This document summarizes developments in fracture control technology for gas pipelines using high strength steels. It discusses how higher strength steels allow for thinner pipe walls and reduced costs but also present challenges for fracture control. Models for predicting crack initiation have been shown to apply to grades up to X100, though more data is still needed for the highest grades. Requirements for seam weld toughness are less stringent than the parent metal since cracks can only propagate one pipe length. Overall fracture control approaches must consider crack initiation and propagation in the pipe material and welds under various operating conditions.
This document summarizes developments in fracture control technology for gas pipelines using high strength steels. It discusses how higher strength steels allow for thinner pipe walls and reduced costs but also present challenges for fracture control. Models for predicting crack initiation have been shown to apply to grades up to X100, though more data is still needed for the highest grades. Requirements for seam weld toughness are less stringent than the parent metal since cracks can only propagate one pipe length. Overall fracture control approaches must consider crack initiation and propagation in the pipe material and welds under various operating conditions.
This document summarizes developments in fracture control technology for gas pipelines using high strength steels. It discusses how higher strength steels allow for thinner pipe walls and reduced costs but also present challenges for fracture control. Models for predicting crack initiation have been shown to apply to grades up to X100, though more data is still needed for the highest grades. Requirements for seam weld toughness are less stringent than the parent metal since cracks can only propagate one pipe length. Overall fracture control approaches must consider crack initiation and propagation in the pipe material and welds under various operating conditions.
This document summarizes developments in fracture control technology for gas pipelines using high strength steels. It discusses how higher strength steels allow for thinner pipe walls and reduced costs but also present challenges for fracture control. Models for predicting crack initiation have been shown to apply to grades up to X100, though more data is still needed for the highest grades. Requirements for seam weld toughness are less stringent than the parent metal since cracks can only propagate one pipe length. Overall fracture control approaches must consider crack initiation and propagation in the pipe material and welds under various operating conditions.
The document discusses the development of higher strength steels for gas pipelines to reduce costs while maintaining throughput. It also discusses the historical understanding of fracture control in pipelines and some recent grades like X100 and X120.
In the 1940s-1950s, long and unstable brittle fractures sometimes kilometers in length occurred in pipelines, prompting collaboration between operators and suppliers to address fracture issues. Some of the early experiments on fracture propagation in pipelines were conducted in 1953.
Pipeline steel grades have increased in strength from the initial X65 up to currently established X80 and being tested grades of X100 and X120. This has been enabled by developments in steelmaking and processing techniques like controlled rolling and accelerated cooling over the last 50 years.
1
DEVELOPMENTS IN FRACTURE CONTROL TECHNOLOGY
FOR GAS PIPELINES UTILISING HIGH STRENGTH STEELS
R M Andrews, Advantica, Loughborough, UK A D Batte, Advantica, Loughborough, UK
1. INTRODUCTION As the worldwide demand for natural gas has continued to grow during recent decades, the challenge to gas companies to transport gas from sources increasingly remote from the centres of population has become more important. Among the options considered to address this issue is the construction of new long-distance high-capacity pipelines, particularly across the more remote areas of Central and south East Asia, Arctic North America, South America and Africa. The economic viability of such major projects is critically dependent on the cost-revenue balance. One of the ways in which the installed cost of a new pipeline can be reduced is to utilise higher strength steels; such an approach offers the possibility of reducing the pipe wall thickness or diameter, hence reducing the cost of materials and construction for a required throughput. Studies relating to conceptual pipelines in North Africa [1], Eastern Europe [2] and Arctic North America [3] have identified potential reductions in installed cost of 5-15% resulting from the adoption of these designs and materials. In response to this challenge, pipe manufacturers and designers have collaborated to introduce a progressive evolution of pipe grades to the industry during the last 50 years. Figure 1, from [4] shows the timescale of these developments and the change in production techniques to use various forms of thermo-mechanical processing such as controlled rolling and accelerated cooling. Much of the existing infrastructure has made use of pipe grades up to X65; many thousands of kilometres of such pipe transport gas across all the inhabited continents. For the pipeline infrastructures in Western Europe and North America the most recent evolutionary step has been grade X80. After a 20 year development period this is now established as a fully satisfactory material technology for constructing pipelines up to 1200 mm diameter operating at up to around 100 bar. However for the next generation of long distance high-capacity pipelines conceptual designs have focused on higher pressures and higher strength materials. Modern grade X100, first produced in trials nearly 10 years ago, has been extensively evaluated and many of the issues connected with the manufacture, design, construction and operational performance have now been addressed, adding confidence to the acceptability of X100 Figure 1 Evolution of linepipe steel grades and production techniques from the initial development stage; from Gray [4]. 2 pipeline technology for future highly-rated applications. More recently a further strength enhancement, Grade X120 has been unveiled and steps are underway to address the same issues for this material. 