You are on page 1of 15

Proceedings of the 2012 9th International Pipeline Conference

IPC2012
September 24-28, 2012, Calgary, Alberta, Canada

IPC2012-90404

EVALUATION OF GIRTH WELD DEFECT ACCEPTANCE CRITERIA FOR GRADE


X100 PIPELINES

Troy Swankie and Vinod Chauhan Ian Wood


GL Noble Denton Electricore, Inc.
Loughborough, LE11 3GR, UK Valencia, CA 91355, USA

Richard Espiner Max Kieba Spencer Quong


BP Exploration US DOT, PHMSA Quong & Associates, Inc.
Sunbury, TW16 7LN, UK Washington DC 20590, USA San Francisco, CA 94115, USA

Noble Denton’s Spadeadam test facility, Cumbria, UK. The


ABSTRACT girth welds were selected to enable the effects of material
There are a number of methods that are commonly used for variability between abutting pipes, different heats and different
the assessment of a girth weld containing a ‘fabrication’ defect. manufacturers (pipe was sourced from two world class pipe
These range from the more generic workmanship limits mills, with the plate supply for one mill coming from two
through to more complex pipeline specific Engineering Critical sources) to be investigated.
Assessment (ECA) methodologies. A substantial test program has been undertaken to fully
The workmanship limits stipulated in pipeline design characterize the mechanical properties of each girth weld,
codes can be very conservative, resulting in un-necessary and comprising curved wide plate (CWP), tensile, Charpy impact
costly repairs. The ECA approach is being increasingly used to and fracture mechanics tests.
derive girth weld defect acceptance limits specific to a pipeline. The results from the CWP tests have been analyzed using
These limits have been derived using either semi-analytical the procedures given in API 1104 (Option 2), EPRG, CSA
methods or from the results of large-scale tests conducted on Z662, BS 7910 and API 579-1/ASME FFS-1.
pipeline girth welds. However, at present there is no one This paper presents an overview of the tests undertaken
standardized method. and a comparison of the actual test results with the predictions
The guidance produced by the European Pipeline Research from the assessment methods.
Group (EPRG) is an example of an established methodology
based on the results of large-scale tests, while commonly used INTRODUCTION
pipeline specific semi-analytical assessment methods include There are a number of methods that are commonly used for
API 1104 and CSA Z662. Other commonly used analytical developing defect acceptance criteria for pipeline girth welds;
methods, which are more generic in application, include API 1104[1], CSA Z662[2] and EPRG[3], which are pipeline
BS 7910 and API 579-1/ASME FFS-1. Application of these specific and BS 7910[4] and API 579-1/ASME FFS-1[5], which
methods to girth welds in grade X100 pipelines has not yet can be applied to a pipeline but are more generic in application.
been validated. The applicability of these methods to girth welds in higher
The US Department of Transportation, Pipeline and strength steel pipelines is not yet verified. The work undertaken
Hazardous Materials Safety Administration (PHMSA) was aimed at identifying whether these methods could
commissioned Electricore, Inc and GL Noble Denton to potentially be used for grade X100 pipelines.
investigate the applicability of these ‘commonly used’ girth To fully assess the limitations of these methods a number
weld assessment procedures to grade X100 pipelines. of variables would need to be investigated, for example, a
To facilitate this project, BP provided 10 girth welds from range of pipe sizes (diameter and wall thickness), different pipe
a full-scale operational trial of two grade X100 48in diameter manufacturing methods, different weld preparations and weld
pipeline test sections, following completion of the trial at GL procedures.

1 Copyright © 2012 by ASME


The project was provided with 10 girth welds from the BP SOURCE OF MATERIAL FOR TESTING
X100 operational trial, on completion of the test. Despite only The girth welds for testing were provided by BP from a
one pipe size being available, 1220x19.8mm the pipe was recently completed full-scale operational field trial of a pipeline
sourced from two different pipe mills with the original plate constructed from grade X100 line pipe. The ‘X100 operational
coming from three different suppliers. Welding of the test trial’ involved laying 800m of 1220m diameter, 19.8mm
pipeline ensured that abutting pipes from different mill/plate nominal wall thickness pipe in two sections; Section A was
sources were welded together. Furthermore, the differences in 600m in length and Section B was 200m.
line pipe properties ensured that a range in weld metal yield A comprehensive overview of the construction of the
strength mismatch conditions was achieved from under to over pipeline sections was presented at IPC2010[7]. The pertinent
matching. Although the majority of the welds tested were points of the construction are summarized below:
produced using a main line tandem GMAW weld procedure, • The pipeline test sections were designed to ASME
the project also benefitted from availability of a weld from a B31.8:2003[8] to a design factor of 0.8, with additional
tie-in weld procedure which also had a different weld guidance from CSA Z662:2003[9].
preparation compared with the main line weld. • Two pipe mills produced the test pipes from three plate
Each weld was subjected to a series of small-scale tests to suppliers (A, B, C) and arrangement of each pipe section
fully characterize its mechanical properties. Based on those took account of the need to ensure that both sources were
results 30 CWP specimens with either a machined defect, a equally subjected to all the test conditions prescribed.
natural welding defect or a deliberate welding defect, were • Pipeline welding was based on API 1104:2005[10], with
prepared and tested. The results of each test were assessed supplementary requirements as specified by BP.
using the different girth weld assessment methods, using the • Section A (58 girth welds) was constructed according to
results of the small-scale test program as inputs. good pipeline practice to simulate a pipeline in normal
For detailed information on the different assessment service with a normal CP protection level.
methods the reader is directed to the appropriate pipeline code, • Section B (21 girth welds) incorporated numerous
guidelines or national standard. This paper focuses on the tests instances of pipeline defects and damage and the CP level
undertaken and the corresponding assessments, full details of was varied along its length;
which are given in the GL Noble Denton project report[6] which o The first third of the test section had no CP applied.
is available from the US Department of Transportation. o The middle third had CP applied at an intermediate
potential of -850 to -950mV.
NOMENCLATURE o The final third had CP applied at a high potential of
AUT - Automated Ultrasonic Testing -1200 to -1300mV.
B - Thickness of the fracture mechanics specimen • Each girth weld was inspected during construction using
CP - Cathodic Protection X-radiography (a single wall single image panoramic
CTOD - Crack tip opening displacement technique) and AUT (using a phased array system).
CWP - Curved wide plate test specimen • The pipeline sections were subjected to a hydrostatic
FAC - Failure assessment curve pressure test on commission, to the requirements of ASME
FAD - Failure assessment diagram B31.8; test pressure of 225 bar, equivalent to 1.25 x design
GMAW - Gas metal arc welding pressure.
HAZ - Heat affected zone • Both pipeline test sections were subjected to accelerated
J - J-Integral pressure cycling for a 2 year period to simulate a 40 year
K - Stress intensity factor design life plus a 20% life extension (i.e., 48 years total
Kr - Ratio of the applied stress intensity factor to the
service).
materials fracture toughness
In total there were 79 girth welds to choose from. The
Lr - Ratio of applied load to yield load
following criteria were used to select the 10 welds considered
Pop-in - A discontinuity in the force versus
to be of most benefit to the project:
displacement record during a fracture
• Weld type (i.e., mainline or tie-in).
mechanics test
• The level of yield strength mismatch between the weld and
RB - Round bar test specimen
abutting pipes.
Rm - Tensile strength
Rp0.2 - Yield strength • Pipe supplier and pipe heat number.
SENB - Single edge notch bend test specimen • Results of the X-radiography and AUT inspections so that
uEL - Uniform elongation consideration could be given to the type and size of defects
WM - Weld metal present.
Y/T - Yield to Tensile strength ratio The 10 girth welds selected for testing are summarized in
Table 1.

