A RATIONAL APPROACH FOR THE EVALUATION OF FATIGUE

STRENGTH OF FPSO STRUCTURES

Gabriel Tam1 and Jer-Fang Wu1

1
American Bureau of Shipping, U.S.A.

ABSTRACT

The ABS SafeHull system for Tankers is an integrated and rational approach for initial design and strength
evaluation and has been widely accepted by the industry since 1993. This paper describes the simplified
fatigue approach used in the SafeHull system and discusses the basis and application of the environmental
severity factors approach to modify the SafeHull/Tanker fatigue process for assessing the fatigue strength of
an FPSO, taking into consideration the site-specific environment and the fatigue damage accumulated during
trading history. The approach is applicable to both FPSO new builds and conversions. Additionally, the
paper describes the ABS standard spectral-based fatigue analysis procedure.

INTRODUCTION operations differ significantly from that of a

trading tanker and each FPSO has its unique site-
FPSO as an offshore hydrocarbon production specific wave environment and loading conditions
facility concept has not only well proven to be a that affect ultimately the fatigue performance of
viable option for marginal fields but is being the structure, a modified approach to the
considered increasingly for development of larger SafeHull/Tanker system is necessary. In this
fields in deeper waters. FPSOs continued success regard, ABS has developed the Environmental
is due mainly to its ability to stay on site for at Severity Factors (ESFs) method, applying the Beta
least a good part of the field’s life without costly (β) ESFs to adjust the dynamic loads and load
periodical dry-docking or shipyard visits. The effects and the Alpha (α) ESFs to adjust the
most important determinant for this success is the expected fatigue damage. This is the basis of the
integrity of the FPSO’s hull structure. Structural SafeHull/FPSO system.
fatigue is by far the most worrisome failure mode
affecting structural integrity and operation. While the above are the minimum requirements for
classification, there is a need to apply more
It is well understood that hull structural design extensive spectral-based fatigue analysis (SFA) to
standards for ship-type FPSOs are based on FPSO structures, particularly when an extra level
classification society rules for oil tankers. In more assurance is required. This paper describes the
recent years, some societies have made SafeHull/FPSO approach and the SFA method of
considerable progress in introducing “design fatigue analysis that ABS uses to assess structural
oriented” fatigue assessment methods (e.g. the integrity of FPSOs against fatigue damage.
permissible stress range approach) to improve
fatigue-resistance of hull structures. At the
American Bureau of Shipping (ABS), the FATIGUE ASSESSMENT IN ABS
SafeHull/Tanker simplified fatigue assessment SAFEHULL/TANKER
procedure employs a two-phase approach: a Phase
A fatigue screening of mainly the longitudinal The approach used to evaluate fatigue strength in
members and a Phase B fine mesh FEM based the ABS SafeHull system [1] is a simplified fatigue
fatigue assessment for other fatigue-critical method based on the Palmgren-Miner cumulative
structural details. The SafeHull system is based on damage theory applied in conjunction with S-N
unrestricted ocean service wave environmental data. The method of development entails the
condition and a 20-year service life. Since FPSOs following elements:
 VBM/ HBM External Internal Wave
1. S-N curve fatigue strength data, SF /SF Pressure Pressure Dir.
LC Max * * * Head
2. Detailed stress analysis to determine the stress
1 Sag
range probability distribution, LC Max * * * Head
3. Spectral analysis in calculating the final 2 Hog
