Given the focus of the Chapter on Materials and Construction (M&C) Defects was features that were introduced prior to service that could become defects, little was said regarding how those features could lead to degradation and cracking as precursors to failure, or degradation kinetics. This QR Segment presents video coverage of some related aspects, in complement to the QR Segment titled Illustrations of Some Pipeline Defects, which via video illustrates some defects that fail other than due to increasing pressure. A video considering degradation and cracking as precursors to failure, with brief discussion of degradation kinetics is below.
video: Failure Analysis – Degradation and Cracking
Quantifying Degradation Rates and Revalidation Intervals: Illustrative Discussion
Pipeline failure occurs when the wall of the pipe is breached by a crack or metal-loss defect. For present purposes a crack is a generic term used to represent sharp features whereas metal-loss is a generic term used to represent usually blunt or rounded features. The crack or metal-loss defect can develop from anomalies that reflect early steel- and/or pipe-making practices, or be introduced during construction, or be due to the effects of pipeline operation and/or the environment that develops along the pipeline right-of-way during service.
Failure at cracks and metal-loss defects occurs by either plastic collapse or fracture mechanisms, although at a sufficiently low wall-stress level corrosion can penetrate the wall without significant evidence of such mechanisms. As necessary the QR Segment titled Collapse and Fracture Controlled Failure provides details concerning these failure processes. Failure occurs when the wall of the pipe is breached by a crack or other defect, leading to a leak or a rupture depending on the length and type of defect, the line-pipe steel, the pipe geometry, and the pressure, temperature and composition of the gas. A leak is identified as a defect whose length does not increase as it fails. A rupture occurs when the length of the defect exceeds a critical value, which leads to axial extension of the breach. Depending on the transported fluid and the operational conditions for the pipeline that extension can become axially unstable and run some distance along the length of the pipeline.
Click here for details concerning Collapse and Fracture Controlled Failure.
Approach. In what follows pipeline integrity and re-verification interval are quantified as a function of wall thickness and stress for a range of parameters including line pipe properties, defect type and size, with consideration given to the effects of differing pipe diameter. For those interested the details can be found in Reference 1, which was completed as background to the development of ASME B31.8S . As B31.8S was specific to natural gas (NG) transmission systems, what follows is specific to such scenarios. In particular, results follow that are specific to an example pipeline representative of the NG infrastructure in the US, which was made of X52 steel whose stress-strain response and toughness were typical of such pipelines.
Degradation and its kinetics are addressed in regard to time-dependent processes due to external corrosion, stress-corrosion cracking (SCC), and to a lesser extent fatigue. The cracking processes are addressed in terms of sharp crack-like features whereas corrosion is treated as a blunt feature. File data were trended as a function of line-pipe vintage and grade. The Regulatory incident database was reviewed to identify commonly occurring pipeline geometries and grades. On this basis, a worst-case pipeline geometry was identified as the focus for what follows. Results characterizing properties as a function of vintage and grade are presented first, followed by generic geometries and the pipeline’s service conditions. Thereafter, the results are presented and discussed.
Property Trends with Vintage and Grade. The evolution of modern line-pipe steels began in the 1960s with the introduction of high-strength low-alloy steel (HSLA) making practices, to avoid the potential toughness and cracking problems that would be associated with the traditional strengthening mechanisms in applications to higher-strength grades. The HSLA practices of the 1960s transitioned through a range of thermal-mechanical processing (TMP), including heavy rolling practices, beginning in the 1970s. The desire to further improve the dynamic-ductile fracture resistance for that class of steels led to further evolution, beginning in the 1980s, which culminated in today’s thermal-mechanical controlled processed (TMCP) line-pipe steels. This class of steels makes use of a range of finishing practices, including accelerated cooling and controlled rolling. It is noteworthy that steel-making and pipe-making practices are not standardized, and it is still possible to purchase line pipe today that will not exceed toughness values typical of 60s-vintage line-pipe steels.
Strength. Whereas much has been done to improve the strength and other properties of line pipe steels, most of what is in service remains the order of X52, such that it was chosen for present purposes. As the resistance to plastic collapse is characterized by the mechanical properties of the line pipe, trending was done to quantify the hoop stress at failure for defect-free pipe, for later use in quantifying crack growth and failure in pipes due to the effects of fatigue, SCC, and corrosion. Reference 3 assembled such data and included consideration of a range of stressing conditions such as stress biaxiality. That trending clearly showed that the ultimate tensile strength (UTS), a parameter that historically has been measured within the pipeline industry, correlated very well with the hoop stress in failures controlled by plastic-collapse in line pipe, as is evident in Figure 1.
Figure 1 shows a tight correlation whose results indicate that the UTS clearly is a better metric to quantify failure for defect-free pipelines in contrast to the specified-minimum yield stress (SMYS). This same conclusion was reached recently in work for the European Pipeline Research Group (EPRG) . This outcome is anticipated for collapse-controlled failures, as becomes evident by review of the QR Segments titled Collapse and Fracture Controlled Failure and Failure Mechanisms and Pipeline Properties.
Click here for more on Failure Mechanisms & Pipeline Properties.
Figure 1. Illustrating UTS as a failure criterion for plastic collapse in steel pipelines.
