1

In the last 12 months — since the last PPIM — many organizations have adapted to a new normal and have had to revisit leadership approaches, operational practices and find creative solutions for new challenges we have not faced before. The pandemic has brought, health, financial and political factors to the surface. At the same time, it has also produced positive surprises, opportunities and learnings. Work-life-balance has taken on a new dynamic, not least in terms of all those new challenges we need to deal with in our personal lives! This keynote will reflect on leadership and uncertainty: its implications in the context of critical essential structure and in terms of lessons learned and opportunities for the pipeline sector in particular.

2

In 2019, PHMSA issued new regulations (49 CFR §192.493) that require natural gas pipeline operators follow API Standard 1163, In-Line Inspection Systems Qualifications. Included in API 1163 are requirements for verifying and validating an ILI survey. Verification refers to ensuring the ILI was conducted according to plan, procedures, and processes and that the inspection conditions are consistent with those used to establish the ILI Service Provider’s Performance Specification(s). Validation refers to an evaluation of the accuracy of the reported anomaly types and characteristics (depth, length, width, etc.). An inspection is validated if the accuracies are consistent with the Performance Specification.

This paper describes the three validation levels given in API 1163:

• Level 1—Applies only to pipelines with anomaly populations that represent low risk in terms of either consequence or failure probability. The ILI Service Provider’s stated tool performance is assumed to be met based on limited comparisons between current and prior ILI run results, in either the same or a similar line.
• Level 2— Applies to lines with higher risk levels and tests whether the actual tool performance did not meet the ILI Service Provider’s Performance Specification. Although it is possible to state with a high degree of confidence whether the tool performance is worse than the specification (and perhaps reject the inspection results as a consequence), the approach does not allow one to state with confidence that the tool performance is within specification.
• Level 3— Applies to lines with higher risk levels and uses extensive validation measurements to estimate the actual tool performance. Some broad guidelines for statistical processing of validation measurements are described.

The paper also outlines the steps needed to validate an ILI’s probability of detection (POD), probability of correct identification (POI), sizing accuracy, false call rate, and false negative rate.

3

In 2018 Louisville Gas & Electric (LG&E) initiated an extensive program across its gas transmission pipeline infrastructure to re-establish the safe MAOP in line with regulatory requirements that became effective in July 2020. LG&E started the program before regulation changes came into force to align with the wider inspection schedule, and ensure the company was prepared for the anticipated requirements that had been under discussion for many years. The program encompassed seven pipeline segments representing a total of X miles. The majority of the pipelines are vintage construction, with one modern line, constructed in 2000, and various replacements across the other pipelines. The foundation is a significant ILI assessment program in combination with an ECA approach to ultimately reconfirm the MAOP on each segment. Within the scope of work were many extensive processes to complete, most notably material property and attribute verification. The end point of the program is approval that MAOP has been successfully re-established through PHMSA. A significant juncture for the process is a review audit by PHMSA, which took place in early 2021. The aim of this paper is to present the data collected, documentation created and procedures developed that were used as the basis for the audit and the ongoing communication process. A significant amount of learning has been gained regarding the process in general and the intricacies of the regulatory requirements, which can be shared to support the industry through the significant tasks facing many operators.

4

Cracking in pipelines can be a significant threat to pipeline integrity, potentially requiring a substantial number of actions and analyses to manage effectively. Many of these actions and analyses were recently codified as part of the Mega Rule Part 1 (§192.493 and §192.917), and operators are now incorporating those elements. Codes and standards typically specify the minimum actions required by operators, with an expectation that operators will go above and beyond using good engineering practices. This paper will document a process – undertaken by an operator – for managing cracks in gas transmission pipelines and addressing regulatory obligations and beyond. The overall process will be framed by risk assessment, starting with prior knowledge about the condition of the pipelines and initial assumptions, through to quantification of risk via in-line and in-field inspections. The analyses culminate in a quantified risk profile, which will document the risk level before, during and after the inspection and assessment campaign. An important aspect of the paper will be a review of the impact and lessons learned of some of the discrete decisions made during the process to manage initial technical and regulatory objectives. These assessments will combine fundamentals, such as inspection and anomaly response, with engineering practices, diagnostic analyses and data management, exemplifying the integrity management process.

5

Traditional code-based approaches for assessing crack-like flaws rely on two key aspects: 1) flaw detection and sizing, and 2) an analytical approach for estimating the stress intensity of a flaw. The estimated stress intensity is then used to predict the pipe performance in terms of burst pressure (static loading consideration) or cycles to failure (cyclic loading consideration) under a set of loading parameters. In following a traditional approach, several idealizations/assumptions are made regarding flaw shape and orientation, material properties, and applicability of the analytical framework used to estimate pipe performance. The uncertainties in these assumptions can, and often do, result in overestimation or underestimation of pipe performance. This paper outlines an experimental approach to quantify the effective stress intensity for crack-like flaws as it correlates to the results from full-scale testing. The proposed full-scale testing approach bridges over several levels of assumptions/idealizations and provides more systemic, direct, and robust solutions for predicting pipeline performance that are better aligned with real-world observations.

By way of examples, supplemented with full-scale testing data generated from fabricated and “natural” crack-like flaws, the paper provides comparisons between analytical calculations (using API 579-1 FAD-based approach) and observed pipe performance data. Gaps in the traditional approach will be discussed, and an improved methodology to empirically-estimate pipe performance in the presence of crack-like flaws will be outlined. The results will be used to estimate an “effective” crack intensity for an idealized flaw shape/orientation.

The approach outlined in the paper will be valuable for addressing and understanding the uncertainties in the performance assessment of pipes with crack-like flaws, and in interpreting the use of experimental data to provide a more robust approach for assessing the severity of crack-like flaws in pipes.

6

The threat of fatigue induced seam weld cracking continues to be a primary issue for pipeline operators. Cyclic pressure loading can cause initial manufacturing flaws in a pipeline seam weld to grow over time. While this behavior is most prevalent in pre-1979 electric resistance welds (ERW) and electric fusion welds (EFW), historical data from PHMSA also shows that submerged arc welds (SAW) have been observed to develop cracks at the toe of the weld, and those cracks have exhibited fatigue growth from transit fatigue, operating pressure cycles, or both.
It is important to prioritize pipelines with respect to fatigue seam weld cracking. While there are industry-accepted methodologies used to prioritize pipelines with respect to seam weld integrity (TTO-5, “Low Frequency ERW and Lap Welded Longitudinal Seam Evaluation and API 1176, “Recommended Practice for Assessment and Management of Cracking in Pipelines” being the most well-known), these methodologies contain some disadvantages when applied specifically to fatigue.

API RP 1176 provides guidance for the prioritization of pipelines with respect to all seam weld cracking threats associated with ERW and EFW pipe. Saudi Aramco developed a methodology to determine a ranking of prioritization for a group of pipelines specifically with respect to fatigue seam weld cracking. The primary metrics used to determine the prioritization ranking are cyclic pressure aggressiveness, predicted remaining life with respect to recent hydrostatic testing, and the API 1176 Annex B prioritization classification. Together, these metrics consider all contributors to the likelihood of a fatigue failure in the seam weld, including vintage, seam weld type, coating and CP performance, maximum and average operating stress levels, hydrostatic test pressure, prior failure history and cyclic pressure aggressiveness. The methodology was implemented into Saudi Aramco’s IMP to prioritize mitigation such as Crack Detection ILI and other mitigation measures such as direct examination.

7

A multi-diameter pipeline, which transports refined products in South America, suffered in-service failures due to circumferentially orientated cracking. The pipeline runs through mountainous terrain. The cracking on this pipeline is Stress Corrosion Cracking (SCC). Areas of cracking were identified by previous in-line inspections. However, the operator was keen to implement new technologies and approaches to best manage the threat of circumferential cracking.

The operator worked closely with ROSEN to define an in-line inspection (ILI) and assessment solution. The result was to provide a single inspection with a combination of technologies capable of detecting and sizing circumferential cracks, detecting metal loss and mid-wall features, detecting areas of bending strain and pipe movement, and detecting deformation such as dents.

In order to assure a successful inspection, the tool was configured based on the pipeline mechanical characteristics and operational conditions. This was further supported by a strong understanding of the integrity threats to this pipeline. A key element in the inspection process was to identify areas of the pipeline susceptible to circumferential cracking, further enhancing data evaluation and providing more accurate results. Following a successful inspection, with 99.99 % of the UT-A / UT-WM inspection data being of high quality for evaluation, 398 planar anomalies (389 pipe body and 9 linear girth weld anomalies) were reported. 50 crack-like anomalies had predicted burst pressures < 1.25 MOP all of which were recommended for repair.

Following in field verifications of 29 of the reported anomalies, it was confirmed that the reported circumferential cracking anomalies met the ILI performance specifications. This paper discusses they key steps that ROSEN and the operator went through together to assure such a success, with the aim of sharing what we have learned so that others may benefit and improve their management of circumferential cracking threats.

8

This paper is the continuation of a 3 year-long study in the use of magnetic based ultra-high sampling density ILI systems (up to 2,000 samples per square inch) to detect and size CSCC. The objective of the study is to enable a shift from the CSCC direct assessment approach to diagnostics by an ILI measurement system.

This phase of the study focuses on the detection of CSCC by ILI systems that integrate Axial and Circumferential magnetic flux leakage (MFL) along with internal depth detection sensor arrays (IDD), a high precision geometry measurement unit (HP-GMU) and inertial mapping data (IMU). The amount of data collected also required the development of sizing algorithms and an analysis process that can categorize and prioritize relevant CSCC amongst the hundreds of thousands of magnetic anomalies the system can record on any given run.