2 Copyright © 2012 by ASME


Weld ID Weld Type Source Comment for detailed tensile testing, using different types of test
A06 Mainline B to B Same heat specimen, at different through wall locations.
A17 Tie-in C to C Same heat At equi-spaced increments around the circumference of
A33 Mainline B to A weld B10, all WM specimens were extracted from the weld
A44 Mainline C to B root (11 RB specimens), weld cap (11 RB specimens) and weld
A46 Mainline C to C mid-thickness (11 RB specimens). At the same circumferential
A50 Mainline B to B positions, 11 prismatic specimens sampling almost the full weld
B03 Mainline C to C thickness were extracted. All testing was undertaken at ambient
laboratory temperature (approximately 20°C).
B06 Mainline A to B
B08 Mainline C to C
TEST RESULTS
B10 Mainline A to C
Weld macro-sections
Table 1: Summary of girth welds selected for testing
In general, three macro-sections, transverse to the girth
weld, were extracted from each weld, adjacent to where each
TEST PROGRAM CWP was to be extracted. Each section was ground, polished
A comprehensive test program has been undertaken, which and etched (2% nital solution) to reveal the weld, HAZ and
comprised of a series of small-scale tests (tensile, Charpy and surrounding microstructure, and then photographed. Typical
fracture mechanics tests) and mid-scale CWP tests to fully weld macro-section photographs are presented in Figure 1.
characterize the grade X100 line pipe and girth weld and enable
an assessment of the validity of the pipeline specific girth weld
assessment procedures in API 1104 (Option 2), CSA Z662 and
EPRG (Tier 2) and the suitability of the more generic fracture
mechanics assessment procedures in BS 7910 and API
579-1/ASME FFS-1. The small-scale test program comprised
the following:
• Weld macro-sections - generally at the weld 12, 3 and 6
o’clock positions.
(a) Mainline weld A33 (b) Tie-in weld A17
• Vickers hardness surveys – three traverses per weld macro-
Figure 1: Examples of the mainline and tie-in weld profiles
section; 1.5mm below the weld cap, at the pipe mid wall
thickness, and 1.5mm up from the weld root.
In addition, for weld B10, macro-sections were extracted at
• Tensile tests – full stress strain curves, measured at
each of the 11 positions around the weld circumference that the
ambient laboratory temperature (approximately 20°C);
tensile specimens were extracted.
o Parent pipe – ‘full thickness’ strap specimens, pipe
The reinforcement of the weld cap and weld root was
longitudinal and transverse directions.
measured and found not to exceed 3mm (the upper limit
o Parent pipe – round bar specimens, pipe transverse
specified in API 1104) and 2mm respectively.
direction.
With the exception of the tie-in weld (A17), weld bead
o All WM – round bar specimens, sampling the weld
height was found to be dependent on sampling position; around
root and cap regions.
the 3 and 9 o’clock positions the weld beads were thinner
o Cross weld full-thickness strap specimens.
towards the outer surface of the pipe with corresponding
• Charpy impact tests – full size specimens, notched through
changes in the amount of grain refined WM. The bead height of
thickness in the weld direction at the WM centerline and
weld A17 showed little variation around the weld
the HAZ (50/50), tested at -20°C.
circumference.
• Fracture toughness tests – rectangular (Bx2B) SENB Vickers hardness surveys
specimens, notched through thickness in the weld direction Hardness measurements were undertaken according to EN
at the WM centerline and the HAZ (50/50), tested at 1043-1:1996[11] and ISO 6507-2[12]. A typical set of hardness
-20°C. measurements are shown in Figure 2; these being taken at the
The type, position (around the pipe circumference) and 12:00 position of weld A06.
number of tests undertaken was not the same for each weld.
Depending on the results of some of the initial tensile tests
undertaken, either 3 or 4 CWP specimens were extracted from
a weld. Consequently, this impacted on the number and
position of the remaining small-scale tests. Furthermore, due to
the variation in all WM tensile properties that was observed
around the circumference of the weld, weld B10 was utilized

3 Copyright © 2012 by ASME


400 to show the variation in the properties around the weld
Pipe 1 HAZ-1 WM HAZ-2 Pipe 2 circumference.
Hardness, HV10

Rp0.2 Rm uEL
300
N/mm² N/mm² Y/T %
Weld cap (round bar specimen)
min (max) 758 (853) 856 (891) 0.89 (0.96) 7.8 (9.8)
avg 819 873 0.94 9.0
STDV 29 13 0.02 0.7
200
0 2 4 6 8 10 12 Weld root (round bar specimen)
Indent min (max) 815 (876) 876 (919) 0.93 (0.96) 5.2 (7.7)
Weld cap Mid thickness Weld root avg 855 899 0.95 6.6
STDV 19 12 0.01 0.7
Figure 2: Example of a Vickers hardness survey (weld A06, Full-thickness (rectangular specimen)
12:00 position). min (max) 811 (886) 873 (914) 0.93 (0.97) 6.3 (7.6)
avg 845 890 0.95 7.0
A summary of the results of the hardness surveys is given STDV 24 14 0.01 0.4
below (values are HV10): Notes: min, max and avg are the minimum, maximum and average value from
• Average pipe metal hardness ranged from 295-297 (Pipe the set of specimens, and STDV is the standard deviation.
A), 272-300 (Pipe B) and 271-291 (Pipe C). Table 2: Weld B10: Summary tensile test results.
• Average HAZ hardness of the mainline welds ranged from
279-282 (Pipe A), 263-292 (Pipe B) and 252-272 (Pipe C). Similar results and observations have been
Tie-in weld A17 was lower, ranging from 248-251. reported[14][15][16], but the location of minimum and maximum
• The average pipe metal hardness was consistently higher strength is seen to vary around the pipe circumference. Hence
than the average HAZ hardness, inferring HAZ softening for these tests it was considered important to characterize the
in all welds. Pipe C was found to have a higher sensitivity circumferential variation in tensile properties specific to the
to HAZ softening. welds being tested.
• Average WM hardness of the mainline welds ranged from The test results showing the variation in yield strength
290-296, which was similar, if not greater than the around the weld circumference are shown in Figure 3. As can
corresponding pipes. be seen, there is a difference in strength of approximately
• The average WM hardness of the tie-in weld (A17) was 100N/mm² around the weld circumference; the lowest values
much lower than the mainline welds. were measured at approximately the weld 12:00 and 06:00
In addition, a through-thickness hardness survey was positions, the highest at approximately the weld 03:00 and
undertaken at the WM centre-line of weld B10; measurements 09:00 positions. Comparing the RB specimen results, the yield
were made at 1.5mm increments, starting 1mm up from the strength of the weld cap region was consistently lower than that
weld root. The hardness ranged from 266-331, the lowest measured in the weld root region. The yield strength measured
values being measured in the weld cap region, and the highest from the rectangular specimen was similar to that measured in
individual peaks occurred at the weld root and near weld mid- the weld root (the results are consistent with the results of the
thickness. It is likely that the reduction in hardness in the near hardness survey; highest hardness/strength in the weld
weld cap region is attributed to the lower cooling rate (or the root/near weld mid-thickness region, lowest in the near weld
increase in welding cooling time) associated with the welding cap region).
process. The variation in tensile strength showed a similar trend;
Tensile tests the weld 12:00 and 06:00 positions giving the lowest values,
The test procedure in BS EN 10002-1[13] was used but the variation in strength was less; approximately 60N/mm².
throughout for each different type and size of specimen tested. As a consequence of this, there was also a sizeable variation in
In total, 217 tests have been undertaken to fully characterize Y/T ratio (0.92 to 0.97) around the weld circumference, as
the stress-strain behavior of the weldments. shown in Figure 4.
The detailed testing of weld B10 showed that there is a
significant variation in strength around the weld circumference.
The test results are summarized in Table 2 in terms of
maximum and minimum measured values, average value and
standard deviation for each specimen type and through wall
position sampled by the specimen. The individual quantities of
yield and tensile strength are presented in Figure 3 and Figure 4

4 Copyright © 2012 by ASME


950
the circumferential position of the specimen. For the WM
specimens (RB specimen type) two types of behavior were
Yield strength, Rp0.2 (N/mm²)

observed; ‘double n’ and discontinuous yielding with a Lüders


plateau was observed. The response of the parent pipe
specimens (full thickness flat strap specimens) was similar,
850 resulting in a smooth stress-strain curve. However, the strain
capacity of the pipes is seen to vary significantly depending on
circumferential position.

