41.1 Introduction

This chapter considers the circumstances and processes involved in managing an aging pipeline infrastructure. Background to this topic is presented in terms of the high-level factors that drive the pipeline industry, the products it transports, and some of the challenges it faces as supply basins become increasingly remote to the markets. Thereafter, the circumstances and evolution of technology that determine the construction processes and define the initial condition of the infrastructure are reviewed. These aspects are then discussed and evaluated with regard to the threats to pipeline integrity. The development and role of codes and standards is outlined, as is the role of the regulatory framework with regard to actions that affect system integrity. Finally, the processes involved in managing an aging pipeline infrastructure are reviewed as the basis to identify process limitations and related technology gaps. The chapter closes with the finding that the current process that prioritizes defect size could make a step improvement through use of local properties, but this is constrained by a technology gap. A near-term stopgap is suggested, as elaborated in the chapter.

41.2 Background

The earliest pipelines transporting oil and natural gas in the United States first moved natural gas from a local source to and between nearby houses through hollow logs in Fredonia, NY circa 1825, while oil was transported first near Oil City, PA through a short 2-in. diameter cast-iron system in the early 1860s [1]. Pipelines came into use in Canada in 1862, moving oil from the Petrolia field near the city of Sarnia, ON into the city, with natural gas being transported first by pipeline a year later moving the gas through a cast-iron system into Trois Rivières, Québec [2]. Pipelines have since been increasingly involved in raw energy transportation. However, when lower volumes were required, petroleum products such as oil were also moved from the wellhead to nearby railheads via horse-drawn carts. In those modes of transportation, the product was contained in large wooden barrels, whose significance remains apparent today as the unit of measure for the quantity of liquid being transported or marketed. However, the use of these transport modes diminished as the volume transported grew with demand and the distance to markets increased.

Demand grew as petroleum-based products were found increasingly merchantable for lighting and other purposes, and as the population expanded, which over time became focused in larger urban centers. As the demand grew, the locally available supply from nearby shallow wells dwindled, which forced deeper wells and eventually led to the discovery and development of major supply basins. With growth in demand, and the easily accessible product exhausted, the supply basins have become increasingly remote to the markets. More critically, in many such cases, the desired resources (i.e., oil and natural gas) often lie in reservoirs associated with other aggressive fluids and gases, with higher temperatures involved for the deeper wells, whereas very low temperatures and ice or permafrost complicate arctic exploration and production (E&P).

While many factors can be involved, where large volumes of oil or gas had to be transported over long distances, pipelines became recognized as the most economical mode of transportation from the wellhead to the market. Such considerations paved the way for pipelines to replace the early modes of shipping, although the barrel remains the metric for the quantity of liquid delivered. Pipelines remain the best method to transport such products when feasible. However, where the geography or other factors make this mode impractical, transport by ship was adopted. As onshore resources near to market became inadequate, E&P moved offshore and into sometimes very remote pristine areas, such as Alaska or further out into the Gulf of Mexico—with safety in those settings being more in jeopardy. This is evident from Figure 41.1 [3], which presents the risk involved in oil and gas E&P and that of other industries in terms of incident likelihood in the y-axis and their dollar consequences in the x-axis, wherein offshore E&P lies along the upper bound of these metrics.

Onshore in the lower 48 US states, pipelines are used to move the imported as well as domestic supplies to market. This is also true onshore in other countries around the world. As the supply basins became increasingly remote, a system of high-capacity pipelines was developed to transport crude and refined products, and natural gas, from the supply points to the market centers, hubs, or terminals. Thus, a cross-country infrastructure of pipelines developed that traverses throughout the United States and adjacent countries, with eventual cross-border interconnects. Such development continues today in the United States, particularly concerning shale E&P and related processing, and also worldwide, wherever demand must be tied into supply.

At present, the natural gas transmission system comprises more than 210 natural gas pipeline systems and in excess of 300,000 mi. of interstate and intrastate pipelines [4]. The system that is committed to moving crude and other hazardous liquids also involves a large number of operators but is somewhat shorter, being in excess of 170,000 mi. [5]. As contrasted elsewhere in detail [6], differences exist in the design and components of gas versus liquids transmission systems because of key differences in the products being transported, and the circumstances involved in maintaining efficient operation. Critical in this context is compressibility and the major differences in the phase boundaries of the transported products in pressure–temperature–volume space. Equally important are the terms and conditions of the tariffs and product quality specifications, and the need to keep liquid water out of the systems to manage internal corrosion (IC).

Transmission pipeline systems exist “midstream” in the oil and gas value chain that begins “upstream” with E&P and ends “downstream” with a distribution network to the end user. As such, a transmission system comprises end-point and distributed storage in various forms, as well as prime movers, pressure controls, and other related support facilities that include a supervisory control and data acquisition (SCADA) system, and measurement and regulator stations. These system elements can be discriminated with regard to their location and function. These elements are (1) either located at discrete sites (e.g., facilities, compressor/pump stations, storage) or (2) part of the often-remote lineal aspects (e.g., pipeline, mainline valves). Regarding function, and its implications for integrity, the elements either do not involve moving parts (i.e., they are static, such as the linepipe) or they involve moving parts (like prime movers and other equipment). This chapter specifically addresses the network of transmission linepipe and how that linepipe was built into a pipeline.

Logically, supply basins were sought as close to the markets as possible. Pipelines constructed between these points tended to develop first along routes that were most easily traversed, which also targeted the largest markets—tracking the so-called path of least resistance. Early pipelines often traversed open country, such as farm fields. Construction made use of ditching machines and hand shovels, with mechanized equipment used as it became available. The pipeline was constructed beside the ditch as a string and then lowered onto the ditch bottom. A-frames were used to lower-in the pipe string, which gave way to side-booms beginning in the 1930s. Where the ditch bottom was smooth, it also served as the foundation for the pipeline, with native soil returned to the ditch as bedding and padding if needed, as well as cover for the pipeline.

Given that the history of pipelines in the United States traces back more than a century, it is instructive to contrast the changes over time, as these affect all aspects of pipelines, from routing and design to construction and operations and maintenance (O&M). Figures 41.2–41.4, adapted from Ref. [7], provide perspective for such changes and indicate the timeline over which these practices were prevalent, specifically with regard to the United States. These figures, respectively, indicate the evolution in the type of steel used to make the body of the linepipe, how that steel was manufactured into linepipe, and how that linepipe has been joined, or otherwise built into pipeline systems. To facilitate later discussion of the evolution over time, each figure has been divided into three timeframes—up through ~1940, which is labeled therein “early,” between 1940 and 1970, which is labeled “vintage,” and beyond 1970, which is labeled “modern.” The last in this sequence of dates, 1970, has been chosen in part because that is when the current regulatory framework was introduced in the United States, and in part because that date can be associated with a transition in steel and pipe making. The first of these dates is somewhat more arbitrary, being chosen to reflect a transition in many aspects of construction, the beginning of a period of change with regard to steel and pipe making, and the onset of a period that opened to the construction of a significant portion of the current pipeline network. Following discussion with regard to Figures 41.2–41.4, this chapter focuses on the pipeline network prior to what has been termed the “modern” era.

41.3 Evolution of Linepipe Steel, Pipe Making, and Pipeline Construction

Figure 41.2, adapted from Ref. [7], illustrates aspects of the evolution of the steel-making processes used to make linepipe, over a timeline from circa 1900 through modern technology. The steel-making process is central to pipeline design and its in-service integrity, as it affects the chemistry of the steel, which coupled with processing affects the microstructure and the presence and location of impurities or undesirable constituents (cleanliness). All such aspects couple to control the steel’s flow and fracture response, details for which can be found in Ref. [8].

Significantly, the processing of the steel with regard to rolling and other thermal mechanical (TM) controlled processing can have a beneficial effect on the flow and fracture response, but also can have a negative effect—depending on the TM conditions and history. What follows is a thumbnail sketch of such aspects, which as mentioned earlier have been addressed elsewhere with regard to pipeline applications in great detail [8].

