIST

where k = wave number

L = structure length

K + k n tan k n h =
0, k n =
ikn

π n − θn
kn = : roots of dispersion equation
h

 ω 2h / g 
θ n = tan −1  
 π n − θ n 

○ For infinite depth case :

∞  (K + τ )eτ ( x3 +ξ3 ) cosτ x1 − ξ1 e −τ 

) ln ( x1 − ξ1 ) 2 + ( x3 − ξ3 ) 2 + P.V .
G ( x1, x3 ; ξ1, ξ3=
∫ 0

 τ (K − τ )
−
τ 
 dτ
 (3)
+ 2π ie K ( x3 +ξ3 ) cos K x1 − ξ1

Boundary integral equation


○ For x on S B ( x3 = 0) :
 
    ∂G ( x , ξ )   ∂φ (ξ ) 
2πφ ( x ) −=
2πφI ( x )
∫ φ (ξ )
S B  ∂n
− G( x,ξ )
∂n 
 dS B (4)

ga cosh[k ( x3 + h)] ikx

where φI = i e : potential of incident wave
ω cosh kh

a= amplitude of incident wave

Using the body boundary condition and free surface boundary condition

∂φ
= iωu3 ,
∂n
(5)
2
∂G ∂G ω
= = G,
∂n ∂x3 g

2
 OSE522, KAIST

equation (4) can be rewritten

    ω2   
2πφ ( x ) − 2πφI=
(x)
∫ φ (ξ )
S B  g
− iωu3  G ( x , ξ ) dS B .

(6)

The linearized Bernoulli equation is

∂Φ P
+ 0,
+ gx3 = (7)
∂t ρw

where ρ w = density of fluid.

For the steady state problem, we use

= , U 3 u3eiω=
x3 U 3= t ωt
, Φ φ ei= , P peiωt ,
(8)

and then equation (7) becomes

p p gu iω p
iωφ + 0, φ=
+ gu3 = − − 3 , − φ−
u3 = . (9)
ρw iωρ w iω g ρw g

Using equation (9) in equation (6) gives

 i  ig    iω   
2π 
ωρ
 w ω  ρ
SB w∫
p ( x ) + u3 ( x )  − 2πφI ( x ) = p (ξ )G ( x , ξ )dS B
g
. (10)

 
Since x = ( x1, 0, 0) T and ξ = (ξ1, 0, 0) T . Equation (10) can be rewritten

 i ig  iω
2π 
ωρ
 w ω  S B
∫
p ( x1 ) + u3 ( x1 )  − 2πφI ( x1 ) = p (ξ1 )G ( x1, ξ1 )dS B .
ρ wg
(11)

3
 OSE522, KAIST

Variational formulation

○ Structure part ( δ u3 = virtual displacement) :

     
−
∫ V
ω 2 ρ su3 ( x )δ u3 ( x ) dV +
∫ V
σ ( x )δε ( x ) dV =
∫
p ( x )δ u3 ( x ) dS B
SB
(12)

where ρ s = density of structure


σ ( x ) = stress


ε ( x ) = strain

In the Euler-Bernoulli beam theory,


σ ( x ) 
→ σ xx ( x1 ) = Eε xx ( x1 ) ,
(13)
2
 ∂ u3 ( x1 )
ε ( x ) 
→ ε xx ( x1 ) = − x3 ,
∂x12

where E = modulus of elasticity.

Using equation (13) in equation (12) gives

∂ 2u3 ( x1 ) ∂ 2δ u3 ( x1 )
−
∫ V
ω 2 ρ su3 ( x1 )δ u3 ( x1 ) dV +
∫ V
Ex32
∂x12 ∂x12
dV =
∫
p ( x1 )δ u3 ( x1 ) dS B .
SB
(14)

○ Fluid part ( δ p = virtual pressure) :

i gi
∫ SB
2π
ωρ w
p ( x1 )δ p ( x1 )dS B +
∫ SB
2π
ω
u3 ( x1 )δ p ( x1 )dS B −
∫ SB
2πφI ( x1 )δ p ( x1 )dS B
(15)
iω
=
∫ SB ρw g ∫ SB
p (ξ1 )G ( x1, ξ1 )δ p ( x1 )dS B dS B

