Proceedings of IPC2010

8th International Pipeline Conference

September 27 – October 1, 2010, Calgary, Alberta, Canada

IPC2010-31526

LIMIT STATE DESIGN BASED ON EXPERIMENTAL METHODS

FOR HIGH PRESSURE SUBSEA PIPELINE DESIGN
Chris Alexander
Stress Engineering Services, Inc.
Houston, Texas
[email protected]

ABSTRACT What has changed over the past 30 years is that deepwater activity
The design of offshore subsea pipelines is facing new challenges around the world has increased significantly. As a result, industry is
as the pipeline industry is moving into environments requiring high now being required to operate at conditions that necessitate a change in
pressure design. Conventional pipeline design codes such as ASME design philosophy. It is appropriate at the present time to reexamine
B31.4 and B31.8 establish pressure limits based on percentage of the conventional design methods in an effort to advance limitations
pipe material’s minimum specified yield strength. While this has currently imposed on modern manufacturing and joining capabilities.
traditionally worked for relatively thin-walled pipe at moderate
pressures, there are concerns that full utilization of the material’s At the core of this discussion is an attempt to more fully utilize
capacity is not being realized when designing for high pressure material capacities available in modern high strength steel, as well as
conditions. Additionally, there are concerns regarding the ability to accounting for the ability that manufacturers have to fabricate a more
achieve high quality manufacturing and consistently fabricate welds in consistent work product with less variance in variables such as wall
thick-wall pipes. thickness. Even though there are still some limitations on the ability to
manufacture high quality, ultra high strength, large diameter thick-wall
This paper presents details on a testing program that incorporated pipes, the steel produced today is far superior to the material produced
full-scale burst testing to qualify the design pressure for an 18-inch x when the prevailing design codes were established decades ago.
0.75-inch, Grade X65 subsea gas pipeline using the methodology of Computer-controlled steel making and heat treatment processes
API RP 1111. A lower bound burst pressure was established based on produce chemically-consistent, clean, and fine grained steel.
the recorded burst pressures to which a design margin of 0.72 was Additionally, pipes having more consistent dimensional characteristics
applied to determine a design pressure. Had the pipeline been are complemented by improved and automated full pipe body
conventionally-designed using ASME B31.8, the design pressure inspections that lead to a final product which is far more reliable than
would have been 3,900 psi. However, using the experimentally-based previously observed. It is more prudent to use a reliable product with a
design option in API RP 1111 the resulting design pressure was 4,448 realistic safety margin, than an uncertain less reliable product with a
psi. This results in a net increase in the design pressure of 14 percent. notionally higher safety margin established years ago based on less
reliable products. As an industry we must be cognizant that requiring
When one considers either the potential cost savings in material thicker and stronger pipes may not necessarily generate a system that is
requirements at construction or the additional throughput associated more conservative and reliable.
with higher design pressures for a given pipeline system, it is not
difficult to demonstrate the economic benefits derived in performing a There is certainly a cost that must be considered when designing
more rigorous material qualification and limit state design process any system, including pipelines, at high stress levels (and potentially
based on experimental methods as presented in API RP 1111. lower design margins). The cost is related to confidence levels in both
the quality of construction materials, as well as confidence in the
expected operating conditions (e.g. cyclic pressure and temperatures).
INTRODUCTION All too often there is a singular focus on the former, while failing to
The concept of designing piping, pipelines, and pressure vessels consider that variability in the latter could have equally-disastrous
using limit state design is not a new. The first edition of API RP 1111, consequences.
Design, Construction, Operation, and Maintenance of Offshore
Hydrocarbon Pipelines (Limit State Design), was first published in Figure 1 is a flowchart showing the conventional design process,
1976. In 1997 a task force was assembled to consider changes to the alongside a modified version of the design process for establishing
existing recommended practice as concerns existed among the pipeline design pressures based on limit state design including experimental
engineering design community that overly-conservative designs were burst test results coupled with more stringent quality control
being generated for pipes having low diameter to wall (D/t) ratios that procedures. As noted in this figure, the conventional design process
are typical for high pressure deep water subsea pipelines. A utilizes minimum material properties. To account for variations in wall
modification to API RP 1111 that resulted from this effort was thickness and other unknowns, safety factors are used. Safety factors
changing the limit state design to be based on the actual burst strength are often more conservative than necessary because of these variations
of the pipe. The change in design was confirmed by more than 250 and unknowns. There is a legitimate basis for reducing safety factors as
burst tests of full-size specimens covering a wide range of pipe grades, greater confidence in material quality (mechanical properties including
diameters, and wall thicknesses [1]. fracture toughness), wall thickness, design conditions, and other factors
is achieved.

