You are on page 1of 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/327873227

Background and New Revision of DNVGL-RP-F108

Conference Paper · June 2018


DOI: 10.1115/OMAE2018-78652

CITATIONS READS

0 2,546

4 authors, including:

Steinar Bjerke Jens Tronskar


DNV GL DNV Singapore Pte.Ltd.
1 PUBLICATION 0 CITATIONS 58 PUBLICATIONS 194 CITATIONS

SEE PROFILE SEE PROFILE

Steven Chong
Det Norske Veritas
10 PUBLICATIONS 44 CITATIONS

SEE PROFILE

All content following this page was uploaded by Steinar Bjerke on 14 September 2020.

The user has requested enhancement of the downloaded file.


Proceedings of the ASME 2018 37th International Conference on Ocean, Offshore and Arctic Engineering
OMAE2018
June 17-22, 2018, Madrid, Spain

OMAE2018-78652

BACKGROUND AND NEW REVISION OF DNVGL-RP-F108


Steinar Lindberg Bjerke Jens P. Tronskar
DNV GL DNV GL
Hovik, Norway Singapore

Steven Chong Asle Venås


DNV GL DNV GL
Singapore Hovik, Norway

ABSTRACT feature of the new RP is the guidance on sour service testing


DNV-RP-F108 [1] was first issued in 2006. The and assessments included in the Appendix C of the document to
Recommended Practice was developed to provide guidance on support pipeline/riser ECAs to develop flaw acceptance criteria
testing and analyses for fracture control of pipeline girth welds for NDT.
subjected to cyclic plastic deformation, e.g. during installation
by the reeling method, but also for other situations where INTRODUCTION
pipelines may be subjected to large plastic strains. The DNV-RP-F108 was first time issued in 2006. The
Recommended Practice was based upon a Project Guideline Recommended Practice was developed to provide guidance
developed within the Joint Industry Project "Fracture Control regarding testing and analyses for fracture control of pipeline
for Installation Methods Introducing Cyclic Plastic Strain - girth welds subjected to cyclic plastic deformation, e.g. during
Development of Guidelines for Reeling of Pipelines". installation by the reeling method, but also for other situations
The new revision is based on the extensive experience and where pipelines may be subjected to large plastic strains. The
knowledge gained over the years use of the previous versions, Recommended Practice was based upon a Project Guideline
as well as new knowledge from recent R&D projects. developed within the Joint Industry Project "Fracture Control
The main content of Appendix A of DNV-OS-F101 (now for Installation Methods Introducing Cyclic Plastic Strain -
DNVGL-ST-F101) [2] have been transferred to DNVGL-RP- Development of Guidelines for Reeling of Pipelines".
F108. Only the requirements relative to ECA and testing have However, when Appendix A of DNV-OS-F101 was introduced
been retained in DNVGL-ST-F101 [2]. in 2007, the fracture mechanics assessment approach in DNV-
The new revision has got a new number and new title, i.e. RP-F108 was superseded and only the procedure for
DNVGL-RP-F108, “Assessment of Flaws in Pipeline and Riser performing fracture toughness testing using SENT specimens
Girth Welds”. was still of value.
This paper lists the fundamental changes made in the new The new DNVGL-RP-F108 is based on extensive
RP from the old Appendix A of the previous DNV-OS-F101 experience and knowledge gained over 12 years of application
and discusses some of the changes, although within this paper of previous versions with both DNV-RP-F108 and later
it is not possible to cover all changes. The focus is on Appendix A of DNV-OS-F101, as well as new knowledge
clarification of use of S-N versus the fracture mechanics gained from recent R&D projects.
approach for fatigue life computation, classification of fatigue Appendix A of DNV-OS-F101 (now DNVGL-ST-F101
sensitive welds, calculations of more accurate crack driving [2]) have been modified, improved and transferred to DNVGL-
force by re-introduction of the plate solution, for which a new RP-F108 and only the requirements relative to ECA and testing
Lr,max (plastic collapse) calculation and a modified way to have been retained in DNVGL-ST-F101 [2]. The recommended
account for residual stresses have been specified. The RP procedure for performing fracture toughness testing on SENT
presents new assessment procedures pertaining to use of finite specimens have been removed because this is now specified in
element analyses for fracture mechanics assessments. A unique BS 8571.