1000 P1 (12:00) - L
750 P1 (03:00) - L
0 3 6 9 12 P1 (06:00) - L

P1 (08:00) - L
Weld circumferential position (o'clock)
P1 (08:30) - L

Engineering stress (N/mm²)


Weld cap (Round Bar) Weld root (Round Bar) Full-thickness (Rectangular) P2 (12:00) - L

Figure 3: Weld B10: Circumferential variation of yield 800


P2 (03:00) - L

P2 (06:00) - L
strength measured using RB specimens (cap and P2 (08:00) - L
root region) and rectangular full thickness P2 (08:30) - L

specimens. WM cap (00:30)


WM cap (02:45)
WM cap (06:15)

1.00 WM cap (08:00)


600
WM root (07:45)
0 2 4 6 8 10 WM root (09:30)
Engineering strain (%)

Notes: P1, P2 - pipe 1, pipe 2


L - pipe longitudinal specimen orientation
Y/T ratio

All parent pipe (P1, P2) specimens were full-thickness strap


0.95
All WM specimens were round bar
Figure 5: Weld A06: Circumferential variation of
engineering stress-strain response (specimen
location referenced to the weld 12:00 position).
0.90
0 3 6 9 12 The corresponding values of Rp0.2 and Rm are presented in
Weld circumferential position (o'clock)
Figure 6. The WM data appear to follow a sinusoidal trend,
comparable to the detailed study of weld B10; Rp0.2 is a
Weld cap (Round Bar) Weld root (Round Bar) Full-thickness (Rectangular)
maximum at near 03:00 and 09:00, and a minimum at near
Figure 4: Weld B10: Circumferential variation of Y/T 12:00 and 06:00.
measured using RB specimens (cap and root From the data presented in Figure 6, it can be deduced that
region) and rectangular full thickness specimens. the Y/T ratios for P1 and P2 were similar, although P2
exhibited a higher degree of scatter; P1 ranged from 0.91 to
The remaining 9 welds were subjected to detailed testing 0.93, and P2 from 0.90 to 0.94. In general the Y/T ratios of the
of the parent pipes and WM to characterize the tensile WM (from 0.91 to 0.97) were found to be higher than either
properties of each material, around the circumference of the pipe.
pipe/weld. For the pipe, tests were undertaken using RB and Based on the data presented in Figure 6, the weld can
full thickness flat strap specimens in both the pipe longitudinal either be considered to under-match or over-match the pipe
and transverse directions. The WM tests were undertaken using material depending on the circumferential position of the
RB specimens, sampling the weld root and weld cap regions. specimen(s). This is seen more clearly in Figure 7 where the
For each test the stress-strain response was measured and the variation in Rp0.2 mismatch between the pipes and WM is
corresponding values of Rp0.2, Rm, Y/T and uEL were presented. The level of mismatch is calculated based on the
determined. minimum, maximum and average values of Rp0.2 from each set
Addition tests were undertaken using cross-weld full of parent pipe (P1 and P2) and WM specimens. For P1 the
thickness flat strap specimens to identify the weakest part of variation in mismatch can be as much as 16% (from -1 to
the weldment and corresponding failure stress. +15%) and 19% for P2 (from -5 to +14%).
Some of the results obtained from weld A06 are presented
in Figure 5 to Figure 7 to provide an illustration of the type of
data and trends observed.
As can be seen in Figure 5 the magnitude and shape of the
individual stress-strain curves was found to differ depending on

5 Copyright © 2012 by ASME


1000 Yield strength Tensile strength measured in the weld root and mid thickness regions,
Pipe 1 Weld metal Pipe 2 lowest at the weld cap.
• The properties of the line pipe varied greatly between the
different pipe manufacturers and plate sources; although
Strength (N/mm²)

800
not one consistently achieved the highest average values of
yield or tensile strength, Y/T ratio or elongation.
• The variation in strength observed between the different
pipe manufacturers and plate sources resulted in a wide
range of weld metal strength mismatch, ranging from 11%
600 WM root (07:45)
under-matching to 26% over-matching. API 1104, CSA

WM root (09:30)
WM cap (00:30)

WM cap (02:45)

WM cap (06:15)

WM cap (08:00)
P1 (12:00) - L

P1 (03:00) - L

P1 (06:00) - L

P1 (08:00) - L

P1 (08:30) - L

P2 (12:00) - L

P2 (03:00) - L

P2 (06:00) - L

P2 (08:00) - L

P2 (08:30) - L
Z662 and EPRG require that the weld metal yield strength
is not less than the yield strength of the line pipe.
Charpy impact energy tests
Notes: P1, P2 - pipe 1, pipe 2 Each specimen was prepared and tested according to BS
L - pipe longitudinal specimen orientation
EN 10045-1[18]. The Charpy impact energy and percentage
All parent pipe (P1, P2) specimens were full-thickness strap
All WM specimens were round bar shear area was measured from each tested specimen. A total of
Figure 6: Weld A06: Circumferential variation of yield 108 tests have been undertaken.
(Rp0.2) and tensile strength (specimen location The results from the specimens notched at the WM centre-
referenced to the weld 12:00 position). line are presented in Figure 8, in relation to the weld ID and the
position around the weld circumference that the specimens
20 were extracted. The data are presented as the minimum,
WM strength over-match maximum and average values from a set of 3 full-sized Charpy
specimens. The results from the specimens notched at the HAZ
(50/50) are presented in an identical manner in Figure 9.
Yield strength mismatch (%)

10

300
min
Full size Charpy energy (J)

0 max
avg
200

WM strength under-match
-10 100
P1(max) : WM(min)

P1(avg) : WM(min)

P2(max) : WM(min)

P2(avg) : WM(min)
P1(min) : WM(min)

P2(min) : WM(min)

P1(avg) : WM(max)
P1(min) : WM(max)

P1(max) : WM(max)

P2(min) : WM(max)

P2(max) : WM(max)

P2(avg) : WM(max)

P1(max) : WM(avg)

P1(avg) : WM(avg)

P2(max) : WM(avg)

P2(avg) : WM(avg)
P1(min) : WM(avg)

P2(min) : WM(avg)

0
B 06 (04:00)

A 06 (06:00)
A 33 (06:00)
A 44 (06:00)
A 46 (06:00)
A 50 (06:00)
B 03 (06:00)
B 08 (06:00)

A 06 (12:00)
A 33 (12:00)
A 44 (12:00)
A 46 (12:00)
A 50 (12:00)
B 03 (12:00)
B 06 (12:00)
B 08 (12:00)

A 17 (06:00)
A 17 (12:00)
Figure 7: Weld A06: Variation in yield strength (Rp0.2)
mismatch.
Figure 8: Charpy impact tests: all WM test results (tested at
A summary of the results and observed trends from the -20°C).
tensile tests undertaken on all 9 girth welds is given below:
300
• All line pipe achieved the specified minimum yield and min
Full size Charpy energy (J)

max
tensile strength requirements of the line pipe specification, avg
ANSI/API 5L[17]. 200

• The stress-strain response of the line pipe in the pipe


longitudinal direction was similar, unlike the response of 100
the line pipe in the transverse direction, where the post
yield behavior was found to vary considerably. 0
• The properties were found to vary significantly depending
B 06 (04:00)

A 06 (06:00)
A 33 (06:00)
A 44 (06:00)
A 46 (06:00)
A 50 (06:00)
B 03 (06:00)
B 08 (06:00)

A 06 (12:00)
A 33 (12:00)
A 44 (12:00)
A 46 (12:00)
A 50 (12:00)
B 03 (12:00)
B 06 (12:00)
B 08 (12:00)

A 17 (06:00)
A 17 (12:00)

on the type of test specimen; round bar or flat strap.