1 © Copyright 2010 ASME

 This concept of reduced conservatism is embodied in the multiple axial strains were recorded during testing. Data were collected using a
divisions of the ASME Boiler & Pressure Vessel Codes (Section VIII, data acquisition system during the burst tests at a rate of 1 scan per
Divisions 1, 2, and 3). Division 3 has lower safety factors than those second. Figure 1 shows the location of the strain gages installed on
permitted by Division 1; however, the former has significantly more Sample #1. As demonstrated, the strain gages were positioned to
stringent design and manufacturing requirements that permit and justify monitor strain not only in the base pipe, but also in close proximity to
the reduced safety factor. API RP 1111 Appendix B also embodies a the welds. Wall thickness measurements were taken prior to testing at
similar concept and permits qualification of design pressures using the locations where strain gages were installed. Table 1 provides a list
insights gained from experimental efforts. of the measured values. During testing 5 minute pressure holds were
made at 3,705 psi (MAOP) and 5,417 psi (SMYS for Grade X65 pipe).
One of the primary issues at stake concerns the margin, or safety After these pressure holds the samples were taken to failure.
factor, used to determine design pressures relative to a limit state
condition such as yield, ultimate, or flow stress for pipelines (as well as
offshore risers and flowlines). A review of the available codes TESTING RESULTS
demonstrates a range of numbers that include the following for offshore Pressure testing was performed on Sample #1 and Sample #2.
pipelines and pressure vessels (not included in the current list is DNV- Testing was done in accordance with the protocol discussed previously
OS-F101:2007, Submarine Pipeline Systems).: with pressure holds at 3,705 psi (MAOP) and 5,417 psi (SMYS).
• Value of 0.72 on yield strength for ASME B31.8 for hoop stress Figure 3 and Figure 4 show the test set-up prior to testing. For safety
• Value of 0.72 on yield strength for ASME B31.4 for hoop stress purposes, during testing each sample was placed inside a thick-wall
• Value of 0.72 on burst strength for API RP 1111 for hoop stress pipe to provide a barrier and contain any fragments released during
• Value of 0.67 on yield strength for ASME Boiler & Pressure rupture.
Vessel Code, Section VIII, Division 3 (Division 3) for hoop stress
Figure 5 and Figure 6 show pressure as a function of elapsed time
In looking forward to advanced design methods it is important to for Sample #1 and Sample #2, respectively. The following burst
select a design factor that is appropriate for the design of high pressure pressures were recorded as noted on the plots.
systems. The methods presented in Appendix B of API RP 1111 and • Sample #1 – 7,473 psi (2.0 times the MAOP for this line)
Division 3 (Parts KD-1253 and KD-1254) should be considered. • Sample #2 – 7,437 psi (2.0 times the MAOP for this line)

