Jordan Journal of Civil Engineering, Volume 6, No.

4, 2012

Dynamic Behavior of Jacket Type Offshore Structure

Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

Department of Civil Engineering, University of Basrah, IRAQ. E-Mail: [email protected]

ABSTRACT
Unlike structures in the air, the vibration analysis of a submerged or floating structure such as offshore structures is possible only
when the fluid-structures is understood, as the whole or part of the structure is in contact with water. Using the commercial F.E.A.
program ANSYS (v.12.0) (to model the Winkler and Brick-full bond models) and program ABAQUS(v.6.9) (to model the Brick-
interface model), the stress matrix considering a dynamic load was superposed on the stiffness matrix of the structure. A time domain
solution is recommended, using the generalized Morison’s equation by FORTRAN90 program to construct a program to calculate the
wave forces, and Airy's linear and Second Order Stoke's wave theories are employed to describe the flow characteristics by using
MAPLE13 program (to solve and apply the boundary conditions of the problems on Laplace's equation), and the results are
compared and discussed. Both free and forced vibration analyses are carried out for two case studies.
KEYWORDS: Natural frequency, Fixed jacket offshore, Finite element method, Validity of wave theory, Dynamic
analysis.

INTRODUCTION Based on this design, we calculated the complex

load on the offshore (fixed jacket platform).
The development of the offshore oil industry led to Calculations in the offshore have to take into account
numerous installations of offshore platforms. The wave load(caused by waves). However, since current
major use of these platforms is in the drilling for oil load is insignificant compared to wave load, it can be
and gas beneath the seafloor. Other uses include, but ignored.
are not limited to, military applications, navigational
aid to ships and generating power from the sea. LOAD CALCULATION IN OFFSHORE
Modal test which is one of the examination
assessments is the method to analyze the dynamic The loads that strongly affect offshore structures
characteristics. It is a method to measure frequencies, can be classified into the following categories;
finding the natural frequency of the fixed jacket 1. Permanent (dead) loads. 2. Operating (live) loads.
platform and forecasting the vibration phenomenon for 3. Environmental loads. 4. Construction-installation
mode shape. In case of domestic applications, the study loads. 5. Accidental loads.
of fixed jacket platform system has been actively Whilst the design of buildings onshore is usually
conducted in some big corporations, small and medium influenced mainly by the permanent and operating
enterprises and national researches. But it was loads, the design of offshore structures is dominated by
impossible to obtain systematic data. environmental loads, especially waves, the impact and
loads of which arise in the various stages of
constructional installation.
Accepted for Publication on 14/6/2012.

- 418 - © 2012 JUST. All Rights Reserved.

Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

Environmental loads are those caused by particles, where (ρ) is the density of water, (D) is the
environmental phenomena such as wind, waves, cylinder diameter, (Cm) and (Cd) are inertia and drag
currents, tides, ice and marine growth. Their coefficients (Madhujit and Sinha, 1988).
characteristic parameters, defining design load values, Various methods exist for the calculation of the
are determined in special studies on the basis of hydrodynamic loads on an arbitrarily oriented cylinder
available data. by using Morison's equation. The method adopted here
In this study, the effect of wave, impact and ground assumes that only the components of water particles
motion will be taken into consideration in the dynamic and accelerations normal to the member produce loads
analysis of the offshore structures. (Qian and Wang, 1992).
For an offshore platform, the most important loads To formulate the hydrodynamic load vector Fw,
are the hydrodynamic loads and impact loads which are consider the single, uniform, cylindrical member (i)
included in this study. These hydrodynamic forces are between nodes I and J as shown in Fig. (1). The forces
governed by sea waves, while impacts usually occur are found by the well-known semi-empirical Morison's
during berthing of ships. The most widely used formula (Equation(2)). It also represents the load
approach to determine the hydrodynamic loads (water exerted on a vertical cylinder, assuming that the total
wave forces) acting on the members of a structure is force on an object in the wave is the sum of drag and
normally the semi-empirical Morison's equation. This inertia force components. This assumption (introduced
equation is originally developed to compute the by Morison) takes the drag term as a function of
hydrodynamic forces acting on a cylinder at a right velocity and the inertia force as a function of
angle to the steady flow, and is given as: acceleration (Zienkiewicz et al., 1978; Dean and
Dalrymple, 1984; McCormick, 1973), so that:
dF = ρπ C m a ds + ρD Cd V
2
D 1
π π
(1)
Fn = ρ . D 2 .Cm .ν n′ − ( Cm − 1 )ρ . D 2 .u′n′ +
V ds
4 2
4 4
(2.a)

ρ . D.Cd .(ν n − un′ ) . ( vn − u′n )

In this equation, it is assumed that the wave force is 1
acting on the vertical distance (ds) of the cylinder due 2

to the velocity (v) and acceleration (a) of the water

Y
J
GLOBAL X L d
COORDINATES
S .
u(s)
D
I u.n(s)

MEMBER i s
Figure 1: Water particle velocities along member i

- 419 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

which can be simplified to: are directional cosines in x, y and z directions,

respectively (Zienkiewicz et al., 1978), and:
π
Fn = ρ . D 2 . Cm .ν n′ + ρ . D . C d . (ν n ) . ( v n ) ⎡v x ( s ) ⎤
1 (2.b)
{V ( s )} = ⎢v y ( s ) ⎥
⎢ ⎥
4 2
(7.a)