1 Copyright © 20xx by ASME


The new revision has got a new number and new title, i.e. Due to the potentially high bending strains that may occur
DNVGL-RP-F108, “Assessment of Flaws in Pipeline and Riser for pipelines during installation or in operation, fracture
Girth Welds”. “Riser” has been included in the title to clarify assessment is an important limit state. The applied fracture
that the approaches are applicable also for girth welds in risers. assessment must also consider crack growth from dynamic
This paper summarizes some of the main changes to the loads, the biaxial stress state that exists in a pipeline during
recommended practice and presents new assessment procedures operation and a certain degree of displacement-controlled
pertaining to use of finite element analyses for fracture conditions. Figure 1, shows different pipeline installation
mechanics assessments. The paper also presents the new methods and HPHT lines subjected to lateral buckling, with
guidelines included in Appendix C covering testing and typical axial/bending strain levels indicated.
assessments for sour service, where fracture toughness may be
adversely affected and the fatigue crack growth rate is
substantially increased compared to in air.

NOMENCLATURE
εUTS: Engineering strain at UTS (uniform elongation)
εYS: Engineering strain at YS (uniform elongation)

CDFWRS: Crack Driving Force due to weld residual


stresses
CDF: Crack Driving Force
CTODmat: Material CTOD fracture toughness
Jmat: Material J-integral fracture toughness
σ: Longitudinal stress
Rp0.2: Yield strength at 0.2% plastic deformation
Rm (or UTS): The ultimate tensile strength representing the
Figure 1: Typical strain levels for different pipeline
material tested (same material condition and
installation methods and lateral buckling in operation.
test temperature).
ReL Lower yield plateau stress
The fracture assessment methodologies described in
YS: The yield stress (Rt0.5 or Rp0.2)
DNVGL-RP-F108 [1] are based on fracture mechanics
Qm: Membrane component of transverse weld
principles and evaluate the criticality of circumferential flaws
residual stress
in metallic pipeline and riser systems. The assessment
tnom: Nominal wall thickness
procedures specified in this RP, except for the FE fracture
tfab: Fabrication tolerances on wall thickness
mechanics assessments, have been used successfully for many
tcorr: Corrosion allowance
years and are based on the Option 2 procedure in BS7910:2013
tc: Characteristic wall thickness. tc=tnom-tfab-tcorr
[3] with tearing (equivalent to Level 3B in BS 7910:2005).
(operation) or tc=tnom-tfab (installation)
It is known that the crack driving force (CDF) for girth
welds is inaccurately calculated (typically under-estimated) in
DNVGL-ST-F101 AND DNV GL-RP-F108
many cases if the strain-based assessment approach using the
The structural utilization of the pipelines is high, with
reference stress solution for cylinders specified in BS 7910 is
nominal circumferential stresses of up to 90% of the yield
applied without implementing weld residual stresses
stress. The utilization in the axial direction is often even higher.
conservatively. Another way would be to assess the crack
Primarily this is due to nominal bending strains during reeling
driving force more accurately (higher) and then to include
installation (approx. 2%) and during operation (typically
residual stresses less conservatively. Such an alternative has
limited to 1%).
been suggested in the new RP.
DNVGL-ST-F101[2] provides firm criteria for the service,
ultimate, accidental and fatigue limit states, but not for fracture
FUNDAMENTAL CHANGES TO RP
limit state.
The fundamental changes from Appendix A of the previous
However, DNVGL-ST-F101 [2] requires fracture
DNV-OS-F101, and which some will be discussed in the
assessment to be performed to establish weld repair criteria
following include:
when strains exceed 0.4% during pipeline installation or
1. Combined fatigue and fracture limit state presented (not
operation, but does not state how this should be performed.
one S-N fatigue and one fracture mechanics which also
DNVGL-ST-F101[2] refers to the new DNVGL-RP-F108 for a
considers fatigue).
detailed recommended procedure for how to perform such
2. Clarification of the S-N approach vs. the fracture
assessments.
mechanics approach

