Report Number: R9017 August 2009

PROJECT #153H

CORROSION ASSESSMENT GUIDANCE FOR

HIGH STRENGTH STEELS
(PHASE 1)
GL Industrial Services Restricted to: US DOT PHMSA, PRCI
Restricted & GL Industrial Services

Prepared for: Prepared by:

R Smith V Chauhan and J Crossley
US DOT PHMSA GL Industrial Services UK Ltd
1200 New Jersey Avenue Holywell Park
SE Building, Second Floor Ashby Road
Washington DC 20590 Loughborough Leicestershire
USA LE11 3GR
United Kingdom
Tel: +44 (0)1509 282363
Fax: +44 (0)1509 283119
E-mail: Vinod.Chauhan@gl-group.com
Website: www.gl-group.com

Customer Reference: DTPH56-05-T0003 - Project #153H

This Report is protected by copyright and may not be reproduced in whole or in part by any means without the approval in writing of GL Industrial Services USA, Inc. No
Person, other than the Customer for whom it has been prepared, may place reliance on its contents and no duty of care is assumed by GL Industrial Services USA Inc
toward any Person other than the Customer.
This Report must be read in its entirety and is subject to any assumptions and qualifications expressed therein. Elements of this Report contain detailed technical data
which is intended for analysis only by persons possessing requisite expertise in its subject matter.

GL Industrial Services USA, Inc. is a company incorporated in Delaware with its headquarters at 600 Bent Creek Blvd., Suite 100, Mechanicsburg, PA17050, USA
Report Number: R9017
Issue: 4.0

Report Issue / Amendment Record

Report Title: PROJECT #153H

CORROSION ASSESSMENT GUIDANCE FOR HIGH STRENGTH STEELS (PHASE 1)

Report Number: R9017 Project Title: Project #153H Corrosion Assessment Guidance for High
Strength Steels (Phase 1)
Project SAP Code: 1/07620

Amendment details
Issue Description of Amendment Originator/Author
1.0 Draft report for customer comment V Chauhan and J Crossley

2.0 Report modified to address comments by PRCI Project Team and to take into V Chauhan and J Crossley
account findings reported in GL Report 6781

3.0 Report amended to take into account further comments from the PRCI Project V Chauhan
Team

4.0 Report modified to address PHMSA comments J Crossley

Report approval
Issue Checked by Approved by Date
1.0 R M Andrews C Ward August 2006

2.0 V Chauhan (amendments to report) P Ingham June 2008

3.0 V Chauhan P Ingham October 2008

4.0 V Chauhan P Ingham August 2009

Previous issues of this document shall be destroyed or marked SUPERSEDED

GL Industrial Services Restricted Page i

Report Number: R9017
Issue: 4.0

Project Code: 1/07620

Distribution
Name Company
R Smith US DOT PHMSA
A Mayberry 1200 New Jersey Avenue
SE Building, Second Floor
S Nanney Washington DC 20590
B Keener USA

J Merritt
M Piazza PRCI, Inc.
R Owen
D Johnson
I Wood Electricore, Inc.

DISCLAIMER
This report is furnished to the U.S. Department of Transportation (DOT), Electricore, Inc. (Electricore) and Pipeline Research Council International, Inc. (PRCI)
under the terms of DOT contract DTPH56-05-T-0003 between DOT and Electricore, Electricore agreement DTPH56-05-T-0003 between Electricore and GL
Industrial Services USA, Inc. (GL), and PRCI contract PR-273-0323 between PRCI and GL. The contents of this report are published as received from GL. The
opinions, findings and conclusions expressed in this report are those of the authors and not necessarily those of DOT, Electricore or PRCI, including PRCI’s
member companies or their representatives. Publication of this report by Electricore or PRCI should not be considered an endorsement by Electricore, PRCI or
GL, or the accuracy or validity of any opinions, findings or conclusions expressed herein.
In publishing this report, Electricore, PRCI and GL make no warranty or representation, expressed or implied, with respect to the accuracy, completeness,
usefulness, or fitness for purpose of the information contained herein, or that the use of any information, method, process, or apparatus disclosed in this report
may not infringe on privately owned rights. Electricore, PRCI and GL assume no liability with respect to the use of, or for damages resulting from the use of, any
information method, process, or apparatus described in this report. The text of this publication, or any part thereof, may not be reproduced or transmitted in any
form by any means, electronic or mechanical, including photocopying, recording, storage in an information retrieval system, or otherwise, without written approval
of Electricore, PRCI and GL.

GL Industrial Services Restricted Page ii

Report Number: R9017
Issue: 4.0

Executive Summary
The continuing worldwide demand for natural gas presents major challenges to pipeline operators. There is
increasing need to construct long distance, high capacity transmission pipelines, particularly in the more
remote areas of Arctic North America, Asia, Africa and South America. To achieve satisfactory economic
returns on the investment, operators are focusing attention on the use of increasing material strength (pipe
grade) that allows for higher operating pressures and smaller pipe diameters, thus reducing the total steel
tonnage, transportation costs and the volume of weld metal needed to be applied during pipe installation.
Steel making and pipe manufacturing developments during the 1970’s and 1980’s resulted in the
progressive evolution of API 5L Grade X65 to X70 and X80. In North America and Europe, Grade X80
pipelines have gained general acceptance. The economic benefits of further increases in strength have
focused attention on the next step increase to Grade X100 and even X120. In the US two major operators
have recently announced a joint venture to build a major pipeline using Grade X100 pipe.
Extensive experimental and numerical work has been undertaken to develop methods for assessing the
remaining strength of corroded transmission pipelines. These methods, embodied in documents such as
ASME B31G, RSTRENG and BS 7910 have, however, only been validated for pipeline materials of grades
up to and including X65. The method detailed in BS 7910 is based on the output of a Group Sponsored
Project (GSP) led by Germanischer Lloyd (hereafter GL, formerly Advantica 1 ) in the late 1990’s. The
method is often referred to as the Line Pipe Corrosion (LPC) method. The output from the GSP also forms
the basis of the assessment method described in DNV RP-F101.
As operators start to use higher material strengths, there will be an increasing need to assess the integrity
of high strength corroded pipelines. Use of existing assessment methods may be inappropriate for higher
strength pipelines. A particular concern is the high yield to tensile (Y/T) ratio of high strength steels; early
development Grade X100 materials had (Y/T) values up to 0.98. Although more recent materials have
reduced this to some extent, there is still a concern that high strength steels may not have sufficient work
hardening capacity, or strain to failure, to ensure that existing assessment methods are appropriate.
This report describes a program of work to extend existing methods to material strengths up to grade X100
using finite element (FE) analyses and validation using full scale testing.

Conclusions
1. For the burst tests on high strength line pipe investigated in this report, standard assessment
methods used by the pipeline industry generally give failure predictions lower than the recorded
burst pressures. For a small number of test points the ASME B31G, Modified ASME B31G and the
LPC-1 methods give failure predictions higher than the recorded burst pressures; these are for tests
on line pipe with defect depths greater than 50% of the pipe wall. It is noted that for machined
defects, particularly those that are rectangular flat bottomed patches the use of ASME B31G and
Modified ASME B31G to predict failure pressures may be inappropriate because the area of metal
loss can be underestimated.
2. The RSTRENG method is the most reliable method for predicting the failure pressure of corroded
pipelines. RSTRENG predicts lower failure pressures than the recorded burst pressures from tests
for defect depths up to 80% of the pipe wall in line pipe of strength grades up to X100.
3. Modifying the flow stress to equal the arithmetic mean of the specified minimum yield strength and
the ultimate tensile strength adds conservatism to the calculated failure predictions.