Included in this study are the CSCC inspection results from over 200 miles of natural gas pipelines in 6, 8 and 10” diameters. These results are utilized to review susceptibility factors for CSCC and discuss best practices for excavation and inspection of CSCC anomalies. Field NDE and metallurgical analysis from identified anomalies have been used to baseline the system’s probability of detection (POD) and probability of identification (POI) of critical and sub-critical CSCC.

9

For energy pipelines that experience cyclic pressures, fatigue is a persistent failure mechanism therefore an important element of a continuous structural integrity management program. Fatigue is a complex process of progressive cumulative damage and is known to initiate and propagate cracks to failure in pipelines due to internal cyclic pressures that are less than the maximum operating pressures. While it is challenging to establish the appropriate conditions to predict the time for cracks to initiate in pipelines, crack growth behavior can be reasonably predicted by measuring the rate of fatigue crack growth experimentally. Fatigue crack growth rate is known to be influenced by the cyclic loading characteristics (magnitude, frequency and stress ratio-R), starting crack size, environment and the test sample material behavior.

Fatigue crack growth rate data of API 5L line pipe steels tested in air laboratory conditions have been published widely, however, there are limited data obtained in corrosive environments. This paper will present corrosion fatigue crack growth rate testing data of two API 5L line pipe steels in 3.5% NaCl environment with a pH level of 11 and measured in accordance with ASTM E647 Standard Test Method for Measurement of Fatigue Crack Growth Rates. Fatigue crack growth rate data will be presented for pipe body and long seam weld toe areas. These data would reaffirm a range of corrosion fatigue crack growth rate parameters to use in pressure cycle induced corrosion fatigue analysis and remaining life assessment of pipelines with planar flaws.

10

The ILI inspection of smaller diameter low-pressure gas lines has historically been a challenge. The predominant technology for the inspection of these lines has been MFL. However, in many cases, the pressure required to propel MFL tools through a line is significantly higher than what is operationally achievable or allowable for the pipeline.

Utilizing a liquid to propel an ILI tool can reduce this pressure requirement, although a liquid medium is often undesirable due to additional project costs, handling, or reluctance to introduce a liquid into a gas pipeline. This option, which would allow for the deployment of UT based ILI tools, is therefore often not considered unless a hydro test is scheduled or required by the pipeline owner.

Combining a system that utilizes both a liquid and UT based tool can actually result in a solution that is operationally achievable and cost-effective for a pipeline owner and the ILI provider. The challenges consist of batch size, slippage of the ILI tool in the batch and the required differential pressure required without exceeding the maximum speed for the ILI tool.

The combination of these challenges has made it very difficult to inspect this type of pipelines in an efficient and effective manner. This paper provides an overview of recently performed ILI inspections using a UT based ILI tool with a large velocity operating window in a low differential pressure batching configuration. The pipelines in a gas-producing field were close to the end of the design life. The high-quality data obtained from these inspections provided the client with valuable information for the integrity management of these pipelines that would have otherwise been considered unpiggable. The full fitness-for-service assessment allowed the client to apply for a life extension of the pipelines.

11

API Standard 1163 defines hard spots as a localized increase in hardness through the thickness of a pipe. Although hard spots can be created by a variety of sources, a predominant concern is hard spots that were created as a result of inadequate manufacturing controls. Inadequate pipe material manufacturing controls related hard spots often occur during hot rolling of a steel plate as a result of unintended localized quenching.

Pipeline operators often consider and treat hard spots as a non-injurious stable pipeline feature. In most cases, hard spots formed as early as the pipe plate manufacturing process will remain as a non-injurious feature throughout the life of pipeline. However, the locally hardened material does pose an increased risk for potential cracking. The risk increases in the presence of high pressures and hydrogen, which is a component of sour natural gas or can be a product of a chemical reaction as a result of the cathodic protection system.

This paper will present opportunities to leverage multiple ILI technologies, including high and low field MFL among others, to assess hard spots beyond length, width, and hardness values. Applications of identification of interacting features and complex hardness mapping will be assessed for enhanced hard spot prioritization.

12

Stress Corrosion Cracking (SCC)is a threat that is well known within the pipeline industry. Traditional industry guidance around SCC has been focused on higher stress pipelines (>60% Specified Minimum Yield Strength (SMYS)). Tool technology for in-line inspection (ILI) to identify SCC for remediation and industry practice have evolved, however, there is increasing focus on SCC management in lower stress pipeline (<60% SMYS). In order to assess the potential need for more aggressive management of SCC on its pipeline system, FortisBC Energy Inc. (FEI) conducted a quantitative risk assessment (QRA) on its mainline transmission pipelines. The QRA assessed on a segment by segment level the risk due to all threats, including SCC. Select pipelines were identified based on the risk assessment and general susceptibility to SCC as candidates for EMAT (Electro-Magnetic Acoustic Transducer) ILI assessment to enable management of SCC (and cracking threats in general). The QRA was also used for prioritization of activities. This paper presents the details of the QRA assessment, the findings and how they were applied to management of SCC with the FEI transmission pipeline system.

13

Magnetic Particle Inspection (MPI) has been the main reference for Stress Corrosion Cracking (SCC) detection in pipeline integrity for years. Although this technique is relatively economical and easy to deploy – thanks to a large pool of certified technicians – it remains time-consuming and highly user dependent. Some of the factors impacting results during SCC Direct Assessment (SCCDA) include the total surface area requiring examination, hard-to-reach positions underneath pipes during inspection, improper surface preparation due to poor sandblast or contrast, condensation on pipes, and operator fatigue.

Recent trials have proved that Eddy Current Array (ECA) technology compares favorably against MPI on many aspects in the field, and that ECA has the potential to become the new standard for SCCDA on pipelines. Offering an impressive speed, combined with a particularly high Probability of Detection (PoD), ECA could transform the work of technicians in ditches and above all, offer greater control over the human factor.

Besides detection, ECA has also proved its reliability for SCC characterization on real SCC colonies in both lab and field environments. Comparisons with metallography cuts, grinding measurements and X-Ray Computed Tomography (XCT) data have greatly contributed to optimized depth sizing algorithms for this new solution, providing accurate SCC depth readings. Although ECA and Phased Array Ultrasonic Testing (PAUT) are often complementary techniques in the field, the main advantage of ECA over PAUT resides in the short amount of time required to locate and size the deepest cracks among colonies containing sometimes thousands of cracks. Within a few minutes, technicians and engineers know where to concentrate and how critical SCC really is so that decisions can be made instantly. Combining ease of use and repeatability (ways to control the human factor) is another key benefit of ECA technology.

This paper provides information about a complete ECA solution for SCC detection and depth sizing on pipelines. It reveals results from the field, comparing ECA with MPI, covering several key points and demonstrating how ECA stands out as improving the overall screening process efficiency during examinations in digs. Furthermore, it also exposes and compares ECA data with both destructive and non-destructive testing performed on test pieces containing real SCC.

Keywords: Non-Destructive Testing, Stress Corrosion Cracking, Direct Assessment, Carbon Steel, Eddy Current Array, Pipeline Integrity, Magnetic Particle Inspection

14

The evolution of EMAT technology, has allowed for the reliable detection, identification and sizing of cracking anomalies and has increasingly provided an effective basis for managing the stress corrosion cracking (SCC) threat to an appropriate safety level. When evaluated against other SCC assessment approaches, EMAT ILI exhibits the distinct advantage of providing information on both critical and sub-critical flaws. This paper focuses on presenting an effective and practical framework for managing SCC threat safety margins through detailed evaluation of EMAT ILI feature response.
Establishing partnerships with EMAT ILI vendors ensures that the required reporting necessary to evaluate EMAT ILI tool performance is provided. Defining detection and sizing requirements at the outset establishes the capability to manage a pipeline to a prescribed safety margin. Based on verification and validation of the EMAT ILI results, operators can apply appropriate crack assessment methodologies, acceptance thresholds and excavation strategies for feature evaluation and response.
Insights are provided on the application of the CorLAS™ tool to develop critical flaw curves at various safety margins as a basis for establishing appropriate consequence-based safety margins as it applies to feature response criteria.
Strategies for incorporating and justifying key SCC analysis assumptions, including crack growth, are presented along with approaches for accounting for the impact of uncertainties and interactions. Methods for determining EMAT ILI re-inspection intervals in application with a “just-missed-flaw” (JMF) ½-life analysis or alternative analysis, in consideration of system performance metrics, are also discussed. The assessment and management of manufacturing related features as collateral findings from EMAT ILI is also addressed including the supplemental analyses needed to support the assessment.
An integrated approach for establishing and developing effective excavation assessment targets using data overlays, historical findings and other ILI data for considering both EMAT ILI feature response and additional EMAT validation interrogations (where warranted), is outlined.