Provided in this paper are results associated with a limited Both hoop and axial strain were recorded using bi-axial strain
program involving full-scale burst tests of pipe for a subsea pipeline. A gages. For pressure testing the strain of primary interest is hoop strain.
review of design codes is also part of the process to assess the validity Figure 7 and Figure 8 provide hoop strain gage results for the two test
and applicability of the above assertions. It should be noted that the samples. Refer to Figure 2 for details on the locations of the respective
contents of this paper are not exhaustive in terms of what is required to strain gages. When the curves associated with the plotted strain gage
actually qualify a designated pipe product for a particular design, but data stop, it indicates that the strain gage disbonded from the pipe and
rather the intent is to initiate discussions on how a more rigorous design was no longer valid. This typically occurs at strains on the order of
process can benefit industry. 10,000 microstrain (e.g. 1 percent strain). Microstrain is a measurement
unit of strain that corresponded to 1 x 10-6 in/in.
The focus of the presentation of this paper is limit state conditions
associated with ductile burst. Other limit state considerations not After testing post-failure inspection work was done. Figure 9
explicitly addressed in this paper include tension, bending, and collapse shows the ductile failure that occurred in Sample #1. As noted, the
due to external pressure. failure did not occur in the longitudinal seam weld. The failure in
Sample #2 is shown in Figure 10 and also shows a ductile failure. What
The sections that follow provide details on these above subjects is significant in this photo is that the failure location was not influenced
and present concepts from API RP 1111 that should be considered by either the longitudinal seam or girth welds.
when designing high pressure pipelines for subsea service. Of
particular note are topics addressed in the Discussion section dealing Figure 11 shows a cross-sectional view of the failure from Sample
with the additional efforts that must be undertaken to qualify a material. #1. What is observed is a classic cup and cone fracture associated with
ductile tensile overload. The 0.75-inch nominal wall necked down to be
less than 50 percent of the original wall thickness, further
TESTING METHODS demonstrating the level of ductility associated with the failure.
Two burst tests were conducted using the 18-inch x 0.75-inch,
Grade X65 pipe material. A greater number of samples should have
been tested in order to achieve a greater level of statistical significance. DISCUSSION
Appendix A of API RP 1111 recommends a mini-mum of six burst One of the primary aims of this study was to evaluate the level of
tests be conducted, although Appendix B of this same document does conservatism present in traditional design methods and determine how
not specify a minimum number of test samples. that compared to a limit state design basis. The advantage in
conducting full-scale burst tests, especially a program involving enough
Prior to testing, end caps were welded to two test samples that test samples to generate statistically-significant answers, is that
were provided by the pipeline operator. One sample, designated as designers are better-positioned to understand the actual behavior of a
Sample #1, only included the base pipe with a longitudinal weld seam. given pipe material. As has been clearly demonstrated, API RP 1111 is
The other sample, designated as Sample #2, included the longitudinal a valid method for designing pipelines. Unlike the ASME B31.4
weld seam, as well as a girth weld centered at the sample. (liquid) and B31.8 (gas) design codes that rely primarily on elastic
design criteria, API RP 1111 is a strain-based design document that is
Bi-axial strain gages were installed at specific locations on each of based on limit state design. As will be demonstrated in this discussion,
the test samples. The bi-axial gages were oriented so that hoop and the design pressure limits associated with ASME B31.8 are less than
the design limits based on API RP 1111.