1
Advantica was acquired by Germanischer Lloyd in 2007.

GL Industrial Services Restricted Page iii

Report Number: R9017
Issue: 4.0

4. The non-linear FE method gives failure predictions within a scatter band of ±10%. This level of
scatter is typical. More accurate modeling of the geometry and material properties, to take into
account of any through wall variation, should reduce the observed scatter.
Recommendations
1. Burst tests on higher strength pipe have to date only been conducted using machined defects to
simulate volumetric corrosion. Predictions of failure pressures using the ASME B31G and Modified
ASME B31G methods for machined defects, particularly those that are rectangular flat bottomed
patches may be inappropriate because the area of metal loss can be underestimated. It is
recommended that a focused program of burst tests are conducted on grade X80 and X100 pipe
with corrosion defects that are more representative of those found in the field. Failure pressure
predictions using ASME B31G, Modified ASME B31G and RSTRENG should then be compared to
the recorded burst test pressures.
2. The results and conclusions described in this report should be reviewed following completion of the
work in Phase 2 of Project #153H and when the results from the BP Exploration X100 Operational
Trial become available.

GL Industrial Services Restricted Page iv

Report Number: R9017
Issue: 4.0

Contents
1 Introduction.................................................................................................................................. 1
2 Assessment Methods Used by the Pipeline Industry ............................................................... 1
3 Approach ...................................................................................................................................... 2
4 Burst Tests on Grade X80 and X100 Line Pipe.......................................................................... 2
5 Failure Predictions Using Non-Linear Finite Element Analysis............................................... 3
5.1 Method.............................................................................................................................................. 3
5.2 Model Generation ............................................................................................................................ 4
5.3 Loading and Boundary Conditions................................................................................................ 4
5.4 Material Properties .......................................................................................................................... 4
5.5 Method of Predicting Failure Pressures........................................................................................ 5
6 Comparison of Test and Failure Pressure Predictions............................................................. 6
6.1 Failure Predictions with Flow Stress Unmodified ........................................................................ 6
6.2 Failure Predictions with Flow Stress Modified ............................................................................. 7
7 Discussion.................................................................................................................................... 8
8 Conclusions ................................................................................................................................. 9
9 Recommendations....................................................................................................................... 9
10 References ................................................................................................................................... 9

GL Industrial Services Restricted Page v

Report Number: R9017
Issue: 4.0

1 Introduction
The continuing worldwide demand for natural gas presents major challenges to pipeline operators. There is
increasing need to construct long distance, high capacity transmission pipelines, particularly in the more
remote areas of Arctic North America, Asia, Africa and South America. To achieve satisfactory economic
returns on the investment, operators are focusing attention on the use of increasing material strength (pipe
grade) that allows for higher operating pressures and smaller pipe diameters, thus reducing the total steel
tonnage, transportation costs and the volume of weld metal needed to be applied during pipe installation.
Steel making and pipe manufacturing developments during the 1970’s and 1980’s resulted in the
progressive evolution of API 5L Grade X65 to X70 and X80. In North America and Europe, Grade X80
pipelines have gained general acceptance. The economic benefits of further increases in strength have
focused attention on the next step increase to Grade X100 and even X120. In the US two major operators
have recently announced a joint venture to build a major pipeline using Grade X100 pipe [1].
Extensive experimental and numerical work has been undertaken to develop methods for assessing the
remaining strength of corroded transmission pipelines. These methods, embodied in documents such as
ASME B31G [2], RSTRENG [3], [4] and BS 7910 [5] have, however, only been validated for pipeline
materials of grades up to and including X65. The method detailed in BS 7910 is based on the output of a
Group Sponsored Project (GSP) led by Germanischer Lloyd (hereafter GL, formerly Advantica 2 ) in the late
1990’s. The method is often referred to as the Line Pipe Corrosion (LPC) method [6]. The output from the
GSP also forms the basis of the assessment method described in DNV RP-F101 [7].
As operators start to use higher material strengths, there will be an increasing need to assess the integrity
of high strength corroded pipelines. Use of existing assessment methods may be inappropriate for higher
strength pipelines. A particular concern is the high yield to tensile (Y/T) ratio of high strength steels. For cold
expanded pipe, API 5L/ISO 3183 [21] states that the (Y/T) ratio should not exceed 0.93 for Grade X80 pipe
and 0.97 for Grade X100 pipe. Early development Grade X100 materials had Y/T values up to 0.98.
Although more recent materials have reduced this to some extent, there is still a concern that high strength
steels may not have sufficient work hardening capacity, or strain to failure, to ensure that existing
assessment methods are appropriate.
This report describes a program of work to extend existing methods to material strengths up to grade X100
using burst tests, ring expansion tests and finite element (FE) analyses.

2 Assessment Methods Used by the Pipeline Industry

Existing assessment methods regularly used by the pipeline industry are ASME B31G, Modified ASME
B31G, RSTRENG, LPC, BS 7910 and DNV RP-F101. The refinery and petrochemical industry also use API
579-1/ASME FFS-1 [8]. These methods have been developed from the results of a large number of full-
scale burst tests on ring expansion and vessel specimens. Some researchers have supplemented their
database of full-scale test results with finite element (FE) analyses. A wide range of material properties and
pipeline geometries has been investigated. Most of the experimental work considered volumetric corrosion
defects, predominantly longitudinally orientated, subject only to internal pressure. Some investigations have
been undertaken to study the effect of in-plane bending and axial loading on pipelines. Some tests have
also been undertaken on pipes with circumferentially or helically orientated corrosion defects. In the US, the
Federal Regulations, CFR 192 [9] and 195 [10] recommend using ASME B31G or RSTRENG.
A brief background to the development of the main assessment methods described above is given in [11].

2
Advantica was acquired by Germanischer Lloyd in 2007.

GL Industrial Services Restricted Page 1

Report Number: R9017
Issue: 4.0

3 Approach
The non-linear finite element (FE) method described in BS 7910 and PRCI’s Guidance Document [12] has
been routinely used by GL to predict the failure pressure of corroded pipelines [13], [14]. The method was
also used to develop the LPC method which forms the basis of the assessment methods described in BS
7910 and DNV RP-F101.
The general approach is consistent with a Level 3 assessment described in API 579-1/ASME FFS-1. In
agreement with the PRCI project team, the approach taken on this project was as follows:
1. Validate the failure pressures predicted from the FE analyses against available burst test data. It is
to be noted that burst test data for Grade X80 and X100 materials is generally not available in the
public domain. GL has undertaken 8 vessel tests on Grade X80 line pipe for a PRCI member
company (National Grid plc). The tests were conducted on 1219.2mm (48-inch) diameter Grade
X80 line pipe. Basic details of the test results are given in [11] and are reproduced in Table 1 of this
report. In addition, GL has undertaken a project on behalf of BP Exploration to investigate the
failure behavior of corroded 1320.8mm (52-inch) diameter Grade X100 line pipe, see section 4
below. BP Exploration has agreed to release the burst test results in support of Project #153H.
Basic details of the test results are given in [11] and are reproduced in Tables 2 and 3 of this report.
2. Compare the burst test data with failure predictions obtained from the FE analyses and existing
methods such as ASME B31G, Modified ASME B31G, RSTRENG and LPC 3 .
3. Based on the above, make recommendations for assessing the remaining strength of corroded pipe
up to grade X100.

4 Burst Tests on Grade X80 and X100 Line Pipe

Both grade X80 and X100 pipe with real corrosion defects was not available for conducting burst tests.
Therefore a series of burst tests were undertaken by GL using line pipe with machined defects to simulate
volumetric corrosion defects. Tests on 1219.2mm (48-inch) diameter grade X80 line pipe were conducted
using 8 full-scale vessels. Two pipe wall thicknesses were tested, 19.89mm (0.783-inch) and 13.79mm
(0.543-inch). The test report and interpretation is described in [15] 4 . Basic details of the tests and the
recorded failure pressures are summarized in Table 1. The numbering terminology used to identify each test
is consistent with that used in [11].
In addition to the burst test program on grade X80 pipe, GL has also completed a series of burst tests for
BP Exploration to investigate the corrosion defect tolerance of 1320.8mm (52-inch) diameter grade X100
line pipe. Two pipe wall thicknesses were tested, 20.6mm (0.811-inch) and 22.9mm (0.902-inch). Tests
were undertaken using both ring expansion specimens and full-scale vessels. The test report and
interpretation is described in [16] 5 . Briefly, the test program comprised 39 ring expansion tests and 4 full-
scale vessel tests. Defects were machined on the external surface of the pipe defects to simulate areas of

3
Failure predictions quoted do not include a safety factor.
4
This is a confidential GL report for National Grid and is not available in the public domain.
5
This is a confidential GL report for BP Exploration report and is not available in the public domain.