15

In this study, the advantages of optimizing a hydrotest pressure level are being explored. Unlike gas lines where the hydrotest may be 90 to 100% SMYS, the liquid line hydro test could be as low as 1.25 times the MOP that could be 40% SMYS. There is considerable range between 40% and 90% SMYS for the hydrotest optimization. One philosophy is to go as high of pressure as possible to remove all flaws, but there can be many blowouts during such a hydrotest, and flaws that survive the hydro test may experience some ductile tearing that could be detrimental to the fatigue life.
In the work being shown, the factors explored were; How much crack growth can occur in hydrotesting during the hold time, how the toughness of the surface-cracked pipe changes with surface crack depth, and how the hydrotest overload of a surface-cracked pipe can cause a delay in crack reinitiation and a retardation of the fatigue growth rate. In one test, an axial surface-cracked 1960 vintage pipe was hydrotested with a long hold time to cause plasticity at the crack front, and the crack delay/retardation resulted in the pipe surviving pressure cycles that would occur in 1,400 years of normal operation. Interestingly, by the new October 2019 DOT rules (if applied to a liquid line), that same flaw would have been required to be taken out of service in 270 days.

16

As the ability to detect and size crack-like flaws with inline inspection (ILI) improves, more operators are exploring the option of managing seam weld integrity primarily with ILI as an alternative to hydrostatic testing. Before an operator makes a decision to proceed in this direction, they must have confidence that they will achieve an equivalent level of reliability compared to hydrostatic testing. While ILI has the benefit of being able to identify specific flaws size and locations, there is still the potential for ILI to miss flaws that are large enough to meet the vendor-specified reporting threshold. Alternatively, hydrostatic testing is only capable of identifying flaws large enough to be unstable when subjected to the test pressure.

Quest Integrity, TL Anderson Consulting, and Marathon Pipe Line have developed a probabilistic framework to quantify the probability of failure over time associated with continued ILI and hydrostatic testing. This probabilistic methodology utilizes Monte Carlo analysis and state of the art fracture mechanics models to accurately predict failure probabilities associated with different assessment scenarios. By accounting for how the ILI tool performed in terms of probability of detection (POD), the results of the probabilistic analysis will identify how many repairs are needed after an ILI to demonstrate equivalent reliability to hydrostatic testing.

17

Strain-based assessment is an important part of integrity management of pipelines located in areas with unstable ground conditions. Strain-based integrity assessment is conducted by comparing pipeline strain capacity with strain demand (the level of elongation or compression produced in the pipe wall as a result of external and internal factors). While the bending component of the longitudinal strain is well understood and can be derived from routine IMU (Inertial Measurement Unit) in-line inspections, the pure axial part of the longitudinal strain has been a recognised gap in the knowledge of the strain condition of a pipeline. Now, the inline axial strain inspection tool (AXISS™) can be used to measure the pure axial strain component. The measured axial strain can originate from many sources, such as geotechnical hazards, temperature effects as well as from the combination of soil restraint conditions and internal pressure effects.

This paper describes an approach to combining bending strain, measured by IMU tool, with axial strain, measured by the AXISS™ tool, in order to determine total longitudinal strain demand. The total strain demand can be determined at the girth welds in the pipeline, and at anomalies, such as metal loss, dents, etc, reported by magnetic, ultrasound and deformation inspections. The strain demand is compared with the strain capacity to determine whether remedial action is required. The tensile and compressive strain capacity will not be constant along the length of a pipeline and is influenced by several factors including material properties and imperfections in the girth welds, corrosion and geometric anomalies such as dents, buckles, wrinkles. A case study is included in the paper showing how the axial and bending strain components are combined to determine the longitudinal strain demand and an approach for evaluating the strain capacity to assess the integrity of the pipeline.

18

Inertial measurement unit (IMU) technology has advanced significantly in the last decade and has become an integral tool in the management of geohazard threats. Initially, operators used IMU technology to improve line location and reduce the size of excavations. Today, many operators use IMU to determine whether the pipeline has been moved from its original shape and, if multiple inspections are available, potentially even the rate at which the pipeline has moved. At present, there have been no significant studies examining the repeatability of IMU technology in identifying and quantifying bending strain locations. This is in part due to the relatively recent development of reliable pipeline movement detection and mandated inspection intervals of five years for liquid and seven years for gas operators. Consequently, only a few operators have multiple sets of IMU data for the same pipeline system. This paper provides an unparalleled comparison of eight different pipeline segments that were inspected between seven and ten times each from May 2018 to October 2020. This paper provides a thorough comparison of multiple bending strain areas following these successive inspections, demonstrating tool repeatability and reliability for managing pipeline integrity in geohazard areas. The paper also investigates how to distinguish noise from reliable data, as well as how to better interpret tool interactions with pipeline features such as girth welds, field bends, dents and wall thickness changes. This paper presents numerous examples of how tool interaction with girth welds and field bends can vary significantly and provides guidance for how operators can properly interpret these signals and discriminate them from actual bending strain areas. The results also compare the performance of IMU technology to published specifications. Finally, this paper identifies avenues for future study, including methods to reduce noise and identifying the impacts of tool type and run conditions.

19

Understanding the strain capacity of girth welds has long been an interest of pipeline operators. Over the years, industry has designed girth welds to be stronger than the pipe material (i.e. overmatch), forcing failures to take place in the base pipe material as opposed to the girth weld. Recently, tensile properties of X70 pipe have increased, resulting in pipe that is stronger than the girth welds produced by traditional weld procedures and creating under-matched welds. This mismatch in strength has been attributed to multiple failures in new X70 pipelines, under nominal loading conditions, via strain accumulation in the girth welds. The performance of a given girth weld design is often assessed using standardized tensile testing, an example being API 1104. However, this testing produces little more than a simple pass/fail result in its most basic form. In this study, digital image correlation (DIC) was used to enhance the results of API 1104 tensile tests by providing full-field strain measurements of girth weld tensile specimens through the point of failure. A benefit of DIC is the ability to place virtual strain gages and extensometers at multiple locations of interest during post-processing of the data, allowing for comparisons of strain data across the weld and base pipe at any point during the test. This resulted in a more detailed assessment of the girth weld design and its behavior under load. The data produced using this method allowed for a better understanding of the strain distribution in a specific weld/pipe combination and can be used to either support or reject the procedure/material for strength mismatch. This paper will provide background information on elements of the new girth weld design, details on incorporating DIC into testing, results from the testing, and discussion on implementing results into girth weld evaluation and qualification plans.

20

Hydrostatic testing of pipelines is a proven method to remove axial flaws at higher than normal operating pressures that may grow to failure during service. After being returned to service, those surviving flaws can continue to grow to a through-wall flaw and cause hazardous leaks or even a rupture. Given crack-growth rates (length and depth rates) ideally based on service experience, it is possible to determine the minimum time after a hydrotest to having a rupture or even the time where a leak of a certain magnitude may occur. In the case of a sour line with H2S, it is desired to keep the leakage to some safety limit while still maximizing the hydrotest interval for cost and safety considerations.
This paper describes; (1) the family of axial surface flaw sizes that can survive a hydrotest, (2) how to determine if a leak or rupture will occur, (3) what size opening area may occur for different leak and rupture cases, and (4) how to optimize the retest time interval based on the known hydrotest pressure history for a maximum allowable leak rate.
It is highly desirable that SCs that grow to “failure” result in TWCs shorter in length than the critical TWC length to avoid ruptures. Furthermore, the shorter that resulting TWC at operating conditions, the smaller the leak-rate will be. As will be shown, there is a minimum time to reach the given leakage rate, after which another hydrotest is needed.
Some examples will be provided, future developments and improvements will be reviewed, and specific differences in determining the “Rupture-Free” time for different crack growth mechanisms like fatigue crack growth from pressure cycling.

21

Hydrostatic test is a powerful tool in managing the integrity of a pipeline. The pipeline holding water at a targeted pressure without leak provides safe margin for operating the pipeline at a lower pressure. If any of the pipe in the test segment is buried, the success of the test needs to be determined through engineering assessment. The pressure change during the test may be due to leak or water temperature changes. This paper presents equations to calculate the adequate pressure changes due to temperature variation. The equations account for the changes of water volume and pipe volume due to temperature and pressure. These equations can be applied on any complex test segment with combined buried and exposed conditions and various pipe dimensions. If a leak is suspected, the amount and rate of the leak can also be estimated through the equations based on the difference between the recorded pressure decrease and the estimated pressure decrease due to temperature change only. These equations can be easily implemented into a spreadsheet or computer code. Working examples are also provided in the paper. From the examples, it shows that considerable amount of pressure changes can result from a small percentage of exposed segment, in which the temperature change during the test is much larger than that in the buried segment. It should be emphasized that the hydrostatic test is a tightness test and has its limitations on detecting tiny leaks in a buried segment. The capability of a test in detecting small leaks relies on the accuracy the model and the accuracy of temperature measurement devices and their placement. Several key contributors to the success of a hydrostatic test and how to control them in the field are discussed at the end the paper.

22

In late 2017, Colonial Pipeline approached NDT Global with a need to inspect a 36” crude oil pipeline known because its high density of dents and metal loss.
Within a 9 months period, NDT Global worked closely with Colonial Pipeline to complete a feasibility study, developing a 36” ultrasonic geometry robot, guaranteeing the high axial and circumferential resolution present in small diameter ultrasonic geometry robots, while ensuring the required 1.5D bend capabilities. This new robot was available in September 2018 for testing and ready for commercial use in October.

In October 2018, NDT Global inspected a 36” x 584 km pipeline with two different robots; Evo Series 1.0 UMp robot with the main objective to locate and measure metal loss anomalies and Evo Series 1.0 Atlas UG to locate and size deformation anomalies.
Due operational conditions of the pipeline and length of the pipeline, in order to inspect the pipeline with both technologies, both robots had to inspect the pipeline with a few hours of difference. As planned, both tools were trapped and recovered successfully, and Data Quality Assessment show no issues with the data.
Three weeks after the robot trap, NDT Global reached out to Colonial Pipeline with a high priority notification describing a buckle type deformation with a 4% OD depth affecting a girth weld. After a meticulous review and correlation with historic ILI results, Colonial Pipeline setup a expedite plan to dig and repair this feature.
This paper will address the challenges faced during the robot development, inspection execution challenges and timely repair of the reported buckle.