2 Copyright © 20xx by ASME


3. New weld type classification “fatigue sensitive welds”
4. R-curves accounting for blunting introduced for situations
without compressive plastic strain
5. More guidance for how to combine loads (not necessary to
include “non-fatigue-sensitive welds” in combined fatigue
and fracture analyses)
6. Simplifications related to “end-of-life” fracture check
(critical flaw size in operation). May reduce testing and
simplify analyses.
7. More guidance on sour service testing and assessments
(Appendix C).
8. Comprehensive comparisons between FEM fracture
mechanics analyses and the FAD approach as specified in
the RP has shown that the CDF is considerably
underestimated considering high plastic strain if the
cylinder reference stress solution is used and that much Figure 2: Illustration of fatigue crack growth and the
better agreement is obtained if the plate reference stress relationship between the S-N and FCG assessment
solution is used. Hence, both the plate solution and the approaches.
cylinder solution are included in the new RP, but this also
implies that different approaches for Lr,max (plastic Special consideration should be given to girth welds
collapse) and residual stresses have been specified. subjected to low-cycle high strain fatigue and special
9. Stricter AUT acceptance criteria specified for “fatigue- consideration should be given to girth welds susceptible to in
sensitive welds”, is included in DNVGL-ST-F101 and in sour service or other environments that may degrade the
this RP. materials properties.
10. It is clarified that ECA is required for C-Mn steels in sour The loading may vary significantly for different girth
service and mode guidance and recommendation for how welds in a pipeline or riser and it is acceptable to divide the
to perform the required material testing under such pipeline or riser into different zones with different weld
conditions are provided. quality requirements and flaw acceptance criteria. If a
pipeline or riser is subjected to longitudinal strain exceeding
As it is not possible to cover all changes listed above in 0.4% and ECA is required per DNVGL-ST-F101. However, if
this paper, focus is on the changes mentioned in bullet points an ECA is required this may also introduce the need for
3, and 6-10. assessing dynamic stresses and loads below 0.4% strain. As
the direct input to the fracture mechanics analyses described
FATIGUE AND FRACTURE LIMIT STATE in the RP is always stresses and not strains, it is necessary to
S-N and fracture mechanics based fatigue assessments distinguish between “strain-based” and “stress-based”
are different approaches with somewhat different loading. In simplicity, “strain-based” loading may be
applications. Fatigue analyses based on S-N curves are most considered as situations where the stress level is above the
applicable when evaluating fatigue initiation from weld toes proportional yield stress, e.g. applied stress level dependent
or other geometrical stress concentrations that are included in of the stress-strain curve, while “stress-based” may be
the actual S-N curve, i.e. certain quality requirements are considered as situations where the applied stress is below the
necessary to meet the dedicated S-N curve. Typically, the S-N proportional yield stress, e.g. applied stress independent of
curves used for pipeline and riser girth welds are applicable the stress-strain curve. If an ECA is performed, all loads that
for welds without significant surface breaking flaws. may contribute to crack growth shall normally be considered.
However, in fracture mechanics based fatigue crack growth The following loads are relevant for ECA:
assessments, significant surface breaking flaws are typically  Maximum applied longitudinal stress or strain
assumed. Hence, the fatigue lives assessed by fatigue crack during installation (reel lay, J-lay or S-lay)
growth analyses are normally significantly shorter than the  Dynamic loads when the pipeline or riser is
fatigue lives calculated using the S-N approach. This means unsupported between the installation vessel and
that it is necessary to make rules for how and when these two touch-down. Maximum longitudinal stress or
different approaches should be used. The differences between strain at the sagbend. Dynamic loads in the
the two approaches are illustrated in Figure 2. temporary phases before production start-up (free
spans, etc.).
 Dynamic loads in the operational phase (free
spans, lateral buckles/expansion loops, etc.).
 Maximum longitudinal stress in operation.