GL Industrial Services Restricted Page 2

Report Number: R9017
Issue: 4.0

metal loss. Patch 6 , groove 7 and slit 8 type defects were investigated. Basic details of the tests and the
recorded failure pressures are summarized in Tables 2 and 3. Once again the numbering terminology used
to identify each test is consistent with that used in [11].

5 Failure Predictions Using Non-Linear Finite Element Analysis

The non-linear FE analysis method described in Annex G of BS 7910 was used to predict the failure
pressure of grade X100 line pipe with a single volumetric corrosion defect. A description of the defect
dimensions and nomenclature is illustrated in Figure 1. To validate the results of the FE analyses, a
selection of the burst tests from the BP Exploration test program described in section 4 was modeled. Due
to budget and time constraints the validation was undertaken using a selection of grade X100 tests,
primarily because they exhibited high yield to tensile (Y/T) ratios. The validation was undertaken using the
results from 4 vessel and 10 ring expansion tests, see Tables 2 and 3. In addition to the comparing failure
pressures obtained using the FE method, a comparison of the predicted failure pressures (without safety
facors) was also made using standard assessment methods such as ASME B31G, Modified ASME B31G,
RSTRENG and LPC.

5.1 Method
Volumetric metal loss corrosion defects in pipelines are generally present as smooth profiled areas with a
reduced ligament of the pipe wall. The failure mechanism of this type of defect is dominated by plastic
collapse at the remaining ligament. The failure pressure of internally pressurized ductile steel pipe with
either local or general metal loss defects, such as corrosion, can be predicted by numerical analysis using
the non-linear FE method and a validated failure criterion. Complex flaw shapes and combined loading
conditions can be considered in the analysis. This method is described in BS 7910 Annex G [5] and the
PRCI Corrosion Assessment Guidance Document [12]. Briefly, the method consists of four major steps as
follows:
 Create a finite element model of the corroded pipe or vessel, using information on the flaws
detected, the measured material properties and the structural constraints and applied loads.
 Perform a non-linear, large deformation stress analysis using an appropriate finite element analysis
software package and a validated analysis procedure.
 Examine analysis results obtained from the stress analysis.
 Determine the failure or critical pressure value based on the variation of local stress or strain states
with reference to a validated failure criterion or test work.
As with any FE simulation, the results obtained are highly dependent upon the assumptions made in the
generation of the model, material properties and the prescribed boundary conditions.

6
Patch defects are defined as areas of general metal loss resulting from corrosion, erosion or a combination of both. The area of metal loss is
uniformly distributed in the axial and circumferential directions.
7
Grooves defects are defined as long elongated areas of metal loss caused by directional corrosion and/or erosion. The length of the groove is
much greater than the width.
8
Slit defects are much narrower than the groove or patch defects. They are machined using a wire feed electro-discharge machine (EDM). The
diameter of the wire used was 0.1mm (~3.9 mil), giving a final slit width of approximately 0.15mm (5.9 mil).

GL Industrial Services Restricted Page 3

Report Number: R9017
Issue: 4.0

5.2 Model Generation

For the vessel models, quarter symmetry, three-dimensional (3D) non-linear FE models were created as
shown in Figure 2. This approach takes advantage of symmetry to reduce the size of the FE model, thereby
reducing computer run/post-processing times. The models were created using the mesh generating
software MSC PATRAN [17] and analyzed using the commercially available finite element code,
ABAQUS/Standard [18]. The 3D models were constructed using 20 noded, reduced integration brick
elements (ABAQUS type C3D20R). As recommended in Annex G of BS 7910 care was taken to ensure that
at least four layers of elements were used through the remaining ligament of each corrosion defect. This
was to ensure that the high stress gradients could be predicted with sufficient accuracy in the areas of
interest. To ensure that the mesh was fine enough a mesh convergence study was conducted to confirm
that the FE model was sufficiently fine. All groove defects were modeled to be round bottomed with
spherical ends, the radius of which is equal to the wall thickness, t, so each defect was modeled with a
width equal to 2t.
The ring expansion specimens were modeled using two-dimensional (2D), 4 noded plane strain solid
elements (ABAQUS type CPE4) with one plane of symmetry, see Figure 3. Patch defects were modeled
with a spherical radius to give a circumferential surface width, W, of approximately 4 times the pipe wall
thickness. The groove defects were modeled to be round bottomed with spherical ends, the radius of which
is equal to the required defect depth, as shown in Figure 1. The slits were modeled with a rounded bottom
of radius equal to half the width.

5.3 Loading and Boundary Conditions

Failure pressures were investigated for internal pressure loading only. For each model the load was applied
as a monotonically increasing internal pressure, where pressure loads remain normal to the pipe surface
throughout the analysis. External loading was not considered.
For the 3D models, symmetry boundary conditions were used to reduce the size of the FE models. Two
axes of symmetry were applied to the quarter models, in the x=0 and z=0 planes (see Figure 2). The model
was not allowed to rotate, or to expand or contract axially. This simulates a buried pipe in which axial
expansion and contraction is restricted by the soil. The model was, however, allowed to expand and
contract radially. Rigid body motion was prevented by restraining nodes in the axial direction at the end of
the cylinder furthest away from the area of interest. The cylindrical shell was extended sufficiently far away
to ensure the application of boundary conditions did not affect stresses in the area of interest.
In order to represent the pipe sections being capped off, as in the full-scale tests, pressure end loads were
applied to the unrestrained end of the model.
For the 2D plane strain models, one axis of symmetry was applied in the x=0 plane (see Figure 3). Rigid
body motion was prevented by restraining one node in the y direction at the bottom center of the ring,
furthest away from the area of interest.

5.4 Material Properties

Stress versus strain curves were obtained for grade X80 and X100 line pipe material. Data from round bar
tests was used in preference to data from flattened strap tests. For FE analyses, data from round bar tests
is considered more reliable as the Bauschinger effect can influence stress versus strain data from flattened
strap tests. The Data for each material grade was obtained as follows:
 Grade X80
FE simulations of the burst tests were not conducted for grade X80 pipe. However, for completeness and to
aid with comparison with grade X100 materials, stress versus strain curves for X80 material was obtained.
Data was obtained from the public domain and from PRCI member companies [15], [20]. The data was

GL Industrial Services Restricted Page 4

Report Number: R9017
Issue: 4.0

obtained from 4 812.8mm (32-inch) diameter by 19.05mm (0.75-inch) thick and four 1219.2mm (48-inch)
diameter by 15.9mm (5/8-inch); 19.89mm (0.783-inch); 13.79mm (0.534-inch) thick line pipe specimens.

 Grade X100
Stress versus strain data for 1320.8mm (52-inch) diameter grade X100 line pipe was available from the BP
Exploration test program, see section 4. Data was also available from a Joint Industry Project (JIP) on X100
[19] that was led by GL and from published work, primarily from the 2004 ASME International Pipeline
Conference (IPC) proceedings [20].
The specified minimum yield strength (SMYS) and tensile strength (SMTS) for Grade X100 pipe is quoted in
[21] as 690MPa (100ksi) and 760MPa (110ksi) respectively, with a maximum yield to tensile strength ratio of
0.93. Yield strength is quoted at a total strain of 0.5%, designated as Rt0.5.
Figure 4 shows a compilation of nominal stress versus strain curves obtained from the sources described
above. When defining plasticity data in FE codes such as ABAQUS/Standard, true stress versus true strain
data must be used, where zero plastic strain corresponds to the yield point of the material. The equations
for true stress and true strain are valid only up to the onset of necking, i.e. the tensile strength of the
material. Hence the engineering stress versus strain data used was truncated at this value before being
converted to true stress versus true strain data. The data is input into ABAQUS/Standard as a piecewise
linear representation. A rate-independent plasticity model using the von Mises yield criterion and isotropic
hardening rule was adopted. An isotropic hardening rule is generally used for assessing structures subject
to a monotonically increasing load. The ABAQUS documentation recommends use of a kinematic hardening
rule when cyclic loading is modeled.
A comparison of the true stress versus true strain curves used for grade X65, X80 and X100 material are
shown in Figure 5.
All the analyses were undertaken using a Young’s Modulus of 210000 MPa (30460 ksi) and a Poisson’s
ratio of 0.3.