2 © Copyright 2010 ASME

 Table 2 presents design calculation results for ASME B31.8, pressure is 7,455 psi and the standard deviation is 25 psi (0.34% of the
ASME B31.4, and API RP 1111. As noted, the basis of design for the mean). Listed below are lower bound burst pressures as functions of
ASME pipeline design codes is yield pressure using the minimum confidence level (shown in %).
specified yield strength, whereas API RP 1111 employs the use of a • 95.45% confidence level (n = 2) PLB = 7,404 psi
flow stress that incorporates the minimum yield and ultimate strengths. • 99.73% confidence level (n = 3) PLB = 7,404 psi
Also included in Table 2 are design pressures calculated using the • 99.99% confidence level (n = 4) PLB = 7,353 psi
experimental burst pressure completed during the course of this study.
Figures 12a and 12b provide the MathCAD sheets used to calculate the Because the standard deviation for this program was relatively
various design pressures, including the methodology embodied in API low, there is not a statistically large difference between the above three
RP 1111. values. For many design applications, a 95.45% confidence level is
sufficient and achieved considering two standard deviations. It should
Note that in Table 2, the design basis for the ASME B31.4 (liquid) be noted that a larger sample size than reported herein is advised and as
and B31.8 (gas) pipeline codes is yield pressure defined using the noted previously, a minimum number of six samples is recommended
following equation. according to Appendix A of API RP 1111.
2St
P= (1) API RP 1111 integrates experimental data into the design process
D
by permitting the specified minimum burst pressure to be a function of
Where: the burst pressure data by using the following relation.
P Yield pressure (psi)
S Minimum Specified Yield Strength (SMYS, psi) P actual
t Nominal wall thickness of pipe (inches) k= (3)
D Outside diameter of pipe (inches) ⎛D ⎞⎛ t min ⎞
(Yactual + U actual ) ln ⎜⎜ ⎟⎜⎜
⎟ t ⎟⎟
⎝ Di ⎠⎝ ⎠
The design basis for API RP 1111 is burst pressure defined using the Where:
following equation (Equation 2a from API RP 1111 that is the Pactual Actual recorded burst pressure for pipe sample (psi)
recommended equation for pipes having D/t ratios less than 15). Yactual Actual material Yield Strength (psi)
Uactual Actual material Ultimate Strength (psi)
⎛D⎞ t Nominal pipe wall thickness (inches)
Pburst = 0.45 (S + U) ln ⎜⎜ ⎟⎟ (2) tmin Minimum measured wall thickness (inches)
⎝ Di ⎠
Where: The specified minimum burst pressure is then calculated as follows as
Pburst Specified minimum burst pressure (psi) specified in Section A.5.1 in API RP 1111.
U Minimum Specified Ultimate Strength (psi)
Di Inside diameter of pipe calculated as D – 2t (inches) ⎛D ⎞
Pburst = k (S + U) ln ⎜⎜ ⎟
⎟ (4)
The design factors for offshore pipelines using each of the three design ⎝ Di ⎠
codes discussed in this paper are provided below.
• ASME B31.8 (Yield strength): 0.72, 0.80, and 0.90 for hoop, Where k is determined as the minimum of the following:
axial, and combined stresses • 0.875 · kaverage
• ASME B31.4 (Yield strength): 0.72, 0.80, and 0.90 for hoop, • 0.9 · kmin
axial, and combined stresses • 0.45
• API RP 1111 (strength as noted): 0.72 (burst), 0.60 (yield), and
0.90 (burst/yield) for hoop, axial, and combined stresses (this For the study at hand, the value of k is determined to be 0.45.
standard also permits strain-based design) Therefore, specified minimum burst pressure is 5,560 psi. This value is
well-below the calculated 99.99% confidence level burst pressure value
If one is to use experimental data to establish a design pressure, it of 7,353 psi calculated previously. Because the actual burst pressures
is critically important that the selected burst pressure used in the are so much higher than the specified minimum burst, a value for k of
calculations truly represent a lower bound condition. The issue of 0.50 can be used in Equation 4 above for calculating the recommended
establishing a confidence level is at the core of this subject. When maximum value of the specified minimum burst pressure according to
multiple burst tests are completed, a mean pressure and standard Section B.1.2 of API RP 1111. This results in a specified minimum
deviation are calculated. To be statistically significant there must be a burst pressure of 6,177 psi (which is still only 85% of the 99.99%
sufficient number of samples. According to Section A.1.2 of API RP confidence level burst pressure value of 7,353 psi).
1111, the procedure recommends a minimum of six burst tests be
conducted (it is recognized that the two burst tests in this program to From the maximum value of the specified minimum burst pressure
not meet this recommendation). The confidence level of the calculated of 6,177 psi, a design pressure of 4,448 psi is calculated. When
lower bound burst pressure is a function of the standard deviation compared to the ASME B31.8 72% SMYS design pressure of 3,900
associated with the burst test results. Figure 13 shows the relationship psi, there is a 14% increase obtained when using the API RP 1111
between confidence level and standard deviation. The lower bound methodology. Regardless of the statistical significance of the burst test
burst pressure is calculated using the following relation, results, the bottom line conclusion is that the burst pressure results
validate the calculated minimum burst pressure based on the API RP
PLB = Pmean − nσ 1111 equation based on minimum specified material properties.
where PLB is the lower bound burst pressure, Pmean is the mean burst
pressure, σ is the standard deviation, and n is the confidence level A final comment concerns the economics associated with
quantifier. For the test data at hand, consider that the mean burst establishing a design pressure based on experimental methods. Any