3 Copyright © 20xx by ASME


In DNVGL-RP-F108 both the plate solution and the
Only stresses in tension need to be considered in cylinder solution are included, but different approaches for
fracture mechanics analyses assessing ductile tearing or estimation of Lr,max (plastic collapse) and residual stresses
unstable fracture while the total stress range, including the have been specified. It is known that the reference stress
compressive part, is normally required in fatigue crack solution for cylinders (Kastner), combined with applied stress
growth analyses. extracted from the engineering stress-strain curve typically
results in a too low crack driving force. This is illustrated in
ASSESSMENT CATEGORIES Figure 3 for a 12” OD, 15.9 mm wall thickness pipe with an
It is recommended to evaluate the limit state and decide external surface breaking circumferentially oriented flaw of 2
on the assessment approach by performing the following mm height and 25 mm length.
steps:
0,20
1) Determine the maximum longitudinal strain levels εl,nom Comparison 3D Abaqus and DNVGL-RP-F108

as described in DNVGL-ST-F101 [2]. 0,18 3D ABAQUS, 2x25

2) Classify the girth welds as either fatigue-sensitive welds 0,16


RP-F108 Plate, 2x25
RP-F108 cylinder, 2x25mm
or non-fatigue-sensitive welds. 0,14

Crack driving force, CTOD [mm]


3) Divide the pipeline or riser into different zones as
applicable, with different maximum longitudinal strain 0,12

levels and, if welds are classified as fatigue-sensitive or 0,10

non-fatigue-sensitive. 0,08
4) Determine the assessment category from Table A1.
0,06
5) Determine the weld quality requirements (NDT
acceptance criteria). 0,04

0,02

Assessment categories I to IV are not applicable for sour 0,00


service or other conditions where the fatigue crack growth rate 0,0 0,5 1,0 1,5
Applied strain, %
2,0 2,5 3,0

is likely to increase or the fracture toughness is likely to be Figure 3: Crack Driving Force computed by 3D FEA
degraded due to the surrounding environment. compared with the FAD approach using the plate and the
If the welds have lower or partially lower tensile properties cylinder reference stress solutions.
than the parent material the fracture mechanics approaches
described in Sec.4 and Sec.5 of DNVGL RP-F108 [1] may not It is seen from Figure 3 that the plate solution and the
be safe because the crack driving force may be underestimated. 3D FEA gives corresponding results, whereas the cylinder
For such situations, FE based fracture mechanics analyses are solution underestimates the CDF for applied strains above
recommended performed to account for strength mismatching. 0.5%.
Assessment category V “Environmental” is recommended Hence, it is necessary to define residual stresses
for welds where the environmental conditions are likely to conservatively if the cylinder solution is used. If the reference
adversely influence the fracture toughness properties or fatigue stress solution for plates is used, it is seen that the calculated
resistance. Guidance for testing and assessments are provided CDF is more correct and the influence of residual stresses can
in Appendix C of the new recommended practice. be reduced. In Figure 4, the total crack driving force including
both the applied strain and the weld residual stresses for the
CRACK DRIVING FORCE AND RESIDUAL STRESS cylinder reference stress solution approach is shown.
Comprehensive comparisons between FEM fracture It is observed from Figure 4 that if a conservative
mechanics analyses and the FAD approach as specified in the assumption about weld residual stress as a uniform residual
RP have shown that the CDF is considerably underestimated stress Qm = YS, where YS is the pipe longitudinal yield stress is
if the cylinder reference stress solution is used and that much made and included for the cylinder case, good correspondence
better agreement is obtained if the plate reference stress is achieved with the 3D FEA results for the CDF when adding
solution is used. This has been known for quite long time and a strain ε = YS/E to the applied strain for the FEA. Because, the
in 2012 it was decided to specify the plate solution instead of plate reference stress solution has proven to give a crack
the cylinder solution in DNV-OS-F101, Appendix A. driving force more in agreement with FE fracture mechanics
However, by a mistake no changes were made to the methods analyses at high applied strain, DNVGL-RP-F108 allows for
for how to include residual stresses and the result was that the including weld residual stress as an increase in the applied
“total” CDF considering both the applied strain and residual strain with YS/E similar as for FE fracture mechanics analyses.
stress increased significantly compared to earlier. The
approach was rejected and in the 2013 version of DNV-OS-
F101 Appendix A, the plate reference stress solution was
again omitted.