5.5 Method of Predicting Failure Pressures

The method of predicting failure pressure of corroded pipelines using FE analysis is described in Annex G
of BS 7910.
For each model analyzed, the von Mises equivalent stress was monitored at three points through the
highest stress portion of the ligament of each defect as the internal pressure in the pipe was increased. As
shown in Figure 6, the stress variation with increasing internal pressure exhibits three distinct stages. The
first stage is a linear response progressing to a point when the elastic limit is reached. As the pressure
continues to increase, a second stage is evident as plasticity spreads through the ligament. The von Mises
equivalent stress increases very slowly because of the constraint provided by the surrounding pipe wall. The
third phase is dominated by material hardening and begins when the von Mises equivalent stress in the
entire ligament exceeds the material’s yield strength. Once this stage is reached, the whole ligament
deforms plastically but failure does not occur immediately due to strain hardening. Figure 7 shows a typical
von Mises equivalent stress contour plot of a pipe with an axially orientated groove defect.
For the analyses described in this report, the failure pressure was determined as that corresponding to the
point at which the average von Mises equivalent stress at the ligament was equal to the true ultimate tensile
strength of the material; this is consistent with the approach described in Annex G of BS 7910.

GL Industrial Services Restricted Page 5

Report Number: R9017
Issue: 4.0

6 Comparison of Test and Failure Pressure Predictions

6.1 Failure Predictions with Flow Stress Unmodified

Figures 8 to 12 show a comparison of the actual failure pressure (PA) versus the predicted failure pressure
(Pf) 9 using the ASME B31G, Modified ASME B31G, RSTRENG, LPC-1 and the non-linear finite element
analysis methods. In each case the flow stress is calculated using the specified minimum yield strength or
the specified ultimate tensile strength as appropriate. The results are presented in a non-dimensional form.
Values of the ratio (PA/Pf) greater than unity indicate that the actual recorded burst pressure is greater than
the predicted failure pressure. Conversely, values of the ratio (PA/Pf) less than unity indicate that the actual
recorded burst pressure is lower than the predicted failure pressure. Tabulated values of the assessment
points are given in Tables 4 to 6.
The following is concluded from the assessments:
1. ASME B31G predicts a lower failure pressure than the recorded burst pressure for 36 out of the 40
valid 10 test points. Out of the remaining 4 test points, a failure prediction higher than the recorded
burst pressures is obtained for a relatively deep defect (d/t=77.5%) in grade X80 pipe. The
remaining 3 failure predictions that are higher than the recorded burst pressure are for defects of
depths 50% and above. One of these test points (INDEX 300) is for a machined slit defect. As
discussed in [11], it is debatable whether standard methods for assessing corrosion damage in
pipelines are suitable for slit type defects. As also discussed in [11], for machined defects,
particularly those that are rectangular flat bottomed patches the use of ASME B31G to predict
failure pressures may be inappropriate because the area of metal loss can be underestimated.
2. Modified ASME B31G predicts a lower failure pressure than the recorded burst pressure for 32 out
of the 40 valid test points. Failure predictions higher than the recorded burst pressures are obtained
for 2 tests on grade X80 pipe. It is also to be noted that for 3 of these points (INDEX 277, 292 and
299), ring expansion testing was used. As discussed in section 4 of [11], the Modified ASME B31G
method uses an arbitrary shape factor of 0.85 for the corrosion defect. For tests conducted using
ring expansion specimens, where the defect length is infinitely long, use of a shape correction is
inappropriate. Therefore, the comparison of burst pressure and predicted failure pressures should
be treated with caution. For the cases where the failure predictions are higher than the actual
recorded burst pressure the defect depth is approximately 50% of the wall or deeper.
3. RSTRENG is predicts a lower failure pressure than the recorded burst pressure for 39 out of 40
valid test points. For the one test point (INDEX 299) where the failure prediction is higher than the
recorded burst pressure, this was for a deep (d/t = 77%) machined slit defect, i.e. similar to a sharp
crack-like defect. It is debatable whether existing assessment methods for assessing corrosion
damage in pipelines are suitable for this type of defect.
4. LPC-1 predicts a lower failure pressure than the recorded burst pressure for 45 out of the 49 valid
test points. For the 4 test points where the predicted burst pressure is higher than the recorded
burst pressure, one is for a relatively deep defect (78.2% of the pipe wall) in grade X80 pipe. Out of
the 3 remaining points, for one case the ratio (PA/Pf) is only marginally below unity; another point is

9
Note the distinction between the predicted failure pressure Pf and the maximum safe operating pressure PMSOP which is calculated using an
appropriate factor of safety.
10
Validity of the test points is based on the defect depth. The ASME B31G, Modified ASME B31G and RSTRENG methods are valid for
assessing defect depths up to 80% of the pipe wall. The LPC-1 method is valid for assessing defect depths up to 85% of the pipe wall.

GL Industrial Services Restricted Page 6

Report Number: R9017
Issue: 4.0

for INDEX 299, see above and should be discounted. The final test point is for a 50% deep groove
defect in grade X100 pipe.
5. The non-linear FE method predicts a lower failure pressure than the recorded burst pressure for 6
out of the 14 tests that were modeled. However, in the majority of cases the failure predictions the
ratio (PA/Pf) is only marginally less than unity. For 3 of these cases where the ratio (PA/Pf) is less
than unity, they are for line pipe with slit defects (INDEX 289, 294 and 298) and as discussed
previously it is debatable whether standard methods for assessing corrosion damage in pipelines
are suitable for slit type defects. The comparison of tests with slit defects should therefore be
discounted. The remaining failure predictions are within ±10% of the actual failure pressure. This is
consistent with the level of scatter observed for lower strength grades and can be explained by the
fact that the FE method is based on an idealized geometry, both of the pipe and the defect. In
reality, there may be some ovality in the test pipe and/or local variation in the wall thickness. There
may also be local variations in material properties, around the circumference and through the pipe
wall. A through wall variation in tensile properties is not unexpected for high strength steels due to
the potential differences in cooling rates during plate manufacture. Variations in strain during the U
and O stages of the pipe forming process may also add to these variations. An investigation of the
variation in tensile properties of grade X100 pipe is the subject of the study in Phase 2 of Project
#153H. Once these factors have been taken into account, it is judged that failure predictions, for
smooth groove and patch like defects, using the FE method will be in very good agreement with
actual burst pressures.

6.2 Failure Predictions with Flow Stress Modified

As discussed in [11] 11 , the flow stress is not a precisely defined parameter; its magnitude lies somewhere
between the yield strength and ultimate tensile strength of the material. Some operators impose additional
requirements on the material properties for higher strength line pipe. For example if the (Y/T) ratio is limited
to 0.9 for grade X80 line pipe, then the minimum tensile strength is equal to 1.11 times the specified
minimum yield strength. Using the definitions of flow stress as appropriate for each assessment method, the
following is obtained for grade X80 line pipe:

Assessment Yield Strength Tensile Strength Flow Stress

Method (ksi) (ksi) (ksi)
ASME B31G 80 89 88
Mod ASME B31G 80 89 90
RSTRENG 80 89 90
LPC-1 80 89 89

In this example, the flow stress is calculated to be greater than the tensile strength for the Modified ASME
B31G and RSTRENG methods. Care is therefore required in how the methods are used when assessing
corrosion defects in higher strength pipelines. In the fitness-for-purpose standard, BS 7910 [5], the flow
stress is defined as the arithmetic mean of the yield strength and tensile strength up to a value of 1.2 times
the yield strength. The effect of the modification is that the flow stress will always be calculated to be less

11
More details regarding the concept of the flow stress and the definitions used for the ASME B31G, Modified ASME B31G, RSTRENG, LPC,
SHELL92 and PCORRC assessment methods is given in [11].