⎣⎢v z ( s ) ⎦⎥
where:
Fn: nodal hydrodynamic force normal to the cylinder.
D: Outer diameter of cylinder. ρ: Sea water density. ⎡ ax (s ) ⎤
Cd: Drag coefficient. ν n′ : Water particle acceleration. {a ( s )} = ⎢ a y ( s ) ⎥
⎢ ⎥
(7.b)

⎣⎢ a z ( s ) ⎦⎥
Cm: Inertia coefficient. vn: Water particle velocity.
u ′n : Structural velocity. u ′n′ : Structural acceleration.
Equation (2.b) neglects the non-linear terms of drag These velocity and acceleration components are
coefficient (Al-Jasim, 2000; Sarpakaya and Issacom, derived in detail in the next section.
1981), water particle velocity and acceleration can be Now, to calculate the load vector in global
evaluated by potential velocity computed from wave coordinates system, the element is divided into two
theories, the absolute value of velocity is needed to parts by using equation (4) distributing the wave
preserve the sign variation of the force. effects on the beam element equally to the end nodes as

{ }
Generalizing the one-dimensional form of nodal forces. Therefore, the element of hydrodynamic
Morison's equation to the three-dimensional from of load vector f e corresponding to the element nodal
the hydrodynamic force per unit length along the beam displacement vector{q} can be expressed as follows:

⎛f ⎞
element at location (s) measured from its end to the
⎜f ⎟
x1

⎜ ⎟
nearest node is given as (Zienkiewicz et al., 1978):

{F ( s )} = ρ . 4 . D 2 .Cm .{ν n′ ( s )} +
π ρ
. D .Cd .{ν n ( s )}.{ν n ( s ) } ⎜f ⎟
y1

⎜ 0 ⎟
⎜ ⎟
z1
2
(3)
⎜ 0 ⎟
⎜ ⎟
The hydrodynamic force per unit length vector is

{f } = ⎜ 0 ⎟
given as:
⎡ Fx ( s )⎤
{Fw ( s ) } = ⎢⎢ Fy ( s )⎥⎥ ⎜f ⎟
e
(8)

⎜ ⎟
(4)
⎢⎣ Fz ( s )⎥⎦
⎜f ⎟
x 2

⎜f ⎟
y 2

⎜ ⎟
and the normal water particle velocity and

⎜ 0 ⎟
z 2
acceleration vectors are given as:

⎜ 0 ⎟
{Vn(s)}=[s]{u(s)} and {an(s)}=[s]{a(s)} (5)

⎜ ⎟
⎡ 1 − S12 − S1S 2 − S1S3 ⎤ ⎝ 0 ⎠
[s ] = I − S .S T = ⎢ − S1S2
⎢ ⎥
1− S 22 − S 2 S3 ⎥ (6)
⎢− S S − S 2 S3 1 − S32 ⎥
⎣ 1 3 ⎦ ENVIRONMENTAL LOADS

where; I: is the (3x3) identity matrix, s : is the unit The external loads include hydrostatic pressure,
directional vector along the member and S1, S2 and S3 wind, waves, currents, tide, ice, earthquakes,

- 420 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

temperature, fouling, marine growth and scouring. appropriate for most of the wave parameter domain
(Demirbilek and Vincent, 2002).
Hydrodynamics of Water Waves Airy linear theory is the common linear wave
Ocean waves are undulations in the water's surface theory. Stoke’s theory assumes that wave motion
resulting in the transfer and movement of energy. The properties such as velocity potential (φ) can be
disturbance is propagated by the interaction of represented by a series of small perturbations. The
disturbing (e.g. wind) and restoring (e.g. gravity) linear Airy wave theory can be used when the wave
forces. The energy in most ocean waves originates height to wave length ( H ) ≤ 1 as given in
L 50
from the wind blowing across the water's surface. (McCormick, 1973).
Large tsunami or seismic sea waves are generated by
earthquakes, space debris, volcanic eruptions or large
marine landslides. On the other hand, tides, the largest
of all ocean waves, result from the combined
gravitational force exerted on the oceans by the sun and
the moon (Glossary, 2002).
Dean (1968, 1974) presented an analysis by
defining the regions of validity of wave theories in
terms of parameters H/T and d/T, since T is
proportional to the wavelength. Le Mehaute (1969)
plotted a figure to aid in the selection. This figure
shows the validity of different water waves based on
Figure 2: Validity of water wave theory (Al-Jasim, 2000)
the water depth(d), the wave height(H) and the wave
length (L) as shown in Fig.(2).
For the sake of comparison, two wave theories are Hydrodynamic Forces
used to evaluate the velocity and acceleration of water In order to calculate the hydrodynamic forces
particles: Airy wave theory (linear) and Stoke's second acting on the members of an offshore structure, it is
order theory (non-linear). The loads resulting from required first to describe the sea state as occurring in
using these theories with Morison's equation are nature which determines the wave surface profile and
applied on the platform and the resulting bending then the characteristics of the water wave particles
moments, axial forces and displacement variations with hitting the structure. It is always assumed that the water
time of the platform (at a point on the top of the waves are represented as two-dimensional plane waves
platform and near the sea bed on the platform) are that propagate over a smooth horizontal bed in water of
plotted to show the differences between the theories. a constant undisturbed depth (h).
Both Le Mehaute and Dean recommend cnoidal It is also assumed that the waves maintain a
theory for shallow-water waves of low steepness and permanent form, which means that there is no
Stoke’s higher order theory for steep waves in deep underlying current and that the free surface is
water. For low steepness waves in transitional and deep uncontaminated. The fluid (water) is taken to be
water, linear theory is adequate but other wave theories incompressible and inviscid and the flow to be
may also be used in this region. Fenton's theory is irrotational (Dean and Dalrymple, 1984).