4
 OSE522, KAIST

Discretization
 
 Unknown DOF : p = pressure, u = displacement
3

 Structure : 2-node Euler-Bernoulli beam element (2 DOF per node)

 Fluid : 2-node fluid element (1 DOF per node)

w(1) w(2) p(1) p(2)

θ(1) θ(2)

r = −1  r =1 r = −1  r =1

< 2-node Euler-Bernoulli beam element > < 2-node fluid element >

○ Hermitian cubic interpolation function ( Hue ) :

H ue =  hw(1) hθ (1) hw(2) hθ (2)  (16)

1 1 1 1
where hw(1) = (1 − r ) 2 (2 + r ) , hθ (1) = (1 − r ) 2 (1 + r ) , hw(2) = (1 + r ) 2 (2 − r ) , hθ (2) = (1 + r ) 2 (1 − r )
4 8 4 8

The superscript e denotes the element.

○ 2-node interpolation function ( H e ) :

H e =  h(1)

h(2) 

(17)

1 1
where h(1) =−(1 r ), h(2) =+(1 r )
2 2

○ Interpolation scheme :

5
 OSE522, KAIST

T
e e e
=x1 H= x1 , x1  x1(1) x1(2) 
 
(18)
T
e e e
=ξ1 H=ξ 1 , ξ1  ξ1(1) ξ1(2) 
 

Note that ξ1(i ) are identical to x1(i ) . The unknown DOF can be defined

 N
T
∑ H p ,
e
p
p H=
= e p e  p (1) p (2)  ,
=
 
e =1

(19)
 N
T
∑
 e e
u3 H=
= u
u 3 H ue u 3 u 3
,=  w(1) θ (1) w(2) θ (2)  .
 
e =1

We can discretize equation (14) and (15) using equations (16) ~ (19).

○ Structure part :

 T  
1 EI 1  ∂ 2Hu   ∂ 2Hu   1  
ω 2ρ H
 s ∫ −1
HTu H u det Jdr −
B ∫ 
−1  ∂r 2




 ∂r 2

 det Jdr  u 3 + 



 
∫ −1
HTu H det Jdr  p =

0 (20)
 

where H = structure thickness

B = structure width

I = moment of inertia

det J = determinant of Jacobian matrix ( = /2)

○ Fluid part :

 1    1 1  


∫−1
HT H u det Jdr  u 3 + 
  ρ w g −1
∫
HT H det Jdr  p

(21)
 ω 2 1  1    1
∫ ∫ 
H( r )G ( x1, ξ1 ) det Jdr  det Jds  p = H exp ( x1 ) det Jdr
∫
T T
− H ( s )  a
 2πρ g 2 −1  −1   −1
 w 

6
 OSE522, KAIST

Final matrix form

We can obtain the final matrix form using equations (20) and (21) :
 
S M - S K SC  u 3   0 
     =    , (22)
 FC FM - FG   p   F
   I


1
where SM = ω 2 ρ s H
∫ −1
HTu H u det Jdr ,

T
EI 1  ∂ 2Hu   ∂ 2Hu 
SK =
B ∫ 
−1  ∂r 2




 ∂r 2

 det Jdr


,

1
∫ HTu H det Jdr ,
T
SC F=
= C
−1

1 1
FM =
ρw g ∫ −1
HT H det Jdr ,

ω2 1  1 
FG =
2πρ w g 2 ∫ −1
HT ( s ) 

∫ −1
H( r )G ( x1, ξ1 ) det Jdr  det Jds ,


 1
I = a
F
∫ −1
HT exp ( x1 ) det Jdr .

Note that the matrix form in the left hand side of equation (22) is symmetric. Now we can
calculate hydrodynamic pressure and displacement. The total pressure can be divided
   
p = p + p + p
d s r , (23)

where p =
d diffraction pressure,

p = hydrostatic pressure,
s


p = radiation pressure.
r


Since diffraction effect happens when the waves encounter the fixed structure, we set uˆ 3 = 0 in
equation (22) and obtain the diffraction pressure
 
p = ( F - F )-1 F
d M G
I . (24)