GL Industrial Services Restricted Page 7

Report Number: R9017
Issue: 4.0

than the tensile strength. To investigate the sensitivity of the failure predictions, the flow stress definition
was modified to that recommended in BS 7910.
Figures 13 to 16 show a comparison of the actual failure pressure (PA) versus the predicted failure pressure
(Pf) using the ASME B31G, Modified ASME B31G, RSTRENG and LPC-1 methods. The main conclusion
drawn from the study is that when the flow stress is modified according to that given in BS 7910, then
additional conservatism is introduced to calculate the predicted failure pressure.

7 Discussion
The results of the study described in this report have shown that for the majority of cases, existing methods
used by the pipeline industry can be used to assess volumetric corrosion defects in pipelines of strength
grades up to X100. However, for a small number of tests, failure predictions higher than the recorded burst
pressure from tests were obtained. In particular, the ASME B31G, Modified ASME B31G and the LPC-1
methods gave failure predictions higher than recorded burst pressures for those tests where defect depths
were greater than 50% of the pipe wall. As discussed in [11], for machined defects, particularly those that
are rectangular flat bottomed patches the use of ASME B31G and Modified ASME B31G to predict failure
pressures may be inappropriate because the area of metal loss can be underestimated. Therefore these
results need to be treated with caution. It is recommended that a focused program of full-scale burst tests
are conducted using high strength pipe (Grade X80 and X100) with simulated defects that represent real
corrosion damage in the field. More realistic corrosion defects could be produced by a number of methods.
Starting with a flat bottomed machined defect, corrosion features could be produced by either treating an
area of the pipe with a mineral acid such as hydrochloric acid (HCl) or by accelerating corrosion by
simulated ground water (e.g. NS4 solution) using electrochemical methods. In either case a realistic
corroded surface would be produced which would better simulate an actual service defect compared to a
machined defect. Failure pressure predictions using ASME B31G, Modified ASME B31G and RSTRENG
should then be compared to the recorded burst test pressures.
The non-linear FE method generally gave failure predictions within a scatter band of ±10%. Failure
predictions higher than the recorded burst pressures were generally obtained for relatively deep slit defects.
It is debatable whether standard methods for assessing corrosion damage in pipelines are suitable for slit
type defects and these results should be discounted.
More accurate modeling of the geometry and material properties, possibly to take into account the through
wall variation, should reduce the observed scatter in the failure predictions. In Phase 2 of Project #153H, the
through wall variation of material properties in grade X100 pipe will be investigated. It is recommended that
once the outcome of this work is known and when the results of the BP X100 Operational Trial [22], [23]
become available, failure predictions described in this report using the FE method are revisited.
A further concern is that for higher strength steels, the (Y/T) ratio starts to rise. API 5L/ISO 3183 stipulates
limits of 0.93 and 0.97 for Grade X80 and X100 respectively. Depending on the assessment method used,
the flow stress definition when applied to assessing higher strength steels can exceed the tensile strength.
When the flow stress is modified to equal the arithmetic mean of the specified minimum yield strength and
the specified minimum tensile strength, then this adds additional conservatism to the predicted failure
pressure.
The RSTRENG method proved to be the most reliable method. This conclusion is consistent with that
obtained for the much larger test database of material grades from A25 to X100 investigated in [11]. As

GL Industrial Services Restricted Page 8

Report Number: R9017
Issue: 4.0

discussed in [11], the SHELL92 method 12 [24], which is a modified version of the ASME B31G method,
conservatively predicts failure pressures for corrosion defects up to 80% deep in line pipe of strength grade
up to X100.

8 Conclusions
1. For the burst tests on high strength line pipe investigated in this report, standard assessment
methods used by the pipeline industry generally give failure predictions lower than the recorded
burst pressures. For a small number of test points the ASME B31G, Modified ASME B31G and the
LPC-1 methods give failure predictions higher than the recorded burst pressures; these are for tests
on line pipe with defect depths greater than 50% of the pipe wall. It is noted that for machined
defects, particularly those that are rectangular flat bottomed patches the use of ASME B31G and
Modified ASME B31G to predict failure pressures may be inappropriate because the area of metal
loss can be underestimated.
2. The RSTRENG method is the most reliable method for predicting the failure pressure of corroded
pipelines. RSTRENG predicts lower failure pressures than the recorded burst pressures from tests
for defect depths up to 80% of the pipe wall in line pipe of strength grades up to X100.
3. Modifying the flow stress to equal the arithmetic mean of the specified minimum yield strength and
the ultimate tensile strength adds conservatism to the calculated failure predictions.
4. The non-linear FE method gives failure predictions within a scatter band of ±10%. This level of
scatter is typical. More accurate modeling of the geometry and material properties, to take into
account of any through wall variation, should reduce the observed scatter.

9 Recommendations
1. Burst tests on higher strength pipe have to date only been conducted using machined defects to
simulate volumetric corrosion. Predictions of failure pressures using the ASME B31G and Modified
ASME B31G methods for machined defects, particularly those that are rectangular flat bottomed
patches may be inappropriate because the area of metal loss can be underestimated. It is
recommended that a focused program of burst tests are conducted on grade X80 and X100 pipe
with corrosion defects that are more representative of those found in the field. Failure pressure
predictions using ASME B31G, Modified ASME B31G and RSTRENG should then be compared to
the recorded burst test pressures.
2. The results and conclusions described in this report should be reviewed following completion of the
work in Phase 2 of Project #153H and when the results from the BP Exploration X100 Operational
Trial become available.

10 References
[1] Anon. ‘Denali – The Alaska Gas Pipeline’, http://www.denali-thealaskagaspipeline.com/, BP and
ConocoPhillips, 2008

12
The SHELL92 method uses the same Folias (bulging correction) factor as that used by ASME B31G. The shape of the defect is modified
from parabolic to rectangular and the flow stress is modified to equal 0.9 times the specified minimum tensile strength.

GL Industrial Services Restricted Page 9

Report Number: R9017
Issue: 4.0

[2] Anon: ASME B31G-1991 (revision of ANSI/ASME B31G-1984): Manual for determining the remaining
strength of corroded pipelines - a supplement to ASME B31 code for pressure piping, The American Society
of Mechanical Engineers, New York, 1991
[3] Kiefner, J.F. and Veith, P.H., ‘A Modified Criterion for Evaluating the Remaining Strength of Corroded
Pipe’, Final Report on PR-3-805 to the Pipeline Research Committee of the American Gas Association,
Battelle, Ohio, 1989
[4] Vieth, P.H. and Kiefner, J.F., ‘RSTRENG2 User’s Manual, Final report on PR-218-9205 to Corrosion
Supervisory Committee, Pipeline Research Committee, American Gas Association, Kiefner & Associates,
Inc., Ohio, 1993
[5] Anon: ‘Guide to Methods for Assessing the Acceptability of Flaws in Metallic Structures’, BS 7910:2005,
Incorporating Amendment No. 1, BSi, September 2007.
[6] Fu, B. and Batte, A.D. ‘Advanced Methods for the Assessment of Corrosion in Linepipe. Summary
Report OTO 97065, UK Health and Safety Executive
[7] Anon: ‘Corroded Pipelines’, Recommended Practice DNV-RP-F101, Including Amendments dated
October 2006, Det Norske Veritas, October 2004
[8] Anon: ‘Fitness-For-Service’, API 579-1/ASME FFS-1, American Petroleum Institute, Second Edition,
2007
[9] Anon. ‘Title 49 – Transportation. Transportation of Natural and Other Gas by Pipeline: Minimum Federal
Safety Standards’, Part 192, Pipeline and Hazardous Materials Safety Administration, US Department of
Transportation,
http://ecfr.gpoaccess.gov/cgi/t/text/textidx?sid=73362a5c11608da8fa62ec9710cb9640&c=ecfr&tpl=/ecfrbro
wse/Title49/49cfrv3_02.tpl
[10] Anon. ‘Title 49 – Transportation. Transportation of Hazardous Liquids by Pipeline’, Part 195, Pipeline
and Hazardous Materials Safety Administration, US Department of Transportation,
http://ecfr.gpoaccess.gov/cgi/t/text/text-
idx?sid=73362a5c11608da8fa62ec9710cb9640&c=ecfr&tpl=/ecfrbrowse/Title49/49cfrv3_02.tpl
[11] Chauhan, V. and Brister, J., ‘A Review of Methods for Assessing the Remaining Strength of Corroded
Pipelines’, Advantica Report Number 6781, Issue 3.0, November 2007
[12] Fu. B, Jones, C.L. and Chauhan, V. ‘Guidance for Assessing the Remaining Strength of Corroded
Pipeline’, Advantica, Inc Report Prepared for the Materials Technical Committee of Pipeline Research
Council International Inc., August 2002 Contract PR-273-9803, Catalogue No. L51958
[13] Chauhan, V. and Wood, A., ‘Experimental Validation of Methods for Assessing Closely Spaced
Corrosion Metal Loss Defects in Pipelines’, Advantica Report Prepared for the PRCI Materials Technical
Committee, Catalog No. 52007, September 2004
[14] Chauhan, V. and Sloterdijk, W., ‘Advances in Interaction Rules for Corrosion Defects in Pipelines’,
Proceedings of the International Gas Research Conference, Vancouver, Canada, November 2004
[15] Fu, B., Franklin, J.G. and Vu, D.Q., ‘A Damage Tolerance Study on 1200mm Grade X80 Line Pipes’,
Advantica Report GRTC R2058 Issue 1, November 1999 (Confidential – Access Restricted)
[16] Wood, A. Morgan, G. and Swankie, T., ‘Ring Tension and Full Scale Burst Testing of Grade X100
Linepipe Material’, Advantica Report R7702, Issue 1.0, November 2005, Prepared for BP Exploration,
November 2005 (Confidential - Restricted to BP Exploration and Advantica)
[17] MSC PATRAN version 2001 r3, MSC.Software Corporation
[18] ABAQUS/Standard, version 6.4.1, Dassault Systèmes