fatigue damage. LC * * * Max Head
3
The simplified method used in the SafeHull LC * * Max * Head
4
considers the expected statistical distribution of
LC * * * Max Beam
stress range over time, the mathematical 5
description of the S-N curve, the applicability of LC * * Max * Beam
the Palmgren-Miner’s Rule, and a target value of 6
fatigue life. The assumption made in this method LC * Max * * Bow
is that the long-term stress range in the target life is 7 Qtr
represented by a Weibull distribution. Thus, if the LC * Max * * Bow
8 Qtr
extreme stress range in the target life is known or
can be calculated, the characterizing Weibull
When performing the Phase B structural analysis
distribution “shape” factor can be used with other
for the eight (8) load cases, the response for the
information to produce a controlling fatigue design
corresponding still-water condition is calculated
parameter. The parameter can be the allowable (or
separately to establish a Reference Point for
permissible) stress range, fatigue life in terms of
determining the cyclic wave-induced stress
time, or fatigue damage ratio. The target fatigue
response. For each load case, the fatigue inducing
life in the SafeHull is a 20-year nominal exposure
dynamic stresses are obtained by deducting the
in the North Atlantic wave environment.
static stress due to the still-water load from the
total stresses due to the total loads.
In the SafeHull Phase A fatigue assessment, Rule
load formulas that are calibrated for the wave
The cyclic fatigue induced stress ranges are
induced load components on a tanker structure are
determined by combining the stresses of the
used in conjunction with beam theory to obtain the
corresponding load case pairs as follows:
nominal stresses, and thus the stress ranges, for the
ship details considered. This calculation addresses
Stress Range, Pair 1 = [Dynamic Stress of LC 1] +
mainly the fatigue performance of longitudinal
[Dynamic Stress of LC 2]
members.
Stress Range, Pair 2 = [Dynamic Stress of LC 3] +
[Dynamic Stress of LC 4]
In the case of a complex stress field, i.e. a stress
Stress Range, Pair 3 = [Dynamic Stress of LC 5] +
field induced by more than one type of load
[Dynamic Stress of LC 6]
component, detailed stresses may need to be
Stress Range, Pair 4 = [Dynamic Stress of LC 7] +
determined by finite element analysis. The
[Dynamic Stress of LC 8]
SafeHull analysis process involves a strength
assessment of the initial design by detailed finite
For a given element of “hot spot” of a detail, the
element analysis using a series of standard load
largest stress range of the four pairs is the most
cases. A special procedure is given in Phase B for
probable extreme stress range and is used for
combining the various standard load cases and the
fatigue damage calculations in the simplified
detailed stress results for calculating the total stress
analysis approach.
ranges for use in connection with the simplified
fatigue assessment procedure. This is known as
the Phase B fatigue process.
ABS SAFEHULL FOR FPSO
There are a total of 12 load cases (LCs) in the
SafeHull/Tanker system [1], and only the first four As a classification organization, ABS requires that
(4) pairs of combined load cases (LCs 1 – 8) are the design and construction of ship type FPSOs are
relevant to fatigue. The table below shows the to be based on all applicable requirements of the
local load effect that is being maximized for each ABS “Rules for Building and Classing Steel
of the load cases. Vessels” (SVR) [1]. In order to address the design
and construction considerations that are unique to
FPSO, ABS has published a “Guide for Building
and Classing Floating Production Installations”
(FPI Guide) [2] in June 2000 with the updated
version expected in 2003.