3 © Copyright 2010 ASME

effort involving supplementary material qualification, inspection, and permit the use of an increased minimum burst pressure. The opening
full-scale testing will require additional up-front costs beyond what paragraph (B.1.1) of this appendix is provided below.
would be expected for a conventional design. However, the potential
benefits in terms of reduced construction costs and greater product Equations 2a and 2b are suitable for estimation of the minimum
throughput far exceed the initial investment associated with the burst pressure for pipe listed in 5.1.2. The coefficients in
additional work. Consider the two example problems that are provided Equations 2a and 2b [see 4.3.1] (0.45 and 0.90, respectively)
in relation to reduced material requirements, as well as additional include considerations of specification requirements, such as
product throughput. minimum wall thickness and mechanical testing frequency.
Improved control of mechanical properties and dimensions can
Example #1 – Reduced Material Requirements produce pipe with improved burst performance. The requirements
One manner of looking at the increase pressure capacity using the in this appendix are intended to permit users to take advantage of
API RP 1111 limit state approach is to consider that less material is improved manufacturing control, to increase the specified
required to achieve the design pressure that would be calculated using minimum burst pressure.
ASME B31.8. Consider the pipe program discussed in this paper. One
could argue that if the target design pressure was 3,900 psi (per ASME From this paragraph and Section B.1.4, as well as other related
B31.8), a reduction of material on the order of 14% could be achieved references [5-7], we extract the following supplementary requirements
if a limit state design was imposed. If the cost for a pipeline is $500 that include the following.
million, a 14% savings is equivalent to $70 million. Considering that a • Full-length helical ultrasonic inspection of each length, including
qualification program of the type discussed herein should not exceed $5 ultrasonic wall thickness measurement with a minimum area
million, there is no question as to the economic merits in conducting coverage of 10%.
such an effort. • Specified minimum wall thickness greater than or equal to 90% of
nominal wall thickness.
Example #2 – Increased Throughput Potential • Inspection efforts to identify the presence of any flaws in the pipe
In addition to the one-time savings associated with reduced material The minimum detectable flaw should serve as the
material requirements as cited in Example #1, even greater returns are minimum assumed flaw size in any subsequent fracture mechanics
likely when considering the additional revenue associated with studies.
operating at higher pressures. Consider a linear relationship between • Mechanical properties, including yield strength and ultimate
increased throughput and increased operating pressure. Assume that a strength, to be tested for compliance using ANSI/ASQCZ1.9-
gas pipeline generates $50 million annually using a design pressure 1993, with an acceptable quality level = 0.10%.
based on ASME B31.8. Assuming that the pipeline company had • Inspection efforts to quantify variations in pipe diameter.
instead designed the pipeline using API RP 1111 limit state methods • Mechanical testing to obtain fracture toughness properties
and the design pressure (and associated throughput) was increased on including KIC value.
the order of 10%, the additional annual revenue would be $5 million. • Consideration of all potential environmental conditions with all
With all things being equal (including the price of natural gas), over a mechanical testing being performed in the identified
20-year period the potential revenue increase would be on the order of environmental conditions. The importance of environmental
$100 million. The suggested $5 million qualification program would testing cannot be overemphasized.
pay for itself in the first year of operation, while the long-term benefits
• Full-scale burst testing.
are even more significant.
While the detailed discussions associated with the burst tests are
pertinent and essential, it is critically important that material
MATERIAL QUALIFICATION
performance and inspection activities be part of the overall system
Probably the most unique difference between limit state design as
qualification process. The capacity of a product (including
presented herein and conventional design processes concerns
experimental burst testing of pipe) is a function of the complete product
supplemental material qualification efforts. For this reason the author
definition.
has elected to include an entire section within the paper to discuss this
particular subject. Although this presentation is not exhaustive, it does
As stated previously, the concepts presented herein are not new. In
highlight some of the more critical aspects associated with material
their landmark paper on Load and Resistance Factor Design (LRFD),
qualification.
based in-part on limit state conditions, Lewis et al [7] make the
following statement in their conclusions.
If one is to operate at elevated stresses levels it is critically
important to understand all aspects of the material’s performance. The
LRFD is a complete and integrated package, covering design,
failure mode of most thin-wall pipelines is assumed to be ductile
material specifications, and quality systems. All parts within
overload. Although the potential for brittle fracture exists in gas
LRFD are based on Industry accepted standards, specifications
pipelines due to the stored energy, the thin-wall nature of most
and practices. LRFD has shown that the tubular costs for wells can
pipelines alleviates the need for concern over brittle fracture due to
be reduced by 15-25% as compared to those based on the
plane strain conditions. However, the tri-axial nature of thick-wall
traditional Working Stress Design. This translates to tens of
pipes requires that a consideration of brittle fracture be made.
millions of dollars per year of savings while at the same time
Consequently, material qualification of thick-wall pipes involves
quantifying the design risks inherent in and economics of the well.
extensive testing to quantify fracture toughness properties.
Additionally, environmental factors such as sour service must be
These comments are consistent with the findings of the basic study
considered and quantified in order to integrate their contribution into
presented and discussed in this paper.
the design process.