- 421 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

Fig.(3) indicates the general form of the xz-plane (b) The pressure is zero and the energy equation must

( ) ( ) ⎤⎥⎦
wave train conforming to these assumptions. Here, the be satisfied, the dynamic condition is:

∂φ ⎡ ∂φ ∂φ
wave is progressive in the positive x-direction and the
+ gη + ⎢ + =f (t ) at z = η
2 2
1
∂t 2⎣ ∂x ∂z
z-axis measured positive upwards from the mean water (10.b)
level, the wave height being H, the wave length L, the
wave period T and η the elevation of the water above 2- At the Sea Floor Where the Vertical Velocity is
the mean water level (Al-Salehy, 2002). Zero
That is:
Direction of

∂φ
Z
propagation

=0 at z = − h
l

∂z
Celerity (10.c)

η a Sea level
The components of water particle velocity can be
X given as (Essa and Al-Janabi, 1997):
∂φ
a

νx =
∂x
h (11.1a)

∂φ
Sea bed

νz =
∂z
Figure 3: Definition of terms used in the wave equations (11.1b)

whereas, the components of the local particle

The surface must satisfy the special linear form of
acceleration, which are only taken into account in the

velocity potential (φ ) and is subjected to the above

the wave equation of Laplace solution to obtain the
computation of the hydrodynamic force are given as:

∂ν
conditions and linearized boundary conditions.
=
∂t
a x
x (11.2a)

∂ φ ∂ φ
2 2

+ = 0
∂ν
∂X ∂Z
(9)
=
2 2

∂t
a z
z (11.2b)

The major problem in solving for (φ ) arises from

To solve Equation (9), some BCs must be satisfied
the boundary conditions to be applied at the air-water
interface, η (t), being itself part of the solution sought.
(Al-Salehy, 2002; Dean and Dalrymple, 1984;
McCormick, 1973):
Therefore, there are several solutions in common use.
1- At the Sea Surface (Air-Water Interface)
These are linear wave theory in deep water, Stoke’s
(a) The velocity of a particle must be tangential to the
higher order wave theories, stream potential function
surface, the kinematic condition is:
∂η ∂φ ∂η ∂φ
theory and Cnoidal theories in shallower water
+ ⋅ − = 0 at z = η
(Sarpakaya and Issacom, 1981).
∂t ∂x ∂x ∂z
(10.a)
In the present study, only the linear (Airy) wave

- 422 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

theory and Stoke’s second order theory are considered - more accurately describes orbital velocity
to compute the characteristics of water particles and the asymmetries.
hydrodynamic forces.
To find velocities and accelerations that are used in
Morison’s equation, Laplace equation must be solved
by considering the BCs at the sea surface η(t), being
itself part of the solution sought. Therefore, different
wave theories are used as mentioned.
The linear wave theory (Airy theory) is used only to STOKE’S WAVES

find velocities and accelerations at different depths,

locations and times. Figure 4: Wave profile shapes of
different progressive gravity waves
Wave Theories (Demirbilek and Vincent, 2002)
All wave theories obey some form of wave
equation in which the dependent variable depends on Airy Theory (Linear Wave Theory)
physical phenomena and boundary conditions (Al- This theory is termed following a linearization of
Salehy, 2002). In general, the wave equation and the the boundary conditions at the air-water interface. In
boundary conditions may be linear or non-linear. this theory, the essential idea or restriction is that the
Some of these theories are shown in profiles wave height (H) must be much smaller than both the
Fig.(4): wave length (L) and the still water depth (d); that is
(H<<L, d). The linear wave theory for two-
1- Airy Wave Theory: (Sinusoidal Waves) dimensional, free, periodic, waves is developed by
This theory: linearizing the equations that define the free surface
- is most accurate for low amplitude waves in deep boundary conditions. With these and the bottom BCs,
water. the periodic velocity potential is sought that satisfies
- is less accurate for predicting wave behavior in the requirements of irrotational flow. The free surface
shallow water. BCs may now be applied directly at the still water level
- is the most commonly used wave theory because it (Dean and Darlymple, 1984).
is the least mathematically complex one. Therefore, the free surface BCs, as expressed in
- does not take into account the effect of wave height Equation (12.a) and Equation (12.b), are reduced to:

∂φ ∂η
in determining wave velocity.
− =0
∂z ∂t
at z=0 (12.a)
Stoke’s Wave Theory (Trochoidal Waves)

∂φ
This theory:
+ g .η = 0
∂t
- can be used for deep, intermediate and shallow at z=0 (12.b)
water waves.

12.a, b), the velocity potential (φ ) can be found as

- is mathematically complex. By using separation of variables and BCs (Equation
- takes into account the effect of wave height on
velocity. (Othman and Dawood, 2002):

- 423 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

π .H cos[ k ( z + h )]
φ = . . sin( k x − Ω .t ) (13) small so that the theory is applicable only in deep water
k .T sin( k . h )
and a portion of the intermediate depth range.
in which: For engineering applications, the second-order and
.
k: wave number ( ); T: wave period; possibly the fifth-order theories are the most commonly
used (Sorensen, 2006).
Ω: wave circular frequency ( Ω 2 .π
=
); Stoke’s wave expansion method is formally valid
T
under the conditions (Iraninejad, 1988):
h: depth of water; H: wave height.
Now, it is simple to obtain the velocities and H/d << (kd)2 for kd < 1 and H/L << 1.
accelerations as:

∂ϕ π .H cosh[k (z + h )]
Stoke’s wave theory is considered most nearly valid
Vx = = . cos( kx − Ω.t )
∂x
. (14.a) in water where the relative depth (D/L) is greater than
T sinh(kh )
about (1/10) (Patal, 1989).