7
 OSE522, KAIST

The hydrostatic pressure can be calculated

 
p = − ρ g u
s w 3
, (25)

and the radiation pressure is

   
p =p − p − p
r d s . (26)

Complex response

The nodal solution û 3 consists of complex numbers and the steady state response is
 
 = u e iωt ,
U (27)
3 3

  
where=
u 3 ( )
re u 3 + i im u 3 ( ),
eiωt cos ωt + i sin ωt .
=

Equation (27) can be rewritten

  

U
= 3
re U( )3

 + i im U
3 ( ), (28)

  
where
= 
re U ( ) re (u ) cosωt − im (u )sin ωt ,
3 3 3

  

im U
= 3 ( ) ( )
im u 3 cos ωt + re u 3 sin ωt . ( )



U 3
is called the complex response. It is understood that the actual steady state motion will be
 
given by either the real part of 
U 3
or the imaginary part of 
U 3
depending on ωt .

8
 OSE522, KAIST

Im

( )

im U 3


U 3

θ

( )
Re

re U 3

< Complex response representation >


The magnitude and phase angle of 
U 3
are

  2  2

=
U 3
   
( ) ( )
 
 re U 3  +  im U 3  ,
   
(29)


θ = tan −1
im U 
( ).

3


re U ( ) 3


Note that the actual steady state motion of 
U 3
will be maximum when the phase θ is zero and
 
it is same with the magnitude of 
U 3
. Mathematically, the magnitude of 
U 3
is identical to

magnitude of u 3 :

  2  2
=
U 3 

( )
3 
 
 ( )
 re u  +  im u 
3 

. (30)

9
 OSE522, KAIST

Example problem

In this section, we report an example. The structure and fluid properties are summarized in Tables 1
and 2. Using these properties, we can calculate the dimensionless parameters. The definitions and
quantities of the dimensionless parameters are summarized in Table 3.

Fig 1 shows the real, imaginary and absolute values of displacement. Fig 2 shows the hydrodynamic
pressure and each pressure component .

10
 OSE522, KAIST

<Table 1> Structure properties

Notation (unit) Quantity

Structure length L (m) 10
Structure width B (m) 1
Structure thickness H (m) 1
2
Modulus of elasticity E (N/m ) 5.79×103
4
Moment of inertia I (m ) 8.33×10-2
3
Density of structure ρ s (kg/m ) 8.57
Draft d (m) 8.36×10-3

<Table 2> Fluid properties

Notation (unit) Quantity

3
Density of fluid ρ w (kg/m ) 1025
Water depth h (m) 1.1
Wave length λ (m) 3.176
Wave number k (rad/m) 1.979
Angular frequency ω (rad/sec) 4.347
Amplitude of incident wave a (m) 0.001

<Table 3> dimensionless parameters

Definition Quantity
h
δ= 0.11
L

K = kL tanh kLδ 19.2823

EI
β4 = 4.8026×10-6
ρ w gL4

11
 OSE522, KAIST

Fig 1. Displacement (heave)

12
 OSE522, KAIST

Fig 2. Total pressure and each pressure component

13
 OSE522, KAIST

Homework #6
due date : December 6, 2011

You can download the MATLAB code for the hydroelastic analysis of floating beam problems
at the course website. However, the code does not have a completed function which calculates
the coupling matrices Sc or FcT in equation (22) as shown in Fig 1.

% A function that computes the A_ele

function A_ele = A_ele(span_xbar)

for i = 1 : 3

r = Gauss_p(i);

(i)*span_xbar/2;

End

Fig 1. A function to calculate the coupling matrix

(1) Fill in the blank of the MATLAB file : A_ele.m.

 
(2) Calculate and plot the displacement û 3 and hydrodynamic pressure p̂ for the example
problem in the lecture note.

(3) The hydroelastic behavior of floating beam structures depends on the dimensionless stiffness

parameter β 4 . Plot the displacement û 3 for the different β 4 given in the Table 1 and discuss your
results.

Table 1. Dimensionless stiffness parameter β4

Case 1 Case 2 Case 3

β4 4.8026×10-6 4.8026×10-4 4.8026×10-2

14