The ratio of the yield stress to the UTS, denoted as Y/T, also was trended. These results are shown here in Figure 2 as a function of grade for a wide variety of steels. Unfortunately this trend is widely scattered and so of limited practical utility.
Figure 2. Y/T as a function of actual yield stress.
Toughness. Resistance to fracture is characterized in terms of fracture toughness, which today is measured according to established standards whose roots lie in fracture mechanics [e.g., see 5]. The development of this technology is recent as compared to the initial use of fracture concepts in the pipeline industry, which many decades ago adopted the Charpy-vee notch (CVN) energy as their measure of apparent toughness. Figure 3 (adapted from ) trends fracture mechanics resistance for pipe steels quantified by strip-yield energy-release rate, Gc, for full-scale tests with axial slits as a function of CVN energy. Inspection of that image indicates that Gc becomes nonlinearly dependent on CVN energy as toughness increases, which is apparent therein for a broad class of line-pipe steels. It is apparent that these full-scale test results fall well above the three published correlations between Gc and CVN due to Barsom and Rolfe  and British Standard BS7910 , which also involve a linear dependence. Figure 3 includes the data originally trended by Maxey et al , who concluded this trend was one to one, whereas those early results do follow the nonlinear best-fit to this dataset. All results presented, which include steels from the 1950s and 1960s through modern production fall along the nonlinear trend noted. The deviation from linearity becomes evident at quite low toughness ~30 ft-lb (40 J) FSE CVN – which is well within the range of the data that underlay Maxey’s work.
Figure 3. Fracture mechanics driving force vs CVN energy.
Full-size area-equivalent (FSE) CVN energy results representing well in excess of 600 joints of line pipe produced from the 1930s through into the new millennium makes it clear that the plateau value of CVN energy (CVP) has increased over time. Trending in Reference 1, shown here in Figure 4 begins to increase sharply circa 1970, much like the evident for Grade. This increased resistance reflects the shift away from carbon-manganese steels to the modern higher strength Grades. Unfortunately, such trends are too scattered to be of practical utility.
Figure 4. Trends in CVN plateau energy over time.
The trending noted above and related work that considered changes in steel chemistry, finishing, and rolling practices that affected changes in strengthening mechanisms and cleanliness led to a grouping of mechanical and fracture properties by decade. Table 1 summarizes those combinations, which were considered representative of the line pipe in the ground in the U.S.
Table 1. Summary of property combinations by decade.
|CVN, ft-lb||20||30||40||65||80 +|
The values in this table suggested that Grade X52 line-pipe be coupled with a typical plateau toughness of 30 ft-lb (40 J) to represent much of the pipeline-miles constructed through the late 1960s. This combination was adopted, along with consideration of a plateau toughness of 20 ft-lb (27 J) to reasonably represent much of the construction through the late 1950s.
Typical Pipeline Geometries and Failure Boundaries for this Typical Pipeline
Typical geometries have been identified based on trending the Regulatory reportable-incident database. That trending pointed to the use of a 30-inch diameter pipe with a 0.250-inch thick wall. Failure boundaries for the typical pipeline (i.e., 30” x 0.250” X52 or 762 x 6.35mm Gr 358) are shown in Figure 6, which presents the failure pressure normalized by that at SMYS on the y-axis as a function of defect length on the x-axis. Part a) of this figure reflects the outcomes for a 30 ft-lb (40 J) steel whereas Part b) represents the same scenario except that the toughness is now 20 ft-lb (27 J). As the format of this chart is quite involved, this format is discussed before moving on to consider the trends.
a) for the 30 ft-lb (40 J) steel
b) for the 20 ft-lb (27 J) steel
The failure boundaries in Figure 5 shown as either red or black lines represent a specific value of flaw depth normalized (i.e., divided) by the wall thickness. Thus, a normalized flaw depth equal to 0.9 represents a flaw that is 90 percent through the wall. The trends in Figure 5 start at a normalized depth of 0.9, which is the lowest, left-most trend in each plot. Each trend above that for 90 percent represents a step reduction in defect depth by one-tenth of the wall thickness.
With this coordinate system used in Figure 5, a value of 0.72 on the y-axis corresponds to a pipeline pressure causing a wall stress equal to 72 percent of SMYS. Likewise, a value of 0.60 corresponds to 60 percent of SMYS, and so on. Flaw sizes causing failure at any normalized wall stress thus can be found by tracing a horizontal line across the chart. Where that line intersects each of the depth contours defines the depth for failure at that normalized pressure. The corresponding initiation length is found by dropping a vertical line from that intersection to the y-axis.
Click here for more on Defect Response to Pressure
The blue trend in this chart corresponds to fracture instability (that is the boundary between leak and rupture), while the green trend signals the onset of stable tearing, which eventually culminates in instability. Combinations of pressure and flaw size that lie above the blue boundary result in ruptures, whereas those below it are stable and so leak. Further details on the underlying concepts and models can be found in the QR Segment titled Defect Response to Pressure.
It can be determined from Figure 5a that operation at 72 percent of SMYS for this steel with FSE CVP toughness of 30 ft-lb leads to ruptures for defects longer than about 5.5 inches (140 mm), for this “typical” pipeline situation. It is also apparent from the data trends that operation at lower stresses (equally pressures) leads to critical sizes that increase significantly as pressure drops. The critical length at 40 percent of SMYS for example is ~9.5 inches (241 mm) long.