∂ϕ π .H sinh[k (z + h )]
Stoke’s theory would be adequate for describing
Vz = = .sin(kx − Ω.t )
∂z
. (14.b) water waves at any depth of water. In shallow water,
T sinh(kh )
the convective terms become relatively large, the series

∂v x ∂ϕ 2.π .H cosh[k (z + h)]

convergence is slow and erratic and a large number of
ax = = = .sin(kx −Ω.t )
2 2

∂t ∂t ∂x
2
. (15.a) terms is required to achieve a uniform accuracy (Muga,
T sinh(kh)
2003).

∂v z ∂ϕ 2.π .H cosh[k (z + h)]

The fluid particle velocities are then given by:
az = = = .sin(kx −Ω.t )
2 2

∂t ∂t ∂z
2
. (15.b)
T sinh(kh) . cosh .
. . cos Ω .
These velocities and accelerations in Equation (14) sinh
and Equation (15) are used in Morison's equation to cosh
calculate load vectors of hydrodynamic loading by . cos Ω .
sinh
using linear Airy wave theory after being transformed
from global coordinates for each member of the
offshore platform. . sinh .
. . sin Ω .
sinh
Stoke’s Second Order Wave Theory
sinh
Stokes (1874) employed perturbation techniques to . sin Ω .
sinh
solve the wave boundary value problem and developed
a theory for finite—amplitude wave that he carried to
The fluid particle accelerations are then given by:
the second order. In this theory, all the wave
characteristics (velocity potential, celerity, surface . .
. . sin Ω + .
profile, particle kinetics…etc) are formulated in terms
of a power series in successively higher orders of the
cosh
wave steepness (H/L). . sin Ω .
sinh
A condition of this theory is that (H/d) should be

- 424 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

− ω 2 [M ]{U }s i n ω ( t − t 0 ) + [K ]{U }s i n ω ( t − t 0 ) = {0}

. sinh
. . cos Ω (20)
sinh
Since the sine term is arbitrary and may be omitted,
the above equation is reduced to the form:

[[k ] − ω ]
[ M ] {U } = {0}
sinh
. . cos Ω .
sinh 2
(21)

These velocities and accelerations in Equation(16) According to the Cramer's rule, the solution of this
and Equation(17) are used in Morison's equation to set of simultaneous equations is of the form:

{U } = {0}
calculate load vectors of hydrodynamic loading by
using Stoke’s wave theory after being transformed
[ K ] − ω2 [ M ]
(22)
from global coordinates for each member of the
offshore platform. Hence, a nontrivial solution is possible only when
the denominator determinant vanishes. In other words,
DYNAMIC ANALYSIS finite amplitude free vibrations are possible only when:

In offshore structures, the applied loads [Κ ] − ω

2
[Μ ] = 0
(23)
(environmental loads) generally have a dynamic nature. In general, equation (23) results in a polynomial
To study the behavior of these structures, free vibration equation of degree N in ω for a system having N
2

and forced vibration must be considered in order to degrees of freedom. This polynomial is known as the
understand the actual (as possible) behavior and characteristic equation of the system. The N roots of
response. this equation, which are called the eigenvalues,
represent the frequencies of the N modes of vibration
Free Vibration Analysis which are possible in the system (Paz, 1980).

{ }
The equations of motion for a freely, undamping Corresponding to each eigenvalue, there will be an
system can be obtained by omitting the damping matrix eigenvector, or natural mode U n , where:
and applied load vector from Equation(2) to get: ⎡U1 ⎤
⎢ ⎥
⎢U 2 ⎥

{U n }
⎢M ⎥
= ⎢ ⎥
(18) n=1,2,3,…N. (24)
⎢M ⎥
⎢ ⎥
⎢U N −1 ⎥
⎢U ⎥
The solution of this equation, which is known as an
eigenproblem, can be postulated to be of the form: ⎣ N ⎦

{U } = {U } s i n ω (t − t 0 )
The modes are determined only within a constant
(19) multiplier. Thus, modes can be normalized in any

where {U } is the vector of amplitudes of motion { }

convenient manner. Therefore, if the value of one of
the elements of the natural mode vector U N is
(which does not change with time). The substitution of assigned as a specified value, say a unity for the first
equation (19) into the equation of free vibration element, then the remaining (N-1) elements are
equilibrium (18) gives: uniquely determined. The process of normalizing a

- 425 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

natural mode is called normalization and the resulting Forced Vibration Analysis
modal vectors are called normal modes. For To understand the response of offshore structures
convenience, these normal mode vectors are usually subjected to a load of dynamic nature, forced vibration
expressed in a dimensionless form by dividing all the analysis is used to get the response of the platforms to
components by one reference component. The resulting these forces.
vector corresponding to the nth eigenvalue is called the There are different methods available for evaluating
nth mode shape { φ n} (Clough and Penzien, 1975), the structural response to dynamic loads, such as
thus: frequency domain analysis, direct integration
⎡φ 1 n ⎤ ⎡1 ⎤
⎢ ⎥ ⎢ ⎥
method,…etc. These are Newmark’s implicit, most
⎢φ 2 n ⎥ ⎢U 2 n ⎥
⎢ ⎥
flexible step-by-step integration methods in time
⎢φ ⎥
{φn } = ⎢ 3n ⎥ = ⎢U 3 n ⎥ (25) domain, presented by Newmark (Zienkiewicz et al.,
⎢M ⎥ Uk n ⎢M ⎥
1

⎢ ⎥ ⎢ ⎥
1978; Bathe and Wilson, 1977; ANSYS, 1997; Vugts
⎢M ⎥ ⎢M ⎥
⎢φ ⎥ ⎢ ⎥
and Hayes, 1979). This method (including an improved
⎣ N n⎦ ⎢⎣U N n ⎥⎦ algorithm called HHT (Chung and Hulbert, 1993) and
using finite difference expansions in the time interval
In which U k n is the reference component. ( ∆ t )) is based on the following expressions for the
The square matrix made up of the N mode shapes velocity and displacement at the end of the time
will be represented by[ φ ]: interval (Bathe and Wilson, 1977).