Steel manufacturing in the United States began with the introduction of the Bessemer process in the mid-1800s, which found use for the production of butt-welded, lap welded, and seamless (SMLS) pipe. The open hearth process quickly followed (1870s–1880s) and evolved into the primary method for steel production in the world by the early 1900s. These developments were followed by the electric furnace (1899), which began production between 1900 and 1910. Subsequently, basic oxygen steel making was implemented by the mid-1950s, and by 1969 accounted for nearly half of the annual steel production. Of these processes, steel produced for linepipe through the 1960s included the open hearth, basic oxygen, and Bessemer processes, with electric furnace steel used only in limited quantities.

Prior to the 1960s, ingot casting and hot rolling were typically used to produce plate and skelp for linepipe production. Both partially deoxidized (semi-killed) and fully deoxidized (killed) steels were used, and depending on the deoxidation practice, ingot structural soundness varied. During this same period, higher yield strength materials were introduced, wherein the primary method used to increase strength was increased alloy content, which tended to increase the hardness and reduce the weldability. Pre-1960 steels have higher residual impurity levels and more frequent internal anomalies as compared to post-1970s production. In many cases, the impurities are strung-out parallel to the pipe surfaces, which while they have limited influence on strength, they can act as initiation sites for some forms of corrosion and/or cracking.

From the late 1960s and into the 1970s, several major improvements were implemented in steel manufacturing and in plate/skelp rolling. Microalloyed steels with additions of niobium, vanadium, and other elements, coupled with improved steel rolling practices (controlled rolling) and improved impurity controls (desulfurization, inclusion shape control, vacuum degassing), resulted in “cleaner” steels with higher yield strengths, increased toughness levels, and improved weldability. Continuous casting (concast) began to be used in the same time period, further improving steel quality, while providing more efficient production. Nevertheless, centerline segregation in the concast strand can remain a concern in demanding applications (e.g., [8]).

Further developments in steel production and rolling practices came in the 1970s and 1980s through control of steel microstructures, additional rolling method improvements (accelerated cooling), and chemical additions. These gave rise to higher yield strengths, quantified by the “grade” of steel, which means stronger pipe for the same wall thickness. For linepipe, grade is denoted by the prefix X followed by the specified minimum yield stress (SMYS) in ksi units. These changes in the steel also led to fewer impurities, and improved weldability. Sophisticated steel manufacturing controls and improved nondestructive inspection (NDI) systems have also resulted in a reduction of material-related anomalies found in modern linepipe, with Grade X100 now considered a commercial product. Key in this context is that steel has evolved to address the needs of the pipeline industry with significant quality and related developments being triggered by the need to manage concerns such as brittle fracture and running ductile fracture.

Development was also spurred by the desire to manage capital expenditures (CAPEX), which is evident in the increase in grade as mentioned earlier, which allows reduced wall thickness and so increased productivity on the spread during construction. It is also evident in changes in the rolling mills to provide wider plate and skelp, which permit producing larger diameter pipelines, which diminishes the time on the spread, as well as the scope of related O&M work, to help manage operating expenditures (OPEX).

Figure 41.3 illustrates the evolution of in-linepipe production. This involves transforming flat stock as plate or skelp into either a cylindrical shape that is closed by a longitudinal seam, which can be straight or helical, or a SMLS cylinder. Depending on how the linepipe is produced, and the need for dimensional control, pipe can be cold expanded up to 1.5% following completion of the pipe-making process. Thereafter, it is tested and accepted, then beveled and otherwise prepared, in accordance with contract and other specifications and regulations. Early pipe-making processes noted in Figure 41.3 include furnace butt-welded, lap and hammer welded, low-frequency (LF) electric resistance welded (ERW), flash weld, and continuous butt-welded, with the latter still in use.

Subsequently used processes included single-submerged-arc welded (SAW), and some early variations on SMLS pipe, with high-frequency ERW (HF-ERW), and double submerged arc welding (DSAW) coming thereafter. The DSAW process has been used to produce both straight seam as well as spiral (helical) seam pipe. Of these, modern construction relies on SMLS pipe and the HF-ERW and DSAW processes. In contrast, the “early” category involves several long-since-abandoned processes, as does the “vintage” category. The vintage category can be viewed as a period of transition away from steel and seam making practices that were prone to process variation and production upsets, which lead to seam-related defects. Thus, as noted with regard to the evolution of steel, changes have come in pipe making to improve quality. Changes also have been made with regard to capacity of the mills to produce larger diameter, and heavier wall pipe, as needed to manage CAPEX and OPEX.

Finally, in analogy to Figures 41.2 and 41.3, Figure 41.4 illustrates the evolution of construction practices that have been used to join the linepipe, and to create the directional changes needed to track a routing via bending. It is the evolution in practices to join linepipe, and to create the needed directional changes that account for the improvements in pipeline construction—much of which has been driven by the desire to control CAPEX upfront and the OPEX over the life cycle of the pipeline network. As early practices were found to be problematic with regard to managing CAPEX or OPEX, they were quickly discontinued, with the related issues promptly resolved. In this context, it is not a surprise that much of the vintage construction remains in service today without issue. For this reason, a later section illustrates novel construction practices, wherein their utility and continued serviceability are evident.

Early joining practices relied on threaded collars, which were used on the earliest crude oil pipelines, with mechanical couplings coming into use as well. The earliest welds were made via the oxy-acetylene process, which as time passed incorporated a backing ring below the girth weld. The use of flanged connections, which were approved by the ASME circa 1894 and require welding to attach the fitting to the linepipe, followed as the early welding technology was more broadly adopted. Of these, only flanged connections continued in use beyond the period denoted “early” in the figure. Figure 41.4 also indicates the role of wrinklebends, which were used with other historical field-bending practices, such as welded segmented (miter) bends, through the 1940s. As evident in this figure, the period denoted “vintage” also was a time of transition from the early practices to those in use today—such as submerged arc and other more advanced methods that in modern applications are used under automated conditions and work in conjunction with line-up clamps. The transition to and through the vintage era on into the modern era also gave rise to the use of machine-made field bends, with such machines emerging in the 1940s and continuing to develop since, for straight as well as helical seamed pipe. Directional drilling also traces back to a similar timeline, and has recently become viable for long-distance drills involving large-diameter construction, providing a viable ecofriendly approach to deal with routings sensitive to such issues. And, where difficult routings move through rocky and hilly terrain, bedding and padding and other trench-bottom technology have evolved largely in the modern era to manage such construction.

The transition through the vintage to the modern era opened the use of cathodic protection (CP), beginning in the 1940s, the use of which expanded until it was eventually mandated in the transition into the modern era. As the threat of corrosion was recognized, over-the-ditch coating methods also went into use in the 1940s. Step improvements in coating technology began with the early versions of today’s resin-based formulations that appeared in the 1970s. Continuous improvement in these technologies, coupled with the realization that appropriate preparation was as critical as the nature of the coating system, and its application, have now effectively controlled external corrosion (EC). Likewise, tools and technologies have evolved to better manage IC, which range from designed inhibitors through internal coatings. Field applied coatings likewise have improved to deal with surfaces left uncoated in the mill to facilitate girth welding.

Finally, the transition through the vintage to the modern era gave rise to significant changes in pipeline system monitoring and control. The first monitoring via in-line inspection (ILI) came late in the transition from the vintage to modern era, in 1965, which since has gone through several generations of development, and expanded its initial role with regard to metal loss to deal selectively with cracking, mechanical damage, alignment, and deformation. SCADA introduced initially to better track product delivery was subsequently coupled with sensors, remotely controlled valves, and system-specific software and found other uses, such as tracking pressure and other operational parameters, which facilitate better control, and also leak detection and outflow control.

