1

Fatigue Analyses in Fixed Offshore Jackets using Spectral Method

Peyman. Moazen
1
, Mohammad Javad Ketabdari
2

1-Hormozgan University, Department of Civil Engineering
2-AmirKabir University of Technology, Faculty of Marine Technology

Abstract
Fatigue, is defined as structural failure under alternative load cycles in a long period
of time. This experience is important when cyclic loads exert to the structures. In this
study a model was developed to estimate fatigue life under irregular waves for fixed
offshore jackets using S-N curves and Palmergen-Miner method. In this study valid
codes and most authentic international regulations, such as ABS and API codes of
practices were employed. The inputs to the model are wind velocity, wave spectral
density and structural details including S-N curves, type of connections and welding
situation in different joints. The model can estimate the fatigue life of the structure
based on the worst environmental condition. A case study carried out for a typical
jacket in Persian Gulf. A dynamic analysis using SAP2000 was undertaken with an
exciting force extracted from a proposed spectrum for sea state of the study area. The
results showed that the spectral analysis of fatigue life is more accurate than time
domain analysis using zero- up crossing technique. The result of the model also
shows promise for optimum design of steel offshore jacket type platforms choosing
proper type of connections and high quality welding in joints.

Keywords: Fatigue life, Spectral analysis, Jacket type platforms, Palmergen-Miner
rule, welded connections

1 Introduction
Wave is the most important phenomenon in the sea environment inducing
considerable forces to marine structures. In fact this is the main source of structural
cyclic stresses leading to fatigue fracture. Therefore, estimating fatigue life for such
structures, have a particular importance for marine engineers. To estimate the fatigue
life it is required to determine several transfer functions. First one is a transform from
wave profile to force in time or frequency domain. Second one is a transform of
dynamic forces to stresses in structural elements. Finally a transform is required to
give fatigue life using concentrated stresses.

2 Wave loading
To determine the fatigue life in a marine structure it is required to estimate the wave
loading to the structure. Therefore providing realistic information about the waves
properties in the desired location is the first stage in this study. Wave spectrum is a
useful tool to simulate the free surface profile for seas and oceans. Different research
groups tried to propose wave spectrums for different seas based on long term field
studies using wave buoys. some of these standard wave spectra are JONSWAP,
Pierson Moskowits and Bretshnieder Spectra. For Persian Gulf in south part of Iran
the available field measurements are not adequate to be able to propose an original
wave spectrum. However a few researchers tried to propose a modified spectrum to
above mentioned spectra using limited available data. A modified model to

1 M.Sc. Graduated in Marine structures
2 Assistant Professor
2
JONSWAP spectrum was proposed for north part of Persian Gulf [1]. This study
suggests the following relation for Persian gulf :
)
2
) (
exp{
4 5 4 2
2 2
2
). ) (
4
5
exp( ) 2 ( . ) (
f
f f
m
f
f
f g f E




= (1)
Where
2.68 , =
m
f 0.2 , =0.07 ( for f 0.2) , =0.09 ( for f 0.2)
Using this spectrum the wave amplitude can some how be found using inverse Fourier
transform. This leads to a series of monochromatic waves with different amplitudes
and frequencies. In this study using the inverse Fourier transform on relevant
spectrum a series of monochromatic waves with different periods and amplitudes was
generated. Then superposition of these individual waves each of which has an
amplitude consistent with the energy in the target spectrum and a random phase as
follows:
( ) cos( ) t c k x t
n n n n
n
N
= +
=

1
(2)
where: c S f df
n n
= 2 ( ) and
n n
f = 2 .
The next stage is transform of free surface profile to force on members of the jacket
platforms. This can be performed using Morison Equatin as:
2
2
2
. .
4
. . .
u
D C
t
u D C
F
d m

+

= (3)
Where D is the member diameter, is the sea water density, u t is the horizontal
acceleration of the water particles and Cm and Cd are the inertia coefficient and drag
coefficient respectively. Different Standards suggested some values for these
coefficients [2].