23

The theft of oil from pipelines by installation of illegal hot taps is a major issue which not only costs oil and gas businesses upwards of millions of dollars but can also lead to dangerous loss of containment. By volume, thefts in Nigeria alone are as high as 100,000 to 400,000 barrels of oil each day, and in Mexico the loss of revenue alone is estimated to be over $1 billion dollars yearly.

Early stage detection of Illegal hot taps is key to ensuring pipeline safety and to avoid costly and environmentally damaging incidents. Ingu’s Pipers® are a multi-sensor miniaturized in-line inspection tool that can be deployed easily and at low cost either free-floating or mounted to a pig. Specifically, the Pipers® onboard magnetometer measures magnetic flux signatures from remnant magnetism in the metallic pipe walls and metallic features installed on pipelines.

Illegal hot taps are external modifications of a pipeline which produce a measurable difference in the magnetic flux in a pipeline. This is due to creation of a hole in the pipe wall itself and from the extra metallic materials that are installed on the pipeline’s exterior. After a baseline screening of a pipeline, future runs with the Pipers® are able to easily and rapidly detect these magnetic signatures by comparing and identifying them as changes to the baseline.

By measuring changes in the magnetic flux between subsequent inspections, external modifications to the pipeline can be identified and localized. This presentation aims to discuss the analysis techniques as well as show examples of how remnant magnetism can be used to rapidly identify new illegal hot taps during repeat screenings of a pipeline.

24

The most important objective that pipeline operators / owners strive to meet daily is transporting oil and gas products safely, while protecting people, environment, and integrity of their assets. How this is accomplished requires extensive planning, risk management, coordination, and strategic partnering with trusted suppliers for needed engineering support, products, and services.

When the 633-mile 40 inch diameter CAPLINE Pipeline was built in 1967 it was the largest crude oil pipeline in the U.S. In 2018 the CAPLINE Owners decided to purge the system, and as the operator, this project fell to Marathon Pipeline (MPL) to implement. The project required extensive planning and risk management to determine whether it was feasible to complete a traditional nitrogen purge of the system. A key factor of the purge strategy was the partnering with strategic providers of engineering, construction, products, and services to support this challenging undertaking. The initial hurdle to cross was how to purge approximately 5.0MMBBL of crude oil from the single segment pipeline safely and with minimum impact to the communities along ROW (right-of-way). The overall project plan required an Engineered to Order (ETO) pigging solution for a purging pig capable to safely traverse the entire pipeline and piping configurations of all the pump stations, while maintaining seal to keep the purging media of nitrogen (N2) from bypassing and creating an “airlock” condition during the operation.

The case presentation describes the pigging solution development, pig design process (Engineered to Order - ETO), and purging execution of the CAPLINE Purge Project, through a strategic partnership between MPL and TDW over a 2-year span to meet the project objectives. Also, the sharing of this experience with other pipeline operators / owners will support their efforts to implement similar strategies to safely transport petroleum products in the years to come.

25

ROSEN’s Deep Field Analysis (DFA) technology turns the traditional MFL data evaluation process on its head and facilitates fully combined evaluation of axial and circumferential MFL technologies. DFA, details of which have been provided in previous publications, uses an iterative approach in combination with finite element analysis to calculate accurate 3D metal loss profiles that can be used to improve burst pressure estimates and increase confidence in repair calls. This paper provides an outline of the technology but focusses on the actions that are required to establish the reliability and accuracy of the results and how they can be used to support integrity management decisions.

Application of the technology is demonstrated through a case study for a liquid pipeline operator in Europe. A recent MFL inspection had reported a significant external corrosion anomaly that traditional MFL sizing and integrity assessment methods had indicated was of a critical size. The next step would have been rapidly scheduled pressure reduction, excavation and repair. The local terrain meant that access to the anomaly was challenging, so high confidence in the excavation decision was required. DFA technology was used to generate a 3D profile for the anomaly, and recent developments in burst pressure estimation were used to improve the overall assessment approach. Historic ILI data was used to investigate corrosion activity, provide a characteristic corrosion rate and establish a safe remaining life, avoiding immediate excavation and allowing planned monitoring. Exploratory excavations were performed on other sections of the pipeline to validate the accuracy of the inspections and the DFA process.

The paper closes by discussing the ongoing activities that are required to further the development of the technology and widen its applicability in terms of numbers and types of anomalies it can address.

26

The primary objective of an inline inspection survey is to understand the condition of the pipeline and, upon an FFS assessment, determine the behavioral of the asset under certain operational requirements. Currently the industry has at your fingertips diverse assessment methodologies that solve the common operators’ necessities with an acceptable safety range. However, the search of conservativism reduction is a constant for the integrity experts. Challenging and complex pipeline anomalies requires the application of powerfully methodologies such as Finite Element Method FEM. Unfortunately, the limited input information about the anomaly’s morphology generate the need of assumptions based in technical judgment. Condition that produce conservatism and under-utilization of the methodology.

FEM analysis based on ultrasonic inspection technology produces an exact geometric model of the anomalies and pipeline. Source of information that results in an accurate estimate of integrity parameters such as; stress, strain and failure pressure. Conducting an FEA with ultrasonic technology, is therefore considered the maximum level of evaluation with the highest sensitivity to input data.

As result of a recent ILI service for key client SONACOL, NDT Global detected a 3-meters section of pipe with four major dents which were assessed applied FEM. To validate the results and the methodology created by NDT Global for FEM analysis. NDT Global arranged a hydrotest test with DNV-GL on the 3-meters pipe section that SONACOL removed from service.

The assessment and the hydrotest results validate the methodology develop by NDT Global, and the effect of uses ultrasonic inspection robots as input source information. Additionally, the case allowed the implementation of improves in the process as the calibration of material behavior in simulation stage. The accuracy and consistency of the obtained results is compelling evidence of the improves achieved for FEM analysis by using of ultrasonic inspection robots and an appropriate assessment process.

27

Stress corrosion cracking (SCC) is an ongoing topic of research in the industry. Developing a threat screening process to be able to identify the most likely areas of where SCC may be occurring based on available large scale data is important to address these issues. A process developed recently between Kiefner and Duke Energy involves the following factors at minimum: age of pipe, coating type and coating condition, maximum allowable operating pressure (MAOP) of the line, soil stresses, magnetic flux leakage (MFL) metal loss and signal assessment, and ongoing dig evaluations. By integrating the review of MFL signals from an ILI, shallow corrosion can be assessed to help identify locations of disbonded cathodic protection (CP) shielding coating at girth welds and other locations along the line. Shallow, large area corrosion is key; looking for corrosion that could be indicative of a coating disbondment rather than isolated pitting corrosion to assist in determining locations of SCC digs for further analysis. Assessment all of these factors on a regular basis provides for a robust program that can be revised and improved over time with the increased knowledge gained from condition monitoring and SCC direct examinations.

28

Composite repairs are increasingly being looked at as an alternative method for the reinforcement of crack and crack-like features identified in transmission pipeline seam welds. Several prior test programs have been conducted utilizing the carbon fiber / epoxy repair systems to reinforce fabricated cracks that were generated through pre-cracking of an EDM notch prior to installation of the repair. The limitations of these test programs include their overall maximum length of the crack like defect, ideal positioning, and focus on single features.

CSNRI was provided an opportunity by Williams to receive several crack-like seam weld features that were identified in the field from a 36-inch electric flash welded (EFW) pipeline segment and a 16-inch low frequency electric resistance welded (LF-ERW) pipeline segment. These samples were removed from the field and ultimately wrapped with a composite and pressurized to failure. The specimens tested had a variety of features including those with crack-lengths up to 10-inches and several interacting cracks. Additionally, length of repair past the edge of the feature was reduced to determine if the design severely impacted test results. The testing occurred in CSNRI’s Florida facility with remote monitoring performed by Stress Engineering for verification and 3rd party authentication.

The test results were compared to a fracture mechanics based model including the effects of a composite repair. The results showed that the model and its built-in assumptions are conservative in terms of burst pressure. Defect parameters such as feature length and depth are included in the model calculations and follow standard fracture mechanics modeling. Additional testing was performed by Stress Engineering to characterize and size the crack-like features under the repair. By using actual seam weld anomalies, this test program was able to help verify additional parameters for composite repairs and reinforces the results found in prior industry testing.

29

This study evaluated two composite repair technologies for the reinforcement of severe corrosion and thru-wall leaking defects in thin-walled pipe materials; conditions where conventional Type B steel sleeves cannot be welded. The test program included reinforcement of simulated 85% corrosion defects in 6.625-inch x 0.157-inch, Grade X52 pipe subjected to cyclic pressure and burst testing. The test program included 81 reinforced defects and evaluated the following elements:

• Corrosion features with a depth of 85% of the pipe’s nominal wall thickness in thin-walled pipe material (i.e., 0.157 inches, or 4 mm).
  
• Thru-wall defects having a diameter of 0.125 inches (3 mm).
  
• Repairs made with leaking defects having 100 psig (690 kPa) internal pressure.
  