⎛ φ 11 φ φ φ ⎞ u&t + ∆ t = u&t + ∆ t (1 − δ )u&&t + ∆ t δ u&&t + ∆ t

⎜ φ φ φ ⎟
12 13
L 1 N

⎜ 21 ⎟
(27)

[φ ] = ⎜ M
O O
⎟
22 2 N

⎜ M ⎟ ( 1 − α ) u&&t + ( ∆ t ) 2 α u&&t + ∆ t
O O O M
ut+∆t= ut + ∆t u& t + ( ∆t )2
⎜ ⎟
(28)
⎜φ φ ⎟
O O O M 2
⎝ N 1 L L L ⎠
where α,δ are selected to produce the desired
N N

These mode shapes have certain special properties, accuracy and stability. One of the most widely used
which are very useful in the structural dynamic methods is the constant average acceleration method
analysis. These properties are called the orthogonallity when (δ=0.5, α=0.25) which is a conditionally stable
relationship (Biggs, 1964), which can be expressed for method without numerical damping. This method is
any system as: called an (implicit integration method), since it satisfies

{φ } [ Μ ] {φ } = fo r m ≠ n
T
the equilibrium equation of motion at time t+∆t, or:
0
m n
(26)
and M u&& t + ∆ t + C u& t + ∆ t + Ku t+ ∆t = Ft+ ∆t (29)

{φ } [ Κ ] {φ } = 0 ≠ n
T
fo r m
m n
This equation can be solved by iteration; however
ANSYS12.0 and ABAQUS6.9 programs use a Equations (27), (28) and (29) can be combined into a
method of subspace iteration, this method requires the step by step algorithm which involves the solution of a
Jackobi method, Ritz reduction functions and iterative set of equations. Each time step is of the form:
procedure as detailed in (Zienkiewicz et al., 1978;
Bathe and Wilson, 1977). K* .U t+∆t = F* (30)

- 426 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

Since K* is not a function of time, it can be four piles embedded to a depth of (60m) below
triangularized only once at the beginning of the mudline in the sea bed that support the platform and
calculation. The computer solution time for this type of (303) nodes.
algorithm is basically proportional to the number of A FORTRAN program is constructed to calculate
time steps required. the water particle velocity and acceleration by using
two theories: Airy theory and Stoke’s second order
LOAD CALCULATION IN OFFSHORE theory. Also, this program calculates the wave forces at
each node in the superstructure embedded in the fluid
In order to investigate the dynamic response medium by using Morison's equation.
method in this study, a jacket type model of an offshore
platform type structure has been considered in this The forces on each member are calculated and then
paper, and for the sake of comparison, two wave distributed on the nodes for x, y and z-directions.
theories are used to evaluate the velocity and
acceleration of the water particles; namely the Airy The deck mass is modeled using the lumped mass
wave theory (linear) and Stoke’s second order theory in five nodes forming a pyramid (Al-Salehy, 2002;
(non-linear). The loads resulting from using these Carpon et al.). This model is the same model adopted
theories with Morison's equation are applied on the by Al-Salehy (2002) taking the Winkler model with
platform and the resulting bending moments, axial isolated springs at nodal points only to represent the
forces and displacement variations with time of the soil resistance taking into account normal and
platform (at a point on the top of the platform and near tangential moduli of subgrade reaction in three
the sea bed on the platform) are plotted to show the directions for each node. Table (1) contains the
differences between the theories. properties of the piles, while the properties of the steel
Finite element method is used for both spatial and and physical properties of the soil are defined in Table
temporal coordinate systems. (2).
To take the effect of soil-structure interaction into
account, the two models are analyzed using Winkler, Table 1. Properties of piles
Interface (F.E.) and Full Bond (F.E.) methods which
have different treatments in representing the soil- Outer
Member Thickness
structure relation as discussed earlier. The length of Shape Dimension
Number (mm)
pile embedment in the soil, the end conditions of the (m)
pile: spring, hinged and fixed, and modeling of inertia Circular
forces as lumped mass are taken into account. 1 2.0 55
Pipe
Circular
Structure Description 2 1.525 38
Pipe
Fig. (5) shows that the fixed jacket offshore Circular
platform model described in (Al-Salehy, 2002; Carpon 3 0.91 25
Pipe
et al.) is adopted. The frame is descritized into (178) Circular
beam elements for superstructure and (240) beam 4 0.75 25
Pipe
elements embedded in elastic soil used to model the

- 427 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

Table 2. Properties of material bending and axial modes due to stiffer model adopted
Modulus of in the present study. But the values of natural
Density Poisson’s
Material elasticity 3
frequencies are close in torsion mode. This is because
2 ( to n / m ) ratio (µ)
(Es) ( kN / m ) the pile legs of the structure prevent the superstructure
Steel 200E6 7.8 0.3 from rotation and there is no effect of the torsional
stiffness of the piles at different lengths of the piles on
Soil 45E3 1.9 0.3
the torsional stiffness of the entire structure.
The main variables adopted in the free vibration
Soil Properties
analysis are soil properties (according to type of soil).
The effect of soil type on the dynamic response is
The first four values of cyclic natural frequencies
studied by using three types of soil in addition to the
obtained from the different three models (Winkler, Full
soil of the case study 1. The soil properties including
bond and Interface) (Al-Salehy, 2002) are shown in
elastic modulus, Poisson's ratio and density are given in
Table (5).
Table (3) (Rubin and Coppolina, 1983).

