Proceedings of The Fifteenth (2005) International Offshore and Polar Engineering Conference

Seoul, Korea, June 19−24, 2005

Copyright © 2005 by The International Society of Offshore and Polar Engineers
ISBN 1-880653-64-8 (Set); ISSN 1098-6189 (Set)

Flexible Riser Fatigue Design and Testing

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

Carlos Alberto Duarte de Lemos Murilo Augusto Vaz
PETROBRAS/CENPES/Subsea Technology UFRJ/COPPE/PENO
Rio de Janeiro, Brazil Rio de Janeiro, Brazil

ABSTRACT INTRODUCTION

The development and qualification of new large diameter flexible riser Since the beginning of the oil production in Campos Basin using
structures with higher strength to allow installation in deep waters is flexible pipes, in 1979 with the Early Production System of Enchova
considered of paramount importance to increase Petrobras oil & gas Leste phase I (Machado Filho, 1982), Petrobras has continuously tried
production. Hence the flexible pipe manufacturers are constantly to motivate the flexible pipes manufacturers to develop new products
motivated to develop new concepts and to extend their design and and extend the knowledge about this product.
construction capabilities.
The necessity to exploit fields in increasingly greater depths motivated
The development of new prototype pipe design requires a myriad of the development of new riser technologies (such as, steel catenary
studies and tests to certify their strength and fitting for purpose. Among risers or hybrid risers) and extended the flexible riser application.
them, the fatigue design and prototype test have increased in Related to the flexible risers, those efforts started at the end of the 80`s
importance due to the ever greater depths of the new fields. with several Technological Cooperation Agreements where the
traditional manufacturer Coflexip (nowadays Technip) was motivated
Once the design methodology depends on the manufacturer and the to extend the applicability of their products to deep water and other
fatigue tests required by the international standards consider only the manufacturers, like Wellstream, Furukawa, Pirelli and NKT Flexibles
design life, the limitations and conservatisms of the design (more recently) were invited to develop their pipes. A frustrated
methodologies are unknown by the operators as well as the real safety incursion as shareholder was also tried, with the bonded pipe
factors considered. manufacturer Pag`o`Flex

It is felt that there is a lack of test results that guarantee the applicability Since the beginning, one of the major concern was the pipe design and
of the actual fatigue design methodologies. Furthermore, more and the correlation with test results. This could be translated in serials of
better information about the flexible pipe structural design is needed. qualification tests that the manufacturers were asked to perform and the
Hence, a new fatigue design specification has been issued by Petrobras requirement of a third part to evaluate the results (normally a
including the directional influence of bi-modal seas and the regular classification society). At the same time, Petrobras was developing its
fatigue test was modified to include a second test step where the fatigue own tests to get more knowledge about the pipe behavior (Lemos, Cruz
life is pursuit. This new step tries to reach the theoretical pipe failure and Meniconi, 1991).
increasing the loads applied during the test until the pipe reaches the
accumulated damage equal to 1 (considering the Miner’s law). Nowadays with the installation of FPSOs (Floating Production Storage
and Offloading) in deep waters, one point in focus is the fatigue design
This paper presents a short description of the Working Programme of flexible risers. This topic is not completely normalized by the
carried out in Petrobras considering the flexible riser fatigue life design international standards and is still relying on internal procedures of the
methodology and testing performed in a sample of the Marlim Sul gas manufacturers. Due to this lack of standardization and knowledge of
pipeline, which is connected to the P-40 platform. The sample was the pipe response, real scale tests are being required in new
tested until the theoretical fatigue life was reached. developments.

Unfortunately, the standard tests have been limited to verify the riser
KEY WORDS: Flexible riser; fatigue; test; design methodology. for service life (usually 20 or 25 years) with no guarantee of the
calibration of design methodologies and the safety factors have been
correct or too conservative since the tests have not been destructive.

166
To fill this lack, the traditional service life fatigue test has been Fatigue Failure Modes
modified to include a second step where the test proceeds until the
theoretical failure is reached (or until the linear cumulated damage, for The fatigue analysis shall be based on the knowledge of the failures
one of the layers, reaches one by the Miners law), called fatigue life modes of each layer together with the pipe behavior. The most
test. important failure modes are:

This fatigue life test was performed by the first time for the 11” Gas • Fatigue associated with the wear of the tensile armors;
Export flexible riser of the P-40 Platform that is installed in 1080 m of • Pure fatigue of the tensile armors;
water depth in the Marlim Sul field in Campos Basin. The riser was • Fatigue associated with the wear of the pressure armors;
tested in COPPE’s (Federal University of Rio de Janeiro) recently built • Corrosion fatigue;
test bench (Fig. 1). • Fatigue of the polymeric layer; and
• Fatigue in the armors at the end fittings.