The next step is calculating the stress concentration in particular elements of the
jacket due to wave attach. Finally considering Palmergen-miner rule and S-N curves,
we are able to calculating the fatigue life for members or connections of the jacket
leading to a fatigue life for whole structure. In this project ABS Standard has been
widely used to calculate fatigue life in structural connections [3]. The out put are
shown by plotting some curves which can display the variation of fatigue with the
kind of connection and welding details. Curve 1 shows the effect of connection class
on fatigue life. This curve can be used as a guide line to select proper connections for
particular parts of the jacket based on the fatigue analysis results. Some tests
performed on the effect of water depth on fatigue life of the jacket. The results are
presented in Curve 2. This curve shows that an increase in water depth around the
marine structure, resulted to an increase in stress leading to a decrease in fatigue.
Similarly, corrosion condition is another parameter which has a significant effect on
the fatigue life and will be considered in calculations. Stress concentration factors
3

and its method of calculation in a welded profile is an important parameter in this
process [4]. Existence of any kind of defect in a welded profile, categorize it in a
lower level of connection classification. This evaluation were set up in ABS standard
but could be developed in related studies. The fatigue life for a sample Jacket type
platform in Forouzan region of Persian gulf was studied. This Jacket is shown in fig.
1. To get a better results the height and the number of the stories of the structure were

3
SCF
3
increased based on assumptions that the water depth is 80m and dimensions of deck
were 10x20 m
2
. Using the available hydrodynamic and structural softwares. The
wave loading is dynamic (see fig. 2) and therefore stresses varies in the structure as
the wave troughs and crests pass the jacket elements. Curve 4 shows the time histories
of stress in a typical jacket element. The most common shape for marine structures
members is pipe element.. The connection of the legs to the sea bed is assumed to be
rigid 2 meters deep under the sea bed. All the bracings have a pinned connection to
the main legs. Therefore they dont induce considerable moment in the jacket leading
to no significant flexural moment in these elements [5]. The output of stress analyzer
software then fed to a developed program in MATLAB to estimate the fatigue. To do
this several S-N cures in different situation based on ABS standard was introduced to
the developed model. In the tubular welded connections, because of the local flexural
moment in the pipe wall, the geometric stress is raised up. This increscent usually
occurs due to fabrication errors , edge deflection in joints and inconsistency between
ring stiffness of pipes [6]. SCF can be determine using table 1 for the worst condition
(larger value for this parameter) [2]. In the marine structures, deterioration starts with
initiation of a crack in a defect or unsymmetrical connection leading to stress
concentration. These cracks will propagation in the elastic materials by the local
plastic strains. Concept of damage accumulation is widely used in assessment of
fatigue and fracture mechanics methods.

3 Fatigue assessment
Fatigue assessment1 denotes a process where the fatigue demand on a structural
element or a connection detail is established and compared to the predicted fatigue
strength of that element. One way to categorize a fatigue assessment technique is to
say that it is based on a direct calculation of fatigue damage or expected fatigue life.
Three important methods of assessment are called the Simplified Method, the Spectral
Method and the Deterministic Method. Alternatively, an indirect fatigue assessment
may be performed by the Simplified Method, based on limiting a predicted
(probabilistically defined) stress range to be at or below a permissible stress range.
There are also assessment techniques that are based on time domain analysis methods
that are especially useful for structural systems that are subjected to non-linear
structural response or non-linear loading. Fatigue Demand is stated in terms of stress
ranges that are produced by the variable loads imposed on the structure. (A stress
range is the absolute sum of stress amplitudes on either side of a steady state mean
stress). The term variable load may be used in preference to cyclic load since the
latter may be taken to imply a uniform frequency content of the load, which may not
be the case. The fatigue inducing loads are the results of actions producing variable
load effects. Most commonly, for ocean based structures, the most influential actions
producing the higher magnitude variable loadings are waves and combinations of
waves with other variable actions such as ocean current, and equipment induced
variable loads. Since the loads being considered are variable with time, it is possible
that they could excite dynamic response in the structure; this in turn will amplify the
acting fatigue inducing stresses. The determination of fatigue demand should be
accomplished by an appropriate structural analysis. The level of sophistication
required in the analysis in terms of structural modeling and boundary conditions (such
as. soil-structure interaction or mooring system restraint), and the considered loads
and load combinations are typically specified in the individual Rules and Guides for
Classification of particular types of Mobile Units and offshore structures. When
considering fatigue inducing stress ranges, one also needs to consider the possible
4
influences of stress concentrations and how these modify the predicted values of the
acting stress (see fig. 3). The model used to analyze the structure may not adequately
account for local conditions that will modify the stress range near the location of the
structural detail subject to the fatigue assessment. In practice this issue is dealt with by
modifying the results of the stress analysis by the application of a stress concentration
factor. The selection of an appropriate geometric SCF may be obtained from
standard references, or by the performance of Finite Element Analysis that will
explicitly compute the geometric SCF [7]. Two often mentioned examples of
geometric SCFs are a circular hole in a flat plate structure, which nominally has the
effect of introducing an SCF of 3.0 at the location on the circle where the direction of
acting longitudinal membrane stress is tangent to the circular hole. The other example
is the case of a transverse ring stiffener on a tubular member where the SCF to be
applied to the tubes axial stress can be less than 1.0.