Table 4. Natural frequencies for basic modes of

Table 3. Physical properties of soils
vibration for different models of soil-structure

(ton / m )
Modulus of

( kN )
Density Poisson’s interaction for case study
Soil elasticity (Es) 3

/m 2 ratio (µ) Present work (ANSYS and Al-Salehy

ABAQUS) Lumped (Hz) (2002)
Soft clay 15E3 1.8 0.3 Mode
Brick (Full Brick (STAADIII)
Winkler
Medium Bonding) (Interface) (Hz)
30E3 1.9 0.4
clay Sway 0.38297 0.37947 0.31137 0.207
Hard
60E3 1.9 0.3 Torsion 0.69980 0.69789 0.57251 0.725
clay
Bending 1.51347 1.47543 1.52865 1.507
Axial 1.53777 1.51501 1.55023 0.940
Free Vibration Analysis
Free vibration analysis is carried out taking into
account the length of pile embedment in the soil, the Table 5. First four natural frequencies for
end conditions of the pile: spring, hinged and fixed, different types of soils (case study)
and modeling of inertia forces as lumped mass. Model
Table (4) compares the fundamental four modes of Brick
Soil Brick (with
natural frequencies of the present work for spring pile Winkler (Full
Type Interface Element)
tip in Winkler model and for fixed pile tip in [Brick (Hz) Bond)
(Hz)
model(Full bonding) and Brick model with interface (Hz)
element] with that reported by Al-Salehy (2002). Al- 0.11534 0.08553 0.06384
Salehy used isolated springs to represent the soil- 0.3401 0.32121 0.2397
Soft clay
structure interaction. The tabulated data shows that the 1.09171 1.05369 1.0756
results obtained by the present work for Winkler model 1.3455 1.19533 1.13723
are higher than those given by Al-Salehy for sway,

- 428 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

0.28723 0.2400 0.19694 taken as (2.0) and (0.8), respectively.

Medium 0.3499 0.34196 0.28053 Fig. (7) shows the force-time curve for node (A)
clay 1.51347 1.3879 1.43796 due to wave loads in three directions using Airy theory
and Morison’s equation {Equation (2.b)}.
1.5378 1.4469 1.48054
Fig. (8) shows the comparison between the
0.62233 0.52192 0.44202
variation of wave forces at node (A) in x-direction with
Hard 0.7155 0.60223 0.50966
time for Airy and Stoke’s theories.
clay 1.9457 1.85244 1.6004
Figs. from (9) to (20) show (for the case of lumped
2.08617 1.9361 1.96207 mass approximation using Airy and Stoke’s theories,
respectively) for Winkler, Brick (full bonding) and Brick
Forced Vibration Analysis (with interface element) models, respectively, and (spring
Different computer programs are used in support for pile tip) with Winkler model, (fixed support
calculating the dynamic response of the adopted case for pile tip) with Brick (full bonding) and Brick (with
studies. Software ANSYS12 for Winkler and Brick interface element) models, the dynamic response for both
(Full Bond) models and software ABAQUS6.9 for the bending moment at deck level (member AB) and axial
Brick (with interface element) model are used. force in seabed level (member CD), respectively. It is
The effect of wave theory used was studied in the shown that the two curves have a similar behavior and
forced vibration analysis of offshore jacket platform both of them reached a steady state condition after one
with soil–structure interaction subjected to wave period of time (only 24 sec), and it is shown that there is
loading only. little increase in (max. value of the bending moment at
To study the action of wave forces on the dynamic deck level and in max. value of the axial force in sea bed
behavior of the offshore platform model shown in Fig. level) when using the Stoke’s theory .
(5), the following wave parameters have been In addition, the structure response in the case of
considered (Al-Salehy, 2002): full bond model is less than that resulting from using
Wave height = 21 m, Wave period = 12 sec, interface model, and the least is for Winkler model.
Wave length = 225 m, Water depth = 115 m, This is because the interface element considers relative
Water density = 1025 kg/m3. displacement or slip movement between soil and
The assumed value of the viscous damping ratio is structure.
(5%) for all modes of vibration considering (2%) as
hydrodynamic damping, whereas the remaining (3%) EFFECT OF WAVE THEORY USED ON THE
simulates energy dissipation from sources other than PLATFORM DISPLACEMENT
hydrodynamic, called structural damping (Al-Salehy, In this case, the platform model is analyzed to
2002; Al-Jasim, 2000). evaluate the effect of using different wave theories on
The inertia coefficient (CM) and drag coefficient the variation of displacement with time at point A (as
(C D) are estimated from laboratory experiments shown in Fig.(5)) on the platform due to the variation
indicating a general range from 1.2 to 2 for C M and of wave forces calculated by the different wave
from 0.6 to 1.2 for C D depending on flow conditions theories.
and surface roughness (Rubin and Cappolina, 1983; Depending on the parameters (D/L) and (H/L) and
Faltinsen, 1990). In this study, these coefficients are from Fig. (2) (Al-Jasim, 2000), the appropriate wave