• Strain gage measurement made in non-leaking 85% corrosion defects; it should be noted that the remaining “15%” ligament was 0.024 inches (0.6 mm); to the author’s knowledge, no high-pressure testing has ever been conducted on such a thin remaining wall.
  
• Long-term 90-day test of thru-wall defects that included pressurization with nitrogen gas, followed by relatively aggressive pressure cycling up to 80% SMYS followed by burst testing.

This is the first comprehensive study conducted by a major transmission pipeline operator evaluating the performance of competing composite technologies used to reinforce severe corrosion features with thru-wall defects. The reinforcement of leaks has not been accepted by regulatory bodies such as the Canadian Energy Regulator (CER), or the U.S. Pipeline and Hazardous Materials Safety Administration (PHMSA). A goal of the current study is to validate composite repair technologies as a precursor to regulatory approval.

The results of this study indicate that viable composite repair technologies exist with capabilities to reinforce leaks in pipelines that experience operating conditions typical for gas transmission systems (i.e., minimal pressure cycling).

30

The ongoing evolution of in-line inspection (ILI) technologies, and the introduction of new ILI technologies and vendors, have elevated the need for a more rigorous and timely assessment of the performance of these ILI systems. The use of repeated inspection of test strings is a means of addressing this need. The assessment results can help operators determine which tool offers the best applicability to the operator’s threat mechanism as well as providing feedback to the ILI vendors to improve their performance. These benefits are more pronounced for cracking inspection, but the principles discussed here are universal in their application.

The purpose of this paper is to outline a robust test and analysis protocol that can be used to evaluate and improve ILI capabilities. A goal of this paper is to encourage the use of a common test and analysis protocol to standardize the manner in which ILI capabilities are demonstrated. The information included in the paper has been reviewed and agreed to by both pipeline operators and ILI service providers.

The paper will review the execution parameters necessary for a successful test including, but not limited to, determining test velocities, the purpose of repeat inspections, and setting reporting guidelines. It will also include details on aligning ILI data to the non-destructive examination data and performing anomaly matching. Lastly, the paper will explain the analyses necessary to fully evaluate the inspection results in terms of detection thresholds, probability of detection (POD), probability of identification (POI), sizing accuracies (including incorporating measurement errors and developing lower and upper bound confidence intervals), false call rates, and repeatability.

31

With mature integrity programs under heightening public scrutiny, the focus of pipeline integrity management is shifting to outlier situations, challenging morphologies and conditions that are not so common but nevertheless remain a challenge for risk mitigation through conventional inspection and integrity assessment. With this change in focus has come a change in expectation of the data collected, inspection capabilities and quality of results. In this paper, the treatment and characterization of complex defect morphologies and integrity conditions, and how advancement of ILI data analysis processes & techniques are specifically targeting these needs is discussed.
It will outline the distinction between the generation of specifications vs validation by real-world sources (such as conventions of POD, POI, POS) and the impact that uncertainty has for integrity assessment at the given defect level as well as pipeline segment level, depending on approach used to define and characterize the defects of interest.
It will outline through examples, the limitations that historic convention (as used to define defect types of inspection specifications) has on the validation process of challenging pipeline anomalies for use in integrity management. Further characterizing and refining integrity objectives and parameters leads to a natural evolution to address new definitions, terminologies, outliers and concerns of pipeline integrity.
Various examples are presented including the successful use of machine learning today and since the 1990’s, the change in perception of tool performance validation on specific morphologies relative to large “Big Data” catalogues of dig data already collected, and furthermore, its use to target these “outliers” issues using real field data to improve anomaly characterization in addition to depth and burst pressure performance.
Within today’s conventions for defining inspection performance, cases are presented where even high-performance specifications are reliably exceeded on a regular basis with beneficial implications to the effectiveness and efficiency of integrity programs.

32

In-line inspections (ILIs) often form the basis for integrity management assessments. The data analysis results from a single ILI can be used to plan integrity management responses that meet operator or regulatory requirements. Combination inspections, such as magnetic flux leakage (MFL) and geometry, are used to detect and quantify interacting threats such as a dent with metal loss. Incremental inspections can be correlated in order to calculate observed corrosion growth rates of metal loss between two or more inspections and across different ILI service providers.

The integration of multiple ILIs and data analysis results can be challenging. These challenges include different technologies, spreadsheet formats, reporting criteria, measurement deviations, and independent data presentation. However, the value of integrating multiple ILIs and data analysis results far outweighs the challenge. Each ILI technology is able to provide valuable information about the integrity condition of a pipeline. Independently, no ILI is able to provide a complete picture of the pipeline. By integrating multiple ILIs and multiple ILI service provider data sets universally, a clearer picture of the current and future integrity of the pipeline is formed.

In the following case study, an aging 10-inch wharf line located over water required inspection. Multiple technology in-line inspections were performed utilizing different ILI service providers. The ILI technologies included MFL, ultrasonic wall thickness measurement (UM), and geometry. Each ILI technology provided unique information about the system. The current ILIs were correlated to a previous ILI in order to further understand piping degradation. The ILI data from different technologies and different ILI service providers was able to be correlated, viewed, and interrogated within a single viewing software platform. Universal presentation of ILIs and data analysis results allowed the operator to comprehensively plan for optimal mitigation and safe continued operation.

33

In-line inspection (ILI) surveys provide valuable insights into the current integrity of a pipeline. As pipelines have subsequent ILI survey results spanning several years, knowledge can be gained and used to project the status of a pipeline’s integrity into the future. However, inconsistencies between subsequent ILI surveys are inevitable, and efforts should be made to understand those inconsistencies in order to maximize the value of run-to-run comparison benefits. A common inconsistency observed in industry among metal loss ILI tools pertains to the classification of anomalies, namely metal loss (corrosion) versus manufacturing related anomalies.

The authors’ ILI run-to-run comparison team has gained experience in comparing a diverse range of ILI-reported anomaly types. As a part of this comparison process, anomalies reported by subsequent aligned ILI surveys are matched based on axial and circumferential positioning within the pipeline (i.e., anomaly-to-anomaly matching). It is common for matched anomalies to be classified as a metal loss anomaly in one survey and a manufacturing related anomaly in the other. So which survey is correct, and how should these scenarios be handled? This paper will address such industry relevant questions.
Additionally, this paper will use case studies to explore a range of other questions regarding metal loss and manufacturing anomaly classification discrepancies including, but not limited to:

• Why are matched anomalies often times classified differently between subsequent ILI surveys?
  
• Should ILI-reported manufacturing and metal loss anomalies be treated the same?
  
• Should ILI-reported manufacturing anomalies be included in establishing corrosion growth rates?
  
• Is it possible that an ILI-reported manufacturing anomaly is actually a metal loss anomaly… and vice versa?
  
• How can raw signal data be leveraged in these situations?

The conclusions of this paper will create a better understanding within industry about the differences between ILI-reported metal loss and manufacturing related anomalies.

34

Selective seam weld corrosion (SSWC) is a time-dependent threat that many operators of liquid and gas pipelines are trying to manage. Due to the ever-aging infrastructure in the U.S., existing regulatory requirements and developing inspection systems, there is renewed attention on how SSWC is managed. Phillips 66 is currently managing the threat of SSWC on a pipeline segment with LF-ERW long seams. Phillips 66 has recently completed an inspection on this pipeline using an ultra-high-resolution MFL-C ILI system. The ILI system incorporates optimized technology and a targeted evaluation process to assess the specific threat of SSWC. This paper will walk through Phillips 66’s approach to managing SSWC. The history of the inspected line and the background that stimulated the Phillips 66 response will be presented, together with an acknowledgement of current industry research that is driving a comprehensive review of SSWC assessment. Each stage of the process will be discussed – from the initial ILI results, prioritization of anomalies, verification results from excavations and iteration of dig results to the evaluation process to, finally, the definition of the response plan. The intent is to present a current best practice approach and share recent industry experiences with a threat that is a significant part of the integrity management plan.

35

Recently, PHMSA promulgated new regulations as part of the “Mega Rule” update to both Parts 192 and 195 of 49 CFR. One significant portion of this rule was the requirement for operators to reconfirm their maximum allowable operating pressure (MAOP) for segments of line that do not have tracible, verifiable and complete (TVC) records for the original MAOP confirmation, or had their MAOP confirmed via historical data vice a pressure test. One of the options that PHMSA provided for MAOP reconfirmation was to perform an Engineering Critical Assessment (ECA), which allows an operator to utilize a technical analysis to determine the failure pressure and MAOP of a segment of line. This analysis will rely on the use of material data that must be validated or obtained prior to the start of the ECA and will allow an operator to analytically determine their MAOP without performing a pressure test, which can be extremely useful. This presentation will provide an overview of the ECA process, highlight required actions and provide guidance on how to correctly execute an ECA for the purposes of MAOP reconfirmation.