ABS has extended its SafeHull/Tanker system,

taking due regard of the main differences between
tanker and FPSO operations, and developed the
SafeHull/FPSO system for scantling evaluation
and fatigue strength assessment of ship-type FPSO
design. The current ABS SafeHull/FPSO system
was developed based on both the ABS SVR and
FPI Guide.
Figure 1 ABS SafeHull/FPSO System
Unlike oil tankers, FPSOs are generally designed
to operate at a specific site throughout its service
life. In order to account for the stationary and ESFs of the Alpha (α) Type
site-specific characteristics, a series of
“Environmental Severity Factors” (ESFs) has been According to ABS FPI Guide, the α–ESFs for
introduced in the ABS FPI Guide. ship-type FPSOs are determined for six different
regions of the vessel structure, namely:
Two sets of ESFs are derived for the FPSO fatigue 1. Deck
design. The α–factors are used to adjust fatigue 2. Side Shell
strength performance expectations between 3. Longitudinal Bulkheads
unrestricted ocean going tankers and the long-term 4. Centerline Bulkheads
site-specific environment. The β–factors are used 5. Inner Bottom
primarily to adjust the dynamic component of 6. Bottom
loads that are used to assess the hull structural and
fatigue strength. These regions follow those defined for tanker hull
structures in the ABS SVR (Part 5, Chapter 1).
The site-specific environmental loads are
accounted for by the ABS Site-Specific For each specific region in a FPSO vessel, the ESF
Environment Assessment System (SEAS), which is factor α is defined as:
a computer software module that has been
integrated into the current ABS SafeHull/FPSO α = Ls/Lu (1)
system. Figure 1 shows the design analysis where Ls is the expected fatigue life at a specific
methodology of the ABS SafeHull/FPSO system site, and Lu is the expected fatigue life of an ocean
and the relationships between individual modules. going tanker that is subject to the unrestricted
A more detail description of the background and route.
development of the SEAS system can be found in
Zhao et al [3] and ABS internal document Therefore, a α of 1.0 corresponds to the
“Reference Manual for SEAS” [4]. unrestricted condition of an ocean going tanker.
Similarly, if the value of α greater than 1.0, it
indicates that the expected fatigue life of the vessel
at the intended site will be greater than the
unrestricted ocean going condition.

ESFs of the Beta (β) Type

Base Wave Phase A
Conditions As mentioned in the previous section, the β-ESFs
are used to account for the differences and factors
Design Wave SEAS Environment Severity
Conditions Factors

Phase-A Phase-A
Load

Phase-A
 between the design of an ocean going tanker and a introduction of new controlling limit states, such as
site-specific FPSO. The differences mainly come unacceptable deflections, vibrations, etc.
from the different design return periods for
environmental loads, the effects of mooring/riser Extending to Conversions
systems, different assumed wave energy spreading
characterization, and different basis of extreme In SafeHull/FPSO Phase A, the ABS SEAS system
design storm characterization. It is ABS’ design has been extended to account for the conversion
methodology that the β-ESFs are to be applied cases [4]. The remaining fatigue life of a
only to the dynamic portion of the load conversion is calculated by the following:
components. The load components that are
 Nr Ns 
considered “static” are not affected by the LR = α 3  LSHα1 − ∑ S ri / α 2ri − ∑ S sj / α 2 sj  (3)
introduction of the β-factors. There are 13  i =1 j =1 
Dynamic Load Parameters (DLPs) identified for
the ship-type FPSOs in the FPI Guide: where:
LR is the remaining fatigue life of the converted
1. Vertical Bending Moment vessel at the intended site,
2. Horizontal Bending Moment LSH is the fatigue life of the new build ocean going
3. External Pressure Starboard tanker,
4. External Pressure Port α3 is the α–factor for the intended site,
5. Vertical Acceleration α1 is a correction factor, currently set to be 1.0,
6. Transverse Acceleration α2ri, α2sj are the α–factors for the i-th historic route
7. Longitudinal Acceleration and j-th historic site, respectively.
8. Relative Vertical Motion at Forepeak Sri, Ssj are the number of service years during i-th
9. Wave Height historic route and during j-th historic site,
10. Pitch Motion respectively,
11. Roll Motion Nr, Ns are the number of historic routes and sites,
12. Vertical Shear Force respectively.
13. Horizontal Shear Force
It is clear that for the case of a new build FPSO,
For each DLP, the β-factor is defined as: both Sri and Ssj are set to zero and Equation (3)
returns to its original form of Equation (1).
β = max{Ls}/max{Lu} (2)
Employing similar procedure in the
where max{Ls} is the maximum of a DLP Ls of the SafeHull/FPSO Phase B fatigue process, it is
FPSO under the intended site wave condition with possible to calculate the cumulative fatigue
100-year return period, and max{Lu} is the damage due to past service history and thus predict
maximum of the same DLP Lu of the vessel under the remaining fatigue life applying the site-specific
unrestricted condition with 20-year return period. β factors for the loads that are used in the Phase B
FEM fatigue analysis.
It is noted in the ABS FPI Guide that the β factors
are subject to a lower limit to keep design
parameters from going “too low”. The limits are
introduced in two ways: SPECTRAL FATIGUE ANALYSIS FOR FPSO