4 Copyright © 20xx by ASME


0,6 Crack driving force with and without residual stress, RP-F108 cylinder vs. Abaqus specified as 0.5tc (no testing for operation required, analyses
RP-F108 cylinder, 3x25mm simplified).
RP-F108 cylinder, Qm=YS, 3x25mm
0,5
3D FE analyses, 3x25mm external surface flaw
Lr,max
Crack driving force, CTOD [mm]

3D FE incl. weld residual stress (YS/E)

0,4 Pipe OD = 323,8mm, WT = 15,9mm The failure assessment diagram (FAD) should not be
extended to arbitrarily large plastic deformations and a cut-off
0,3 limit (referred as Lr cut-off or Lr,max) for the Lr (Lr = σref /YS)
axis is recommended. Depending on the reference stress
0,2 solution used, the following Lr,max values are recommended:
1) Reference stress solution for surface breaking flaw in
0,1 cylinder: it is acceptable for strain-based assessments to
define the Lr,max value as the engineering UTS/YS ration of
0,0 the parent pipe representing the characteristic high stress-
0,0 % 0,5 % 1,0 % 1,5 % 2,0 % 2,5 % 3,0 % 3,5 % 4,0 %
Applied strain strain curve for the relevant material condition. The Lr,max
Figure 4: Crack Driving Force computed by 3D FEA and value should not exceed 1.5 for C-Mn materials.
per the cylinder reference stress solution with and without 2) Reference stress solution for surface flaw in plate: the true
weld residual stresses. values of YS and UTS are used instead.

“END-OF-LIFE” FRACTURE CHECK YS may be defined as Rt0.5 or Rp0.2. If the stress-strain


A small flaw that starts to grow by fatigue will grow very curve has a yield plateau, the yield stress should be defined as
fast towards the “end of life” and the difference in time for this ReL, see Figure 6.
flaw to grow through thickness or for example halfway through
the thickness, is quite insignificant, as illustrated in Figure 5.
Unstable fracture is in such cases not important as long as the
maximum stress is not far above the yield stress, the material is
not brittle or a combination of both, i.e. the critical flaw height
is not very small.
Previously it was required to calculate and test the critical
crack size at “end of life” even if there was no plastic
deformation. However, it was realized that very little error is
made if an allowable height is set to 50% of the wall thickness.

Figure 6: Illustration of the tensile stress-strain curves and


definitions.

For strain-based analyses where the reference stress


formula for cylinder is used Lr,max is defined as:

(1)

Figure 5: Schematic illustration of a growing fatigue crack


If reference stress formula for plate is used for strain-based
where towards the ”end of life” growth is accelerated.
analyses:
Hence, for fatigue-sensitive welds in pipelines that have
(2)
only elastic utilization it is not necessary to determine the
critical flaw size at “end-of-life” for the following situations:
Longitudinal stress below 0.5YS and embrittlement not
likely. Critical flaw height may be specified as 0.75tc (no FRACTURE TOUGHNESS
testing for operation required, analyses simplified) The fracture toughness may be expressed in terms of the
Longitudinal stress between 0.5YS and YS and where Jmat value or the crack tip opening displacement (CTODmat)
embrittlement is not likely. Critical flaw height may be value. The procedure for testing and calculating CTOD and J as

5 Copyright © 20xx by ASME


relevant should be in accordance with BS 8571 [4]. It is in toughness test specimens as specified in BS 8571. The
general not recommended to convert CTOD values to J or vice microstructures in the root and hot-pass of CRA girth welds are
versa. If J is calculated from fracture toughness testing, J examples where standard testing is difficult to execute. One
should also be used in the FAD assessment, and vice versa. possible way may be to test smaller specimens with notching as
However, if it for some reason is necessary to convert Jmat illustrated in Figure 7.
values to CTODmat or CTODmat values to Jmat it is acceptable to
use:

(3)
Figure 7: Suggested method for fracture toughness testing
of microstructure in the root of CRA girth welds in lined or
If Jmat is converted from CTODmat values, low Rp0.2 and Rm clad pipes.
values should be used to give conservative results. If CTODmat
values are converted from Jmat, high Rp0.2 and Rm values should ASSESSMENTS FOR SOUR SERVICE
be used to ensure sufficiently conservative results.
Introduction
RESIDUAL STRESS
Stress-based loading will normally not be evaluated by FE Exploration and production of oil and gas resources is
fracture mechanics analyses unless the material has low moving into deeper water as most of the “easy-oil” fields have
fracture toughness properties or large detected flaws need to be been depleted. Such wells pose large technical challenges as
assessed. While a procedure for implementing weld residual they tend to produce fluids of high pressure (HP) and high
stresses for strain-based loading has existed, no such procedure temperature (HT) while also containing different mixtures of
has been given if FE fracture mechanics analyses are performed H2S, H2O and CO2. Sour environments are known to have a
for stress-based loading. Per the new DNVGL-RP-F108, a detrimental effect on the fracture and fatigue performance of
procedure for including weld residual stresses in FE fracture welded C-Mn line pipe steels. However, methodologies to
mechanics analyses considering stress-based loading have been evaluate fracture toughness and fatigue crack growth rate
suggested as an increase in the crack driving force (CDF) as parameters in sour service are not well established. Therefore
follows: in DNVGL-RP-F108 [1], a separate appendix i.e. Appendix C
has been dedicated to address these technical concerns by
(4) providing sour service testing guidelines for the fracture and
fatigue limit state. The fracture toughness (FT) and fatigue
crack growth rate (FCGR) testing of welded C-Mn line pipe
If the girth welds have been subjected to plastic strain, the steel in sour environments will ensure reliable and conservative
increase in the CDF due to residual stresses in subsequent material properties are developed for ECA calculations.
stress-based load cases may be calculated as follows:
Sour service testing is the subject of ongoing research in
(5) the oil and gas industry. It is not the intention of this Appendix
to provide a comprehensive review of all sour service test
methods that exist or that are under development, rather to
Here, the CDF is the crack driving force assessed by FE provide helpful baseline information considering the current
fracture mechanics analyses without considering state of knowledge in this area, to support the design and
residual stresses and CDFWRS and CDFRS are the crack installation of welded C-Mn steel pipelines and risers that will
driving force including weld residual stresses (WRS) be exposed to sour service conditions. It is anticipated that this
and residual stresses after plastic deformation (RS) Appendix will be updated and expanded in future editions of
respectively. , is the applied longitudinal stress and is DNVGL-RP-F108 [1] as more information becomes available
always below YS. These formulas are simply based on and more consensus on test procedures is reached in the
curve fitting to the results obtained using a traditional industry.
FAD based approach for stress-based cases and the In the context of Appendix C, a sour environment is
reference stress solution for surface breaking flaw in defined as having a partial pressure of H2S (pH2S) above
cylinder allowing for relaxation. 0.05psia (0.003bara), in accordance with standard NACE/ISO
criteria.
SPECIAL CONSIDERATIONS FOR CRA CLAD PIPE
If different material combinations and weld deposits are
used, it may be difficult to test the fracture toughness properties
of all representative microstructures using standard fracture

6 Copyright © 20xx by ASME


7.0 Sample Details
Region 0 Region 1
The sample details section provides guidelines for the user in
6.0 Region 2 selecting the appropriate specimen geometry and configuration
to be used for the FT and FCGR tests in the specific
environment/condition. These include the following:
5.0
pH

Geometry: SENB, CT or SENT specimens are typically used


for sour FT testing. SENB or CT specimens are typically used
4.0
for FCGR testing.
Region 3
3.0 Notching: L-C orientation is more consistent and representative
of the overall pipe properties; L-R orientation more closely
matches the geometry of the surface breaking flaw as used for
0.001 0.01 0.1 1 10 SENT samples.
H2S partial pressure (bar)
Figure 8: NACE domain diagram which depicts regions of Coating Configuration: non-metallic/non-brittle coating; 5/6
environmental severity with respect to SSC for carbon and sides to simulate one-sided diffusion.
low alloy steels (Source: NACE MR0175/ISO 15156-2)

The following sub sections provide a brief outline of the


contents covered in Appendix C of the RP.

Test Environment

The following key environmental variables should be


reproduced accurately and conservatively based on available
project information and considering the discussion below.