The failure boundaries in Figure 5a change as toughness changes, so long as failure is controlled by fracture rather than by plastic collapse. For example, consider Figure 5b that deals with the same geometry pipeline made with X52, but having a FSE CVP toughness of 20 ft-lb. Comparing the trends in both images shows that for the lower toughness the initiation boundaries shift slightly down and to the left. For the reduced toughness case, operation at 72 percent of SMYS for this steel now leads to ruptures for defects slightly longer than 5.3 inches (135 mm). References 10 and 11 demonstrate that inherent toughness tends to have the greatest influence on the initiation length of the various pipe property and geometry related parameters. Reference 1 also addresses critical defect sizes, leak versus rupture, and fracture arrest, which are not essential to the purposes of the QR Segment.
Usual Gas Transmission Pipeline Operating Conditions. Natural-gas transmission pipelines operating at near-capacity typically experience relatively small and infrequent pressure cycles as compared to hazardous liquid systems. Because of their minimum regulatory-based burial depth, thermal cycling likewise is neither frequent nor severe. In spite of the observation that incidents due to fatigue caused by in-service cycling have not been a major concern, consideration follows for the possible fatigue-sharpening of defects, and their subsequent growth due to pressure cycles, with a daily demand-induced cycle from the maximum allowable operating pressure (MAOP) to 70 percent of MAOP considered subsequently. As SCC is also considered subsequently it is appropriate as well to consider the losses that occur from outlet to suction on a typical long distances gas-transmission pipeline. The heat of compression increases the gas temperature, the extent to which depends on the degree of compression. It follows that the gas at discharge is warmer than at suction at the next compressor station, as is shown for typical behavior in Figure 6. There is an associated pressure drop, which also reflects frictional losses in-route. These aspects have been dealt with as detailed in Reference 1. In addition, the effects of a wall thickness that is increased relative to the circumstances addressed in Figure 5 have been addressed, for Class Locations 2, 3, and 4. Interested readers should consult the above-noted citation for details.
Degradation Rates and Re-verification Intervals. All mechanisms leading to in-service degradation exhibit a dependence on wall thickness. This dependence develops directly through the added thickness that must be penetrated, and/or indirectly through the wall-stress dependence of the critical-flaw size, as was illustrated for example for sharp flaws in Figure 5. The crack-driving force for sharp, crack like features leading to fracture depends on hoop stress, as does the net-section stress leading to plastic collapse. The increased critical length and the corresponding depth can be estimated as a function of stress (equally pressure) for sharp, crack-like defects from the trends shown in Figure 5. These results provide a conservative basis to assess the behavior of blunt metal-loss and corrosion defects. The results for sharp, crack-like features will be adopted hereafter to define critical length and depth for the sake of simplicity, in lieu of comparable calculations and further discussion associated with criteria for ductile failure at blunt metal loss defects(e.g., see ).
The following sections address the kinetics of the in-service degradation mechanisms potentially active in transmission pipelines and evaluate the corresponding effect of increased wall thickness on re-verification interval. Fatigue is briefly considered first in relative terms, then SCC and external corrosion are addressed in detail. While mechanical damage, and material and construction defects were addressed in Reference 1 they are omitted here for the sake of brevity. The increased wall thickness in Class 3 and 4 designs as compared to Class 1 is addressed from two perspectives. First, the kinetics of both fatigue and SCC decrease with decreasing stress, which means a reduced rate of degradation and so longer re-verification interval as compared to Class 1 designs, all else being equal. Second, the increased wall thickness leads to more time before the degradation breaches the wall causing an incident, as compared to Class 1 designs. This also leads to an increased re-verification interval, and develops for all mechanisms considered. Finally, for the stress-driven mechanisms (i.e., fatigue and SCC), this effect of increased wall increases nonlinearly, as it also leads to a relative increase in critical-flaw size. This gives rise to a further increment in the re-verification interval for Class 3 and 4 designs as compared to Class 1.
Fatigue Degradation Rate and Re-verification Intervals. Fatigue damage per cycle develops in response to the maximum stress in a given cycle and the ranges of the stress and the strain, which are proportional if the response is locally elastic. Details as to how the effects of that cycling are quantified for fatigue crack initiation and propagation can be found in a pipeline-specific context in Reference 1, and/or Reference 13, along with benchmark degradation rates. Textbooks such as References 3 and 5 provide the basics in a generic context.
Rosenfeld  has broadly quantified the effects of fatigue for gas-transmission pipelines considering among other circumstances cases involving cross-country systems, concluding that degradation due to fatigue typically was not a primary threat. This section reviews the outcomes of a conservative, relative assessment of its potential effects as a function of the increased wall thickness affected by changes in Class Location relative to the cross-country scenario. Given the initial focus of ASME B31.8S was high-consequence areas (HCAs) this analysis contrasted a Class Location 3 to the corresponding cross-country case. The results indicated that if the wall thickness was increased in proportion to the decrease in stress for locations in Class 3 Locations then the crack initiation kinetics decreased dramatically, as the life is about 80 times longer for the lower stress case. As this result was determined without specific reference to the line-pipe steel, and pipeline geometry and operation, this proportional decrease can be applied without regard to those circumstances. For completeness it is noted that this result ignored the somewhat smaller gain in life associated with fatigue crack propagation phase, and the increased distance for propagation resulting from the increased wall thickness and the deeper critical crack size.