• The properties of the weld metal varied significantly
Figure 9: Charpy impact tests: all HAZ (50/50) test results
around the pipe circumference, showing a sinusoidal trend;
(tested at -20°C).
yield strength was lowest at approximately the weld 06:00
and 12:00 positions and highest at approximately the 03:00
To summarize, each weld achieves the performance
and 09:00 positions. The strength was also observed to
criterion specified in API 1104, CSA Z662 and EPRG, where
vary through the weld thickness; the highest strength
the minimum and average Charpy impact energy of the

6 Copyright © 2012 by ASME


weldment is to be greater than 30(40)J respectively (note, CSA Two of the tests failed to achieve the minimum CTOD
Z662 specifies 40J, but does not state whether this should be a requirement of API 1104.
minimum or average value). Although not a requirement of
CSA Z662 and EPRG, API 1104 requires that the percentage 0.40
shear area should be greater than 50%. Although the shear area Type 'c'
data are not presented here, only one specimen from the 108 Type 'u'
tests undertaken failed this requirement (a value of 45% was 0.30 Type 'm'
estimated from the fracture surface examination).

CTOD (mm)
Fracture mechanics tests
0.20
Each specimen was prepared and tested in three-point-
bend loading according to BS 7448:Part 2[19]. To ensure
accuracy of notch placement, the sides of each specimen were 0.10
ground to a fine finish and etched (5% nital) to reveal the weld
profile/microstructure. After notching, each specimen was
locally compressed and fatigue pre-cracked at ambient 0.00
laboratory temperature to produce a sharp crack of depth Figure 11: Fracture tests: all HAZ (50/50) test results (tested
approximately equal to half of the specimen thickness. at -20°C)
On completion of each test the fracture faces were
measured to confirm the initial crack depth (i.e., depth of the As can be seen in Figure 11 the fracture toughness of the
fatigue pre-crack) and final crack depth. HAZ exhibited a high degree of scatter, CTOD ranging from
Single point values of fracture toughness (CTOD, J and K 0.03 to 0.38mm. Of the 27 specimens tested, 11 resulted in a
at the fracture point) were calculated for each specimen tested. type ‘c’ failure, 11 type ‘u’ and 5 type ‘m’. The fracture
The results, in terms of CTOD, are presented in Figure 10 toughness of the HAZ region was low; 9 tests failed to achieve
(WM tests) and Figure 11 (HAZ, 50/50). Type ‘c’, ‘u’ and ‘m’ the minimum CTOD requirement of API 1104 (CSA Z662 does
are used to describe the failure behavior of each specimen: not specify a minimum requirement, the minimum measured
Type ‘c’ - the critical value of fracture toughness at the value is used in the assessment, and EPRG does not require
onset of brittle crack extension (or ‘pop-in’) fracture mechanics testing to be undertaken).
when the average stable crack extension is less The high degree of scatter in the HAZ results is likely to be
than 0.2mm. due to the abutting pipes being from 3 different sources and/or
Type ‘u’ - the critical value of fracture toughness at the different production heats. Those data from different welds, but
onset of brittle crack extension (or ‘pop-in’) from pipe from the same source and production heat are
when the average stable crack extension is described in more detail below:
equal to or greater than 0.2mm. • Three of the welds tested were from Source B (A06, A33
Type ‘m’ - the value of fracture toughness at the first and A50), and from the same production heat. CTOD
attainment of a maximum force plateau for varied from 0.04 to 0.19mm, the average being 0.10mm
fully ductile behavior. with a standard deviation of 0.05mm. One specimen failed
to achieve the minimum CTOD requirement of API 1104.
0.30 The data set consisted of 5 type ‘c’ failures (2 ‘pop-in’
Type 'c' events) and 4 type ‘u’.
Type 'u'
Type 'm'
• Four of the welds tested were from Source C (A17, A44,
A46 and B03), and from the same production heat. CTOD
0.20
CTOD (mm)

varied from 0.06 to 0.36mm, the average being 0.18mm


with a standard deviation of 0.09mm. One specimen failed
to achieve the minimum CTOD requirement of API 1104.
0.10 The data set consisted of 3 type ‘c’ failures (1 ‘pop-in’
event), 4 type ‘u’ and 5 type ‘m’. The 3 specimens from
weld A17 (Tie-in weld) failed in a fully ductile manner;
0.00
CTOD ranging from 0.17 to 0.30mm.
CWP tests
Figure 10: Fracture tests: all WM test results (tested at -20°C) Preparation and testing of a CWP specimen is not covered
by an international test standard. The CWP specimens were
As can be seen in Figure 10, of the 27 specimens tested, 4 prepared and tested according to in-house procedures that have
resulted in a type ‘c’ failure, each failing due to a significant been developed by Laboratory Soete (University of Gent) since
‘pop-in’ event. Two tests resulted in a type ‘u’ failure and the 1979.
remaining 21 tests exhibited fully ductile behavior (type ‘m’).