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

When an unbonded flexible riser is subjected to dynamic loads each
layer can move in relation to the others. This movement together with
the contact pressures between the layers creates the fatigue failure
mode associated with wear. The evaluation of this failure mode is
strongly based on wear models and tests. The reduction of the wire’s
sectional area shall be evaluated and considered in the stress analysis
(Out, 1989). This failure mode is more prone to become important in
risers that do not consider in their construction anti-wear layers
between metallic ones (old risers for shallow waters, for instance).

The current deepwater risers utilize anti-wear layers, and then the pure
fatigue is the failure mode to be evaluated. For the Petrobras operation
conditions, deepwater, mild internal pressure and catenary
configurations, this is the most important failure mode.

Fatigue associated with wear is the failure mode to be evaluated for the
pressure armors, but the stress concentration due to the profiling and in
contact areas (Fig. 2) shall be considered in the design (Zhang, Chen, et
al., 2003). Due to triaxial stress state that exists in this layer, the fatigue
design shall consider equivalent stresses using the Sines criteria
(Coflexip, 1989).

Fig. 1 – Test bench with pipe specimen (COPPE - Brazil) Wearing and stress
concentration points
Some fatigue design highlights and experimental work developed by
Petrobras, such as the fatigue life test, will be presented in this work.

FLEXIBLE RISERS FATIGUE DESIGN

Flexible pipes are complex structures, particularly from the fatigue Fig. 2 – Wearing and stress concentration points at the pressure armor
design point of view. There may be several fatigue failure mechanisms
that could become critical. For each flexible riser type or application a The corrosion can act as a crack initiation point, and then the material
variety of failure modes may take place, which may affect one or behavior when subjected to a corrosive medium has to be considered in
several layers. Furthermore, there isn’t a single fatigue curve for the the S-N curve.
whole pipe, but curves for the materials that compose each layer.
The polymeric layers seldom present problems due to fatigue, except
The fatigue procedure for flexible risers shall take into account the the internal pressure barrier that can exhibit problems inside the end
following steps: fitting due to pressure or temperature cycles (specially when PVDF is
used), but this is often identified by lose of tightness.
• Knowledge of the fatigue failure modes;
• Knowledge of annulus environment to enable a correct material The fatigue in the armors inside the end fitting is related to stress
selection and definition of fatigue curves; concentration in wire bends, used to improve the bonding to the epoxy,
• Global dynamic analysis (interaction with bend stiffener, soil and and marks in the wires due to the use of knifes to cut the plastic layers.
floatation if existent) with a procedure for taking into account
Metocean conditions and production unit response; Material Behavior Definition
• Local stress analysis (considering friction and wear);
• Fatigue evaluation using Haigh diagrams and linear Miner’s law; The selection of materials is of paramount importance for the fatigue
design. The correct specification of the internal conveyed fluid, not
• Tests.
only at the beginning of the field life but during all its life, together
with a careful integrity control, will be the input for a correct evaluation
Some points of these steps will be discussed in the following topics.

167
of the pipe annulus environment which the wires will be subject to. The which is a simulator for anchored FPSO, generating a wave scatter
effect of the annulus environment (that could evolve from sweet to sour diagram related to the ship, this heading distribution can be used in
during pipe’s service life) can be schematically seen in the S-N curves uncoupled time domain programs (Lemos, 2005). Fig. 4 shows two
shown at the Fig. 3 (Berge, Bendiksen at al., 2003). For a deepwater different distributions obtained using different environmental data
riser, the effect of the mean stress level shall be also taken into account. bases for Campos Basin.

Log ∆σ 0

22.5
25%
-22.5
In air
20%
45 -45

15%

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

67.5 10% -67.5

5%
Corrosive medium
90 0% -90

Log N 112.5 -112.5

5 6 7

Fig. 3 – S-N Schematic curves for corrosive and air environment

135 -135

Global Dynamics Analysis

157.5 -157.5
SIMO
Full datacompleto
base
The dynamic analysis to evaluate tensions and curvatures in flexible 180 Conjunto
Reduced dataresumido
base

risers can follow two different procedures:

Fig. 4 – Long term heading distribution using bi-direction sea
• Regular wave approach; or
• Irregular wave approach (stochastic analysis). The frequency to be considered for fatigue load cases can be chosen
using frequency screening procedures (as proposed by MCS, 2004) or
The regular wave approach is performed in a time domain program and using techniques of Response Based Design (Lemos, 2005) as ship
can consider an individual wave scatter diagram or maximum expected motion peaks can tremendously influence the riser response.
waves of a spectral scatter diagram.
The current profile can change the riser mean tension and final
Stochastic analyses are seldom performed for fatigue evaluation of configuration, it can be modeled considering the current profile that
flexible risers primarily due to interface limitations between the occur simultaneously to the waves, if simultaneous environmental data
programs used for global and local (or stress) analysis and is available, or in a simplified way considering the most expected
computational costs. Nevertheless, they are the best way to model the surface current and its associated profile. Normally, VIV is not a major
behavior of offshore structures subject to wave loads and can be concern for flexible risers, due to the small stress variation associated to
performed in the time or frequency domain. the small vibration amplitude. However, if the wear is the failure mode,
it can become important.
The time domain procedure is the best way to perform the analysis due
the possibility to model properly the geometric and hydrodynamic non- Stress Analysis
linearities. By the other side, frequency domain analysis allows a great
number of analyses in a project time frame. The stress analysis to evaluate fatigue loads using tension and
curvatures in flexible pipes shall consider the change in the riser
Several spectra can be fitted to field wave measurements, for the behavior due to internal and external pressures and temperatures. This
Campos Basin area that presents bi-modal and bi-direction seas, the stick slip behavior is considered with changes of the resultant curve
Ochi-Hubble Spectrum was tested but the best fit was found using the (loxodromic or geodesic) and in the bending stiffness.
JONSWAP modeled in two different directions (Seixas, 1997).
Fatigue Tests
The spectra data are gathered and presented in wave scatter diagrams,
which need besides the wave height and period present also its Due to the number of unknowns or simplifications performed during
direction and, if possible, simultaneously sea and swell data. Due to the the fatigue evaluation of flexible risers, prototype tests are still required
large number of load cases that this data can represent, it is a normal and constitute a very important tool to check the applicability of the
procedure to group the waves in classes. In this case caution should be riser to the real field conditions and to check the calibration of design
taken to correctly consider the wave heading related to the ship (for a tools and procedures.
FPSO with bow turret, for example, using only bow waves is normally
unconservative), and to choose the wave period in order to correctly With these goals, several tests were developed trying to reproduce the
account for the influence of frequency into the response of the riser. different fatigue modes and location along the riser subjected to fatigue
loads. The main tests are:
The heading distribution can be evaluated using coupled analysis
programs such as Dynasim, (Nishimoto, Fucatu and Masetti, 2001)

168
• Fatigue tests with tension and bending at the riser top connection,
(Fig. 5) and vertical (Fig. 6);
• Fatigue test at the sag bend, for wave or S configurations (Fig. 7);
• Rotational fatigue test (Fig. 8);
• Fatigue during laying operation (Fig. 9);
• Tension to tension fatigue test (to check the wires inside end
fittings)
• Pressure or temperature cycling test; (among others).

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

Fig. 8 – Rotational fatigue (NKT, 2001)

Fig. 5 – Fatigue test at the top connection, horizontal (Wellstream

International, 2005)

Fig. 9 – Fatigue during laying (USP – Brazil)

FATIGUE TEST

The new fatigue test procedure encompasses two steps, the first one
named service life fatigue test, aims to test the duration that the pipe
will be in service, and the second, the fatigue life test, seeks to reach
the pipe failure, trying to evaluate the actual fatigue safety factor. Once
the manufacturers design procedures considers high degree of
conservatisms the fatigue life test can be too long, so to reduce the test
duration the loads are increased and the theoretical fatigue life limit is
pursued (this means, until the linear cumulated damage, for one of the
layers, reaches one by the Miner’s law).

This test was performed by the first time for the 11” Gas Export
Fig. 6 – Fatigue test at the top connection, horizontal (C-FER)
flexible riser of the P-40 Platform that is installed in 1080 m of water
depth in the Marlim Sul field in Campos Basin. The service life step of
the test was performed at the C-Fer fatigue test bench (Fig. 6) and the
fatigue life step was performed at COPPE (Fig. 1).

Service Life Fatigue Test

The pipe sample was pressurized up its design pressure (3000 psi) and
the service life test program presented in the Table 1 was applied. By
the theoretical calculations no fatigue damage was expected during this
phase. After this test phase the pipe was non-destructively inspected
using gamma ray (Fig. 10). No wires failure was found and the test
proceeded to the second stage (samples of the gamma ray inspection
from Lemos and Costa, 2002, can be found at Fig. 11 for the pipe area
Fig. 7 – Fatigue test at the wave section (Technip, 1996) under the bend stiffener and at Fig. 12 for the end fitting).