- 429 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

theory for the cases studied is given in Table (6). Brick (full Fixed 15.0972
Table (7) shows the maximum deck displacements Airy bonding) Hinge 15.0972
for all end conditions of pile tip (spring, hinge and Brick Fixed 19.9381
fixed) and for all soil-structure interaction (with
representations (Winkler, Brick (full bonding) and Hinge 19.9381
interface)
Brick (with interface) models).
Spring 9.654
The comparison of the results of using other
Winkler Fixed 8.8568
theories to using the most suitable theories is given in
Hinge 8.8568
Table (8) for all end conditions of pile tip and for all
Brick Fixed 16.0332
soil-structure interaction representations.
(full
The comparison between Airy and Stoke’s theories Hinge 16.0407
Stoke’s bonding)
by the plot of variation of displacements at node A
with time for spring end condition of pile tip in Brick Fixed 21.1942

Winkler model, and for fixed end condition of pile tip (with
Hinge 21.1942
in Brick (full bond) and Brick (with interface element) interface)

models and for all soil-structure interaction

representations is given in Fig. (21). It is obvious that Table 8. Comparison of maximum deck
there is a little difference in time variation of the displacement with end conditions and soil
displacement between the two theories for all soil- representations from different theories
structure interaction representations. Winkler Model
Case of comparison Displacement ratio
SA T
Table 6. The most suitable wave theory for SST 0.943
waves used FA T
FS T 0.943
Case Wave Wave
D H H A T
Study length L L Theory H S T 0.943
(m) Brick (full bonding) model
F A T
1 225 0.5 0.09 Stoke’s F S T 0.942
2 261 0.08 0.04 Airy H A T
H S T 0.942

Table 7. Maximum deck displacement with end Brick ( with interface) Model
FA T
conditions and soil representations FS T 0.941
H A T
H S T 0.941
Soil-Structure End
Max.displacement
Theory interaction conditions SAT: Spring end condition of pile tip, Airy theory.
(mm)
model of pile tip FAT: Fixed end condition of pile tip, Airy theory.
Spring 9.1075 HAT: Hinge end condition of pile tip, Airy theory.
Winkler Fixed 8.3555 SST: Spring end condition of pile tip, Stoke’s theory.
Hinge 8.3555 FST: Fixed end condition of pile tip, Stoke’s theory.

- 430 -
Jordan Journ
nal of Civil Eng
ngineering, Vollume 6, No. 4, 2012
2

HST: Hinge end conditionn of pile tip, Sttoke’s theory.. stuudy. The varriation in the resulting disp placements iss
less than 6% forf the ratio bbetween Airy y and Stoke’ss
th
heories, respectively, thereefore, with su uch range off
diifference Airyy wave theoryy can be used to reduce thee
co
omputation effffort, especiallyy in the first design
d stages.

Figure 7: Wave
W forces aat node (A) in
n x, y and
z--directions wiith time for A
Airy theory inn case study 1
Airy Theo
ory Stokke Theory
200.00
Figurre 5: Geometry and dimen nsions of 160.00
120.00
jaacket type of offshore
o platform 80.00
40.00
0.00
‐40.00
‐80.00
‐120.00
‐160.00
‐200.00
0 5 10 15 20 25 30 35 40 45

Figure 8: Commparison in vaariation of wave

w forces at
node
n (A) in X-
X direction w
with time for case
c study 1
betweeen Airy and Stoke’s theoories

Figure 6: Model
M of the sttructure with
h soil-structurre-
interactionn (3-D) (a,b,cc)
Based onn Table (6), thhe appropriatte selection off a Figure
F 9: Timme variation foor the maximmum bending
wave theoryy for the casse study is Stoke’s theoory. moment
m at th
he level of decck due to wav
ve forces for
However, Airy
A theory caan be applied for this caase (Win nkler) model by Airy theo ory

- 431 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

(Winkler) model by Stoke’s theory

Figure 10: Time variation for the maximum Figure 13: Time variation for the maximum
bending moment at the level of deck due to wave bending moment at the level of deck due to wave
forces for (Winkler) model by Stoke’s theory forces for Brick (full bond) model by Airy theory

Figure 11: Time variation for the maximum axial Figure 14: Time variation for the maximum
force at the level of sea bed due to wave forces for bending moment at the level of deck due to wave
(Winkler) model by Airy theory forces for Brick (full bond) model by Stoke’s theory

Figure 12: Time variation for the maximum axial Figure 15: Time variation for the maximum axial
force at the level of sea bed due to wave forces for force at the level of sea bed due to wave forces for

- 432 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

Brick (full bond) model by Airy theory forces for Brick (interface) model by Stoke’s theory

Figure 16: Time variation for the maximum axial

force at the level of sea bed due to wave forces for
Brick (full bond) model by Stoke’s theory Figure 19: Time variation for the maximum axial
force at the level of sea bed due to wave forces for
Brick (interface) model by Airy theory

Figure 17: Time variation for the maximum

bending moment at the level of deck due to wave
forces for Brick (interface) model by Airy theory

Figure 20: Time variation for the maximum axial

force at the level of sea bed due to wave forces for
Brick (interface) model by Stoke’s theory

Figure 18: Time variation for the maximum

bending moment at the level of deck due to wave

- 433 -
Dynamic Behavior of… Anis A. Mohamad Ali, Ahmed Al-Kadhimi and Majed Shaker