7 Copyright © 2012 by ASME


The nominal dimensions for the CWP specimens are specimens targeting the HAZ, the position of the notch tip in
presented in Figure 12 (figure provided courtesy of the relation to the fusion boundary was measured to confirm
University of Gent). accuracy of notch placement.
The plastic straining capacity and defect tolerance were
quantified by means of the remote (pipe metal) failure strains.
A pipe metal failure strain of 0.5% was used as a performance
requirement (i.e., the Gross Section Yielding or pipe yielding
criterion):
• Local Collapse LC - collapse of the remaining
ligament below the surface breaking defect
• Net Section Yield NSY - collapse of the section
containing the defect without significant straining of
the parent material
• Gross Section Yield GSY - collapse by gross straining
Note: W - width of the CWP specimen, approximately 300mm remote from the defect. A pipe metal strain of 0.5% is
Figure 12: Curved wide plate test specimen dimensions required to consider GSY has been achieved
An overview of each CWP specimen defect geometry and
All of the CWP specimens from the A-series welds had a corresponding test result is presented in Table 3.
surface breaking defect machined at the weld root, either
targeting the fusion line or at the WM centre-line. The CWP Defect Dimensions
specimens from the B-series welds contained either deliberate
GW ID Type d l ρ Sfail Mode
or natural welding defects that were either surface breaking or
embedded within the weldment. mm mm mm N/mm2
The machined surface breaking defects were produced Machined defects
using a 0.15mm wide chevron cutting wheel. The notch A06 H1 SD 3.0 50 - 865 GSY
sampling the fusion-line was positioned on the weaker side of
the weldment, based on the results of the small-scale tests. A06 H2 SD 3.0 100 - 810 GSY
The position of the CWP specimens from the B-series A06 H3 SD 4.0 100 - 787 NSY
welds was determined from the results of X-radiography and A06 WM SD 3.0 50 - 857 GSY
AUT, which was undertaken prior to commencement of the BP A17 H1 SD 3.0 50 - 829 GSY
X100 operational trial.
A17 H2 SD 3.0 100 - 795 GSY
During the test, linear variable differential transformers
were used to measure the elongation (hence strain) of each A17 H3 SD 3.0 100 - 782 GSY
pipe, and the weldment. In addition, a clip gauge was used to A17 WM SD 3.0 50 - 817 GSY
measure the crack mouth opening displacement of the defect. A33 H1 SD 3.0 50 - 848 GSY
The load applied during the tests was measured by the test
A33 H2 SD 3.0 100 - 811 GSY
machine load cell.
After positioning the CWP specimen in the test machine A33 H3 SD 4.0 100 - 791 GSY
and attaching all instrumentation, the specimen was cooled to A33 WM SD 3.0 50 - 862 GSY
-20°C using refrigerated methanol (the methanol was circulated A46 H1 SD 3.0 50 - 823 GSY
around curved cooling boxes that were clamped either side of
A46 H2 SD 3.0 75 - 793 GSY
the specimen). After allowing the temperature to stabilize for
not less than 1 hour, testing commenced. A46 H3 SD 3.0 100 - 783 GSY
Each test was undertaken under displacement control, at a A46 H4 SD 4.0 100 - 750 NSY
constant rate of 1 mm/min. The test was stopped when either A50 H1 SD 3.0 50 - 838 GSY
failure occurred or a maximum load was achieved during A50 H2 SD 4.0 100 - 797 GSY
loading.
On completion of the test, photographs were taken of the A50 WM SD 3.0 50 - 820 GSY
tested specimen and the fracture surfaces were then cut from Natural and deliberate welding defects
the specimen. The fracture surface features were subsequently B03 WP1 SD 4.5 151 - 821 GSY
photographed and the dimensions of the defect(s) were
B03 WP2 ED 1.7 202 7.3 O 852 GSY
measured.
For those specimens that fractured during the test a B03 WP3 ED 6.5 116 5.4 I 836 GSY
stereoscope was used to identify the fracture initiation point or B03 WP4 ED 4.4 147 7.6 O 774 GSY
the deepest point of the notch, as appropriate. For those B06 WP1 SD 5.4 309 818 GSY

8 Copyright © 2012 by ASME


Defect Dimensions CWP specimen details Assessment failure ratios
GW ID Type d l ρ Sfail Mode GW ID Sfail API CSA BS API
mm mm mm N/mm2 N/mm2 1104 Z662 7910 579-1
B06 WP2 No defect found 765 * GSY A46 H2 793 1.00 1.04 2.96 2.82
B06 WP3 ED 6.0 145 5.8 O 799 * GSY A46 H3 783 1.00 1.04 2.99 2.88
B06 WP4 ED 5.0 144 7.0 O 773 NSY A46 H4 750 0.97 1.02 6.42 5.62
B08 WP1 ED 9.0 173 0.5 I 617 LC A50 H1 838 1.09 1.07 19.83 11.42
B08 WP2 No defect found 837 * GSY A50 H2 797 1.14 1.38 NAP NAP
B08 WP3 ED 10.8 145 4.2 O 627 LC A50 WM 820 1.05 1.05 1.17 1.52
Notes: GW - girth weld
Natural and deliberate welding defects
H - defect located in the HAZ (followed by specimen number)
WM - defect located in the weld metal B03 WP1 821 1.40 2.31 11.50 11.35
WP - wide plate (followed by specimen number)
SD - surface breaking defect B03 WP2 852 1.17 1.18 1.15 1.24
ED - embedded (buried) defect B03 WP3 836 1.56 1.24 2.00 2.92
d, l, ρ - defect depth, length and ligament dimensions respectively
Sfail - CWP specimen failure stress B03 WP4 774 1.30 1.13 1.32 1.75
O - outer pipe surface ligament dimension B06 WP1 818 1.60 1.20 NAP NAP
I - inner pipe surface ligament dimension
* - test terminated without failure of the specimen B06 WP2 765 * No defect found
Table 3: Summary of CWP test results (tested at -20°C) B06 WP3 799 * >1.45 >1.22 >1.55 >2.13
B06 WP4 773 1.32 1.15 1.29 1.73
EVALUATION OF GIRTH WELD DEFECT
B08 WP1 617 1.22 1.14 NAP NAP
ACCEPTANCE CRITERIA
The results of the analyses are summarized in Table 4 B08 WP2 837 * No defect found
showing the CWP test specimen failure (or maximum) stress B08 WP3 627 1.30 1.05 3.88 3.64
compared with the predicted failure stress (presented as a Notes: GW - girth weld
failure stress ratio) using the different assessment criteria. H - defect located in the HAZ (followed by specimen number)
WM - defect located in the weld metal
The failure stress ratios presented in Table 4 correspond to WP - wide plate (followed by specimen number)
the ratio of the actual to predicted failure stresses; a failure * - test terminated without failure of the specimen
stress ratio greater than 1.0 means that the predicted failure Sfail - CWP specimen failure stress
stress is less than the actual failure stress. NAP - No assessment possible (limitations of assessment method)
Table 4: Summary of CWP analyses.
CWP specimen details Assessment failure ratios
An overview now follows of the pertinent points of the
GW ID Sfail API CSA BS API different analyses undertaken.
N/mm2 1104 Z662 7910 579-1 API 1104 (Option 2)
Machined defects To verify the applicability of the API 1104 Option 2
A06 H1 865 1.11 1.08 4.19 3.77 assessment procedures the following analyses were undertaken:
1. Construct a locus of critical defect height versus length for
A06 H2 810 1.08 1.17 4.61 4.15 the weldment and compare this with the CWP specimen
A06 H3 787 1.11 1.33 22.98 13.87 defects. The locus was based on the minimum measured
A06 WM 857 1.05 1.07 1.34 1.77 tensile properties and fracture toughness for the weldment,
A17 H1 829 1.12 1.13 1.20 1.48 and the applied longitudinal stress in the analysis was set
equal to the SMYS of grade X100 line pipe, 690N/mm²
A17 H2 795 1.04 1.11 1.18 1.48 (the results of this analysis are shown in Figure 13 for weld
A17 H3 782 1.08 1.09 1.33 1.67 A06).
A17 WM 817 1.10 1.11 1.32 1.82 2. Construct a material specific FAD (based on material
A33 H1 848 1.03 1.06 1.58 2.1 properties adjacent to the CWP location). Analyze the
CWP tests undertaken to determine their corresponding
A33 H2 811 1.01 1.04 1.66 2.16
assessment points (Kr, Lr), and compare them with the
A33 H3 791 1.02 1.15 3.88 3.61 FAC (the results of this analysis are shown in Figure 14 for
A33 WM 862 1.03 1.08 1.47 1.94 weld A06 specimen HAZ (1)).
A46 H1 823 1.04 1.07 2.99 2.81 3. Similar to (2.), but the FAD is based on the minimum
measured tensile properties for the weldment. The

9 Copyright © 2012 by ASME


objective was to determine the critical stress for failure of 1.2
each specimen by varying the applied longitudinal stress 1.0

Brittle Fracture, Kr
until the corresponding assessment point (Kr, Lr) was
coincident with the FAC (the results of this analysis are 0.8
presented in Table 4). 0.6
0.4
10
0.2
8 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Defect height (mm)