GL Industrial Services Restricted Page 10

Report Number: R9017
Issue: 4.0

[19] Andrews R.M, Edwards A.M, Espiner R.J., Yates J.R., Howard I.C., Hashemi S.H., ‘Damage mechanics
studies on high strength linepipe’. 15th Joint Technical Meeting on Linepipe Research, Orlando, May 16-20
2005. Arlington, Virginia: Pipeline Research Council International; 2005. Paper 10 (Confidential to PRCI,
EPRG and APIA)
[20] Sadasue, T. et al, ‘Ductile Cracking Evaluation of X80/X100 High Strength Pipelines’, IPC04-0249,
Proceedings of the 2004 ASME, International Pipeline Conference, Calgary, Alberta, Canada. October 4-8,
2004
[21] Anon. ‘Specification for Linepipe’, ANSI/API Specification 5L, Forty-Fourth Edition/ISO 3183:2007
(Modified), Petroleum and Natural Gas Industries – Steel Pipe for Pipeline Transportation, American
Petroleum Institute, October 2007
[22] Andrews, R.M., ‘Operational Trial for X100 Pipelines – Technical Description for DOT OPS’, Advantica
Technical Memorandum, Ref 1009323 DoT OPS, 2006
[23] Andrews, R.M. and Crossley, J., ’Corrosion Defect Design for the BP X100 Operational Trial’, Advantica
Report R9245, Issue 2, December 2006
[24] Ritchie, D. and Last, S., ‘Burst Criteria of Corroded Pipelines – Defect Acceptance Criteria’, Paper 32,
Proceedings of the EPRG/PRCI 10th Biennial Joint Technical Meeting on Line Pipe Research, Cambridge,
UK. 18-21 April 1995

GL Industrial Services Restricted Page 11

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect Type L d YS UTS YS Failure Failure
Mode Pressure
Dt t SMYS SMTS UTS (MPa)
INDEX 255 GL P1V1A X80 60.1 Machined (AG) 3.890 0.775 1.060 1.166 0.808 R 7.6
INDEX 256 GL P1V1B X80 60.1 Machined (AS) 3.877 0.207 1.060 1.166 0.808 R 21.4
INDEX 257 GL P1V2A X80 60.1 Machined (AG) 3.890 0.374 1.073 1.179 0.809 R 17.7
INDEX 258 GL P1V2B X80 60.1 Machined (AG) 3.903 0.089 1.073 1.179 0.809 R 23.3
INDEX 259 GL P2V1A X80 81.8 Machined (AG) 4.538 0.782 1.030 1.149 0.797 R 4.7
INDEX 260 GL P2V1B X80 81.8 Machined (AS) 4.450 0.167 1.030 1.149 0.797 R 15.3
INDEX 261 GL P2V2A X80 81.8 Machined (AG) 4.546 0.395 1.068 1.191 0.797 R 12.0
INDEX 262 GL P2V2B X80 81.8 Machined (AG) 4.523 0.112 1.068 1.191 0.797 R 16.1

Table 1. Test Results on Grade X80 Line Pipe

Notes
1. INDEX numbers are consistent with those used in Reference [11].
2. AG = axial groove and AS = axial slit

GL Industrial Services Restricted Page 12

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect Type L d YS UTS YS Failure Failure
Mode Pressure
Dt t SMYS SMTS UTS (MPa)
INDEX 263 GL HKL-R03 X100 57.7 Machined (P) 146.332 0.111 1.134 1.057 0.976 R 27.0
INDEX 264 GL HKL-R04 X100 57.7 Machined (P) 146.332 0.099 1.134 1.057 0.976 N/A 27.7
INDEX 265 GL HKL-R05 X100 57.7 Machined (P) 146.396 0.101 1.134 1.057 0.976 R 27.5
INDEX 266 GL HKL-R06 X100 57.6 Machined (P) 146.300 0.294 1.134 1.057 0.976 R 21.3
INDEX 267 GL HKL-R07 X100 57.9 Machined (P) 146.588 0.294 1.134 1.057 0.976 R 21.8
INDEX 268 GL HKL-R08 X100 57.9 Machined (P) 146.588 0.287 1.134 1.057 0.976 R 22.0
INDEX 269 GL HKL-R09 X100 57.8 Machined (P) 146.372 0.502 1.134 1.057 0.976 R 15.9
INDEX 270 GL HKL-R10 X100 57.8 Machined (P) 146.404 0.497 1.134 1.057 0.976 R 15.7
INDEX 271 GL HKL-R11 X100 57.8 Machined (P) 146.460 0.502 1.134 1.057 0.976 R 15.9
INDEX 272 GL HKL-R12 X100 57.7 Machined (P) 146.308 0.809 1.134 1.057 0.976 R 6.2
INDEX 273 GL HKL-R13 X100 57.8 Machined (P) 146.492 0.833 1.134 1.057 0.976 R 5.5
INDEX 274 GL HKL-R14 X100 57.8 Machined (P) 146.372 0.814 1.134 1.057 0.976 R 6.4
INDEX 275 GL HKB-R01 X100 63.9 Machined (P) 154.075 0.102 1.134 1.057 0.976 R 23.2
INDEX 276 GL HKB-R02 X100 63.9 Machined (P) 154.171 0.286 1.134 1.057 0.976 R 18.9
INDEX 277 GL HKB-R03 X100 63.9 Machined (P) 154.075 0.503 1.134 1.057 0.976 R 13.2
INDEX 278 GL HKB-R04 X100 63.9 Machined (P) 154.075 0.807 1.134 1.057 0.976 R 5.1
INDEX 279 GL HKL-R15 X100 57.9 Machined (AG) 146.620 0.204 1.134 1.057 0.976 R 25.0
INDEX 280 GL HKL-R16 X100 58.0 Machined (AG) 146.597 0.204 1.134 1.057 0.976 R 25.7
INDEX 281 GL HKL-R17 X100 57.8 Machined (AG) 146.492 0.508 1.134 1.057 0.976 R 16.0
INDEX 282 GL HKL-R18 X100 57.9 Machined (AG) 146.588 0.499 1.134 1.057 0.976 R 16.2

Table 2. Test Results on Grade X100 Line Pipe

Notes
1. INDEX numbers are consistent with those used in Reference [11].
2. P = Patch and AG = axial groove.
3. All results obtained using ring expansion testing.