SCC Degradation Rate and Re-verification Intervals. For SCC to occur, three concurrent conditions must be mutually satisfied:
- the operating environment must promote the formation of a cracking environment;
- the pipe steel must be susceptible to continued cracking in that environment; and,
- a tensile stress (applied or residual) must exist that exceeds the threshold for crack initiation and continued cracking in that environment.
For SCC to pose a practical threat the process must initiate and grow the cracking sufficiently within the initiated cracking to support coalescence between adjacent cracks, which can lead to rupture, or it must support continued cracking through wall, which generally results in a leak.
Two environments have been identified as cracking environments within the soil-ground-water operating conditions of a pipeline that have initiated and support continued SCC for the stressing conditions and steels typically encountered in gas-transmission pipelines. As these reflect extremes in the associate value of the pH, one has been termed high pH SCC, with the second being termed low pH SCC, which also is termed near-neutral (NN) pH SCC. References 15 and 16 respectively identify the traits of these environments, with reference 15 also characterizing the high pH mechanism. References 17 and 18 detail the mechanism for NN pH SCC, and indicate that these environments can be seasonally active or dormant. As well it should be noted that the content of Reference 17 indicates that the NN-pH and the high-pH environments are reversible between their limits, with a spectrum of environments between them. As such what appears to be NN-pH cracking can be developing from what was originally a high-pH crack, and vice-versa. The mutual presence of both types of cracking has been observed, which coupled with the above observation of reversibility indicates that work such as what follows concerning a high-pH cracking threshold stress is relevant only in situations absent any hint of a NN-pH environment. It must be noted that over roughly the last 20 years cracking identified as SCC has nucleated and grown to cause incidents in portions of pipelines operating at nominal stress levels well below the often cited and codified threshold of 60% SMYS.
References 19 and 20 are good resources providing practical perspective for external SCC on pipelines for applications involving highly susceptible systems. In balance, Reference 21 provides guidance for systems with low susceptibility. Those interested in susceptibility and the factors controlling SCC kinetics should consult this background, which summarizes many of these aspects.
Against the above background, what follows was developed as background to aspects of what became ASME B31.8S circa 2001, and borrows heavily from Reference 1 in that context. It reflects trends and data specific to high-pH cracking. As much has been learned since then concerning the environmental spectrum between high-pH and NN-pH SCC and the reversibility between these limits, such results should be considered only in applications that reflect the upper-end of this spectrum. In parallel, care should likewise be exercised in regard to guidance that has largely been formulated and since codified based on information in hand circa 2001, such as that found in Reference 20.
SCC Thresholds – Implications for Class Locations 3 and 4 Re-Verification Intervals
The above-noted literature indicates that high-pH SCC on pipelines exhibits a threshold-stress below which the kinetics slow to a rate so low that it is no longer practically significant. References 20 and 22 discuss thresholds for smooth-bar data for line-pipe steels, which reflect initiation conditions for SCC, as well as data for pre-cracked specimens that reflect cracking kinetics for longer cracks. They also consider the transition from no-crack to a crack that lies between these limiting states.
The wall stress in pipelines operating at pressures associated with design factors for Class Locations 3 and 4 fall below high-pH thresholds known for pipeline steels in accelerated high pH cracking environments (e.g., see ). Figure 6 (adapted from ) presents some of the data that underlie the existence of a threshold for high pH SCC.
Figure 6. Typical laboratory data for high pH SCC under highly accelerated conditions.
Figure 6 plots crack depth on the y-axis and maximum stress on the x-axis for results reported in Reference 23. These results reflect the effects of a slowly varying cycle designed to simulate the behavior of gas- transmission service. Results are included for cycling conditions that cover the reasonably expected operation for such lines, as values of the stress ratio, R, defined as the algebraic ratio of the minimum to maximum stress in a pressure cycle are included from ~0.7 up to ~0.9. As the operating stresses for Class Locations 3 and 4 fall marginally below the worst-case threshold for high pH SCC initiation, such cracking is highly unlikely on those pipeline segments. This is particularly so given that the data in this figure represent testing done in a very concentrated (aggressive) cracking environment, at the optimum electrochemical potential for cracking (i.e., the worst-case), and a temperature of 167°F (75°C) that is far above the discharge temperatures for pipelines where SCC is a consideration.
With reference to the data shown in Figure 6, there is no evidence of cracking or grain-boundary (GB) etching below about 70% of SMYS for X52, which is ~35 ksi (241 MPa). On this basis, a value of 35 ksi (241 MPa) could be termed the threshold for high pH SCC specifically for that Grade. However, it is evident from these data that both the incidence and depth of cracking depend on stress. A more conservative threshold in that context is determined by extrapolating the trend through the upper bound cracking results back to zero depth. With this more conservative threshold definition, the results in Figure 6 indicate the value of the threshold stress is slightly above the maximum stress corresponding to Class 2 Locations.