6
Plastic Collapse, Lr
4 FAC HAZ (1) HAZ (2) HAZ (3) WM

2 Figure 15: API 1104 (Option 2) analysis: Critical stress


predictions based on the FAD (example shown:
0 weld A06).
0 50 100 150 200 250 300
Defect length (mm)
API 1104 does not explicitly state an upper limit on pipe
Critical defect size locus HAZ (1) HAZ (2) HAZ (3) WM
grade, inferring that the Option 2 procedure is suitable
Figure 13: API 1104 (Option 2) analysis: Critical defect size regardless of strength. However, it is recognized in the
locus of weld A06 compared with the CWP defect procedure that the equation provided for estimating Y/T is only
sizes tested. valid for material grades up to X80. An alternative equation is
not provided for pipe greater than grade X80. Despite this, the
Figure 13 is an example of a critical defect size locus, Option 2 method performed well, giving conservative
which was calculated for weld A06. As can be seen no predictions of failure stress for all, except one CWP test (weld
specimen would be predicted to fail at an applied longitudinal A46, specimen H4, the actual failure stress being 3% less than
stress less than SMYS, as the CWP defect sizes are below the that predicted). In many cases, the predicted failure stress was
allowable defect size locus. This was confirmed by the CWP very close to the actual failure stress; ratio of actual to
test results; each specimen failed at a stress in excess of SMYS. predicted failure stress being 1.0. The accuracy of the predicted
The method used to undertake an individual assessment of failure stresses varied, the least accurate being 60% higher than
each CWP specimen is presented in Figure 14. The FAD was the actual failure stress. The least accurate predictions generally
constructed using Rp0.2 and Rm measured closest to where the came from the B-series welds, the CWP specimens mainly had
CWP specimen was extracted. This analysis shows that the deliberate or natural welding defects embedded within the
failure stress of the CWP specimen would be greater than weldment.
SMYS as the assessment point lays outside the FAC. The results of the analysis agree with the observations of
Wang et al.,[20] who carried out 20 CWP tests towards
1.2 validation of API 1104 to grade X100. Each test specimen had
1.0
a surface breaking defect up to 50mm in length (1/3 of the
specimen width) and 6mm in height (18 tests with 3mm deep
Brittle Fracture, Kr

0.8 defects, 2 tests with 6mm deep defects) and was tested at either
0.6 +20°C or -20°C. Although the analysis resulted in conservative
0.4 predictions, a reduction in conservatism was highlighted when
comparing the predictions for a lower strength material (20
0.2 CWP tests undertaken on grade X70 material).
0.0 CSA Z662
0.0 0.2 0.4 0.6 0.8 1.0 1.2 To verify the applicability of the CSA Z662 assessment
Plastic Collapse, Lr procedures the following analyses were undertaken:
FAC HAZ (1) 1. Construct loci of critical defect height as a function of
Figure 14: API 1104 (Option 2) analysis: Critical stress defect length for the prevention of brittle fracture and
predictions based on the FAD (example shown: plastic collapse respectively, and compare these loci with
weld A06). the CWP specimen defects. Each locus was constructed
based on the minimum measured mechanical properties of
The method used to predict the critical (failure) stress for the weldment, and the applied longitudinal tensile bending
each CWP specimen and the corresponding failure stress ratio stress used in the analysis was limited to SMYS.
(results presented in Table 4) is shown in Figure 15. 2. For each CWP specimen tested, determine the theoretical
maximum longitudinal bending stress for failure. The

10 Copyright © 2012 by ASME


limiting maximum stress was then compared with the the Users interpretation of log-log graphs, and a goodness of fit
actual test failure stress to determine a failure stress ratio. of a power law to the data.
Figure 16 is an example of the critical defect size loci EPRG (Tier 2)
(brittle fracture and plastic collapse), which were calculated for Use of the EPRG guidelines[3] is restricted to line pipe up
weld A06. As can be seen, critical defect sizes are limited by to and including grade X70 and a defect height not greater than
the brittle fracture curve. The defect sizes of three of the CWP 3mm. Recommendations have been made to extend the range
specimens fall below the limiting curve and would therefore be of applicability of the guidelines to grade X80 line pipe and
expected to fail at a tensile bending stress greater than SMYS. defect heights up to 5mm, based on the results of a significant
In contrast the defect dimensions of CWP specimen HAZ (3) number of CWP tests[21][22]. However, a formal update to the
lay outside the limiting curve, meaning that the CWP specimen original guidelines is yet to be published.
would fail at a stress less than SMYS. The same methods used to verify use of the EPRG
guidelines to grade X80 line pipe have been used in this work
200 to investigate their applicability to grade X100.
For the CWP tests undertaken in this work, each weldment
achieved the EPRG Charpy impact energy requirements of
30(40)J min(avg). However, EPRG also requires that the WM
Defect length, Lmax (mm)

yield strength is not less than the line pipe and the pipe’s Y/T
100
ratio in the axial direction is not greater than 0.9. The results of
the tensile tests showed that the tensile properties varied
significantly; resulting in WM yield strength mismatch ranging
from -11% (under-match) up to +26% (over-match), and Y/T
ratios ranging from 0.86 to 0.99.
0
0.0 0.2 0.4 0.6 0.8 1.0 An example of the analysis undertaken of each weld and
Defect depth ratio, d/t individual CWP specimen to verify the applicability of EPRG
Plastic collapse Brittle fracture HAZ (1) HAZ (2) HAZ (3) WM to grade X100 pipelines is summarized below.
Figure 16: CSA Z662 analysis: Critical defect size loci of As can be seen from Figure 17, each CWP specimen from
weld A06 compared with the CWP defect sizes weld A06 exceeded the EPRG performance criterion of 0.5%
tested. remote strain

The same analysis procedures were used to determine the 3


Remote strain at failure (%)

maximum applied tensile bending stress for failure by brittle


fracture and plastic collapse, respectively for the actual CWP
specimen and defect dimensions. To summarize, the following 2
observations were made from the analysis of weld A06:
• The predicted critical stress for CWP specimens HAZ (1),
HAZ (2) and WM was not less than SMYS. 1
• The predicted critical stress for CWP specimen HAZ (3)
was approximately 14% less than SMYS.
• The predicted critical stress for CWP specimens HAZ (1) 0
and WM was limited by the plastic collapse equation, and 0 2 4 6 8
HAZ (2) and HAZ (3) were limited by the brittle fracture Defect area ratio, ld/Wt (%)
equation, these being the predicted failure modes. HAZ (1) HAZ (2) HAZ (3) WM 0.5% strain
• The actual stress at failure of each CWP specimen was Figure 17: EPRG assessment – Comparison of the CWP
between 7 and 33% greater than the predicted critical specimen geometry and test results with the EPRG
stress for each CWP specimen (see Table 4). performance criterion (example shown: weld A06)
The procedure was successfully used to predict a
conservative value of failure stress for each CWP specimen. The CWP specimens were then individually assessed to
Each predicted failure stress was less than the actual test failure compare the defect sizes and test results for the corresponding
stress by 2% or more. Calculation of the critical defect size predictions of failure mode and stress.
locus for the prevention of failure by plastic collapse was As can be seen from Figure 18, CWP specimen HAZ (1)
straightforward, unlike the analysis procedure for the would be predicted to fail by GSY. Furthermore, the actual
prevention of brittle fracture. Whether predicting the critical failure stress of the CWP specimen was greater than the
value of axial bending stress for known defect dimensions or materials measured yield strength and the failure stress
constructing a critical defect size locus, the method is reliant on predicted using the net-section collapse solution adopted by

11 Copyright © 2012 by ASME


EPRG (the locus is based on Rp0.2 measured closest to where limits calculated for grade X80 pipelines, which were seen to
the CWP specimen was extracted). be conservative for the X100 CWP tests undertaken;
Remote stress at max load (N/mm2)

1000
Defect height
h≤3mm 3<h≤4mm 4<h≤5mm
Defect length 5.3t 3.9t 3.2t
Notes: h - Defect height
t - Pipe wall thickness
800 Table 5: Theoretical limits (grade X80 line pipe).