36

There has been a recent push in the oil and gas industry to use conservative toughness values for pipeline steel if no direct data exists, and this can result in pipes being classified as brittle when they are not. PHMSA has recently issued new rules to 49 CFR Part 192.712 that specify certain toughness values be used in predicted burst pressure analyses absent of measured results. These values can be below the threshold to where commonly used fracture mechanics methods are accurate and/or conservative. A typical solution has been to utilize a purely elastic model such as Raju-Newman; however, this can produce overly conservative results compared to actual failure behavior. There have been questions as to whether other methods that assume some degree of ductility are appropriate for low toughness pipe. Using databases that were comprised of in-service and hydrostatic ruptures of pipe segments due to crack-like defects, a comparison study of the Raju-Newman (brittle), CorLAS™ (ductile), and MAT-8 (brittle and ductile) methods was conducted. The databases included burst pressures that were compared with the output from the models for fracture-controlled failures. Additional material properties such as measured tensile strength and toughness were also included for most of the failures in the databases. A second study investigated the degree of conservatism that would be produced using the lower-bound toughness values directed in 49 CFR Part 192.712 (5 ft-lbs in the body and 1 ft-lb in the seam) and the three fracture mechanics models. Overall, the MAT-8 program and Raju-Newman method predicted more conservative (sometimes to a great degree) burst pressures of low toughness failures compared to CorLAS™. However, if forced to use the low toughness requirements in 49 CFR Part 192.712, CorLAS™ is more accurate in predicting actual burst pressures than MAT-8 and Raju-Newman.

37

Accurate prediction of burst pressure of crack-like flaws is essential for pipeline operators to effectively manage integrity. Industry-accepted models predict burst pressure for a single planar crack in a pipe, but these predictions can be very conservative when applied to adjacent cracks subject to required crack interaction criteria. This study evaluated the interaction of stacked crack-like features, i.e. two cracks located at the same axial position on the inside and outside surfaces of the pipe with varying distance between the cracks within the wall thickness.

Elastic-plastic FEA was executed to evaluate the crack driving force and compute the stress intensity of stacked crack-like features where the distance between the cracks was varied. Three-dimensional crack meshes were used to compute the crack front J-integral and stress intensity values versus increasing internal pressure for stacked crack models and single crack models. The stress results show the development of a plastic zone along the crack front that can contribute to the crack interaction. Comparison of the stacked crack results to single crack results show the effect of crack interaction. The results of this analysis were compared to the API 579 crack interaction criteria to determine the instances where these criteria were observed to be over or under-conservative. The crack interaction results depended on two factors: crack spacing and pressure load magnitude. For closely spaced cracks, interaction begins at low pressure, and at higher pressure loads, interaction was observed even for widely spaced cracks. These results of this study provided a basis for a method to determine when it is appropriate to apply the API 579 crack interaction criteria.

38

Performing in-line inspection (ILI) run-to-run comparisons is a vital component of any pipeline integrity management program. As subsequent ILI surveys are performed on a given pipeline segment, the large data sets collected can often present challenges in accurately determining locations of potential integrity threats. The data obtained must be strategically analyzed, the results of which are key in making informed integrity management decisions. A primary objective of ILI run-to-run comparison analyses is to establish corrosion growth rates along the length of a given pipeline segment that are defensible, justifiable and realistic without being overly conservative. Overly conservative corrosion growth rates lead to:

1) Shortened reassessment intervals and/or
2) Performing unnecessary excavations.

The authors have derived a unique methodology of determining corrosion growth rates as a part of recent integrity analysis efforts. This methodology entails calculating corrosion growth rates for all individual pits reported in the most recent ILI survey. Representative individual pit rates are then utilized to establish corrosion growth rates on a per joint basis. The resulting calculated corrosion growth rates have compared favorably to rates observed in raw ILI signal data. The calculated corrosion growth rates are generally slightly higher than the raw ILI signal observed corrosion growth rates, indicating that the calculated rates are realistic and not overly conservative.

This methodology of calculating corrosion growth rates can be useful for all ILI run-to-run comparisons, particularly when performing ILI raw signal review on all joints is not feasible.

This methodology can be used as a screening tool to establish a relative ranking of anomalies and joints along a given pipeline segment. Alternatively, or in addition, when properly benchmarked, the methodology can be utilized to quantify corrosion growth rates, optimize field inspection programs and establish meaningful and data driven reassessment intervals.

39

Over the past 30 years, huge volumes of in line inspection (ILI) data have been collected for pipelines all around the world. Although these pipelines are diverse in their characteristics, many share similar risk profiles for common pipeline threats such as external corrosion, internal corrosion and cracking. This has led the industry naturally towards supervised machine learning (SML) as a complementary monitoring solution for pipelines that cannot be inspected using ILI. For a given uninspected pipeline, we can source data for similar inspected pipelines and use SML algorithms to generate predictive models.

At PPIM 2020, the authors exemplified this approach with an SML model trained on ILI findings and pipeline information for over 5,000 assets. Using basic design and construction variables alone (such as age, coating type, pipe grade and diameter), the model was able to predict the number of external corrosion anomalies in pipelines, within one order of magnitude of the true value, 90% of the time. Although simple, these per pipeline predictions offer tangible benefits with respect to screening and prioritization of pipeline networks.

At PPIM 2021, the authors will present two further models that incorporate more detailed predictor variables (such as soil properties, cathodic protection potentials, terrain, land use, climate and socioeconomics) and make predictions at pipe joint granularity. One of these higher resolution models is trained on a large, diverse dataset (over 3 million onshore pipe joints selected from a global data repository of almost 10,000 assets), while the other is trained on a smaller, but more refined dataset for a pipeline network in North America.

The enhanced models prove that in addition to facilitating prioritization between pipelines, SML can be used to identify high-risk sites within individual assets and support decisions on mitigation and in field investigation.

40

Dents can present potential threats to the integrity of a pipeline. As inline inspection technologies improve, the resolution allows for the detection of once-undetectable dents, typically with smaller depths. Operators are challenged to identify at what point a dent becomes significant, to optimize reportable dent size thresholds. Furthermore, when a dent is on or near a region of metal loss or a stress riser, regulatory requirements indicate that the pipeline is classified as an immediate repair condition even if the dent is inconsequential.

This paper evaluates current industry practices, regulations, and technical standards for the evaluation of dents with and without general metal loss and localized metal loss. This paper provides a technical justification for the minimum threshold of what may be considered a dent including in the presence of coincidental features (primarily metal loss). The effect of coincident metal loss with a dent is quantified and presented as an alternative to immediate repair, provided that the integrity of a pipeline maintains sufficient margin for safe operation.

41

Traditionally validation of in-line inspection (ILI) results are conducted by physically examining the pipeline by excavation of anomalies found by ILI inspection results. Excavation is costly and, in some cases, could be avoided if better method exists to validate the ILI results. Pipeline operators are seeing the value of using a correlation spool to assist in the evaluation/validation of the in-line inspection results. The correlation spools can be placed in-line with the pipe or at the ILI tool launch site. Correlation spools can have features such as dents, gouges, corrosion, etc., along with artificial flaws extending from these features.

Therefore, due to these features and cracks complexity, accurate measurement of the correlation spool flaw geometry is important when evaluating the ILI results. Direct-current electrical potential (d-c EP) technique (also known as Electrical Potential Drop, EPD) was developed by researchers in the 1960s and used to measure the length and depth of cracks, as well as crack size changes during testing. Recent researchers have demonstrated d-c EP can be used to accurately characterize surface-crack profile in a pipe. Furthermore, this technique can measure tight cracks such as fatigue and thermally induced cracks where other methodologies (i.e., 3D laser scan, and direct physical measurement, etc.) have difficulty.

In this investigation, artificial dents and gouges were created in a section of pipe. Electric discharge machining (EDM) was used to create surface flaws in the pipe and in the artificial damage features. The profile of the surface flaws were mapped using the d-c EP method. Then, the flaws were sectioned to verify the actual depth of the surface flaws.

42

Operators of energy pipeline systems face a never-ending challenge to manage and balance integrity data and risk. Anomalies such as dents, gouges and corrosion can develop throughout the life of a pipeline system, including as early as the pipe manufacturing process, when seam anomalies and mid-wall laminations can occur. Girth weld anomalies and dents can be the result of handling during shipping, the process of laying the pipe in the ditch and field welding processes. Even after the pipe system is built and installed, additional metal loss can occur from corrosion, dents due to settling and mechanical damage.

To mitigate risk, operators rely on in-line inspection (ILI) systems to accurately detect, classify and size pipeline anomalies. While there are various systems available, two of the most common ILI technologies are variations of magnetic flux leakage (MFL) for metal loss and geometry (GEO) for dent features. While extensive research and data collection have been completed for MFL and GEO technologies individually, there are no specific ILI specification for mechanical damage as a class.

This presentation will review the results of a recent industry-funded research project where T.D. Williamson leveraged the data from the Multiple Data Set (MDS) system to provide accurate detection and sizing of mechanical damage defects. Results will be presented in terms of the MDS ILI system capability in detecting and characterizing deformations and in discriminating coincident features within a dent when present. The presentation will also address how these results can be applied to the most recent assessment methods.

43

Despite the frequent occurrence of landslides in Mexico, there is no existing publicly available study of the landslide threat to pipelines in this region. Consequently, the principal aim of this project is to develop an assessment methodology that operators can use to determine a preliminary estimate of the varying risk of landslide threats along pipeline networks. The method is intended to be practical, with clearly defined steps, and based on input available in the public domain. One of the key inputs is mapping data on landslide susceptibility. This input describes in qualitative terms the relative likelihood that an area may be affected by landsliding. By itself, however, it is not sufficient to quantify the probability of pipeline failure because it does not define the nature of the landslide or the pipeline response. The method described in this project develops the landslide susceptibility data into a landslide incident rate along the pipeline route based on the landslide dimensions and frequency of occurrence. This is combined with pipeline vulnerability to landslide loading in order to define an estimate of the pipeline failure frequency. Validation of the process has been carried out against pipeline operational data that includes failures. However, this data is limited and typically confidential – for understandable reasons. The methodology has been applied to an 8,000 km refined-products pipeline network in Mexico operated by Pemex Logistics, which represents approximately one-third of the transmission pipeline system in Mexico. A case study that includes a 1,000 km pipeline crossing different terrain types in Mexico was used to demonstrate the application of the methodology.