Inspection of the scale on the y-axis in Figure 6 shows that these data reflect very shallow crack depths, which must grow significantly in depth and length size, and thereafter coalesce axially to form a critical length flaw before such cracking poses an integrity threat. The data shown in Figure 6 lie toward the lower-end of the laboratory database that determine the threshold for high-pH SCC in this grade of line-pipe steel. Service experience in the U.S. as well as elsewhere in the world supports the view that high-pH SCC does not threaten the integrity of pipelines operated at lower stresses. Experience with pipelines in the UK , which typically operate at lower wall stresses then showed no evidence that high-pH SCC was occurring on these lines. For such reasons, it has been argued at times that high pH SCC was not an integrity threat in HCAs comprised of pipelines operating at stress levels corresponding to Classes 3 and 4.
SCC – Effects of Temperature and Pressure Cycling. References 24 and 25 present and validate phenomenological models of the high-pH SCC process, which provide the means to simulate its kinetics on operating pipelines, and to map kinetics developed in accelerated testing data onto pipeline operating conditions. Validation presented in these references involves demonstrating that predictions match the characteristic features of field cracking. Such demonstrations include matching crack length, depth, and aspect ratio, and their statistical distributions in SCC colonies. They also have addressed colony sizes and the distributions of cracks and crack sizes in colonies, as well as populations of colonies along pipelines. Finally, the time frame for the appearance of cracking and its growth to critical proportions has been demonstrated, as have matches for crack sizes exposed by hydrostatic re-testing, and the effectiveness of re-testing in controlling high-pH SCC.
Of the first-order factors that influence the kinetics of high-pH SCC (e.g., see [1,15-21]), three are associated with pipeline operation – gas temperature, maximum stress, and stress range. The field-proven models noted above have been used to evaluate how high-pH SCC kinetics might influence re-verification intervals, as a function of the stressing conditions and temperatures experienced in pipeline segments in Class Location 1, with a similar relative effect in the other Class Locations. Results have been generated for a range of operating scenarios covering a broad window of temperature, maximum pressure, and pressure range. These results have been trended to quantify the relative differences in high pH SCC kinetics, as presented in Figures 7 and 8. As these are relative comparisons, they reflect situations where all other parameters are held constant.
Figure 7 illustrates the relative dependence on gas temperature, under the reasonable assumption that the steady-state temperature of the steel in the pipe wall during operation is very close to this temperature. The trends shown in this figure are relative to the behavior for Class 1 operation at 110 F and a 50 percent probability of failure. Thus, the data point corresponding to 110 F and a 50 percent probability of failure lies at a value of one on the y-axis. Regardless of the probability chosen, the relative effect of temperature is found to be roughly the same, which reflects the nearly deterministic effect of temperature on the kinetics for high-pH SCC. This implies that the trends evident at 110 F and a 50 percent probability of failure are indicative of what occurs at other probability values.
The results in Figure 7 indicate that the kinetics decrease as temperature decreases – either through changes in compression ratio, or through the use of after-coolers. These results show that pipelines experiencing recurrent high-pH SCC problems might achieve further control through temperature reduction. The relative effect of temperature shown in Figure 7 is indicative of the relative kinetics anywhere along a pipeline. Thus, as the temperature at discharge begins to drop as the gas moves down the pipeline, the kinetics decrease as indicated in this figure. Data reported in Reference 1 indicate that the gas temperature can fall on the order of 35°F (19°C) or more over a distance the order of 100 miles (161 km). As such, the temperature at the suction downstream can be much less than at the discharge. It follows that, depending on the downstream location a somewhat higher threshold can be anticipated in the context of Figure 7 due to the combined effects of lower stress and temperature.
Figure 7. Relative kinetics of high pH SCC as a function of temperature
Figure 8 presents the relative mean-life-to-failure on the y-axis as a function of the pressure loading presented in terms of the magnitude of the pressure cycle on the x-axis. Data are included for three values of the maximum allowable operating pressure (MAOP) in the cycle. This variable is included through use of different symbols, which as shown in the figure reflect 72, 65, or 60 percent of SMYS. With reference to these values of MAOP, this figure addresses stress ratios (i.e., R values) from near zero through 0.7.
Figure 8. Relative kinetics of high pH SCC as a function of pressure conditions.
Figure 8 presents the relative dependence of high-pH SCC kinetics referenced to service lives that reflect operation at MAOP with a near-zero pressure cycle. With this normalized format the trends for each of value of MAOP appear to follow similar trends. This does not mean that operation at 72 percent of SMYS leads to the same kinetics as operation at 60 percent of SMYS, as the absolute lives for each of these differs significantly due to the combined effects of lower stress and temperature. In absolute terms, the maximum spread between results at 72 percent and 60 percent of SMYS is on the order of a factor of two. Thus, the near coincidence of these relative trends indicates that the relative effect of pressure cycling on the kinetics is largely independent of the maximum pressure – at least for these finite-life conditions. These results show little difference in kinetics develops for cycles up to about five-percent of MAOP. Thereafter, increasing the range of pressure cycling increases the high-pH SCC kinetics, through pressure cycling on the order of 20 percent of MAOP. It is likely that as the pressure range gets larger there will be an increase in corrosion-fatigue kinetics, which then would become the dominant degradation process.