The effect of yield strength mismatch is not clear, except


for the larger defects tested where under-matched welds
exhibited slightly lower strains to failure than similar sized
600 defects in overmatched welds.
0 50 100 150 200 The CWP test data from the B-series welds demonstrate
Defect length (mm) that a girth weld is more tolerant to larger embedded defects
Collapse locus Yield at -20°C HAZ (1)
than surface breaking defects, upon which the EPRG limits are
based.
Figure 18: EPRG assessment – Comparison of the CWP BS 7910
specimen test results with the materials measured To verify the applicability of the BS 7910 fracture
yield strength and corresponding collapse locus mechanics assessment procedures, the following Level 2A
(example shown: weld A06) analyses have been undertaken:
1. Construct a locus of critical defect height versus length for
The CWP specimen dimensions for test HAZ (1) are the weldment and compare this with the CWP specimen
compared with the EPRG collapse locus in Figure 19. As defects. The locus was based on the minimum measured
shown, the defect dimensions lay within the locus, inferring tensile properties and fracture toughness of the weldment,
that failure would occur at an applied stress greater than the and the applied longitudinal stress in the analysis was set
predicted collapse stress. equal to the SMYS of grade X100 line pipe.
2. For each individual CWP test;
6 a. Undertake an assessment based on the test failure
conditions to confirm whether the assessment point
Defect height (mm)

lies inside (safe) or outside (potentially unsafe) the


4 FAC.
b. Determine the critical failure stress for the assessment
2 point to lay on the FAC and compare this with the
actual failure stress to determine a failure stress ratio
(the results of this analysis are presented in Table 4).
0 The analyses for (2.) were based on the tensile properties
0 50 100 150 200 measured adjacent to the CWP location. Due to the limited
Defect length (mm) number of fracture mechanics tests undertaken, the minimum
value of toughness from each set of three specimens was used,
Collapse locus (Yield) HAZ (1)
depending on whether the defect was located within the HAZ
Figure 19: EPRG assessment – Comparison of CWP specimen or WM.
defect dimensions with the defect sizes predicted to To calculate the critical defect size locus for each
fail by plastic collapse (example shown: weld weldment the analyses were undertaken using the standard
A06). solutions for a ‘curved plate’; the input pipe geometry being the
nominal pipe diameter and wall thickness.
The results of the CWP tests undertaken on the A-series To simulate the CWP test conditions, the analyses were
welds provided good supporting evidence of the potential undertaken using the standard solutions for a ‘flat plate’; the
extension of the EPRG guidelines to grade X100 pipelines. dimensions of which were obtained from detailed metrology of
Unfortunately the defect sizes tested were not sufficient to each tested CWP specimen.
verify use of the recommended EPRG limits for grade X80
pipelines; the defect length ratio only extended to 5t (t is pipe
wall thickness) rather than the 7t limit. Nevertheless, as
discussed, the X100 data do support use of the theoretical

12 Copyright © 2012 by ASME


5.0 1.2
Critical defect height (mm)

4.0 1.0

Brittle fracture, Kr
3.0 0.8

2.0 0.6

1.0 0.4

0.0 0.2
0 50 100 150 200
Defect length (mm) 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
FAC HAZ (1) HAZ (2) HAZ (3) WM Plastic collapse, Lr
Figure 20: BS 7910 Level 2A analysis: Critical ‘surface FAC CWP test, HAZ (1) Critical stress, HAZ (1)
breaking’ defect size locus of weld A06, compared Figure 21: BS 7910 Level 2A analysis: assessment of CWP
with the CWP defect sizes tested. test HAZ (1) from weld A06 and the corresponding
assessment of a critical failure stress.
Figure 20 is an example of a critical ‘surface breaking’
defect size locus, which was calculated for weld A06. As can An additional series of assessments were undertaken to
be seen, each CWP specimen would be predicted to fail at an investigate the effect of material toughness. For each
applied longitudinal stress less than SMYS, as the CWP defect assessment, if the input value of fracture toughness was
sizes are above the critical defect size locus. However, each sufficiently large to ensure that failure was controlled by the
CWP specimen failed at a stress greater than SMYS; the tensile properties of the weldment (i.e., collapse dominated),
minimum recorded failure stress was from specimen HAZ (3), the corresponding predicted failure stresses were found to be in
787 N/mm². much better agreement with the actual CWP test results which
The method used to undertake an individual assessment of failed by plastic collapse; failure stress ratios ranging from 1.07
each CWP specimen is presented in Figure 21. The FAD was to 1.22.
constructed using Rp0.2 and Rm measured closest to where the The fracture mechanics assessment procedures in BS 7910
CWP specimen was extracted. This analysis shows that the are very versatile, enabling a number of different types of
actual failure stress of the CWP specimen was greater than assessment to be undertaken to fully describe the behavior of
SMYS as the assessment point lays outside the FAC. Also the weldment. The method is dependent on three factors;
shown on Figure 21 is the result of the assessment to determine applied loading, material properties and defect geometry.
the critical stress for failure which was predicted to be Although measured tensile properties have been obtained
206.3 N/mm², giving a failure stress ratio of 4.19 (see Table 4). for each weldment, the fracture toughness properties were
As can be seen from the position of the critical stress result on measured using highly constrained fracture mechanics
the FAC, failure at such a low stress is due to the relatively low specimens which do not necessarily reflect the true material
fracture toughness measured for the weldment. This is contrary toughness of the pipe geometry. Essentially, the lower the value
to the test result where failure was controlled by the tensile of toughness used in the assessment, the greater the likelihood
properties across the weldment. of failure in a brittle manner. This effect was illustrated in the
As can be seen from the results presented in Table 4 a additional analyses undertaken where the input value of
conservative prediction of failure stress was determined for fracture toughness was increased to ensure failure was collapse
each specimen, as the margin between actual and predicted dominated.
stress was greater than 1.0; ranging from 1.18 up to 23.0. This API 579-1/ASME FFS-1
large variation between actual and predicted failure stress was To verify the applicability of the API 579-1/ASME FFS-1
found to be due to the input value of fracture toughness for fracture mechanics assessment procedures, the same type of
each assessment. assessments were undertaken as those described for the
BS 7910 assessments.
As with the BS 7910 assessment, the methodology and
results are presented below for weld A06 in Figure 22
(comparison of the CWP specimen defect sizes with the critical
surface breaking defect locus) and CWP specimen HAZ (1) in
Figure 23 (CWP test assessment and prediction of the critical
failure stress).

13 Copyright © 2012 by ASME


5.0 assessment methods to grade X100 pipelines is based on
the performance of CWP tests undertaken on one pipe size,
Critical defect height (mm)

4.0 1220x19.8mm.
2. API 1104 (Option 2): The procedure gave conservative
3.0
predictions of failure stress for all, except one CWP
2.0 specimen (predicted failure stress was 3% lower than
actual). In many cases the ratio of predicted to actual
1.0 failure stress was close to 1.0.
3. CSA Z662: The procedure gave conservative predictions
0.0
0 50 100 150 200 of failure stress, greater than 2%, when compared with the
Defect length (mm) actual CWP test data.
FAC HAZ (1) HAZ (2) HAZ (3) WM 4. EPRG: The defect size limits recommended for inclusion
Figure 22: API 579-1/ASME FFS-1 analysis: Critical ‘surface in the EPRG guidance document for X80 pipelines appear
breaking’ defect size locus of weld A06, compared suitable for grade X100 pipelines; however the length of
with the CWP defect sizes tested. the defects tested did not extend to the 7t limit proposed.
The defect size limits are applicable to surface breaking
1.4 defects and embedded defects of an equivalent size.
5. BS 7910: The failure stress of each CWP specimen was
1.2
conservatively predicted; ratio of actual to predicted failure
1.0 stress ranging from 1.15 to 6.5, except for three CWP
Brittle fracture, Kr