1. The β values taken is not to be less than 0.5, The fatigue assessment method used in the
and SafeHull/FPSO system is a design-oriented,
2. The result of an application of a β factor, such permissible fatigue stress range approach.
as the calculated required scantling, is not to Through both the Phase A fatigue by Rule load
be less than 85% of the unrestricted service formulas and the Phase B fatigue by a detail finite
element stress analysis, it readily covers a large
The reasons for introducing these limits are to: portion of the fatigue critical structural details of
reflect successful experience, a desire not to an FPSO hull structure in so far as the cargo region
inadvertently create a reordering of the dominant is concerned. However, the FPI Guide does not
structural failure modes, and to avoid the preclude the imposition of requirements by ABS to
demonstrate the adequacy of fatigue strength of
structural components by additional fatigue 2. FEM Structural analysis for stress range
analysis. There is also a common need to perform transfer functions
spectral based fatigue analysis of FPSO hull 3. Calculate fatigue damage for a selected base
structure details, which are beyond the range of vessel load case using spectral method
applicability of the SafeHull permissible fatigue 4. Calculate total (or combined) fatigue damage
stress range approach. The most notable of these based on the multiple base vessel load cases
areas are, as applicable, the hull region that
supports the mooring turret, critical intersections Seakeeping Analysis
between the deck mounted equipment supports and
the hull’s deck structure, and the connections of In the context of spectral based fatigue analysis,
“FPSO specific structural details” that require a the main objective of the seakeeping analysis is to
higher fatigue damage factor because of determine the motion and load Response
“criticality” and “inspectability” considerations for Amplitude Operators (RAOs) of the FPSO for a
some areas of the structure. range of wave frequencies and headings, as well as
for each of the base vessel load cases identified.
The overall process of ABS standard spectral The motions are the six-degree-of-freedom rigid
fatigue analysis is published in [5] and outlined in body vessel motions, and the wave load RAOs
Figure 2, and it contains the following three (3) include the hull girder loads, external wave
key elements: pressures, internal tank pressures due to
accelerations, inertial forces on the masses of
1. Determine fatigue demand structural components and significant items of
2. Establish fatigue strength or capacity equipment, and mooring loads. In performing such
3. Calculate fatigue damage or fatigue life analysis, an appropriate seakeeping program that is
capable of modeling the FPSOs mooring system
Before embarking on the analysis, it is important to and including the shallow water effect, if
first obtain the environmental conditions of the appropriate, is to be used.
intended FPSO site. The ocean wave data that is
used in the spectral based fatigue analysis are to be For a single point moored FPSO, a minimum of 3
in a “wave scatter” diagram format. The wave data headings, including he head-on and 30 degree off
consists of a number of “cells” that represent the either side of the head-on, will be considered
probability of occurrence of specific sea states. sufficient, when justified by the submitted
Another key information required for the spectral environmental data.
fatigue analysis is the base vessel load cases. The
number of base vessel load cases is to be selected FEM Analysis for Stress Range Transfer
based on the FPSO’s operational profile or its Functions
loading manual. As a minimum, two cases are to
be modeled and used in the spectral fatigue For each heading angle and wave frequency at
analysis and they are the ones resulting from and which the FEM structural analysis is performed,
representing, the probable deepest and shallowest two load cases corresponding to the real and
drafts, respectively that the FPSO is expected to be imaginary parts of the frequency regime wave
operate during its on-site service life. induced load components are to be analysed.
Then, for each heading angle and wave frequency,
Where applicable, the fatigue demand arising from the wave induced cyclic stress range transfer
the planned FPSO transit case (usually only the function is obtained for a based vessel loading
voyage to the installation site) is to be determined condition.
separately and added to the total fatigue demand
calculation. The FEM structural analysis process in this task
involves a 3-D global model analysis of the entire