44

Regulatory integrity management program (IMP) is often an operator’s first introduction to formal risk management processes. While minimum actions necessary for compliance are defined in the regulations, corporate risk strategies are not. An operator must supplement regulatory guidance with understanding of and reactions to the total risk landscape. As a further challenge, regulatory IMP continues to evolve. With the additional requirements added by the rules published in 2019, it is imperative to develop relevant processes and methodologies. The prudent operator will strive to avoid being purely reactive to risk issues. Risk assessment is an essential precursor to risk management and, when done properly, provides the means for proactive risk management. PHMSA has stated that, a ”comprehensive and systematic risk analysis is a valuable tool to help operators identify and determine the significance of previously unrecognized threats”.

Improved understanding of the process of risk management is emerging to help the industry improve safety in more efficient ways. Part of this improved understanding is the realization that there is uncertainty in any risk analysis, and they are a snapshot of your understanding at that time. This uncertainty can be reduced through the calibration of the analysis with observations made during the direct inspection of the pipeline and the incorporation of new information that is acquired from other sources. This paper will outline several new and innovative risk analysis and management techniques to help move an integrity program from reactive to predictive and allow for better planning for threat mitigation.

45

In preparation for the upcoming Recommended Practice (RP) on Dent Assessment and Management (API 1183), Explorer Pipeline Company, Inc. (Explorer) has performed an internal procedural review to determine how to effectively implement the methodologies into their Integrity Management Program (IMP). Explorer’s pipeline system transports hazardous liquids and is comprised of over 1,800 miles of pipeline ranging in diameter from 6 to 28 inch. The majority of the system was installed in the 1970s, but parts of the system were also installed as early as the 1940s. The primary focus of this review and implementation into the IMP is in regard to performing and responding to in-line inspection (ILI) based integrity assessments. Prior to the development of API 1183, dent assessment and management consisted of following a set of prescriptive condition assessments outlined in the Code of Federal Regulations (CFR) Title 49, Part 195.452. In order to do this, pipeline operators required basic information, such as dent depth, orientation, and interaction with potential stress risers such as metal loss, cracks, gouges, welds, etc. However, in order to fully implement API 1183, additional parameters are needed to define the dent shape, restraint condition, defect interaction, and pipeline operating conditions. Many new and necessary parameters were identified throughout the IMP, from the very initial pre-assessment stage (new ILI vendor requirements as part of the tool/vendor selection process) all the way to defining an appropriate reassessment interval (new process of analyzing dent fatigue life). This presentation summarizes the parameters of API 1183 that were not part of Explorer’s current IMP. The parameters are identified, and comments are provided to rank the level of necessity. Comments are also provided to explain the impact of applying assumptions in place of parameters.

46

The pipeline industry is operating more safely and efficiently than ever thanks in part to a higher in-line inspection data quality. While ILI data quality and accuracy clearly shows substantial advancements in recent years, there are still opportunities for improvement. Recent events sparked the need for innovation to verify and validate the quality of ILI data at all stages. The Pipeline and Hazardous Materials Safety Administration (PHMSA) included sections §192.493 and §195.191, In-line Inspection (ILI) of pipelines, into the Federal Pipeline Safety Regulations in 2020 and 2017, respectively. This means that when conducting in-line inspection of pipelines, each operator must comply with the requirements of several documents incorporated by reference in §192.7 and §195.3.

Addressing in detail these requirements, with checks and balances, can be overwhelming and difficult for a pipeline operator. This paper provides an easy-to-implement methodology to verify the proper execution of an ILI survey, from the selection of the ILI to the actual validation of the results. The paper also describes the entire process from the perspective of a pipeline operator; including the experiences of maintenance personnel, and the challenges faced by field operations during the implementation and use of this methodology.

47

Maintaining and continually improving upon safety and reliability performance requires that activities are conducted in a systematic, comprehensive, and proactive manner that manages risks and prevents incidents. The prevention of low frequency and high consequence incidents is crucial and requires the application of methodical risk management to assure safety critical elements are recognized and managed appropriately. Enbridge analyzed current global best practices in other regulatory regimes and developed a methodology to strengthen the management of major accident hazards.
A major accident is an occurrence (including major emission, fire or explosion) resulting from uncontrolled developments during the operation of any establishment, leading to serious harm to human health or the environment. Major Accident Hazard (MAH) assessment is the process of analyzing MAHs in a methodical and comprehensive way. Looking at major accidents allows an examination of potential consequences and the safeguards which may or may not be in place to prevent or mitigate a minor loss of containment from developing into a catastrophic incident.
The purpose of the Enbridge MAH assessment methodology is to identify potential major accident hazards, associated safeguards, and resulting baseline risk levels at company’s existing facility and mainline right of way. The MAH assessment methodology integrated existing related mainline and facilities risk management activities by applying techniques which enhance the focus on major incidents in order to prevent and limit the potential for catastrophic consequences.
This paper provides a summary of approaches leveraged from other regulatory regimes, a high-level description of the methodology used in facility and mainline case studies and observed benefits and hurdles in applying the methodology to an already mature risk management system.

48

Pipeline operators use different approaches to ensure operational safety of pipelines. One strategy is relying on rigorous In Line Inspection (ILI) campaigns which target the detection of multiple types of integrity threats, such as: crack-like anomalies present in the base material and long seams of the pipe.

As with all sophisticated mechanical instrumentation, ILI tools are subject to sizing tolerances and detection limits (random and systematic components). These values are quantified by vendors using advanced laboratory testing and a series of pull-through or pump tests.

NDT Global was tasked by the operator to investigate, document and obtain the most precise crack-like depth measurements. A high resolution ILI robot engineered to detect and size axial, tilted and skewed crack anomalies was introduced twice in the same pipeline in an interval of 11 months with no changes in regimen or hydrostatic testing.

The main objective was to quantify the ILI robot variance, cumulative tolerance and repeatability of detection of all potentially injurious features in real operational conditions.

The procedure follows statistical analysis guidelines outlined in API RP 1163 by incorporating the known robot tolerance. Additionally, and more importantly, a signal to signal comparison was performed. Ultrasonic Crack Detection (UTCD) relies on amplitudes and a logarithmic function to calculate an absolute depth value of a feature. These amplitudes can be compared to understand variances between inspections. In this particular case, some features are known to have complex morphologies (Ex. Hook Cracks). While such anomalies cause significant amplitude deviations (sizing error) if utilizing conventional ultrasonic crack detection systems, the newly developed ILI robot was specifically designed for the detection and sizing of complex anomalies.

This paper is a summary of the research and collaboration between NDT Global and Marathon.

49

Industry standards such as API 1163 emphasize the role of performance validation in successful in-line inspection (ILI) practice. The standards increasingly suggest the use of statistical methods as part of the validation process. For many, statistical methods can be a significant source of confusion and even frustration. Furthermore, the broad array of specialized statistical terms used in the description of statistical methods, such as certainty, confidence, confidence interval and statistical tolerance interval, can feel bewildering.

The goal of this paper is to help the reader develop a better intuitive feel for the practical meaning of performance validation statistics and associated terminology though the use of numerical simulations and visualizations. To demystify statistical language and explain the relationship between what is observed in a small sample from a dig campaign and what can be stated about the broader population of reported anomalies, this paper presents hypothetical ILI results, simulated potential field dig campaigns, and visualization of the resulting performance statistics. The meaning of certainty and confidence are specifically addressed.

Various statistical approaches, as outlined in API 1163, are treated. A Level 2 validation approach utilizing a confidence interval on the binary proportion of successes and failures is examined. A Level 3 validation approach employing statistical tolerance intervals is also demonstrated and discussed.

50

Several manually operated non-destructive evaluation (NDE) systems are routinely used to gather critical data on pipelines that feed into the decision-making process for safe operations (e.g., hardness measurements, flaw measurements, etc.). The quality of the decisions that are derived from such data is proportional to the quality of the underlying data/measurements gathered. Consequently, it is important to ascertain the quality of the measurements - and more so the quality of the measurement systems that gather the data – before using it. The aim of the paper is to highlight a simplified approach to evaluating the quality of NDE measurement systems by way of estimating repeatability and reproducibility associated with the systems.

The approach will be demonstrated with practical, real-world examples using two NDE systems: 1) in-the-ditch hardness measurements on pipelines, and 2) sizing of flaws using ultrasonic-based techniques that require manual interpretation. A methodology will be outlined to calculate the repeatability and reproducibility of the two systems and detailed interpretation of the results of such calculations will be discussed. The paper will demonstrate that when appropriately applied, the methodology can be effective in identifying high (or low) quality data. Furthermore, when low quality data are identified, the methodology can provide insights into whether the quality issues are a result of operator influence (e.g. lack of training) or instrument influence (e.g. measurement drift, or calibration issues).

The results and the discussion in the paper can potentially provide a valuable framework for developing similar evaluation programs for other measurement systems used for NDE of pipelines. Greater understanding of the repeatability and reproducibility of the NDE measurement systems will not only identify aspects of the systems that need improvement but will also result in the usage of better quality data for decision-making processes.