0.8
specimens which had a very low predicted failure stress
(corresponding ratios of actual to predicted failure stress
0.6 ranging from 11.5 to 23.0).
0.4 6. API 579-1/ASME FFS-1: The failure stress of each CWP
specimen was conservatively predicted; ratio of actual to
0.2
predicted failure stress ranging from 1.48 to 5.6, except for
0.0 four CWP specimens which had a very low predicted
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Plastic collapse, Lr
failure stress (corresponding ratios of actual to predicted
failure stress ranging from 11.3 to 19.9).
FAC CWP test, HAZ (1) Critical stress, HAZ (1)

Figure 23: API 579-1/ASME FFS-1 analysis: assessment of RECOMMENDATIONS


CWP test HAZ (1) from weld A06 and the 1. To fully verify the applicability of the assessment
corresponding assessment of a critical failure procedures to pipelines constructed from grade X100 line
stress. pipe, further testing is required to investigate the influence
of pipe diameter, wall thickness and diameter to thickness
As can be seen from the results presented in Table 4 a ratio.
conservative prediction of failure stress was determined for 2. More detailed testing should be undertaken than currently
each specimen, as the margin between actual and predicted specified to fully characterize the behavior of the
stress was greater than 1.0; ranging from 1.48 up to 13.9. Like weldment.
the BS 7910 assessments, this large variation between actual 3. The API 1104 method requires an estimation of the Y/T
and predicted failure stress was found to be due to the input ratio, but the equation provided is valid for pipe grades up
value of fracture toughness for each assessment. Assuming to X80. No alternative method is provided for pipe greater
failure to be controlled by the tensile properties of the than grade X80. The validity of this equation to X100
weldment, the corresponding predicted failure stresses were needs to be assessed, and/or the procedure should be
found to be in much better agreement with the actual CWP test updated with a more appropriate model for pipelines
results and failure mode; failure stress ratios ranging from 1.06 constructed from line pipe greater than grade X80.
to 1.21.
ACKNOWLEDGMENTS
CONCLUSIONS Electricore, Inc. and GL Noble Denton would like to
The main conclusions from the work undertaken to acknowledge the following contributions to the project:
evaluate the applicability of the different girth weld defect
• US DOT, PHMSA for funding the project.
acceptance criteria for pipelines constructed from grade X100
• BP for providing the girth welds for testing.
line pipe are:
• Professor Rudi Denys and Antoon Lefevre (University of
1. Verification of the applicability of API 1104 (Option 2),
Gent, Belgium) for undertaking the test program and their
CSA Z662, EPRG, BS 7910 and API 579-1/ASME FFS-1
continued assistance in interpreting the test results.

14 Copyright © 2012 by ASME


DISCLAIMER [14] E Osterby, M Hauge , E Levold , A Sandvik , N Bård and
This research was funded in part under the Department of C Thaulow, ‘Strain capacity of SENT specimens –
Transportation, Pipeline and Hazardous Materials Safety Influence of weld metal mismatch and ductile tearing
Administration’s Pipeline Safety Research and Development resistance’, International Offshore and Polar Engineering
Program. The views and conclusions contained in this Conference, Vancouver, Canada, 2008.
document are those of the authors and should not be interpreted [15] Y Y Wang, M Liu and D Rudland, ‘Strain based design of
as representing the policies, either expressed or implied, of the high strength pipelines’ International Offshore and Polar
Pipeline and Hazardous Materials Safety Administration, or the Engineering Conference, Lisbon, Portugal, 2007.
U.S. Government. [16] J A Gianetto, J T Bowker, D V Dorling and D Horsley,
‘Structure and properties of X80 and X100 pipeline girth
REFERENCES welds’, Proceedings of the International Pipeline
[1] API 1104, ‘Welding of pipelines and related facilities’, Conference (IPC2004), 04/10-08/10 2004, Calgary,
American Petroleum Institute, Twentieth Edition (October Alberta, Canada
2005), Errata/Addendum (July 2007), Errata (December [17] ANSI/API Specification 5L, ‘Specification for line pipe’,
2008). American National Standards Institute and the American
[2] CSA Z662, ‘Oil and gas pipeline systems’, Canadian Petroleum Institute, Forty-fourth edition, 2010.
Standards Association, June 2007 – Superseded. [18] BS EN 10045-1, ‘Charpy impact tests on metallic
[3] G Knauf and P Hopkins, ‘The EPRG guidelines on the materials. Part 1. Test method (V and U notches)’, British
assessment of defects in transmission pipeline girth Standards Institution, London, UK, 1990.
welds’, 3R International, Volume 35, pp.620-624, 1996. [19] BS 7448-2, ‘Fracture mechanics toughness tests. Part 2.
[4] BS 7910:2005, ‘Guide to methods for assessing the Method for determination of KIc, critical CTOD and
acceptability of flaws in metallic structures’, British critical J values of welds in metallic materials’, British
Standards Institution, Incorporating amendment No.1, Standards Institution, London, UK, 1997.
September 2007. [20] Y Y Wang, M Liu, D Rudland and Y Chen, ‘A
[5] API 579-1/ASME FFS-1, ‘Fitness-for-service’, API 579 comprehensive update in the evaluation of pipeline weld
second edition, American Petroleum Institute and defects’, Engineering Mechanics Corporation of
American Society for Mechanical Engineers, June 2007. Columbus, Ohio, USA, Report 03-G78-20, May 2007.
[6] T Swankie, ‘Validation of assessment methods for [21] R M Denys, R M Andrews, M Zarea and G Knauf,
production scale girth welding of high strength steel ‘Recommended revisions to the EPRG Tier 2 Guidelines
pipelines with multiple pipe sources’, GL Noble Denton, for the assessment of defects in transmission pipeline
report 10361, March 2012. girth welds’, 6th International Pipeline Technology
[7] N A Millwood, J Johnson, M Hudson and K Armstrong, Conference, Ostend, Belgium, October 2009.
‘Construction of the X100 Operational Trial pipeline at [22] R M Denys, R M Andrews, M Zarea and G Knauf,
Spadeadam, Cumbria, UK’, Proceedings of the 8th ‘EPRG Tier 2 Guidelines for the assessment of defects in
International Pipeline Conference (IPC2010), 27/09- transmission pipeline girth welds’, Proceedings of the 8th
01/10 2010, Calgary, Alberta, Canada. International Pipeline Conference (IPC 2010), Calgary,
[8] ASME B31.8:2003, ‘Gas transmission and distribution Canada, 27 Sept 2010 – 01 Oct 2010.
piping systems’, The American Society of Mechanical
Engineers, New York, USA.
[9] Z662:2003, ‘Oil and gas pipeline systems’, Canadian
Standards Association, Ontario, Canada, June 2003.
[10] API Standard 1104, ‘Welding of pipelines and related
facilities’, American Petroleum Institute, Washington,
USA, Twentieth Edition, 2005.
[11] EN 1043-1:1996, ‘Destructive testing on welds in
metallic materials. Hardness testing. Hardness test on arc
welded joints’, British Standards Institution, London, UK,
August 2005 – Superseded.
[12] ISO 6507-2:2005, ‘Metallic materials. Vickers hardness
test. Verification and calibration of testing machines’,
British Standards Institution, London, UK, January 2006.
[13] BS EN 10002-1:2001, ‘Tensile testing of metallic
materials. Method of test at ambient temperature’, British
Standards Institution, London, UK, September 2001.

15 Copyright © 2012 by ASME