41.4 Pipeline System Expansion and the Implications for “Older” Pipelines

The evolution in the aspects of pipeline systems that were apparent in Figures 41.2–41.4 has been “fueled” by the competitive need to manage both CAPEX and OPEX as pipeline companies expanded their capacity in the competition to meet the ever-growing demand for energy. In this context, that competition to get off the construction spread and into production led to improved steels that could be produced in stronger grades that also were more weldable. Such steels also were tougher and had lower transition temperatures, whose use opened to mechanized more controllable welding processes. The push to build more efficient systems over the life cycle led to improved operational controls as well as better coating processes, and improved control of CP. As such, overall system safety has the potential to improve as a consequence of the evolution evident in Figures 41.2–41.4.

41.4.1 System Expansion and Construction Era

The extent of the expansion to meet the market demands is evident in Figure 41.5, wherein the incremental mileage constructed is shown in the left-y-axis as a function of the period of construction over the interval from before 1920 into the new millennium, which is shown in the x-axis.

The right-y-axis presents the cumulative mileage constructed over this same interval. Figure 41.5a shows these data for interstate natural gas transmission pipelines, whereas Figure 41.5b shows the corresponding data for interstate hazardous liquid transmission systems. The timeline in these figures denoted >2000 includes results up through 2010, while the timeline used in Figures 41.2–41.4 ends at the start of the new millennia. Little is lost in this context as that decade saw little technology growth—except for developments in inspection capabilities. Differences also exist between Figures 41.2–41.5 with regard to the start of the trending. However, again little is lost because the accumulated mileage prior to the 1920s represents a trivial fraction of the total natural gas and hazardous liquid mileage thereafter. Inspection of Figure 41.5a and b indicates that the “early” era comprises a nominal fraction of both the networks.

It is apparent from Figure 41.5a that the natural gas system experienced strong growth beginning in the 1920s, which culminated in the fastest growth in the 1950s and 1960s at a relative rate that has since not been matched. The trends in Figure 41.5b indicate that the hazardous liquid system also experienced steady incremental growth since the 1920s, which led to the fastest growth in the 1950s and 1960s, that like the natural gas system its growth has been leveled and steady since the 1960s. As such, it is not surprising that the amount of mileage split between the periods designated in prior discussion as early, vintage, and modern, is comparable. This is evident by comparing the split mileage built within these three timeframes, which is apparent in the corresponding sets of numbers added into Figure 41.5a and b. For the natural gas system this split is 6% prior to 1940, with 52% from 1940 to 1970, and 42% since 1970, while for the hazardous liquid system construction prior to 1940 involved 7%, with 49% in the interval from 1940 to 1970 and 44% thereafter.

The horizontal lines shown in Figure 41.5a and b correspond to one-half of the current total system mileage. The vertical lines that drop from its intersection with the cumulative mileage trend to the timeline shown in the x-axis indicate that for the hazardous liquid and natural gas transmission systems roughly one-half of those networks were constructed prior to ~1961, with about two-thirds of these networks being constructed prior to 1970. Recognizing that the amount of the network built in the “early” era is but a nominal fraction of the total network, it is evident that roughly one-half of the US transmission network was built in the era designated as “vintage” in the context of Figures 41.1–41.3.

Figure 41.5 indicates that the step increase in mileage for the natural gas system began in the 1950s, which is paralleled by the growth in the liquids mileage. While length provides an indication of transmission pipeline system expansion, the volume transported provides a better measure of its growth.

Volume or capacity is controlled in part by linepipe diameter, which is dictated by mill capacity to produce wider skelp and plate. It is also controlled in part by pressure, which under working stress design (WSD) (e.g., [10]) is controlled by SMYS, or equally the pipe grade. Suffice it here to note that by the start of the 1960s, the mills were broadly capable of producing pipe in Grade X52 for which SMYS is roughly double that for Grade A, with diameters in the order of 30 in. (or more)—much larger than several decades earlier. In this regard, the capacity per unit length increases with the square of diameter and linearly with pressure. Thus, even though the incremental length per decade remains flat for these transportation systems since the 1970s, the network capacity has increased. Considering that diameters up to the order of 48 in. have entered service in the United States in grades up to X80 since then, an almost fourfold capacity increase has occurred relative to a 30-in. X52 pipeline, which is otherwise transparent when capacity is assessed in terms of mileage.

Pipe diameter affects the wetted area on the inside diameter (ID) surface and so increases the threat of ID corrosion. It also affects the outside diameter (OD) surface area that must be coated and protected to control the threat of OD corrosion. Accounting for this difference in surface area over time should be considered as improvements in the Pipeline and Hazardous Materials Safety Administration (PHMSA) database lead to a more comprehensive framework and better quality results free of nulls. Until then, mileage remains the only viable metric to assess expansion and normalize trends over time.

It follows from Figure 41.5 that roughly one-half of the total transportation network has capitalized on the evolution in steel and linepipe production, and the technology and schemes used to build them in what has been called the modern system. What does this imply for the remainder of the transmission systems constructed prior to 1970? There are two basic approaches to address this question. The first is by analogy to the circumstances evident for parallel applications, such as other transportation infrastructure, systems, and components. The second is by direct analysis of the oil and gas transportation infrastructure and systems. Both approaches will be considered in turn.

41.4.2 Qualitative Assessment of Construction Era and Incident Frequency

Consider next parallel transport systems whose infrastructure involves components whose age spans many decades, as occurs for some commercial aircraft and ground vehicles and some pipelines, and portions of other transportation systems, such as bridges and rail transport, whose age approaches that of early pipelines. Some parts of the transportation infrastructure, such as bridges, share the same design basis that is used for pipelines. Bridges, pipelines, and many other structures, initially relied on WSD, wherein a factor of safety was applied to the allowable stress to achieve a safe margin against uncertainty. As time passed, WSD gave way to plastic design or ultimate strength design (USD) for many new bridges and other structures (e.g., [11]). The virtue of USD was that it capitalized on the significant reserve strength evident in the postyield response of typical early-vintage construction steels. It was accepted into practice, even though it offered less reserve capacity as compared to WSD, and is accepted also as a design basis for pipelines in a few countries. Reliability (risk)-based design (RBD, e.g., [12]), which relies on a probabilistic view of structural loading and resistance, also has found favor for structural design in some countries. Canada and a few other countries also allow RBD for pipelines. However, in the United States, pipeline design remains tied to the WSD concept, as does general boiler and pressure vessel design.

Other transportation systems (such as the rail and air transport systems) operating in most countries worldwide, also parallel key aspects of pipeline systems in many ways. Like pipelines, such systems provide for long-distance transport of commodities, but unlike pipelines also provide for public transportation. Similar to a pipeline, the paired ribbons of rail that are the backbone of railroad systems are not structurally redundant—if one of the pair of rails fails it is quite possible that the train will derail. And as for a pipeline, such failures can be as catastrophic depending on where the failure occurs and what commodity was being transported. In contrast to rail and pipeline systems, designs underlying air transport seek structural and subsystem redundancy. Yet even though redundant in concept, related failures occur in air transport causing casualties. Human factors likewise contribute to such failures, as they have in some pipeline incidents.

As for pipeline systems, both the rail and air transport systems rely on periodic inspection to detect system faults prior to their becoming significant. Both rail and air transport systems are “damage tolerant” (e.g., [13]) in that they continue to operate even though cracks have formed and are growing in structurally critical elements. Both rail and air transport systems are subject to periodic reinspection and condition assessment. Finally, both rail and air transport systems continue to operate equipment and facilities provided, they are maintained and so “fit-for-service” (e.g., [14]) and also “fit for their intended purpose” (e.g., [15]) in that system. However, both rail and air transport systems also consider functionality and efficiency and in some cases will “retire assets for cause” (e.g., [16])—where that action can be motivated by brand concerns, inefficiency, and other considerations, including structural reasons. Competitive drivers also can lead to the sale of assets and to modernization in the rail and air transport industries, just as they have for pipeline systems—or for any other business.