 pH
 Partial pressure of H2S (pH2S) Figure 9: Proposed coating configuration: 5/6 sides to
 Salt concentration simulate one-sided diffusion
 Temperature
 Inhibitor
Pre-soaking: 4 days to reach steady state based on H-flux
For C-Mn steels exposed to sour environments, hydrogen measurements and simulation is sufficient for wall thickness
embrittlement effects are known to be most pronounced at 20 mm. For larger specimens, a longer soaking time may be
room temperature (RT). Therefore, conducting FT and FCGR necessary.
tests at RT with relevant partial pressure of H2S should provide
conservative properties for ECA calculations. Fracture Toughness Test Methodology (Rising
Corrosion inhibitors have a dual influence on sour service Displacement)
fracture and fatigue behavior. The presence of inhibitors lowers
the corrosion rate and the associated hydrogen pick up, which Many FT test methods exist and for ECA of pipeline girth
would be expected to reduce the fracture and fatigue welds purposes, and Section B.2.8.3 of the DNVGL-ST-F101
susceptibility of C-Mn line pipe steels in sour environments. states that the SENT specimen is generally recommended since
However, in some instances, such as fatigue crack growth at it is designed to provide a crack tip constraint close to that of
very low values of K or ΔK, the reduction in corrosion rate can an actual pipe in bending containing a flaw. Furthermore, the
have an apparently detrimental effect on the measured crack SENT specimen as per DNV-RP-F108 allows for realistic crack
growth rate, as it may reduce crack tip blunting or crack closure depths of a/W = 0.2-0.5. Other fracture toughness test methods
effects. When performing FCGR tests, it is recommended to and specimen geometries may be used if it can be demonstrated
use a representative concentration of inhibitor. However, for that the results will be sufficiently conservative in relation to
FT tests, it is recommended to use the worst-case environment the situation assessed. DNV GL Singapore and Columbus are
because a fracture event could occur at any time during currently running a Joint Industry Project “Standardization of
operation due process upsets. SENT Test for Sour Service Environments” and different
testing methods in sour service environment are investigated.

7 Copyright © 20xx by ASME


It should be noted that the FT in sour environments is
dependent on the loading method (dynamic vs. static) and the
loading rate. In particular, for the loading rate or K-rate to be
used for ECA related tests, it needs to be carefully selected.
When carrying out ECAs for sour service applications, the
threshold or plateau K-rate should be conservatively use to
establish the fracture toughness for the ECA when no project
specific loading rates are available. It should be noted as this
will most likely result in stringent flaw acceptance criteria and
high repair rates, project specific loading rates (or K-rates)
which are intermediate and more realistic may be considered,
and this approach needs to be accepted by all parties. The main
objectives of the FT tests (Rising Displacement) is to establish
the J value associated with crack initiation and to develop a
complete J R-curve in environment to quantify resistance to
crack propagation. A typical set-up for single specimen SENT
FT test (rising displacement) in sour service environment is
illustrated in Figure 10. Some other key aspects related to the
test are highlighted below:

 Single specimen (DCPD) is recommended since it is


more economical.
 It is recommended to perform minimum three K-rate
sensitivity studies to obtain the plateau K-rate.
 The K-rate to be used should be representative of the
pipeline in operation. Plateau K-rate can be used when
no project specific loading rates are available. Figure 10: Set-up for single specimen SENT FT test (rising
 FT tests should be performed on all three displacement) in sour service environment performed by
microstructures; WCL, HAZ and PP. DNV GL Singapore Laboratory.
 For single specimen approach, a minimum of one test
per microstructure is required. Duplicate tests per Fatigue Crack Growth Rate Tests (Frequency Scan & Paris
microstructure are recommended. Curve)
 Depending on the ECA requirements, J0.2mm or the full
J R-curve can be used for the ECA (if the material In addition to the environmental variables discussed
exhibits stable tearing resistance). above, FCGR in sour environments is dependent on ∆K and
loading frequency.
As such, it is recommended to firstly perform frequency
scan FCGR tests at constant K. The objectives of such a test is
to evaluate the effect of cyclic loading frequency over a range
of frequencies and determine the increase in FCGR with
respect to performance in air. Such a test also helps to
determine the plateau frequency (defined as the frequency
below which there is no further increase in FCGR). The test
also allows one to determine the microstructure (notch
location) that exhibits the highest FCGR.
Some other recommendations related to the test are
highlighted below:
 Perform the frequency scan tests under decreasing
frequency from 1 Hz to 10mHz.
 Perform the tests on all three microstructures; WCL,
HAZ and PP, with a minimum of one test per
microstructure.