GL Industrial Services Restricted Page 13

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect Type Failure Failure

L d YS UTS YS Mode Pressure
Dt t SMYS SMTS UTS (MPa)
INDEX 283 GL HKL-R19 X100 57.8 Machined(AG) 146.524 0.810 1.134 1.057 0.976 R 6.3
INDEX 284 GL HKL-R20 X100 57.9 Machined(AG) 146.468 0.811 1.134 1.057 0.976 R 6.3
INDEX 285 GL HKB-R05 X100 63.8 Machined(AG) 154.096 0.207 1.134 1.057 0.976 R 21.8
INDEX 286 GL HKB-R06 X100 63.8 Machined(AG) 153.888 0.504 1.134 1.057 0.976 R 14.3
INDEX 287 GL HKB-R07 X100 63.9 Machined(AG) 154.075 0.818 1.134 1.057 0.976 R 5.1
INDEX 288 GL HKL-R21 X100 57.7 Machined(AS) 146.276 0.099 1.134 1.057 0.976 L 28.6
INDEX 289 GL HKL-R22 X100 57.7 Machined(AS) 146.340 0.102 1.134 1.057 0.976 L 28.2
INDEX 290 GL HKL-R23 X100 57.7 Machined(AS) 146.332 0.301 1.134 1.057 0.976 L 22.5
INDEX 291 GL HKL-R24 X100 57.7 Machined(AS) 146.396 0.306 1.134 1.057 0.976 L 22.1
INDEX 292 GL HKL-R25 X100 57.7 Machined(AS) 146.332 0.488 1.134 1.057 0.976 R 15.1
INDEX 293 GL HKL-R26 X100 57.8 Machined(AS) 146.492 0.507 1.134 1.057 0.976 R 15.5
INDEX 294 GL HKL-R27 X100 57.7 Machined(AS) 146.308 0.804 1.134 1.057 0.976 R 5.6
INDEX 295 GL HKL-R28 X100 57.7 Machined(AS) 146.244 0.808 1.134 1.057 0.976 R 5.7
INDEX 296 GL HKB-R08 X100 63.7 Machined(AS) 153.851 0.111 1.134 1.057 0.976 L 24.6
INDEX 297 GL HKB-R09 X100 63.8 Machined(AS) 154.059 0.309 1.134 1.057 0.976 L 19.4
INDEX 298 GL HKB-R10 X100 63.4 Machined(AS) 153.444 0.493 1.134 1.057 0.976 R 14.2
INDEX 299 GL HKB-R11 X100 63.8 Machined(AS) 153.888 0.769 1.134 1.057 0.976 R 5.1
INDEX 300 GL HKL V01 X100 57.9 Machined (P) 3.503 0.496 1.134 1.057 0.976 R 18.1
INDEX 301 GL HKK V01 X100 57.9 Machined (P) 6.384 0.500 1.134 1.057 0.976 R 15.4
INDEX 302 GL HKL V02 X100 57.9 Machined(AG) 2.962 0.503 1.134 1.057 0.976 R 17.9
INDEX 303 GL HKK V02 X100 57.8 Machined(AG) 5.825 0.500 1.134 1.057 0.976 R 15.0

Table 3. Test Results on Grade X100 Line Pipe

Notes
1. INDEX numbers are consistent with those used in Reference [11].
2. P = Patch; AG = axial groove; and AS = axial slit.
3. INDEX 300, 301, 302 and 303 results obtained using vessel tests. The remaining results obtained using ring expansion testing.

GL Industrial Services Restricted Page 14

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect L d/t ASME B31G Mod ASME RSTRENG LPC-1 FE
Type B31G
Dt (actual) PA/Pf PA/Pf PA/Pf PA/Pf PA/Pf
INDEX 255 GL P1V1A X80 60.1 Machined 3.890 0.775 0.670 0.853 1.232 1.088 -
INDEX 256 GL P1V1B X80 60.1 Machined 3.877 0.207 1.183 1.186 1.220 1.173 -
INDEX 257 GL P1V2A X80 60.1 Machined 3.890 0.374 1.090 1.131 1.208 1.138 -
INDEX 258 GL P1V2B X80 60.1 Machined 3.903 0.089 1.210 1.195 1.207 1.176 -
INDEX 259 GL P2V1A X80 81.8 Machined 4.538 0.782 1.443 0.745 1.099 0.993 -
INDEX 260 GL P2V1B X80 81.8 Machined 4.450 0.167 1.128 1.127 1.152 1.120 -
INDEX 261 GL P2V2A X80 81.8 Machined 4.546 0.395 1.340 1.080 1.164 1.106 -
INDEX 262 GL P2V2B X80 81.8 Machined 4.523 0.112 1.221 1.139 1.155 1.130 -
Table 4. Comparison of Failure Predictions for the Grade X80 Tests
Notes
1. INDEX numbers are consistent with those used in Reference [11].

GL Industrial Services Restricted Page 15

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect L d/t ASME B31G Mod ASME RSTRENG LPC-1 FE
Type B31G
Dt (actual) PA/Pf PA/Pf PA/Pf PA/Pf PA/Pf
INDEX 263 GL HKL-R03 X100 57.7 Machined 146.332 0.111 1.164 1.141 1.162 1.142 -
INDEX 264 GL HKL-R04 X100 57.7 Machined 146.332 0.099 1.168 1.147 1.166 1.146 -
INDEX 265 GL HKL-R05 X100 57.7 Machined 146.396 0.101 1.165 1.144 1.164 1.144 -
INDEX 266 GL HKL-R06 X100 57.6 Machined 146.300 0.294 1.146 1.077 1.145 1.122 0.922
INDEX 267 GL HKL-R07 X100 57.9 Machined 146.588 0.294 1.179 1.108 1.177 1.154 -
INDEX 268 GL HKL-R08 X100 57.9 Machined 146.588 0.287 1.178 1.109 1.176 1.153 -
INDEX 269 GL HKL-R09 X100 57.8 Machined 146.372 0.502 1.217 1.056 1.216 1.188 -
INDEX 270 GL HKL-R10 X100 57.8 Machined 146.404 0.497 1.192 1.037 1.191 1.164 -
INDEX 271 GL HKL-R11 X100 57.8 Machined 146.460 0.502 1.215 1.055 1.215 1.187 -
INDEX 272 GL HKL-R12 X100 57.7 Machined 146.308 0.809 Invalid Invalid Invalid 1.196 1.028
INDEX 273 GL HKL-R13 X100 57.8 Machined 146.492 0.833 Invalid Invalid Invalid 1.220 -
INDEX 274 GL HKL-R14 X100 57.8 Machined 146.372 0.814 Invalid Invalid Invalid 1.265 -
INDEX 275 GL HKB-R01 X100 63.9 Machined 154.075 0.102 1.091 1.071 1.089 1.072 0.989
INDEX 276 GL HKB-R02 X100 63.9 Machined 154.171 0.286 1.114 1.050 1.113 1.093 -
INDEX 277 GL HKB-R03 X100 63.9 Machined 154.075 0.503 1.118 0.970 1.117 1.094 0.950
INDEX 278 GL HKB-R04 X100 63.9 Machined 154.075 0.807 Invalid Invalid Invalid 1.086 -
INDEX 279 GL HKL-R15 X100 57.9 Machined 146.620 0.204 1.199 1.153 1.198 1.175 0.951
INDEX 280 GL HKL-R16 X100 58.0 Machined 146.597 0.204 1.235 1.188 1.234 1.211 -
INDEX 281 GL HKL-R17 X100 57.8 Machined 146.492 0.508 1.238 1.071 1.237 1.209 -
INDEX 282 GL HKL-R18 X100 57.9 Machined 146.588 0.499 1.230 1.069 1.229 1.201 -
Table 5. Comparison of Failure Predictions for the Grade X100 Tests
Notes
1. INDEX numbers are consistent with those used in Reference [11].