It follows that pipelines built under conditions prevalent in the eras defined as “early” or “vintage” can be as viable as those built in the “modern” era as long as they are fit-for-service and for purpose. This is clear in the observation that some cars and big-rig trucks continue to operate for many hundreds of thousands of miles, whereas absent adequate maintenance they could fall short of their expected design life. Because a vehicle is old does not mean it is unsafe, but it can mean that it could require more care, or might not be as fast as its modern counterpart. This also applies in the context of vintage aircraft, many of which are still in service, but for brand reasons, economics, and other considerations they do not operate in the fleet of a major US carrier. This same brand-motivated decision is made when a two-year-old car is sold solely because the owner does not want to be seen in an “older” vehicle.

41.4.3 Quantitative Assessment of Construction Era and Incident Frequency

While early or vintage pipelines are analogous in most ways to other early transportation infrastructure that remains in daily use today, whose serviceability depends on adequate maintenance, the best metric of serviceability over time is a quantitative analysis of incident statistics. For pipelines, the most comprehensive basis for statistical analysis and incident trending is the pipeline incident database that has been developing since 1971 under the management of the PHMSA. The PHMSA functions under the US Department of Transportation (DoT) [17], which is responsible for all modes of transportation, such as motor carriers (under PHMSA) and air transport (under the Federal Aviation Administration). PHMSA’s pipeline incident database is publicly available and can be downloaded directly from the PHMSA website [9].

Fundamental to any statistical analysis is the quality and consistency of the data, and which data should be pooled or parsed for the purposes of a given analysis. A check of the PHMSA website indicates that their database has been packaged over time intervals during which the definition of a reportable incident (e.g., see [9]) and their input form and scope (e.g., see [18]) remained constant. As pipelines became increasingly encroached and other factors such as the growth of social media came into play, the sensitivity to the consequences of an incident has increased. Such factors have prompted changes over time in (1) the threat categories used, (2) the criteria used to define a reportable incident, and (3) the scope of data that must be reported. It must be noted in this context that as reporting criteria become more stringent, an event that might not have been reportable in past becomes reportable. More stringent criteria open to a bias toward increased incident frequency as time passes. Because the scope of the data required to offset this bias was not always a required input, and so is often not available, either this bias remains or the database must be culled to exclude it—which precludes trending over time. Over the period since 1971, three major changes can be identified in reporting structure and requirements, which can complicate trending, and attempts to offset any bias due to increasingly stringent requirements. Data quality issues also exist, as do many nulls (or empty reporting cells), which also can affect the utility of the trending. As nothing can be done to resolve the data quality concerns, the results are used as is without regard to the possible effects of the nulls.

Because the focus is a possible effect of the era of pipeline construction on the incident rate for the onshore pipeline system, the construction or manufacturing data listed in the PHMSA’s database for onshore interstate pipelines has been pooled and then sorted by construction era as “early” (<1940), “vintage” (1940–1970), and “modern” (>1970), specifically for the liquid pipeline system. The raw outcomes were then normalized relative to the amount of mileage constructed in each of those eras, as reported in Figure 41.5b. Such analysis has been done with regard to the incident data for the modern era up through 2010, as well as up through 2002, because at this time there was a marked shift in the reporting requirements for the liquid system. This analysis was done with regard to the elements of a pipeline system that differ with regard to location and function. That breakdown with regard to location parsed discrete elements (facilities such as stations, tanks, etc.) and the lineal element (including the pipeline and its mainline valves (MLVs)). The discrete elements were further sorted by function with regard to static, such as tanks, versus dynamic, which covered equipment and related components.

Because there were no clear trends across these parsed elements, only the results for aggregate system are presented in Figure 41.6. The y-axis in this figure is the mileage-normalized incident frequency for each of the construction eras, which are parsed on the x-axis. Each era in this bar graph includes two results—the first represents the data evaluated through 2010, while the second represents the data evaluated through 2002, when much more stringent reporting criteria for leaks in hazardous liquid lines were introduced.

The results of the quantitative analysis of the hydrocarbon transportation infrastructure shown in Figure 41.6 do not indicate that the portions of the pipeline system constructed in the “vintage” or the “early” era are more prone to incidents. It is evident in Figure 41.6 that the difference in the relative fractions of pipelines constructed in the vintage and early eras is small, with the incident fraction for the vintage system being ~5% less than for the early construction. While there is a modest decrease in the relative fraction of reportable incidents moving forward in construction era from early to vintage, Figure 41.6 indicates that the relative fraction of reportable incidents increases sharply moving into the modern era of pipeline construction. Contrasting the results for the database before and after the transition to more stringent reporting requirements circa 2002 accounts for roughly one-third of this increase. This outcome opens to the question: what accounts for the remaining upswing in relative incident frequency for modern construction? The results in Figure 41.5b do not indicate that this can be traced to a surge in construction, which could open to the need for contractors and so introduce a less experienced workforce. Further trending of the modern construction era indicates that while other changes leading to more stringent reporting occurred within this era, they played a minor role if any in the upswing in relative incident frequency for the modern construction. Rather, the results of trending for the modern era in 10-year increments suggest there was a period of learning and/or adaptation as the US industry incorporated the many improvements in steel making that were phased in circa 1970 (see Figure 41.2). For example, the transition from LF-ERW to HF-ERW longitudinal seam production that began in the early 1960s and was completed in the United States in the early 1970s (see Figure 41.3) experienced some developmental issues into the 1970s. Likewise, as the use of emerging construction practices broadened, there is evidence in trending the data that this aspect also experienced some developmental issues in the 1970s. If the trend evident for the 1970s was more like that of the 1980s or the 1990s, then the upswing evident in relative incident frequency for the modern construction era is diminished—with little difference evident over time.

It follows that parallel transportation systems share comparable traits with no clear evidence that era of construction or age affects the mileage-normalized incident frequency for the hydrocarbon pipeline system assessed with regard to Figure 41.6. In contrast, a major failure within the pipeline industry and the parallel transportations systems considered earlier often opens to the discussion of a system’s or component’s age. Although the risk to the public or the environment associated with the trends in Figure 41.6 is not quantitatively out of line with other day-to-day risk exposure, it is often involuntary exposure. While such events are infrequent, such exposure can lead to environmental issues and loss of life, whose cost to the pipeline operator is not trivial. Recognizing this, since the 1930s the industry has worked to develop consensus standards and company specifications and protocols for O&M. And, as the public and environment can be adversely affected, a regulatory framework was introduced circa 1970. These aspects and their implications with regard to older pipelines (i.e., pipelines constructed in the vintage and early eras) are considered next.

41.5 The Evolution of Pipeline Codes and Standards, and Regulations

It is evident from Figures 41.2–41.4 that today’s modern pipeline systems have evolved greatly, from routing and design, on through construction, and into O&M. The design codes, related specifications, and regulations have likewise evolved. What began in the 1800s based on rudimentary schemes has evolved and now is directed at system safety over the life cycle of the system. Steel has replaced wood and cast iron, and couplings have been replaced by automatic welding, with inspection and controls now in use with a view to ensure the product remains in the pipeline, and moves in a safe, serviceable, modernized pipeline network. Yet, as just discussed, incidents occur on new as well as the vintage or early aspects of this transmission network—so factors other than age are involved.

41.6 Some Unique Aspects of Early and Vintage Pipelines

The background and related discussion of the evolutionary timeline for steel, linepipe, and construction methods identified eras in the timeline for pipeline construction. The term “vintage” covered the period between 1940 and 1970, while “early” applied <1940 and modern applied to after 1970 to coincide with the introduction of the Federal Regulations. As discussed with regard to Figure 41.6, this date also reflects a time when many major changes occurred in steel and pipe making, as well as in construction practices. The transition date from “early” to “vintage” was similarly coupled to the start of a period of change regarding to steel, pipe making, and in construction practices, and the onset of the period when a significant portion of the current pipeline was constructed. This section presents illustrations for a range of construction features that are unique to early and vintage construction eras yet remain serviceable and safe by virtue of the operator’s O&M protocols and management plans, and diligence in their implementation. While the focus here is on “older” pipelines, serviceability and the level of safety it ensures are essential regardless of the era of construction.