8 Copyright © 20xx by ASME


 One microstructure be selected for subsequent Paris ** For very low cyclic loading frequencies, e.g. associated with
curve testing; typically the notch location exhibiting lateral buckling, FCGRs in sweet environments are typically around
the highest FCGRs will be selected. 10 time faster than in air, but can be as much as 20 times faster than
in air. For typical wave loading frequencies around 0.1Hz, FCGRs are
After the frequency scans are completed, it is proposed to typically 3-6 times higher than in air.
validate and supplement the frequency scan tests by measuring *** The units of pH2S should be mbar (not specified in DNVGL-RP-
FCGRs over a wider range of ∆K. The loading frequency for F108, but will be corrected in next revision). In the presence of a high
Paris curve tests should be conservative and representative of concentration of inhibitors, FCGRs at very low cyclic frequencies may
be significantly higher than 40 times faster than in air.
the loading frequency for the fatigue cycles in question. For
flowline assessments, it is impractical to test at loading
frequencies associated with startup and shutdown cycles, so
tests should be carried out at the plateau frequency. CONCLUSIONS
The new DNVGL-RP-F108 covers the same fracture
It is important to note that the use of the in-air Kth
assessment procedures as previously specified in DNV-OS-
recommended in BS 7910 may not always be conservative for
F101, Appendix A which are based on BS7910 Option 2 and
sour environments. A conservative approach is to not include a
tearing analyses with modifications to accommodate strain
FCG threshold.
based analyses. Some changes have been made and the plate
Other Notes reference stress solution is re-introduced as an alternative but
with new procedures for handling of residual stress and the Lr
In general, sour FCGRs are not expected to be influenced by cut-off.
cyclic plastic straining associated with reeling installation. The new RP provides procedures for FEA based fracture
However, even lower fracture toughness properties have been mechanics that may be required to deal with strength
experienced for materials in sour environment if the material mismatching for CRA or CRA clad pipeline girth welds and to
has been plastically deformed prior to the testing. Therefore, handle biaxial loading associated with internal pressure and
for reeling installation, in addition to testing “as-welded” (i.e. applied axial or bending strain.
unstrained) samples, it is recommended to test strained and A unique feature of the new RP is that it includes
aged samples extracted from girth welds that have been guidelines for sour service fracture toughness/resistance and
subjected to representative reeling simulation with the last fatigue crack growth testing to support sour service
strain increment in tension. Engineering Critical Assessments to develop flaw acceptance
criteria for NDT of pipeline and riser girth welds.
In SAFEBUCK III JIP, it is stated that the FCGR in sour
environment appears to follow the same trends as the S-N
ACKNWLEDGMENTS
endurance behaviour. However, it should be noted that while
The authors are pleased to acknowledge the permission
FCGR shows higher rates of crack growth, the S-N tests may
given by the DNV GL management to publish this paper.
not always show the same level of reduction in life.
The authors would also like to acknowledge and express
Some recommendations for environmental FCG acceleration their sincere gratitude to our colleagues; Dr. Colum Holtam, Dr.
factors with respect to air are provided in the table below: Thodla Ramgopal, Dr. Feng Gui, Dr. David Baxter, Dr. Erling
Ostby and Dr. Rikard Tornqvist for their contributions to
preparation of the revised recommended practice.
Table 1. Environmental FCG acceleration factors for use in
the early stages of design.
Environmen FCG Acceleration Factor
t REFERENCES
Seawater 10*
with [1] Recommended Practice, DNVGL-RP-F108:,“Assessment
cathodic of Flaws in Pipeline and Riser Girth Welds”, October 2017
protection [2] Offshore Standard, DNVGL-ST-F101:”Submarine Pipeline
Sweet 10** Systems”, DNVGL, October 2017
Sour [3] British Standard, BS***7910:2013: “Guide to methods for
assessing the acceptability of flaws in metallic structures”,
BSI, London, 2013
* For very low cyclic loading frequencies, e.g. associated with lateral
buckling, a factor of 10 should be used. For typical wave loading [4] British Standard, BS 8571:2014: “Method of test for
frequencies, the marine environment curves in BS 7910 may be used determination of fracture toughness in metallic materials
(these are typically a factor of 2-3 higher than in air). using single edge notched tension (SENT) specimens”.

9 Copyright © 20xx by ASME


ANNEX A

ASSESSMENT CATEGORIES

Table A1: Classification of assessment categories and recommended assessment approaches.

10 Copyright © 20xx by ASME

View publication stats

You might also like