4 Conclusions
Whatever will differs this study from the other is using the spectral analyze instead of
the usual fatigue assessment. In statistical analysis of fatigue just using the minimum
and maximum of the design wave in an irregular wave the fatigue life is estimated.
Hence this metod is not accurate enough in fatigue assessments and the theorem of
reliability of fatigue surveying is often proposed [8]. This method is not able to
generate irregular waves through out an inverse process, whereas all the waves with
their own height and period are participant in creating cyclic stress in a structural
element. As the basic rules in fatigue assessment, the small stresses which cause the
tiny crack or damage in an element are added together making the final deterioration
in an structure , and this is the Palmergen-Miner rule which is directly used in this
study instead of any secondary relation which are suggested in the regulations or
similar papers.
As a conclusion, we may say that adopting the above process to estimate fatigue life
of a marine structure is accurate enough to be used in engineering projects of larger
scale and real modules of jacket structures.


References

[1] Shafeifar, M., Hamedi, A., (2002): "A spectral model for north part of Percian
Gulf using available data", Icopmas 2002,
[2] API "Recommended Practice 2A-WSD ( RP 2A-WSD )", Recommended Practice
for planning , Desigining and Constructing Fixed, Offshore Platforms-Working
Stress Design, 21
st
edition December 2000
[3] ABS, "Guide for the Fatigue Assessment of Offshore Structures", April 2003,
American Bureau of Shiping , Incorporated by Act of Legislature of the State of New
York 1862.
[4] Martinsson.J., (2004): Fatigue Assessment of Complex Welded Steel
Structure, Division of Lightweight Structures, Department Aeronautical and
Vehicle Engineering, Royal Institute of Technology, SE-100 44 Stockholm.
[5] Dal, K., (1997): "Innovative Tubular Connection", Department of Civil
Engineering, Monash University, Melbourne, 7
th
international offshore & Polar
Engineering conference, Honolulu,USA,1997.
5
[6] Jang C. D., Song, H.C. & Jo, Y.C., (2004): "Fatigue Life Assessment of Fillet
Welded Joint Considering Stress Concentration", Seoul National University, Korea,
Proceedings of OMAE 2004
23rd International Conference on Offshore Mechanics and Arctic Engineering, June
20-25, 2004, Vancouver, British Columbia, Canada.
[7]-Halfpenny, A., (1999): "A Frequency domain approach for fatigue life estimation
from Finite Element Analyses", nCode International Ltd., Sheffield UK, International
Conference on Damage Assessment of Structures (DAMAS 99) Dublin.
[8] Assakkaf, I.A. & Ayyub, B. M., (2004): "Reliability-Based Design for Fatigue
of Marine Structure"
University of Maryland , College Park, Proceedings of OMAE 2004, 23rd
International Conference on Offshore Mechanics and Arctic Engineering June 20-25,
2004, Vancouver, British Columbia, Canada.


Table 1 Estimation of Stress Concentration Factor (SCF)

Comments Equation Status Safety
Factor
1
3
1
T
e
SCF + =
Tube-A High T: member thickness
2 1
T T

e: Distance between midline
Tube 1 to Tube 2

(
(
(
(
(

|
|
.
|

\
|
+
+ =
5 . 1
1
2
1
1
1 6
1
T
T
T
e
SCF

Tube-B Low






6








7