169
 Table 1 – Test program – Service Life (C-FER)
Angular Variation
Load Tension (kN) Number of
(degrees)
Case. cycles
min max min max
1 1070 1168 3.1 4.5 1 000 000
2 1051 1185 2.6 5.0 396 000
3 1039 1199 2.1 5.6 75 400
4 955 1283 0.6 7.9 20 500
Fig. 12 – End fitting gamma ray film
Total 1 491 900
Fatigue Life Test

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

The fatigue life test was performed considering the load case A
presented in Table 2. This load case was applied to the pipe and no
damage was found using a gamma ray inspection.

Table 2 – Test program – fatigue life test (COPPE)

Load Number Angular
Case of cycles Tension (kN) Variation
(degrees)
Min. Max. Min. Max.
A 100 000 1614 2031 0.4 5.9
B 100 000 1614 2031 -7.0 6.0
Fig. 10 – Gamma ray test setup
Afterward this load case was repeated but at this time an acoustic
monitoring apparatus was used to help identifying failure during the
test. After the test the pipe did not again present any sign of failure but
the acoustic monitoring signaled areas of high energy. Those areas
were inspected by gamma ray but nothing was found.

By the third time, the test was repeated but using load case B from
Table 2 and the acoustic monitoring system was still used and again the
pipe did not present any failure despite the suspicious intensive signs
from the acoustic system (presented in Fig. 13).
Fig. 11 – Pipe gamma ray film

Areas found during acoustic monitoring with high

energy signal.

• Sensors
• Areas to be inspected during dissection

Fig. 13 – Sections selected for inspection due to acoustic monitoring

170
PIPE DISSECTION Fig. 15 – End fitting dissection

Despite of the test run out, the pipe was taken off the bench, visually
inspected (Fig. 14), and hydrostatically tested with no sign of failure.
The end fitting was cut off the pipe and dissected (Fig. 15) with no wire
broken and all holding devices in place. The bend stiffener was also
dissected (Fig. 16) and no debonded or cracks in critical areas were
found.

In the pipe dissection (Fig. 17), the tensile and pressure armors did not
show any failure signs and the same happened with the pressure barrier.
The holding bandage Kevlar tape layer showed moderate wear (Fig. 18)

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

at positions distant between 2.0 and 2.5 m from the connector Fig. 16 – Bend stiffener dissection
(positions indicated for inspection by the acoustic monitoring system
Fig. 13).

Following the pipe dissection the last layer, the carcass, showed 2 large
open cracks (Fig. 19) and a third one at the beginning and several light
wear signs. The rupture points were located at 2.0 and 2.5 m from the
connector (this later was positioned exactly at the point of the worst
wear at the holding bandage layer).

Fig. 17 – Pipe dissection – first layer

Fig. 14 – Pipe visual inspection

Fig. 18 – Holding bandage tape (Kevlar® layer)

Fig. 19 – Damaged carcass

171
FAILURE ANALYSIS gamma-ray inspected with no failure, the sample was then dissected.

The damaged carcass was analyzed based in the macroscopic and During the dissection no failure was found in the armors layer, where
microscopic crack characteristics (using optical and scanning electron the failure was theoretically foreseen, therefore indicating higher safety
microscopes). The Fig. 20 shows the typical aspect of the crack surface factors in the analysis for these layers. On the other hand carcass
present in the carcass, with several initiation points and in the Fig. 21 failures were found where they were not expected. It is worthwhile to
the metallografic aspect of the crack before its completely opening. The inform that the acoustic monitoring system used to identify the failure,
following points can be highlighted: actually indicated the failure points to be further inspected, but nothing
• The crack surface indicates fatigue with several initiation points was found in the gamma-ray inspection.
suggesting elevated stress concentration;
• The crack aspect showed that the crack initiated from the internal The failure analysis of the crack surface indicated a fatigue failure
side of the carcass at the region of high hardening; associated to corrosion under tension at the higher hardening point of

Downloaded from http://onepetro.org/ISOPEIOPEC/proceedings-pdf/ISOPE05/All-ISOPE05/1851592/isope-i-05-144.pdf by Repsol Services Company user on 18 March 2025

• The presence of secondary cracks with intergranular aspect, the carcass (inner radius of the bent steel), the corrosion was not
suggest the action of other mechanism during the cyclic completely explained since no high chlorides content were found in the
solicitation, e. g. fatigue associated to corrosion under tension. water used internally. The fatigue crack could be explained due the
completely extension of the carcass in so high tension and indicates that
Outer side not all failure modes are completed covered by theoretical calculations.