GL Industrial Services Restricted Page 16

Report Number: R9017
Issue: 4.0

INDEX Source Reference Grade D/t Defect L d/t ASME B31G Mod ASME RSTRENG LPC-1 FE
Type B31G
Dt (actual) PA/Pf PA/Pf PA/Pf PA/Pf PA/Pf
INDEX 283 GL HKL-R19 X100 57.8 Machined 146.524 0.810 Invalid Invalid Invalid 1.235 1.005
INDEX 284 GL HKL-R20 X100 57.9 Machined 146.468 0.811 Invalid Invalid Invalid 1.235 -
INDEX 285 GL HKB-R05 X100 63.8 Machined 154.096 0.207 1.156 1.111 1.154 1.135 -
INDEX 286 GL HKB-R06 X100 63.8 Machined 153.888 0.504 1.212 1.051 1.212 1.186 0.984
INDEX 287 GL HKB-R07 X100 63.9 Machined 154.075 0.818 Invalid Invalid Invalid 1.155 -
INDEX 288 GL HKL-R21 X100 57.7 Machined 146.276 0.099 1.206 1.185 1.205 1.184 -
INDEX 289 GL HKL-R22 X100 57.7 Machined 146.340 0.102 1.196 1.174 1.194 1.174 0.966
INDEX 290 GL HKL-R23 X100 57.7 Machined 146.332 0.301 1.223 1.148 1.222 1.198 -
INDEX 291 GL HKL-R24 X100 57.7 Machined 146.396 0.306 1.213 1.136 1.212 1.187 -
INDEX 292 GL HKL-R25 X100 57.7 Machined 146.332 0.488 1.118 0.978 1.118 1.092 -
INDEX 293 GL HKL-R26 X100 57.8 Machined 146.492 0.507 1.197 1.036 1.196 1.169 -
INDEX 294 GL HKL-R27 X100 57.7 Machined 146.308 0.804 Invalid Invalid Invalid 1.058 0.863
INDEX 295 GL HKL-R28 X100 57.7 Machined 146.244 0.808 Invalid Invalid Invalid 1.090 -
INDEX 296 GL HKB-R08 X100 63.7 Machined 153.851 0.111 1.163 1.140 1.161 1.143 -
INDEX 297 GL HKB-R09 X100 63.8 Machined 154.059 0.309 1.182 1.107 1.181 1.159 -
INDEX 298 GL HKB-R10 X100 63.4 Machined 153.444 0.493 1.169 1.020 1.169 1.144 0.896
INDEX 299 GL HKB-R11 X100 63.8 Machined 153.888 0.769 0.931 0.621 0.931 0.909 -
INDEX 300 GL HKL V01 X100 57.9 Machined 3.503 0.496 0.931 1.021 1.136 1.045 1.027
INDEX 301 GL HKK V01 X100 57.9 Machined 6.384 0.500 1.175 0.927 1.047 0.999 1.048
INDEX 302 GL HKL V02 X100 57.9 Machined 2.962 0.503 0.909 0.992 1.101 1.001 1.299
INDEX 303 GL HKK V02 X100 57.8 Machined 5.825 0.500 1.145 0.897 1.012 0.960 1.087

Table 6. Comparison of Failure Predictions for the Grade X100 Tests

Notes
1. INDEX numbers are consistent with those used in Reference [11].

GL Industrial Services Restricted Page 17

Report Number: R9017
Issue: 4.0

W L
d
r

r is the groove
blend radius
t

Figure 1. Defect Dimensions

GL Industrial Services Restricted Page 18

Report Number: R9017
Issue: 4.0

Figure 2. Typical 3D Quarter Symmetry FE Model of a Pipeline with an Axial Groove Defect

GL Industrial Services Restricted Page 19

Report Number: R9017
Issue: 4.0

Groove

Patch

X
Slit

Figure 3. 2D Plane Strain FE Models of the Ring Expansion Tests

GL Industrial Services Restricted Page 20

Report Number: R9017
Issue: 4.0

1000
140

900

120
800

700 100
Nominal Stress (MPa)

Nominal Stress (ksi)

600
80
500
X65 32in x 19.05mm RB
400 X80 36in x 20mm RB (T) 60
X80 36in x 20mm RB (T)
X80 36in x 20mm RB (L)
300 X80 36in x 20mm RB (L)
X100 36in x 14.9mm RB (T) 40
X100 36in x 14.9mm RB (T)
200 X100 36in x 19mm RB (T)
X80 IPC 2004
X100 IPC 2004 20
100

0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
Nominal S train (%)

Figure 4. Stress versus Strain Curves for Grade X65, X80 and X100 Line Pipe

900
140

800
120
700
Nominal Stress (MPa)

100

Nominal Stress (ksi)

600

500 80

400
60

300
40
X65 32in x 19.05mm RB
200 X80 36in x 20mm RB (L)
X100 36in x 14.9mm RB (T)
20
100 X100 52in x 20.6mm RB (T)

0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
Nom inal Strain (%)

Figure 5. True Stress versus True Strain Curves for Grade X65, X80 and X100 Line Pipe

GL Industrial Services Restricted Page 21

Report Number: R9017
Issue: 4.0

numerical instability
1200

1100 Inner Surface

M id Surface
von Mises Equivalent Stress (MPa)

1000 Outer Surface

900

800

700 true ultimate tensile strength (UTS)

600 Stage 2 - Plasticity Spreading
500

400

300

200

100 Stage 1 - Elastic Deformation

0
0 5 10 15 20 25 30 35

Internal Pressure (MPa)

Figure 6. von Mises Equivalent Stress Variation Through Ligament with Increasing Internal Pressure

GL Industrial Services Restricted Page 22

Report Number: R9017
Issue: 4.0

Figure 7. Typical von Mises Equivalent Stress Contour Plot for a Pipe with an Axially Orientated Groove Defect

GL Industrial Services Restricted Page 23

Report Number: R9017
Issue: 4.0

1.6
Actual Failure Pressure/Predicted Failure

1.4

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 8. Comparison of Actual and Predicted Failure Pressures using the ASME B31G Method

GL Industrial Services Restricted Page 24

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 9. Comparison of Actual and Predicted Failure Pressures using the Modified ASME B31G Method

GL Industrial Services Restricted Page 25

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 10. Comparison of Actual and Predicted Failure Pressures using the RSTRENG Method

GL Industrial Services Restricted Page 26

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 11. Comparison of Actual and Predicted Failure Pressures using the LPC-1 Method

GL Industrial Services Restricted Page 27

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4

0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 12. Comparison of Actual and Predicted Failure Pressures using the Non-Linear Finite Element Method (Grade X100 Test Points)

GL Industrial Services Restricted Page 28

Report Number: R9017
Issue: 4.0

1.6
Actual Failure Pressure/Predicted Failure

1.4

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 13. Comparison of Actual and Predicted Failure Pressures using the ASME B31G Method (Flow Stress Modified to Equal the Mean of the Specified Minimum Yield Strength and the Specified
Minimum Ultimate Tensile Strength)

GL Industrial Services Restricted Page 29

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 14. Comparison of Actual and Predicted Failure Pressures using the Modified ASME B31G Method (Flow Stress Modified to Equal the Mean of the Specified Minimum Yield Strength and the
Specified Minimum Ultimate Tensile Strength)

GL Industrial Services Restricted Page 30

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 15. Comparison of Actual and Predicted Failure Pressures using the RSTRENG Method (Flow Stress Modified to Equal the Mean of the Specified Minimum Yield Strength and the Specified
Minimum Ultimate Tensile Strength)

GL Industrial Services Restricted Page 31

Report Number: R9017
Issue: 4.0

1.4
Actual Failure Pressure/Predicted Failure

1.2

1
Pressure

0.8

0.6

0.4
X80
0.2 X100
1.0 Line
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Defect Depth (d/t)

Figure 16. Comparison of Actual and Predicted Failure Pressures using the LPC-1 Method (Flow Stress Modified to Equal the Mean of the Specified Minimum Yield Strength and the Specified Minimum
Ultimate Tensile Strength)

GL Industrial Services Restricted Page 32