The process in determining the fatigue demand for length of the FPSO and fine mesh local FEM
the selected structural locations involves a series of model analyses of the selected structural details.
computationally intensive analyses, including the The global structural model and load modeling
following: should be as detailed and complete as possible.
The global model is analysed to yield nodal
1. Direct seakeeping analysis of vessel motions displacements, which are then used as boundary
and wave loads
conditions for the stress calculations of the fine damage resulting from each considered sea state.
mesh local structural details. The cumulative fatigue damage resulting from
combining the damage from each of the short-term
Fine mesh 3-D FEM models are developed for conditions is determined based on the Palmgren-
selected critical joints and details within the hull Miner’s rule.
girder and the interface structures to the mooring
turret and topside process modules. Very fine Total Fatigue Damage
mesh elements with size equal to the plate
thickness are required to be used in order to Cumulative fatigue damage is separately calculated
determine the stresses at the regions of stress for each of the vessel loading conditions identified,
concentration, and hence the site of the potential then the total or combined fatigue life is calculated
fatigue problems. Selection of such critical as a weighted average of the reciprocals of the
locations should be based on experience, detailed lives resulting from considering each vessel load
strength analysis, and/or quantitative fatigue case separately.
screening process.
SUMMARY
Subsequent to the FEM stress analyses above, the
stress range transfer function for each fatigue The ABS SafeHull fatigue assessment procedures
critical detail under examination is constructed by by using Phase A Rule load formulas and also the
post-processing of the stress results based on the Phase B FEM analysis have been presented. The
stress value corresponding to the maximum wave extension of this analysis method to the site-
load identified, and the corresponding shape of the specific FPSOs is achieved through a process of
transfer function determined in the seakeeping environmental assessment from which the ESFs
analysis. are calculated and used for modifying the dynamic
loads and the expected fatigue strength. The
Spectral Analysis spectral-based fatigue analysis procedure provides
a methodology to address details that are beyond
The spectral based fatigue analysis begins after the the range of applicability of the SafeHull fatigue
stress transfer functions described above have been approach or for the purpose of meeting a higher
calculated. The method employed in ABS standard fatigue damage factor because of the criticality and
SFA is based on “short term closed form” inspectability considerations for some specific
integration. Wave data are represented in terms of locations of the structure.
a wave scatter diagram. The wave scatter diagram
consists of a number of sea states, which are short- REFERENCES
term descriptions of the sea in terms of joint
probability of occurrence of wave height and wave 1. American Bureau of Shipping, 2002. Rules for
period. For each combinations of significant wave Building and Classing Steel Vessels.
height and wave period the spectral density of 2. American Bureau of Shipping, June 2000.
stress range is computed by integrating the product Guide for Building and Classing Floating
of wave engery spectrum and the modulus of the Production Installations.
stress range spectrum. This yields the variance of 3. Zhao, Chuntian, Wu, Jer-Fang, Huang, Ken
stress range for the “short term” sea state. and Shin, Yung, May 2002. Environmental
Assuming the underlying process is Gaussian, the Severity factors and Their Application in FPSO
stress range distribution can be shown to be Hull Strength Assessment, Offshore Technology
Rayleigh-distributed. This procedure is repeated to Conference, OTC14231.
yield a series of “short term” responses. These are 4. American Bureau of Shipping, June 2002.
combined to yield a total stress range distribution Reference Manual for SEAS (Internal).
taking due account of the lifetime number of cycles 5. American Bureau of Shipping, January 2002.
for each combination of wave height and period. Guidance Notes on Spectral-based Fatigue
ABS uses the closed form spectral integration Analysis for Floating Production, Storage and
method, in conjunction with the appropriate S-N Offloading (FPSO) Systems.
curve for the location, to establish the fatigue

Establish Fatigue
Demand

Obtain and Verify Site

Environmental Data
Figure 2 SFA Procedure for FPSO