ACKNOWLEDGEMENTS
Inner side
The authors would like to thank PETROBRAS for providing the
Fig. 20 – Typical fatigue crack surface with several initiation points opportunity to publish this work.

REFERENCES

Berge, S., Bendiksen E., Gudme, J. and Clements R.; “Corrosion Fatigue
Testing of Flexible Riser Armour – Procedures for Testing and
Assessment of Design Criteria”, OMAE 37327, 22nd International
Conference on Offshore Mechanics and Arctic Engineering, 2003.
Coflexip, “Coflexip Flexible Pipe Design Description - Fatigue”,
Coflexip Report, 1989.
Coflexip, Coflexip - Test Facilities, Coflexip Report, 1996.
Lemos, C. A .D., Cruz, F. T. L. and Meniconi, L. C. M; “Avaliação da
Fig. 21 – Cracks not completely open
resistência mecânica de linhas flexíveis”; 3º Encontro Técnico
Interdepartamental sobre Explotação em Águas Profundas - ETIEAP,
CONCLUSIONS Rio de Janeiro, TT-40, p. 323-327, (in Portuguese), 1991.
Lemos, C. A. D. and Costa, L. C. S. “Inspeção do Gasoduto Flexível após
The necessity to exploit fields in increasingly greater water depths
Teste Dinâmico”, Petrobras Internal Report, CT TS 039/2002, (in
leads PETROBRAS to be always working together with the flexible
Portuguese), July de 2002.
pipe manufacturers to develop new structures to fulfill its needs.
Lemos, C. A. D.; “Análise de Fadiga em Risers Flexíveis”, D. Sc. Thesis,
Oceanic Engineering, COPPE/UFRJ, (in Portuguese), January 2005.
During those developments the pipe design and the correlation with
Machado Filho, Z. L.; “Sistemas de Produção Antecipada na Plataforma
qualification tests were the major concerns to prove the applicability of
Continental Brasileira”; 2° Congresso Brasileiro de Petróleo; Instituto
the new flexible pipe structures to the operation requirements. One
Brasileiro do Petróleo – IBP, (in Portuguese), 1982.
point not completed fulfilled and still needs further verification is the
MCS, “Real Life JIP - Design Basis”, Galway, Ireland, February 2004.
fatigue analysis.
Nishimoto, K., Fucatu, C.H. and Masetti, I.Q. 2001, “DYNASIM - A
Tine Domain Simulator of Anchored FPSO”, Proceedings of the 20th
With this in mind, the standard fatigue test, previously required with
International Conference on Offshore Mechanics and Arctic
the objective of service life verification, was modified to check the
Engineering, OMAE’01/OFT-1304.
fatigue procedure applied by the manufacturers. This new fatigue test
NKT Flexibles, NKT Design and test Capabilities, NKT Report, 2001
besides checking pipe service life it encompasses a second step to
Out, J. M. M., “On the Prediction of the Endurance Strength of Flexible
check its fatigue life, extending the test until one of its layers reaches
Pipe” OTC 6165, 21st Offshore Technology Conference, Houston,
the theoretical fatigue damage (e. g. damage equal to 1 by the Miner’s
1989
law). With this approach the design safety factors can be evaluated.
Seixas, N. A. B.; “Clima de Ondas na Bacia de Campos: Análise dos
Dados e Proposta de Parametrização”, D. Sc Thesis, Observatório
To verify the applicability of this new fatigue test procedure, the 11”
Nacional, Departamento de Geofísica, Brazil, October 1997
flexible pipe used as gas export riser in the Marlim Sul, P-40 platform
Wellstream International, Testing Statements, Wellstream International
in 1080 m of water depth was tested. The first phase, service life phase,
2005
was performed at the C-FER test rig (in Canada) and the second phase,
Zhang, Y.; Chen, B.; Qiu, L.; Hill, T.; and Case, M., “State of the Art
fatigue life phase, was performed at the COPPE test rig where the test
Analytical Tools Improve Optimization of Unbonded Flexible Pipes
was extended trying to reach the pipe failure.
for Deepwater Environments”, OTC 15169, Offshore Technology
Conference, Houston, 2003.
After the execution of 3 times the number of cycles expected to failure
the test was interrupted, the sample was hydrostatically tested and

172