In closure it is noted that the kinetics for NN pH SCC differ as compared to the scenario presented above for high pH SCC. In particular, NN pH SCC arrears to be much less selective regarding the temperature and potential. More importantly, whereas the high pH mechanism shows a threshold that traces to the active-passive nature of the process, NN pH SCC does not show this sensitivity to the stress needed to fail a passivating film. Such cracking has been found in service at quite low stresses. For these reasons the NN pH process is much less selective as to when and where is occurs along a pipeline. Care should therefore be taken to characterize the field cracking environments, to understand its practical implications for both operations and mitigation.
Corrosion Degradation Rate and Re-verification Intervals.
Corrosion kinetics can be evaluated in regard to the usual process leading to local areas of metal loss on the exterior of pipelines. Much has been done to quantify the kinetics of this process, which reflects the electrochemical reaction between steel and ground water, in a process often referred to as dissolution. Many textbooks and handbooks detail this process, as for example Reference 26. This reaction occurs on bare pipe, as well as on pipe that is under CP, where this protection system fails because of local shielding, or other complexities beyond the scope here. While other forms of corrosion such as microbiologically influenced corrosion (MIC) also occur on pipelines, as does internal corrosion, the scope here is limited to the above-noted electrochemical process. Eventually the corrosion process penetrates the pipe wall, leading to a leak or rupture depending on the size and shape of the defect, the pipeline pressure, and the other factors noted in the earlier discussion on this topic.
Metal-Loss Kinetics. The kinetics that underlie the just-noted external corrosion process depend on a host of parameters such as temperature, local potential, the presence of oxygen, and local corrosivity. As such, corrosion kinetics could in concept be presented on a case-specific basis across the range of such parameters, which has limited practical utility. In lieu of that scope, Figure 9 presents two sets of corrosion rate data for steels observed under field conditions.
Figure 9a presents trends from results of field tests done with bare, unprotected coupons, with a view to simulate general corrosion rates for buried steel structures . These results reflect upper and lower bound trends to the data reported therein. In turn, these bounds have been reinforced with dashed red lines that reflect the time-averaged corrosion “rate”. Rates as presented are really average speeds as they were calculated by dividing corrosion weight-loss averaged over a time interval on the order of a few days. Figure 9a shows the lower-bound rate is ~0.001 inch per year (0.025 mm/year), while the upper-bound rate is ~0.003 inch per year (0.076 mm/year). For the higher-rate data the results imply the corrosion process was initially much more reactive, as compared to the stable behavior that developed after a few days. The factor of three difference between these bounding rates seems large until the variability for this process is compared, for example, to fatigue, where a factor of ten might be viewed as “low” scatter.
Data such as that in Figure 9a have practical utility only if they reflect the actual behavior of in-service pipelines. Accordingly, these coupon data have been supplemented by the results shown in Figure 9b. This figure reflects the time to failure for incidents in the Pipeline and Hazardous Materials Safety Administration (PHMSA) database  circa 1999 for bare but protected gas transmission pipelines. Because data for bare pipelines have been considered, no “adjustment” is required to account for the time for coating failure. Rate has been determined in the context noted above by dividing the full wall thickness by the pipeline’s time in service prior to failing. These incidents are split almost evenly between leaks and ruptures, with no pattern evident in rate as a function of failure mode. The available results have been culled to exclude pipe with diameters smaller than 12 inches (305 mm), and cases involving SCC. The small diameter pipe has been excluded, because steel and pipe making for smaller diameter pipe can be much different than for larger-diameter transmission line pipe. SCC has culled from the data as this process involves cracks, as opposed to metal loss corrosion that typically involves much blunter features.
a) buried, unprotected coupons
b) bare, protected pipelines
Figure 9. Corrosion rate trends for buried bare steel coupons and pipelines.
Corrosion kinetics derived relative to an incident database means that the kinetics reflect worst-case field corrosion conditions as compared to circumstances elsewhere in the gas-transmission pipeline system, where degradation must be occurring at a slower rate. Because the kinetics reflect the leading edge of the population of corrosion rates, these data have been analyzed to quantify three specific corrosion rates. A worst-case rate has been quantified that reflects the fastest rate for all of the incidents on bare, protected pipe. Rates also have been quantified that reflect average kinetics for the bare, protected pipe incident population, as has a rate that reflects the 90th percentile for the population. Results for the pipe population also have been parsed in regard to general versus pitting corrosion. As noted for Figure 9a, there is scatter in the PHMSA data set. For the general corrosion data the rate-data scatter by about a factor of two, while the extremes for the pitting data differ by about a factor of five.
The worst-case corrosion rate for pitting is found to be ~0.022 inch/year (0.559 mm/year), while that for general corrosion is about half the pitting rate, at ~0.012 inch/year (0.559 mm/year). The corresponding 90th percentile rate is found to be ~0.012 inch/year (0.305 mm/year) for both pitting and general corrosion, while the average rate for the worst-case situations is found to be 0.009 inch/year (0.229 mm/year). The similarity between the average and 90th percentile rates for general corrosion and pitting corrosion is expected in light of the trends shown in Figure 9b, as is the difference in the upper-bound rates which reflects the differences in scatter at the extremes of these two populations.