Hints regarding design and construction aspects that are unique to “older” pipelines—the focus of this chapter—can be identified with regard to Figures 41.2–41.4, as practices that have been displaced over time. Not surprisingly, some of these aspects were identified in the NOPR for parts of the then pending Federal Regulations (e.g., [26]) circa 1970. Notable concerns identified then included branch connections, bends, elbows, wrinklebends, miter bends, and dents. It is noteworthy that except for dents all of these concerns involve a change in the direction of the pipeline. Subsequently, the construction features covered in the early notices were augmented in reference to different long seam types, and couplings. The regulatory structure dealing with such aspects developed via a “grandfather clause” that as presented circa 1970 permitted the continued operation of the pipeline under its recent operating conditions without a pressure (proof) test. Where uncertainty existed, some types of linepipe long seams were derated, while others were qualified based on the results of full-scale testing (e.g., [34]). Differences between liquids transportation practices and those for natural gas, which have prompted concern for fatigue, led to some key differences between these vintage networks. It seems likely that many of these aspects will be a focus of regulatory action in the wake of the public workshop on integrity verification process (IVP) (e.g., [35]), which currently considers the gas transmission system.

In general, a process or technology is displaced over time because it is inefficient or ineffective, where such metrics (inefficient or ineffective) are generic with regard to time, cost, fitness for service, fitness for purpose, and so on. What is possible in terms of these metrics is conditioned by the materials, the manufacturing and production capabilities, and the technology available. The routing and related terrain (topography, soil, moisture/water, ice, permafrost, etc.) further condition what is feasible/plausible. Terrain and in particular the type of soil (hardpan, sands, silts, etc.), the presence of water (river, stream, creak, swamp), and the surface topography are major factors coupled with the routing, as they influence the need for vertical and horizontal bends. Topography and local conditions also determine the angle that must be achieved. Against this background Figures 41.7–41.21 introduce views of the construction process as the basis to better understand its evolution over time—and plausible integrity concerns that need to be considered. Of those, Figures 41.7–41.15 characterize the “early” process, while Figures 41.16–41.20 characterize changes as the construction process transitioned into the “vintage” era, with just one committed to illustrate the challenges addressed in the modern era. These figures and the related discussion lay the foundation to outline integrity management (IM) practices for such pipelines.

41.6.1 “Early” Construction Practices

Figures 41.7–41.9 present very early views of the construction process circa 1913. It is apparent in those figures that the linepipe is being transported to the spread by wagons pulled by a hitch of mules. While comparable images along the spread are not available, recognizing that the automobile appeared in the United States circa 1903, it is likely that this same mode was used to string the pipe where feasible.

It is apparent from the scale of Figure 41.7 that these pipes are about 20 ft. long and so are single-random joints about 12 in. in diameter. Loading and handling that size and weight posed a challenge in 1913. Although just a single hitch of mules was used, it is clear from this photo that a large number of joints could be moved along a dry road over flat terrain. However, where hills had to be negotiated or the soil was soft along the right-of-way, the number of joints transported had to be decreased. Alternatively, it is also plausible that self-propelled equipment comparable to the ditching machine used on this 1913 construction project could have been used. As Figure 41.8 shows, the design of that machine incorporates much wider wheels and includes bladed chain-driven wheels, which could facilitate moving heavier loads under adverse conditions.

While electric arc welding had emerged in the United States prior circa 1907, oxyacetylene welding was available earlier (circa 1903 in the United States), and was better developed as compared to arc welding at that time [36, 37]. However, the limited ability to weld uphill with the oxyacetylene process meant that adjacent joints of pipe had to be rolled in sync to complete a girth weld, often termed a “rolled-weld.” Because longer sections of jointed pipe opened the door to handling issues and the stresses induced by handling, the use of welding was limited to double jointing. Thus, completing a pipeline using larger diameter (plain-ended) pipe required couplings to join the double-jointed segments. To minimize handling issues, doubled jointing and coupling were done over the length of a string of joints adjacent to the ditch. (Construction in that era also made use of single joints with couplings used between each joint.) Ditching was done by hand, and facilitated by machines such as that in Figure 41.8 which pulled a ripper tooth/plow that broke the surface cover and opened the upper part of the ditch. Thereafter, the trenching was finished by powered shovels or by hand, and timbers were laid across the ditch. The string of coupled pipe segments was then rolled over the ditch onto the timbers, after which the string was lowered-in manually, with the assistance of leverage and an A-frame, as is evident in Figure 41.9.

As is evident in Figures 41.9 and 41.10, lowering-in was accomplished with the help of mechanical advantage gained through a long lever supported by a stout timber fulcrum that was set on A-frames either side of the trench. This lever-fulcrum scheme provided lift for the pipe string. The string was lifted joint by joint, such that the local cross-ditch timber could be removed with the pipe string incrementally settling into the ditch as this process was repeated stepwise along the trench. This process was continued sequentially along the ditch with subsequent strings coupled onto the prior string that remained at ground level on the timbers—such as that evident in the foreground of Figure 41.10.

The overview of the lowering-in process in Figure 41.10 clearly shows the cross-ditch support timbers, the A-frame supported fulcrum, and the long lever used to lift the pipe string. The pipe string evident in the foreground can be seen fully settled into the ditch at the left margin of this figure, with the transition into the ditch evident behind the crew working the leverage. Two of the couplings used to connect the single or double-jointed pipe segments can be seen in the pipe string in the foreground of this figure. In particular, 1913-vintage Style 38 Dresser couplings were used for this construction. Figure 41.11 illustrates this style and its function in reference to the current form of this coupling. While the passage of time has brought minor changes in its design and production, its function and overall appearance remain the same today as they did a century ago.

As evident in Figure 41.11, installation of Style 38 couplings began by running the slip-fit collar onto one of the pipes, after which the second pipe was brought into position and the collar then shifted back centrally over the ends of the joints being connected. A seal between the pipes was created through elastomeric gaskets that were located toward the ends of the coupling, as can be seen in Figure 41.11a. The schematic in Figure 41.11b indicates that tightening the bolts across the flanges of the coupling created a compressive wedge load on the elastomer, with the radial component of that load compressing the gasket between the pipe and the coupling, which locked the coupling into position and sealed the joint. Depending on the length of the “middle ring” (which is dimension B in Figure 41.11b), company construction specifications and rehabilitation protocols typically limit the total angle formed across the coupling to the order of 4° or less.

It is apparent from Figure 41.10 that such couplings could support the shifting and resultant angular distortion that formed in association with this lowering-in process. As such, they were also capable of accompanying the shifting and distortion as the coupled pipe joints that lay strung along the ditch were rolled out over the ditch onto the cross-trench timbers. Couplings also were developed to provide for larger-angled bends, as shown in Figure 41.12. These special-order Style 56 angled Dresser couplings were based on their Style 38 coupling with a mitered (sometimes spelled mitred) fabrication (Figure 41.12a) or a smooth wrinklebend in place of their usually straight middle ring. Based on Ref. [39], such mitered and smooth wrinklebend fabrications became possible in the 1930s—or toward the end of the early era.

Because the rubber/elastomeric gaskets of a half-century or more ago were prone to degrade (dry out) as time passed, or they were exposed to certain natural gas liquids or tramp species in the product stream, some joints tended to leak. Soil stresses due to shifting soils on slumping slopes could likewise cause the pipe to move relative to the coupling and/or pullout in part or completely, which also led to leaks. Tabs welded across the coupled joint to electrically “bond” the upstream and downstream pipe for purposes of CP continuity, as illustrated in Figure 41.13, also could help to resist pullout. In contrast, leaks due to aging of the early rubber gaskets could not be easily managed given the seal materials then available. Although timely maintenance could offset such concerns, as time passed some operators found it cost effective to abandon their very early coupled high-pressure lines. Other operators dealing with lines constructed toward the end of the early era determined that they could be effectively reconditioned by a process that replaced the couplings with girth welds, and periodically inserting pups to make up lost length (e.g., [40]). Thereafter, the line could be recoated as needed.