Appendix B of API RP 1111, Qualification of Increased Minimum

Burst Pressure, identifies additional efforts that must be undertaken to

4 © Copyright 2010 ASME

CONCLUSIONS REFERENCES
This paper has provided details on a study conducted to evaluate 1. Design, Construction, Operation, and Maintenance of Offshore
the benefits associated with establishing a limit state design pressure Hydrocarbon Pipelines (Limit State Design) API Recommended
for subsea pipelines based on the experimental qualification process Practice 1111, Third Edition. Washington, D.C.: American
outlined in Appendix B of API RP 1111. Although the focus of this Petroleum Institute; 1999.
paper’s discussion was on offshore pipelines, an equally-compelling 2. ASME B31.4. Liquid Transportation System for Hydrocarbons,
argument could be made in the design of offshore risers (although Liquid Petroleum Gas, Anhydrous Ammonia and Alcohols. New
different safety factors are utilized). The potential for increased design York: American Society of Mechanical Engineers; 2003.
pressure is even greater when considering thick-wall high yield strength 3. ASME B31.8. Gas Transmission and Distribution Piping Systems.
pipes generally required for HPHT fields. It is expected that additional New York: American Society of Mechanical Engineers; 2003.
programs in the future will generate subsequent data to support the 4. ASME Boiler and Pressure Vessel Code, Section VIII, Division 3:
position presented herein. Alternative Rules for Construction of High Pressure Vessels. New
York: American Society of Mechanical Engineers; 2004.
The predominant conclusion from this study is that additional 5. Terada, Susumu, Development of Alternate Methods for
efforts are warranted in qualifying a pipeline material for a specific Establishing Design Margins for Section VIII Division 3 (Part 2),
design for three reasons. The first concerns safety. By conducting Proceedings of the ASME 2009 Pressure Vessels and Piping
additional efforts that are more rigorous than those associated with Division Conference, PVP2009-7209, Prague, Czech Republic,
conventional pipeline design, greater confidence is achieved in July 26-30, 2009.
understanding how the pipeline will actually perform in service based 6. Greer, J.B., G. Elliott, and R.M. Shivers, Prototype Sulfide Stress
on the limit state design process. Secondly, because of better Cracking Testing of High Strength Oil Country Tubular Goods,
understanding about future service and the associated operating NACE Technical Paper, 1996.
conditions (including environmental), the added benefit in reduced 7. Lewis, D.B., P.R. Brand, W.S. Whitney, M.G. Hood, and M.A.
downtime is derived. This is critical when considering life-cycle costs. Maes, Load and Resistance Factor Design for Oil Country
Finally, the economic benefits are significant when considering the Tubular Goods, Proceedings of the 1995 Offshore Technology
rewards associated with a rigorous pipeline qualification process. The Conference, OTC 7936, Houston, Texas, May 1-4, 1995.
long-term benefits far outweigh the initial upfront costs. 8. DNV-OS-F101:2007, Submarine Pipeline Systems, 2007 edition.

ACKNOWLEDGMENTS
The author would like to thank Dr. Dave Garrett of Stress
Engineering Services, Inc. for his contributions and insights, including
several of the referenced documents.

5 © Copyright 2010 ASME

Design Methodology Options for the
Design of High Pressure Pipelines
Conventional Design Limit State Design
(includes a performance-based testing
approach, along with supplemental material
Pipe purchase qualification and inspection activities)

and fabrication
Pipe Inspection
(wall thickness, flaw detection
and pipe diameters)
Pipe Inspection
(Typically done in the pipe mill)

Material Qualification
Supplemental Activities

(environmental testing, fracture

toughness, greater sampling)
Acquire Material
Properties
(supplemental to MTRs)
Select Pipe Samples

Design Perform Burst Tests

Standard (Sufficient number of samples
Option to achieve statistical significance)

Working Stress Design

(e.g. ASME B31.8) Limit State Design
(e.g. API RP 1111)

Calculate Pdesign
(based on minimum pipe properties
that is typically Yield Strength) Calculate Pdesign
(function of tmin, burst test results,
and minimum and measured
material properties.