Based on these kinetics Reference 1 presents an extensive analysis of the implications for re-inspection intervals – those interested in those details should review that work. That same citation also considers kinetics and re-inspection intervals for internal corrosion, and other forms of accelerated corrosion such as MIC, for material and construction defects not found by hydrotesting, such as ERW seam defects and hard-spots, and for mechanical damage defects.
Summary and Conclusions
This QR Segment evaluated factors that control threats to gas-transmission pipeline integrity and on that basis determined the frequency at which pipeline condition must be re-verified to ensure integrity in high-consequence areas. Because the initial margin of safety provided by the design factors for pipelines is degraded in service, this evaluation was in terms of time-dependent material and construction anomalies, and defects developed in service. Specifically, the time increment due to the change in stress and wall thickness in Class Location 3 and 4 designs was determined under the assumption that these areas overlay Class 3 and 4 Locations. While an implementation plan that begins with a baseline inspection that reflects current condition, and defines the scope and procedures involved in subsequent periodic inspections also was developed, but was reported independently (see Reference 1 for citations).
Kinetics were discussed in support of quantifying re-verification intervals as detailed in Reference 1. These kinetics and intervals were determined under the assumption that they would become part of a prescriptive integrity management plan destined to deal with wide range of regulated pipelines. For this reason, these intervals were developed within a “one-size-fits-all” framework. While developed for use in a prescriptive framework, the processes used are the same as would be involved in developing performance-based integrity management plans.
Using kinetics for the very worst corrosion rates, all of the incidents considered in the PHMSA database could be avoided using a periodic re-verification interval of:
- 16 years to avoid leaks or ruptures for Class 3 and 4 designs; and,
- 11.4 years to avoid leaks or ruptures for Class 1 designs.
Other potential causes of in-service degradation, and historical causes of incidents, also were evaluated, leading to the following conclusions:
- Fatigue; is an unlikely cause for defect initiation or growth in HCAs, because of the increased wall thickness and reduced wall stress in Class 3 and 4 designs.
- Fatigue kinetics are reduced by at least a factor of 80 as compared to Class 1 designs. This conclusion is supported by full-scale test results and the historic absence of gas-transmission pipeline incidents due to fatigue.
- High pH SCC; is effectively precluded as a threat in well maintained pipelines in Class 3 and 4 Locations because of the increased wall thickness and reduced wall stress in these pipelines.
- This conclusion is supported by laboratory evidence, developed under accelerated worst-case conditions, and by the absence of high-pH SCC-related incidents in such pipelines.
- Weld-seam defects and hard spots, which survived mill or pre-service hydrotesting, will survive longer in Class 3 and 4 pipelines as compared to Class 1 pipelines, all else being equal. However, as for internal and external corrosion when they are driven by locally unique factors, a prescriptive re-verification interval is generally inappropriate for seam defects and hard-spots.
- Technologies exist to support the determination of a re-verification interval for weld-seam defects and for hard spots where these conditions have been identified. Those technologies are uniquely appropriate to each feature and its local circumstances. As above, it is expected that operators will be required to employ these technologies as appropriate, but their description and resulting interval analyses are not within the scope of this work.
- Third-party mechanical damage is a random occurrence in both location and time and therefore cannot be effectively managed by periodic inspection or re-verification.
- Contact could occur minutes after an inspection and cause immediate failure, or a delayed failure at any time in the future. That time can range from minutes or hours to years. Management of third-party damage requires technology that limits future contact, and in the event contact does occur identifies the point of contact. Near-term reduction of delayed incidents due to existing damage might better be achieved using existing direct-assessment technologies that detect the coating damage accompanying mechanical damage. In-line inspection (ILI) tools under development to detect and assess mechanical-damage severity may offer the same potential in the future, for portions of the pipeline system that are piggable.
The central conclusion is that the time interval for re-verification of integrity in Class 3 and 4 designs is significantly longer than that appropriate for Class 1 designs due to the difference in stress associated with increased wall thickness.
Because re-verification intervals were evaluated a “one-size-fits-all” framework, there are exclusions to the above-noted “one-size” conclusions. Significant exclusions involve portions of pipelines exposed to other than sweet, dry gas, stray currents, shielding, as well as pipelines that do not meet the provisions of usual Codes for Class 3 and 4 Locations via increased wall thickness. Portions of the pipeline that involve field situations like stray currents fall outside the current scope. Other exceptions involve designs that might have changed subsequently through maintenance and/or rehabilitation.
- Leis, B. N., and Bubenik, T. A., “Periodic Re-Verification Intervals for High-Consequence Areas,” Gas Research Institute, GRI Report-00/0230, January 2001.
- Anon., ASME B31.8S “Managing System Integrity of Gas Pipelines” ASME, January 2014
- Leis, B. N., and Stephens, D. R., “An Alternative Approach to Assess the Integrity of Corroded Line Pipe – Part One: Current Status and Part Two: Alternative Criterion”, 7th International Conference on Offshore Pipelines and Polar Engineering, Vol. 3, pp 624 – 634 and pp 635 – 641, 1997,
- Zimmermann, S., “Modeling of ultimate limit states: Burst pressure and circumferential elongation of flawless pipes”, SZMF Report D032, (EPRG Research Project 132), 2006.
- Broek, D., Elementary Engineering Fracture Mechanics, Noordhoff, 1974 (& subsequent editions).