While shop fabricated miters and smooth wrinklebends could be reliably produced toward the end of the early era, the result of their use as field construction methods could vary significantly even then, depending on the experience of the crew involved, the tools available, and the techniques used. This is illustrated in Figure 41.14, which contrasts two wrinklebends that were reportedly made in the 1930s within a few years of each other. The wrinkle in Figure 41.14a (reportedly to have formed free of field-induced loads) appears to have collapsed into itself, forming a tight kinked axially asymmetric feature that, while not captured in this image, runs to varying depths the full circumference of the pipe. The aspect ratio for this bend—which is quantified by its height to its length and is one metric of wrinkle severity (e.g., [41])—is about 4. In contrast to this high-aspect-ratio bend, the “smooth” wrinklebend shown in Figure 41.14b has an aspect ratio of ~0.2. Comparing these results indicates that wrinklebends that differ in terms of aspect ratio as a metric of severity by a factor of 20 could be encountered in bends produced within a few years of each another. If pressure cycling typical of some pipeline operations controlled failure for these bends, and both were cold bends made on the same linepipe steel and were free IC and/ or EC, then their service lives are estimated according to Ref. [41] to differ by about 5 orders of magnitude. In absolute terms, the severe bend has near zero fatigue life, whereas for typical historic gas pipeline operation the life of the smooth bend is many times the system’s book life.

It is known from historical archives [42] that wrinklebends could be produced under well-controlled conditions that used a stiff strong back in combination with winches and leverage about a fulcrum to produce such bends. That and other operator documentation indicates that this and similar techniques were used in 1931, and possibly earlier, to produce hot bends as well as cold bends, in construction in Ohio and elsewhere in the mid-West. It is also known that much lower stiffness systems that stored significant energy in an A-frame setup were used to make wrinklebends in much earlier construction. As the wrinkle began to form in these more flexible schemes, that stored energy was unloaded leading to a much less controlled process and the likelihood of creating collapsed bends, like that in Figure 41.14a. Figure 41.15a and b illustrates each of these practices.

Where higher angle bends were needed beyond what could be achieved with one smooth wrinklebend, a sequence of such bends was introduced in the pipe, typically separated axially by a diameter or more. Shop fabricated miters are known to the author that were made during the 1930s as special-order items. It is also known that miter bends were broadly used in some field construction during the 1930s in pipelines that function at pressures up to 400 psi.

As experience along with the tools and techniques developed as time passed, some of the early construction practices used to bend pipe or fabricate fittings evolved greatly within the early era, while others were discarded. Still others, such as the design of the Dresser coupling, have remained largely as devised initially, as is evident in comparing the images and descriptions between Ref. [38, 39]. While the concept is largely unchanged, the early rubber gasket has been replaced by a modern elastomeric material, which underlies its continued use today.

41.7 Management Approach and Challenges

As noted in the last subsection, the IM approach depends on (1) diligent response to the regulatory requirements and (2) the operator’s involvement in programs and best practices that carry beyond those directives. Within the current regulatory framework and the technology that has been developed to implement it and using operator-specific protocols and practices, the operator develops an integrity/change management and continuous improvement plan. Implementing that IM plan involves three steps each of which in balance supports IM and the safety and serviceability it ensures. The three steps in IM are (1) identify the potential threats to asset integrity; (2) assess asset condition periodically over the scope of the relevant threats; and (3) respond on a timely basis to remediate or replace the asset as appropriate. The regulatory framework requires the operator to identify and put into play ways to improve their processes and how they apply them as the operator deploys this plan. It also requires that operators implement this plan as part of a cyclical process. The requirements include a maximum interval for this revaluation that is specified as five years for hazardous liquid systems and seven years for natural gas systems, with this process initially focused on high-consequence areas (HCAs) relative to environmental and public impacts. Some perspective on reevaluation intervals and the related aspects during the time that ASME B31.8S was developing can be found in Refs [58, 59].

Because pipeline systems often incorporate aspects from the early, vintage, and modern eras, which through timely inspection and maintenance are equally fit-for-service, each pipeline system is unique. Through mergers and acquisitions, a larger operator might be running five or more systems each of which is unique. Accordingly, the regulatory framework must be generic; its requirements can be prescriptive at a higher level, but it equally must be adaptable on a case-specific level in the details of their use.

Because of this uniqueness and the case-specific nature of the day-to-day aspects of IM, this section outlines the elements of the steps involved in IM as opposed to providing case-specific illustrations of the implementation of an IM plan, as a book dedicated to that topic would otherwise be needed. Key high-level documents that more broadly define these steps from a regulatory perspective are Parts 192 [32] and 195 [33] of Title 49 of the CFR, while ASME B31.8S [29] and API 1160 [30] present the codes and standards equivalent to the regulatory framework. The CFR, ASME B31.8S, and API 1160 all incorporate by reference a host of other codes, standards, and recommended practices and procedures, some of which are included in the list of references that closes this chapter.

41.8 Closure

In view of the evolution indicated in the bar graphs of Figures 41.2–41.4 and the photographic illustrations of the transformation of the industry to meet the challenges posed from the early days, evident in Figures 41.7–41.14, to that shown in Figure 41.21, it is clear that the industry has come a long way. However, as was apparent in the discussion of the inspection technologies required to quantify the pipeline system’s condition, the industry still has a way to go in order to achieve its stated goal of zero incidents.

While the capability to quantify threat severity has improved significantly through improvements in the inspection arena, this capability has evolved more in response to how a sensor can be improved or an interpretive algorithm can be streamlined than it has been driven by the data needs of the analyst who requires inputs on size and shape (and location if in a seam) to quantify anomaly severity. Critical in this context is the realization that severity assessed in terms of defect length and depth or defect area often suffices where plastic collapse controls failure, whereas shape becomes more important where fracture controls failure. Another key in this context is the observation that metrics of fracture resistance are often uncertain, and in the worst case are unknown. This uncertainty, when combined with the fact that defect shape is ill characterized, means that potential failures controlled by fracture are in double jeopardy. Fracture controlled failure occurs at a pressure less than for collapse control, so this double jeopardy develops on the lower-bound for failure, and thus is an issue for pipelines built of steels with limited fracture resistance. It follows that those developing the inspection tools need to collaborate with those that use their data to more effectively quantify pipeline condition. This collaboration with regard to (1) anomaly size, shape, and location if in a seam and (2) properties local to the anomaly is essential if we are to achieve the stated industry goal of zero incidents [19].

The second related concern involves defect severity assessment and prioritization, as follows. For the last several decades, we initially sought to detect anomalies, and thereafter sought to size them so their severity could be assessed. There are two fundamental assumptions in this context. First, this assumes that schemes that would detect a given type of defect would be equally useful and effective in sizing it. Second, this assumes that defect size correlated directly with severity throughout the length of the pipeline. There is no question that size (including the effects of shape) correlate directly with severity and priority if the loads (used generically) are the same and the properties are the same. Reality is that a pipeline has a distribution of properties that also influence failure, which is overlaid on a distribution of loads, which is overlaid on a population of defects that are represented by populations of lengths, depths, and shapes, where the last three also share distributions of correlation coefficients. Therefore, length and depth do not uniquely determine severity or priority.

If the aforementioned distributions were all known in balance for a given pipeline, then the analytical tools of RBD could couple and interpret them, to assess how they combine to control failure. However, that technology cannot identify where failure will occur along the pipeline, much less when. Therefore, it cannot determine if the potential failure will occur in or near an HCA, and/or in which HCA.