51

Under high magnification, the structure of a steel can reveal information about the manufacturing process, its “quality,” the origination of flaws, and the extent to which properties are likely to be consistent across the wall thickness. Microstructural features, such as grain size, are known to influence strength and toughness. The types and amounts of different microstructural phases can be related to hardness, strength, and toughness. Toughness can also be influenced by the amount and distribution of inclusions. To assess these attributes, “in-situ metallography” encompassing specimen preparation and visual examination with related photo-documentation can be performed nondestructively in the field.

In relation to natural gas pipeline integrity management, the October 2019 revisions to US federal rules governing natural gas pipelines introduces requirements for Operators to consider pipeline vintage, manufacturing process, and steel when performing maximum allowable operating pressure (MAOP) reconfirmation, verification of pipeline materials, engineering critical assessments, and fracture and fatigue modeling. The revised federal rules also instruct Operators to establish populations of similar pipe. In-situ metallography, supported by nondestructive examination results of composition, strength, and hardness, can be used to confirm pipeline populations.

This work explores how in-situ metallography can support pipeline integrity management and addresses practical aspects of performing in-situ metallography in the field. Experimental work was conducted to compare heat-induced degradation of acetate tape and silicone replicas. Additionally, a quantitative comparison was performed on the results of grain size, percent phase (ferrite/pearlite ratio), and inclusion percentage from surface replicas versus standard, destructive-laboratory testing, cross-sections of pipe. Finally, a sensitivity analysis was performed to explore how minor differences in these microstructural characteristics from replicas and cross-sections may influence the quantification results.

52

Under the latest regulations released by PHMSA, operators of Gas Transmission pipelines must now implement Material Verification for undocumented pipe populations and components. For pipe populations more than a few miles in aggregate length, the cost and effort associated with testing at a frequency of one dig per mile becomes significant. In fact, continuing to sample material properties beyond a certain number of digs in each population offers quickly diminishing returns in terms of statistical confidence and ultimately wasted effort.

The requirements for Material Verification within regulation permit operators to implement alternative statistical sampling approaches for Material Property Verification. As no specific methods are set out in regulation, the industry must define and reach consensus on such approaches to strike an appropriate balance between cost and conservatism.

In this paper alternative sampling strategies are proposed to address a full range of different situations and population sizes, including both piggable and non-piggable pipelines, for line pipe and other component types. This will include direct use of in-line inspection (ILI) data with validation per API 1163, as well as statistical approaches using destructive or non-destructive techniques. A novel Bayesian model developed by ROSEN will also be presented, leveraging the benefits of combining ILI data with in-situ or destructive data to achieve higher confidence with reduced sampling.

53

In October 2019, PHMSA amended the existing rules governing natural gas transmission and gathering pipelines. These amendments require Operators perform destructive or non-destructive testing (NDT) where traceable, verifiable, and complete (TVC) records for pipeline material properties and attributes do not exist. The Pacific Gas and Electric Company’s (PG&E’s) position on materials verification has been to estimate a feature’s most probable grade. Since 2016, PG&E has been collaborating with Kiefner and Associates, Inc. (Kiefner) towards developing a machine learning based pipe grade estimation algorithm. The algorithm takes in-situ NDT results for strength and chemical composition as inputs to predict most probable grade. The training and development of this algorithm relies on Kiefner’s material testing database of approximately 1900 unique laboratory samples. Presentations on the proof of concept of the method and database enhancement were made at previous meetings and this paper presents further algorithmic enhancements to the predictive model.
In this paper we will focus on key developments of the machine learning algorithm, specifically: feature selection, algorithm optimization, and establishing a robust independent validation dataset. Feature selection is the process to rank and prioritize features (or inputs) from the original database to include when developing the machine learning algorithm. Features in this context refer to the model inputs considered such as strength (yield and ultimate tensile), vintage, and chemical composition. Optimizing the features considered by the model can potentially: (1) reduce algorithm training time, (2) increase model interpretability, (3) reduce over-fitting, and (4) improve model accuracy. An often overlooked component of machine learning algorithm development is the distinction between validation and testing of the algorithm. In this work the model validation and testing was performed using a subset from the Kiefner database excluded from training and a model-blind dataset established by PG&E.

54

In-Line Inspection (ILI) has traditionally been used to detect, identify and size pipeline anomalies. However, new ILI technology focused on Material Property Verification can now reveal underlying threats from rogue pipes or pipe populations, and eliminate inaccurate property assumptions for undocumented pipe sections. This offers a step change in information to support robust Integrity Management processes through data gathering and integration. Material properties are crucial in order to assess any type of anomaly, from a corrosion defect to a crack or crack-like feature.

In its latest rulemaking, PHMSA, as the regulatory entity in the US, directed operators with incomplete records to verify or establish the material properties of their pipelines to support safe maximum allowable operating pressures (MAOP) and Integrity Management processes. Other regulatory bodies around the world are following. In 2010, the Argentinian standard NAG 100 was updated based on Subpart O of 49 CFR 192. Since then, TGN has taken the lead role in Argentina in implementing verification of pipeline materials and validation of historical data associated with construction records.

This paper presents a collaborative case study between ROSEN and TGN in which a combination of ILI, field verification data and existing records was implemented. The main goal for TGN was to corroborate the information available in the pipeline construction records and confirm the location of the different manufacturers for a crucial 22” high-pressure gas transmission line with a length of 155 km. These confirmed records will support a robust pipeline Integrity Management Program for the lifetime of the asset.

55

The October 2019 federal rules governing pipelines allow operators to utilize nondestructive testing (NDT) technologies to verify material properties for in-service pipes and fittings where traceable, verifiable, and complete records do not exist. Among the suite of NDT methods available, field hardness testing using ultrasonic contact impedance (UCI) can be used to both characterize pipe hardness and predict ultimate tensile strength (UTS). This UTS prediction is important because it provides an additional piece of information about the material to the verification process—for example, when analyzed in conjunction with chemical composition to estimate grade. However, in order to accurately interpret UCI estimates of UTS, validation is necessary.

The Pacific Gas and Electric Company (PG&E) as part of its materials verification program, has collected, approximately 1700 UCI measurements from over 80 pipes and fittings ranging from 4 to 36” in diameter and from 0.2 to 1.0” in wall thickness. This paper explores the impact of variables such as pipe diameter, vintage, UCI vendor, and pipe operational status on the observed variation in measurements, as well as what these variations may tell us about the accuracy of the UCI estimated UTS. Based on this analysis, approaches for identifying outliers and potentially unreliable measurements will be described. In addition, examples where pipe inhomogeneity may be responsible for the measured variations will be highlighted, and in-situ measurement limitations such as pipe vibration will be discussed. PG&E’s efforts to predict UTS using UCI hardness will be discussed, including the evaluation of different conversion standards such as ISO 18265, ASTM A370, and ASME CRTD volume 57. Finally, our predicted UTS values will be compared to destructive UTS measurements performed on the same pipe features for validation.

56

During routine materials verification (49CFR192.607) nondestructive examination (NDE) of station features without traceable, verifiable, or complete (TVC) materials records, the Pacific Gas and Electric Company (PG&E) identified a population of lap-welded pipe at a gas transmission (GT) station. NDE consisted of long seam verification, hardness testing, chemical composition, and instrumented indentation testing (IIT) to estimate strength. A section of the lap-welded pipe population was cut-out (extracted) from service and subjected to destructive laboratory testing including metallography, tensile testing, and test of its chemical composition.

There were then two goals for the lap-welded pipe population at the station. The first goal was to compare the in-situ NDE for the extracted section of pipe to the laboratory destructive testing. This would serve to validate the NDE results. The second goal was to use the NDE results to establish if all the lap-welded pipe found at the GT station belonged to the same heat. If a conclusive determination could be made that all the lap-welded pipe in the population belonged to the same heat, the destructive test results could then be reasonably assigned to the population.

This paper discusses the methodologies developed to achieve these two goals and the outcomes. The results from destructive testing determined that the pipe had insufficient strength to reconfirm design maximum allowable operating pressure. As a result, PG&E removed the entire population of lap-welded pipe at the station. Nevertheless, PG&E continued to analyze the NDE and laboratory testing results. NDE results were found to be in very good agreement with laboratory results. This reinforced the validity of the NDE examinations for strength, composition, and hardness. Comparing the NDE results across the population of lap-welded pipe, PG&E concluded that the pipes were likely from two separate manufacturing heats and not a single heat as initially assumed.

57

Revisions to federal natural gas regulations introduced the process of Maximum Allowable Operating Pressure (MAOP) reconfirmation using the method of Engineering Critical Assessment (ECA). ECA requires Operators to perform defect and remaining life assessments that require data regarding pipeline feature toughness. Traditionally, toughness has been estimated using laboratory Charpy V-notch (CVN) testing. In the absence of laboratory toughness test data operators may use appropriate toughness values that demonstrably “provide conservative [CVN] toughness values of crack-related conditions of the pipeline segment” per 49 CFR § 192.712(e)(2)(i)(E). Steel pipe manufacturers utilize composition and microstructure to achieve their targeted strength and toughness values. Pacific Gas and Electric Company (PG&E) coordinated the collection of destructive test data from two vendors that had performed laboratory composition, microstructure, and CVN testing on steel pipes and fittings that varied in terms of size, vintage, and grade. Utilizing these data, in addition to detailed knowledge of the evolution of historical pipe manufacturing processes, an exploratory study has been performed into the viability of utilizing NDE chemical composition and in-situ metallography to estimate toughness. Several models to predict CVN upper shelf energy and the CVN ductile to brittle transition temperature were constructed so that the most appropriate model can be selected from among the available models based on compositional elements and microstructural data inputs.