Figure 1 – Proposed design method flow chart

6 © Copyright 2010 ASME

 Gage #1

Longitudinal weld seam Gage #2 and #4

Gage #5 Gage #4 Gage #2

Gage #3

Four (4) bi-axial strain gage rosettes

positioned at 0, 90, 180, and 270 degrees
at longitudinal center of pipe

Figure 2 – Location of strain gages installed on burst test sample

Figure 3 – Burst Test Set-Up

Figure 4 – Sample in Enclosure

7 © Copyright 2010 ASME

 8,000
Burst @ 7,473 psi

7,000

6,000

5,000
Pressure (psi)

4,000

3,000

2,000

1,000

0
0 200 400 600 800 1,000 1,200 1,400
Time (seconds)

Figure 5 – Pressure test results for Sample #1 (base pipe)

8,000
Burst @ 7,437 psi

7,000

6,000

5,000
Pressure (psi)

4,000

3,000

2,000

1,000

0
0 500 1,000 1,500 2,000 2,500
Time (seconds)

Figure 6 – Pressure test results for Sample #2 (girth weld pipe)

8 © Copyright 2010 ASME

 8,000

7,000

6,000

5,000
Pressure (psi)

4,000

3,000

2,000

1,000
1H 2H 3H 4H 5H

0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000
Hoop Strain
(µstrain, 10,000 µstrain = 1% strain)

Figure 7 – Hoop Strain for Sample #1 (Base Pipe)

8,000

7,000

6,000

5,000
Pressure (psi)

4,000

3,000

2,000

1,000

1H 2H 3H 4H 5H
0
0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000
Hoop Strain
(µstrain, 10,000 µstrain = 1% strain)

Figure 8 – Hoop Strain for Sample #2 (Girth Weld)

9 © Copyright 2010 ASME

Figure 9 – Failure in Sample #1 (Base Pipe) indicating ductile failure mode

Figure 10 – Failure in Sample #2 (Girth Weld) indicating ductile failure mode

10 © Copyright 2010 ASME

Figure 11 – Cross-sectional view of Sample #1 after failure

11 © Copyright 2010 ASME

Calculating Burst Pressure of Pipe
(based on methods in API RP 1111 Appendix A)
Nomenclature
D = Outside diameter of pipe (inches)
t = Nominal wall thickness of pipe (inches)
E = Modulus of elasticity (psi)
Sy = Specified minimum yield strength of pipe (SMYS, psi)
U = Specified minimum tensile strength of pipe (psi)
Pe = Elastic collapse pressure of pipe (psi)
Py = Yield pressure at collapse (psi)
Pc = Collapse pressure of pipe (psi)

6
D := 18.0⋅ in E := 30⋅ 10 ⋅ psi

t := 0.75⋅ in ν := 0.3

S := 65000psi
⋅ U := 77000psi
⋅

Internal diameter of pipe (inches) Di := D − 2⋅ t

3
⎛t⎞
⎜ ⎟
Pe := 2⋅ E⋅ ⎝ ⎠
D
Elastic collapse pressure Pe = 4770psi
⋅
2
1− ν

Py := 2⋅ S⋅ ⎛⎜ ⎟⎞
t
Yield pressure at collapse Py = 5417⋅ psi
⎝ ⎠D

Py⋅ Pe
Pc := Collapse pressure of pipe Pc = 3580psi
2 2 ⋅
Py + Pe

Additional input values

Average measured yield strength of pipe (psi) Yactual := 77447⋅ psi

Average measured ultimate tensile strength of pipe (psi) S := 89630⋅ psi

u

Minimum measured wall thickness (inches) tmin := 0.746in

⋅

Figure 12a – Page 1 of 2 of MathCAD Sheet for API RP 1111 Calculations

12 © Copyright 2010 ASME

Calculated Values for Appendix A

π
⋅ ⎛ D − Di
2 2 2⎞ 2
Cross-sectional area of pipe steel (in ) A := A = 41⋅ in
4 ⎝ ⎠
2 π
External cross-sectional area of pipe (in ) A := ⋅ D2 Ao = 254⋅ in
2
o 4
Calculated Test Pressure Values

S⋅ A ⎛ Yactual ⎞ ⎛ tmin ⎞
Capped end yield pressure (psi) CEYP:= ⋅⎜ ⎟⋅⎜ CEYP= 7104⋅ psi
3⋅ Ao ⎝ S ⎠ ⎝ t ⎠