Because the linepipe’s properties along the pipeline and sites of unique loadings overlay the defect sizes and shapes found by ILI, the “worst” of the defects quantified by the largest feature is not necessarily the first to fail. Those that have done full-scale tests with a range of defect sizes in the pipe only to see it fail in the pipe body remote to the defect have seen graphic evidence of this reality. It follows that we need to manage integrity in a framework that addresses likely failure sites based on defect sizes, local properties, and locally unique loads. This could be accomplished through an adaptation of the approach and processes used today, wherein the “where” and “when” become the focal point in lieu of the largest defect. Until we can identify the most likely locations of the combination of the worst of the loads, the properties, and the defects, you cannot effectively avoid the next major occurrence in an area that impacts the public or the environment.

At present, defect severity and management response tend to be based solely on defect length and depth—because that is what we can measure today. At present, a tool to practically quantify or even infer the fracture properties, as they are distributed in a piece of linepipe steel, or other such structural material, does not exist. While concepts exist among some thought leaders regarding the basis for such measurements at a laboratory scale—the need to quantify fracture resistance persists as a critical technology gap. This technology gap affects integrity for many aboveground structural applications: it is made more complex for buried structures such as pipelines. Thus, many would benefit by bridging this gap.

The situation is not so bleak regarding cases where failure is controlled by plastic collapse, as tools exist to quantify the UTS in-the-ditch—so for such pipelines, it is a matter of practical deployment more than a technology gap. The situation regarding IM for pipelines subjected to unique loads is also well developed, as many in the industry today can identify sites where the threat of unique loading exists. Indeed, this aspect is already being dealt with by some operators as a layer in their risk models, with adequate numerical modeling capable of assessing integrity.

Until “where” failure is most likely better known in terms of local properties, quantifying “when” can be quite uncertain. As long as these aspects are poorly addressed, the goal of zero incidents in HCAs remains elusive. Realizing that today’s “smart” pigs have evolved over more than 50 years since their first use, the first field run for a true “mensa” pig, which will report likely failure hotspots determined by use of local measurements and analysis, is well down the road. But there is an upside: a near-term technology bridge can be built that quantifies the necessary resistance metrics through correlations based on (1) the fundamental “structure–property” relationships of materials science [63] and (2) data that characterize the pipe steel in terms of steel chemistry and processing harvested from “Moody reports” in the operator’s archives. Clearly, this approach cannot quantify “where” to the same extent that local properties could. However, it does open to a near-term stopgap that could identify hotspots and rank severity relative to HCAs based on parameters that can be harvested from the archives held by many operators. As such, this approach has much potential to move the bar closer to the industry’s goal of zero incidents.

Acknowledgments

Many individuals and operators have contributed a host of photographs and, in some cases, also documentation essential to build the brief history presented and to document the evolution of the tools and technology underling that evolution. Others have contributed thoughtful dialog. With missing someone is always a concern in citing names, a few individuals critical to this process and supporting discussions include Bill Amend, Ted Clark, Rick Gailing, and Jerry Rau. Their operating companies, past and present, including Southern California Gas Company, Panhandle (Energy Transfer), and Columbia (NiSource), provided some of the archival information that underlie this work, as did Petrobras. These contributions are gratefully acknowledged.

References

1. 1 Pees, S.T. Oil history: pipelines, failure and success, 1861–1864. Available at http://www.petroleumhistory.org/OilHistory/pages/Pipelines/failure.html (last accessed November 2, 2013).
   
2. 2 Anon. History of pipelines, CEPA. Available at http://www.cepa.com/about-pipelines/history-of-pipelines (last accessed October 2, 2013).
   
3. 3 Stiff, J., private communication, ABS Consulting, February 12, 2014.
   
4. 4 EIA (2007) About U.S. Natural Gas Pipelines, Energy Information Administration, June 2007.
   
5. 5 PHMSA (2013) Annual Report Mileage for Hazardous Liquid or Carbon Dioxide Systems. PHMSA, January 31, 2013.
   
6. 6 EIA Energy information administration. Available at http://www.eia.gov/. Search site resources on liquid versus natural gas and contrast the networks and their traits.
   
7. 7 Clark, E.B., Leis, B.N., and Eiber, R.J. (2004) Integrity Characteristics of Vintage Pipelines. Battelle Report to INGAA Foundation, October, 2004. Available at http://primis.phmsa.dot.gov/gasimp/docs/IntegrityCharacteristicsOfVintagePipelinesLBCover.pdf. (last accessed October 2013).
   
8. 8 Gray, J.M. (2002) An Independent View of Linepipe and Line-Pipe Steel for High Strength Pipelines. API X-80 Pipeline Cost Workshop, Hobart, AU, October 30, 2002.
   
9. 9 Anon. Available at http://primis.phmsa.dot.gov/comm/reports/safety/psi.html (last accessed January 24, 2014.
   
10. 10 Beedle, L.S. (1964) Steel Design, Ronald Press Co.
    
11. 11 Beedle, L.S. (1960) On the Application of Plastic Design, Fritz Engineering Laboratory (Lehigh University).
    
12. 12 Turkstra, C. (1962, 1970), Theory of Structural Design Decisions, University of Waterloo Press. See also Hasofer, A.M. and Lind, N.C. (1973) An Exact and Invariant First-Order Reliability Format, University of Waterloo Press (Solid Mechanics Division).
    
13. 13 Gallagher, J.P., Giessler, F.J., Berens, A.P., and Engle, J.M., Jr. (1984) USAF Damage Tolerant Design Handbook: Guidelines for the Analysis and Design of Damage Tolerant Aircraft Structures: Revision B. UDRI Final Report, AD-A 153 161, May, 1984.
    
14. 14 Marshall, P.W. (1979) Strategy for monitoring, inspection and repair for fixed offshore platforms. Structural Integrity Technology, ASME, pp. 97–110. See also Pelleni, W.S. (1976) Principles of Structural Integrity, Office of Naval Research.
    
15. 15 Hopkins, P. (2003) The structural integrity of oil and gas transmission pipelines. In: Milne, I., Ritchie, R.O., and Karihaloo, B. (editors), Comprehensive Structural Integrity, Vol. 1, Elsevier.
    
16. 16 Harris, J.A., Jr. (1987) Engine Component Retirement for Cause. United Technologies Report to AFWAL, AFWAL-TR-87-4069, AD-A 192 730, August, 1987.
    
17. 17 Anon. Available at http://www.dot.gov/, opens to United States Department of Transportation website that opens to websites for the many entities it administers.
    
18. 18 Anon. Available at http://ops.dot.gov/forms.htm#7000-1/2.
    
19. 19 Anon. Goal is zero incidents. Available at http://www.mgaa.org/6211/11460.aspx (last accessed August, 2011).
    
20. 20 ASME (1995) ASME B31.8: Gas Transmission and Distribution Piping Systems, American Society of Mechanical Engineers.
    
21. 21 ASME (1995) ASME B31.4: Pipeline Transportation Systems for Liquid Hydrocarbons and Other Liquids, American Society of Mechanical Engineers.
    
22. 22 API API 5L: API Specification for Line Pipe.
    
23. 23 API. API 5LX: API Specification for High-Test Line Pipe.
    
24. 24 Anon. Available at http://www.phmsa.dot.gov/pipeline/historical-rulemaking (last accessed November 2, 2013).
    
25. 25 Anon. Available at http://www.phmsa.dot.gov/hazmat/regs/rulemaking/archive (last accessed November 2, 2013).
    
26. 26 Anon. (1970) Minimum Federal Safety Standards for Gas Pipelines, 35 FR 3237, Notice 70-2; Docket No. OPS-3B, 49 CFR Part 192.
    