- Leis, B. N., “Arresting Propagating Shear in Pipelines: Part I and Part II,” Steel in Translation, Volume 45, Issue 1 (Part I), January 2015, pp, 5-20, and Issue 3 (Part II), March 2015, pp 161-174.
- Barsom, J. M. and Rolfe, S. T., Fatigue and Fracture Control, Butterworth-Heinemann, 1977 (& subsequent editions).
- Anon., “Guide to methods for assessing the acceptability of flaws in metallic structures,” BS 7910, BSI, 2015.
- Maxey, W. A., 1974, “Fracture Initiation, Propagation, and Arrest”, Paper J, 5th Symposium on Line Pipe Research, American Gas Association Catalog No. L30174
- Leis, B. N., Brust, F. W., and Scott, P. M., “Development and Validation of a Ductile Flaw Growth Analysis for Gas Transmission Line Pipe”, NG-18 Report No. 193, A.G.A. Cat. No. L51543, June 1991.
- Leis, B. N., and Rudland, D. L., “Toward Alternatives to Hydrotesting to Ensure Safe Serviceability of Pipelines,” 11th PRCI/EPRG Biennial Joint Technical Meeting on Line Pipe Research, Arlington, 1997, pp 28.1-20.
- Stephens, D. R., Leis, B. N., Kurre, M. D., and Rudland, D. L. “Development of an Alternative Failure Criterion for Residual Strength of Corrosion Defects in Moderate- to High-Toughness Pipe” PRCI Catalog Number L51794, January 1999.
- Leis, B. N., and Bubenik, T. A., “Primer on Design to Avoid Failure in Steel Transmission Pipelines,” Gas Research Institute, GRI 00/0229, January, 2001.
- Rosenfeld, M. J. and Kiefner, J. F., “Basics of Metal Fatigue in Natural Gas Pipeline Systems – A Primer for Gas Pipeline Operators,” Contract PR-302-03152, PRCI Catalog No. L52270, June, 2006.
- Parkins, R. N., “The Controlling Parameters in Stress-Corrosion Cracking,” 5th Symposium on Line Pipe Research, Pipeline Research Committee, AGA, Paper U, pp. U.1- 39, 1974
- Parkins, R. N., “Stress Corrosion Cracking of Pipelines in Contact with Near Neutral pH Solutions,” AGA PRCI Report 232-9501, 1995: see also R. N. Parkins, “Transgranular Stress Corrosion Cracking of High-Pressure Pipelines in Contact with Near Neutral pH Solutions,” 8th Symposium on Line Pipe Research, Catalog No. L51680, American Gas Association, September 1993, p. 16-1.
- Colwell, J. A., Leis, B. N., and Singh, P., “Crack Initiation in Near-Neutral pH Environments,” Environmentally Induced Cracking in Pipeline Systems, Vol. II, Elsevier, 2004, pp. 223-242.
- Asher, S. L., Leis, B. N., Colwell, J. A., and Singh, P. M., “Investigating a Mechanism for Transgranular Stress Corrosion Cracking on Buried Pipelines in Near-Neutral pH Environments,” Corrosion, Vol 63, 2007, pp. 932-939.
- Anon., Report of the Public Inquiry Concerning Stress Corrosion Cracking on Canadian Oil and Gas Pipelines, National Energy Board (Canada), Proceeding MH-2-95, November 1996, pp. 135-139.
- Fessler, R. R., Batte, A. D., and Hereth, M., “Integrity Management of Stress Corrosion Cracking in High Consequence Areas,” ASME STP-PT-011, ASME, October, 2008.
- UKOPA, “Industry Good Practice Guide: Near Neutral pH and High pH Stress Corrosion Cracking”, GP/009 Edition 1, 2015
- Leis, B. N., “Analysis Methods For Stress-Corrosion Cracking”, Pipeline Technology Conference, Part B, Oostende, pp. 18.1-11, 1990.
- Leis, B. N., “Characterization of Axial Flaws in Pipelines, With a Focus on Stress-Corrosion Cracking”, PRC NG-18 Report No. 212, PRCI, 1997.
- Leis, B. N., and Parkins, R. N., “Mechanics and Material Aspects in Predicting Serviceability Limited by Stress-Corrosion Cracking”, Fatigue and Fracture of Engineering Materials & Structures, Vol. 21,1998, pp 583-601.
- Kurth, R. E., and Leis, B. N., “Probabilistic Modeling of Stress-Corrosion Cracking: Part One – Model Development and Part Two –Validation and Implications for Control”, Probabilistic and Environmental Aspects of Fracture, ASME PVP Volume 386, 1999, pp3-12 and pp13-23.
- Jones, D. A., Principles and Prevention of Corrosion, Macmillian, 1991
- Romanoff, M., “Underground Corrosion”, NBS Circular 579, 1957: see also Flitt, H. J., Spero, C., and Ferris, P.M., “Prediction of Mild Steel Corrosion Rate in Clay Soil”, Proceedings of Life Prediction of Corrodible Structures, Hawaii, Vol. 1, pp 320-333, 1991.
- Anon., PHMSA Data and Statistics Overview, https://www.phmsa.dot.gov/data-and-statistics/pipeline/data-and-statistics-overview.