2⋅ Yactual ⎛ Yactual ⎞ ⎛ tmin ⎞

⋅ ln⎛⎜
D⎞
Capped end burst pressure (psi) CEBP := ⎟⋅⎜ ⋅⎜ CEBP = 9222⋅ psi
3 ⎝ Di ⎠ ⎝ S ⎠ ⎝ t ⎠

Experimental burst pressure and std. dev. (input data here): Stdv := 25⋅ psi Pburs t_avg := 7455psi
⋅

Pburst_avg
kavg := As long as k is not less than 0.50,
kavg = 0.516
⎛t ⎞ a value of 0.50 may be used in
( Yactual + Su )⋅ln⎛⎜ DD ⎞⎟ ⋅ ⎜ min Equation B-1 below.
⎝ i⎠ ⎝ t ⎠

Burst pressure per Equation B-1: PB1 := 0.50⋅ ( S + U) ⋅ ln⎛⎜ ⎟⎞

D
PB1 = 6178psi
⋅
Di
⎝ ⎠

Burst pressure per Equation B-2: PB2 := 1.0⋅ (S + U)⋅ ⎛⎜

t ⎞
⎟ PB2 = 6174psi
⋅
D−t ⎝ ⎠
Design Pressure Equations

2⋅ S⋅ t
Yield pressure (SMYS) per Barlow's equation: PSMYS := PSMYS = 5417psi
⋅
D

ASME B31.4 Design pressure (72% SMYS): PB31.8 := 0.72⋅ PSMYS PB31.8 = 3900psi
⋅

API RP1111 Design pressure for PIPELINES: PAPI_pipe := 0.72⋅ PB1 PAPI_pipe = 4448psi
⋅

Calculate the ratio of design pressures to determine the derived benefit in using limit state
design methods based on API RP 1111 relative to ASME B31.8.

PAPI_pipe
Ratio of API RP1111 and ASME B31.8: R := R = 114⋅ %
PB31.8

Figure 12b – Page 2 of 2 of MathCAD Sheet for API RP 1111 Calculations

13 © Copyright 2010 ASME

 Calculating the Mean value of X

Calculating the Standard Deviation of X

Normal Distribution of Data
Standard Deviation and Confidence Level

σ 68.26894921371%

2σ 95.44997361036%

3σ 99.73002039367%
Dark blue is less than one standard deviation from the 4σ 99.99366575163%
mean. For the normal distribution, this accounts for
5σ 99.99994266969%
68.27 % of the set; while two standard deviations from
the mean (medium and dark blue) account for 95.45 %; 6σ 99.99999980268%
and three standard deviations (light, medium, and dark 7σ 99.99999999974%
blue) account for 99.73 %.

Figure 13 – Statistical overview showing normal distribution

Table 1 – Wall Thickness Measurements

Test Sample Gage #1 Gage #2 Gage #3 Gage #4 Gage #5
Designation (inches) (inches) (inches) (inches) (inches)

Sample #1 N/A
0.758 0.746 0.752 0.756
(Base Pipe) (see note)

Sample #2 N/A
0.750 0.756 0.751 0.741
(Girth Weld) (see note)
Note: Measurements at this location were inconsistent and are not included.

Table 2 – Comparison of Calculated Design Pressures for 18-in x 0.75-in Grade X65 pipe

Calculated Pressures
Design Code Stress State
(Design, Yield, or Burst Pressures)
ASME Pipeline Codes
ASME B31 Codes PSMYS 5,417 psi
ASME B31.4 0.72*PSMYS 3,900 psi
ASME B31.8 0.72*PSMYS 3,900 psi
API RP 1111 (Limit State Design)
Specified Minimum 5,560 psi (k = 0.45)
Pb
Burst Pressure, Pb 6,178 psi (k = 0.50)
Design Pressure (k = 0.45) 0.72*Pb 4,003 psi
Design Pressure (k = 0.50) 0.72*Pb 4,448 psi
Notes:
1. SMYS corresponds to the Specified Minimum Yield Strength and PSMYS is the pressure at which this stress state occurs (i.e. 2Syt/D).
2. Note that all calculations presented above are for pipelines, all of the references standards and codes (ASME B31.4, B31.8, and API RP
1111) have different design margins (i.e. safety factors) for risers.

14 © Copyright 2010 ASME