27. 27 Anon. (2004) A Brief History of Federal Pipeline Safety Laws, The Pipeline Safety Trust, Undated But Circa 2004. Available at http://pstrust.org/about-pipelines1/regulators-regulations/a-brief-history-of-federal-pipeline-safety-laws. For details see Gas Pipeline Safety Act of 1968, Public Law 90–481, August 1968 and Hazardous Liquid Pipeline Safety Act of 1979, 49 U.S.C. § 60101, et seq, 1979.
    
28. 28 Anon. (2002) Pipeline Safety Act of 2002, Public Law 107–355, December 17, 2002.
    
29. 29 Anon. (2002) ASME Code Supplement on Integrity Management for Pressure Piping, B31.8S, Revision 1, ASME.
    
30. 30 Anon. API (2001) API Standard 1160: Managing System Integrity for Hazardous Liquid Pipelines, American Petroleum Institute, November 2001.
    
31. 31 Anon. PHMSA R&D website. Available at http://primis.phmsa.dot.gov/matrix/.
    
32. 32 Anon. Transportation of natural and other gas by pipeline, 49 CRF Part 192. Available at http://www.ecfr.gov/cgi-bin/text-idx?SID=661b3e5c845115babe0ece993cb1ffeb&tpl=/ecfrbrowse/Title49/49cfr192_main_02.tpl.
    
33. 33 Anon. Transportation of hazardous liquids by pipeline, 49 CFR Part 195. Available at http://www.ecfr.gov/cgi-bin/text-idx?SID=661b3e5c845115babe0ece993cb1ffeb&tpl=/ecfrbrowse/Title49/49cfr195_main_02.tpl.
    
34. 34 Anon. (1964) Panhandle Lines, Panhandle Monthly Employee Publication, June, 1964, Vol. 21, No. 11, Kansas City, MO.
    
35. 35 Anon. (2013) Public workshop on integrity verification process: pipeline and hazardous materials safety administration, Docket No. PHMSA–2013–0119, August 7, 2013. Available at https://primis.phmsa.dot.gov/meetings/MtgHome.mtg?mtg=91.
    
36. 36 Anon. http://www.weldinghistory.org/whistoryfolder/welding/wh_1900-1950.html. (last accessed 13 June 2012).
    
37. 37 Amend, B. Early Pipeline Construction, Maintenance, and Condition Monitoring: The Evolution of Pipeline Integrity Technologies in the Early to Mid-Twentieth Century. DNV GL White Paper, February 18, 2014.
    
38. 38 Anon. Dresser Style 38 Coupling. http://www.wosupply.com/pdf/Dresser_Products_1.pdf.
    
39. 39 Anon. “Dresser Pipe Couplings General Catalog No. 36,” S.R. Dresser Manufacturing Company, 1936
    
40. 40 Anon. Keeping Line 100 Fit, Panhandle monthly employee publication.
    
41. 41 Leis, B.N., Zhu, X.K., and Clark, E.B. (2004) Criteria to Assess Wrinkle Bend Severity for Use in Pipeline Integrity Management. Battelle Final Report on Project PR3-03113 to the Pipeline Research Committee International, August, 2004.
    
42. 42 Chambers, A.A. Archived private communications to G. A. Reinhardt reporting on the characteristics of construction using ERW line pipe produced by Youngstown Sheet & Tube versus the FW pipe made by A.O. Smith, Youngstown Sheet & Tube Company, 1931 – Report 29, dated April 1931, and Reports 34 and 36, dated August 1931.
    
43. 43 Bell, H. C. Current Trends in the Multi-Flame Welding and Wrinkle Bending of Overland Pipe Lines, Proceedings of the Pacific Coast Gas Association, Vol. 30, p.94–96, 1939.
    
44. 44 Anon. Images scanned from advertisements found commonly in the 1955 issues of the Oil and Gas Journal.
    
45. 45 Binckley, M.J. (1952) Which bending method. Gas, May, 122–125.
    
46. 46 Feier, I.I., Leis, B.N., and Zhu, X.-K. (2010) National Grid Miter Joint Testing. Battelle Report to National Grid, June, 2009: Appendix A, Waiver/SP Petition to PHMSA, May 19, 2010.
    
47. 47 Weise, J.D. (2010) Correspondence concerning case 09-G-0552 (national grid), September 14, 2010. Available at http://phmsa.dot.gov/staticfiles/PHMSA/DownloadableFiles/Files/Pipeline/SP/Brooklyn%20Union%20Gas%20Company%20National%20Grid%20Complete.pdf (last accessed January 13, 2014).
    
48. 48 Anon. (2012) ASME B16.9: Factory-Made Wrought Buttweld-ing Fittings, American Society of Mechanical Engineers.
    
49. 49 MSS MSS SP-75: Specification for High-Test Wrought Butt Welding Fittings, Manufacturers Standardization Society (MSS) of the Valve and Fittings Industry, Inc.; see also API Spec6D: Specification for Pipeline Valves.
    
50. 50 Anon. Recommended Practices for Nondestructive Testing, American Society of Nondestructive Testing (ASNT), ASNT SNT-TC-1A.
    
51. 51 Anon. (2003) ASTM E164: Standard Practice for Ultrasonic Contact Examination of Weldments, American Society for Testing and Materials.
    
52. 52 Anon. (2010) ASTM E273: Standard Practice for Ultrasonic Examination of Longitudinal Welded Pipe and Tubing.
    
53. 53 Anon. (2013) API 1104: Standard for Welding Pipelines and Related Facilities.
    
54. 54 Burdekin, F.M. and Dawes, M.G. (1971) Practical use of linear elastic and yielding fracture mechanics with particular reference to pressure vessels. Proceedings of IME Conference, May 1971, London, pp. 28–37.
    
55. 55 Leis, B.N., Forte, T.P., Flamberg, S., and Clark, E.B. (2005) Emerging Bedding, Padding and Related Pipeline Construction Practices. Battelle Report to RSPA, Contract DTRS56-03-T-0005, and Interstate Natural Gas Association of America, January 2005.
    
56. 56 George, H.H. and Rodabaugh, E.C. (1959) Tests of Pups Support, Bridging Effect, Pipe Line Industry.
    
57. 57 Zhu, X.K. and Leis, B.N. (2004) Plastic collapse assessment of unequal wall joints in pipeline transitions. Proceedings of the Storage Tank Integrity and Materials Evaluation, ASME PVP, 490, 253–260.
    
58. 58 Leis, B.N. and Bubenik, T.A. (2001) Periodic Re-Verification Intervals for High-Consequence Areas. GRI Report-00/0230.
    
59. 59 Leis, B.N. and Bubenik, T.A. (2000) Implementation Plan for Periodic Re-Verification Intervals for High-Consequence Areas. GRI Report-00/0246.
    
60. 60 Kiefner, J.F., Kolovich, K.M., Macrory-Dalton, C.J., Johnston, D.C., Maier, C.J., and Beavers, J.A. (2012) Track Record of Inline Inspection as a Means of ERW Seam Integrity Assessment. Subtask 1.3 Report, US DoT Contract No. DTPH56-11-T-000003, August 22, 2012.
    
61. 61 Leis, B.N. (2013) Compare-Contrast Analysis of Inspection Data and Failure Predictions Versus Burst-Test Outcomes for ERW-Seam Defects. Subtask 4.1 Report, US DoT Contract No. DTPH56-11-T-000003, June 23, 2013.
    
62. 62 Leis, B.N. (2013) Time-Trending and Like-Similar Analysis, for ERW-Seam Failures Final Interim Report. Subtask 4.2, US DoT Contract No. DTPH56-11-T-000003, June 30, 2013.
    
63. 63 Moffatt, W.G., Pearsall, G.W., and Wulff, J. (1964–1966) Structure and Properties of Materials: Structure, John Wiley & Sons, Inc., New York, NY, Vol. I. Also see Hayden, H.W., Moffatt, L.G., and Wulff, J. (1964–1966) Structure and Properties of Materials: Mechanical Behavior, John Wiley & Sons, Inc., New York, NY, Vol. III.