This research shows the concept of dynamic design

about two types of wave theories. Therefore, the
following concluding remarks can be drawn:
1. The overall dynamic response of the structure
(offshore structure) is sensitive to the soil-structure
interaction model. The natural frequencies of the
interface model adopted in this work are less than
natural frequencies of other models considered in
this study, because the interface model considers
the soil block in addition to the pile foundation
leading to an increase in the model mass and a
decrease in the natural frequencies.
2. The interface element model gave more acceptable
and reasonable results than the other models
considered in this study.
3. The interface model gives more realistic results in
comparison with the Winkler model. This is
because the Winkler model concentrates the
physical properties of soil at the locations of
springs, while in the interface model the soil
parameters are distributed among the block
elements, thus giving the true representation of soil
behavior.
Figure 21: Time variation of deck displacement for 4. Wave characteristics represented by wave theories
different theories (Airy and Stoke’s) and for used in the present work have a smaller effect on
different soil-structure interaction models behavior and response of the offshore platform.
5. The results showed a variation in maximum deck
CONCLUSIONS displacement of about (6%). For the case of
different wave theories on case study 1, therefore,
This research compared the results of forced with such range of difference, Airy wave theory
vibration analysis that applied the periodic load to can be used to reduce the computation effort,
exciting force expressed by natural frequency results. especially in the first design stages.

REFERENCES Al-Salehy, S.M.S. 2002. Dynamic Analysis of Offshore

Structure Using Finite Element Method, MSc Thesis,
Al-Jasim, S.A.J. 2000. Dynamic Analysis of Offshore Basrah University.
Template Structures with Soil-Structure Interaction. ANSYS Software User Manual. Copy Right 7, 1997.
PhDThesis, University of Basrah, March. Bathe, K. J. and Wilson E. L. 1977. Numerical Methods in
Finite Element Analysis. Prentice-Hall, Inc.

- 434 -
Jordan Journal of Civil Engineering, Volume 6, No. 4, 2012

Biggs, J.M. 1964. Introduction to Structural Dynamics, Ocean Engineering, 15 (6).

McGraw-Hill, Inc. McCormick, M. E. 1973. Ocean Engineering Wave
Carpon, F., Williams, W. and Symons M.V. A Parametric Mechanics, John Wiley and Sons, New York.
Study of the Free Vibration of an Offshore Structure Muga, B.J. 2003. Deterministic Description of Offshore
with Piled Foundations. Journal of Waterway, Port, Waves-Dynamic of Offshores Structure. Edited by
Coastal and Ocean Engineering. James F.Wilson, John Wiley and Sons.
Chung, J. and Hulbert, G.M. 1993. A Time Integration Othman, R.A. and Dawood, A.O. 2002. Two-and Three-
Algorithm for Structural Dynamics with Improved Dimensional Dynamic Analysis of Submarine Pipelines
Numerical Dissipation: The Generalized-α Method, under the Action of Wave Forces. Proceedings of the
Journal of Applied Mechanics, 60: 371. Second Minia International Conference for Advanced
Clough, R.W. and Penzien, J. 1975. Dynamics of Trends in Engineering, 7-9 April, Minia- Egypt.
Structures. McGraw-Hill, Inc. Patal, M. 1989. Dynamics of Offshore Structures, Buttler
Dean, R.G. and Dalrymple, R.A. 1984. Water Wave and Tanner Co., England.
Mechanics for Engineers and Scientists, Prentice-Hall, Paz, M. 1980. Structural Dynamics Theory and
Inc. Computation. Van Nostrand Reinhold Company.
Demirbilek, Z. and Vincent, C. L. 2002.Water Waves Qian, J. and Wang, X. 1992. Three-Dimensional Stochastic
Mechanics. http://www.usace. army. mil/ publications/ Response of Offshore Towers to Random Sea Waves.
engmanuals/em1110-2-1100/Partll/Part-ll-chap1entire. International Journal of Computers and Structures, 43
pdf. 30 April. (2).
Essa, M.J.K. and Al-Janabi, A.S.I. 1997. Dynamic Rubin, S. and Coppolina, R.N. 1983. Flexibility
Analysis of Plane Frames Partially Embedded in Monitoring of Offshore Jacket Platforms, Copyright:
Winkler Elastic Foundation. Al-Muhandis Iraqi Offshore Technology Conference.
Journal, 130. Sarpakava, T. and Issacom, M. 1981. Mechanics of Wave
Faltinsen, O.M. 1990. Sea Loads on Ship and Offshore Forces on Offshore Structures, John Wiley and Sons.
Structures, Cambridge Ocean Technology Series Sorensen, R.M. 2006. Basic Coastal Engineering, Springer
Cambridge University Press. Science and Business Media, Inc.
Glossary. 2002. Waves. Practical Ocean Energy Vugts, J.H. and Hayes, D.J. 1979. Dynamic Analysis of
Management Systems, Inc. Fixed Offshore Structures: A Review of Some Basic
Iraninejad, B. 1988. Dynamic Analysis of Fixed Offshore Aspects of the Problem. Engineering Structures, 1 (3).
Platforms Subjected to Non-Linear Hydrodynamic Zienkiewicz, O.C., Lewis, R.W. and Stagg, K.G. 1978.
Loading, PhD Thesis, Illinois Institute of Technology. Numerical Methods in Offshore Engineering, John
Madhujit, M. and Sinha, S.K. 1988. Modeling of Fixed Wiley and Sons.
Offshore Towers in Dynamic Analysis. Journal of the

- 435 -