2. THE HISTORICAL PERSPECTIVE One of the key concerns in the design of any pipeline is the avoidance of propagating fractures. The issue of low toughness and propensity to fracture has been known for centuries, firstly for cast irons and later for steels. However with the development of welding technology for large steel structures, the possibility of fracture of an entire structure rather than a single riveted plate became a reality. In the ship-building industry it was highlighted by the catastrophic failures of the Liberty ships in the 1940s; in the pipeline industry it was highlighted by the occurrence of long, unstable, brittle fractures in the 1940s and 1950s, sometimes several kilometres in length. In the pipeline industry these occurrences prompted collaboration between the pipeline operators and pipe suppliers, particularly in North America and Europe, to address the issue. Indeed the first full-scale fracture propagation experiments on line pipe were conducted for the industry by the US Battelle Memorial Institute in 1953, 50 years ago this year. These early tests, followed by substantial programmes of experimental and analytical work in parallel with steelmaking and pipe manufacturing development in many countries throughout the world, have led to the present overall understanding of Fracture Control; the means whereby the threats of fracture initiation and fracture propagation are eliminated or effectively mitigated through attention to pipeline design, construction and operation. This is not to imply that all the issues concerning Fracture Control have been identified and addressed. New issues arise from two directions. Firstly, new pipeline projects are continually pushing the boundaries in terms of higher design pressures, lower operating temperatures and richer gas compositions (more heavy hydrocarbons), resulting in new uncertainties concerning pipeline and gas behaviour. Secondly, the new higher strength steels have inherently less resistance to deformation and damage prior to fracturing and the means whereby these parameters can be measured are becoming more difficult, introducing further potential for uncertainty in establishing Fracture Control plans. This paper explores these new challenges facing Fracture Control for next-generation pipelines, particularly the issues concerning the utilisation of high strength steels. New developments in small-scale testing, large-scale testing and analytical understanding are examined in the context of fracture initiation and propagation criteria and their application within an overall Fracture Control methodology. 3. MODERN FRACTURE CONTROL: ISSUES AND REQUIREMENTS An overall approach to fracture control will consider a range of issues including crack initiation and propagation in the parent metal, seam welds, and girth welds. Such events may be associated with pre-existing flaws associated with material manufacture and construction or they may result from in-service degradation due to corrosion, cracking or mechanical damage. Many of these issues have been discussed and reviewed extensively elsewhere [5,6]. For high strength steels there are particular challenges because the stresses are inevitably higher, which increases the driving force for fracture whilst the toughness, a measure of the fracture resistance, is unlikely to increase to the required extent. Similarly, low temperature may increase the fracture resistance requirements as well as necessitating the achievement of the required material properties at lower temperatures. 3.1. Fracture Initiation To control fracture initiation in the parent metal it is necessary to ensure that the material has sufficient toughness for failure to be dominated by plastic collapse. Once this level is reached, the critical defect size is controlled by the material strength properties and is no longer affected by toughness. The most widely used fracture initiation models were developed in the early 1970s by Battelle [7] and Shannon [8]. The influence of pipe grade on the tolerable flaw size is illustrated in Figure 2; it can be seen that there is a small decrease in tolerable flaw size as the grade increases. However, increasing the pipe grade has a bigger influence on the toughness required to ensure that flow stress dependent behaviour is obtained. Figure 2 also shows the Charpy toughness required to ensure flow stress dependent behaviour of a through-wall crack increases significantly as the grade increases. In practice, for gas pipelines these effects will be over-ridden by the more severe requirements for ductile crack 3 arrest (see below); however for liquid pipelines where ductile crack propagation is not a concern, these requirements will determine the required toughness. The Battelle/Shannon approach is semi-empirical and the models were calibrated against relatively low toughness and low strength materials. While the approach has been used for recent highly-rated pipelines constructed in X70 material [5] there has been some concern as to whether it is applicable to higher grades. Work sponsored by the European Coal and Steel Community [9, 10] on trial production Grade X100 pipes showed that the failure of part-through-wall flaws in 914 mm and 1422 mm diameter pipe was satisfactorily predicted, though without as much conservatism as for lower strength materials. Also, ring expansion and vessel tests undertaken by Advantica [11] on 762mm diameter X100 pipes showed that the similar approach set out in Annex G of BS7910 [12] could be used for both crack-like and volumetric defects. The original model had already been shown to be appropriate for grades up to X80. Hence although the amount of information on crack initiation resistance is still limited compared to that for lower grades, the results show that the flow stress dependent models are likely to be applicable up to grade X100 for materials with the toughness models required for gas transmission. 3.2 Seam weld toughness It is neither practical nor necessary to require the seam weld toughness to match that of the parent metal. Gas transmission pipelines are laid with the seam welds offset so that a crack initiating in the seam weld will only propagate for one pipe length. However, some level of toughness is required in the seam weld to give resistance to crack initiation, and this can be set to be equivalent to the level required for flow stress dependent behaviour in the parent pipe. The requirements for seam weld HAZ toughness are less clear cut. Low Charpy and CTOD toughness have been measured in the seam welds of X100 material [13] but the failure pressures in ring expansion tests with artificial defects in the HAZ were predicted by a ductile collapse analysis [11]. Analysis using a conventional fracture mechanics assessment had predicted much lower failure pressures. However by applying constraint based fracture mechanics procedures it was shown [14] that the actual loading conditions on a seam weld HAZ defect represent a much lower constraint condition than that in standard fracture test specimens. This conclusion is consistent with experience in lower grade pipelines where low-toughness heat affected zones have occurred but there is no evidence of service failures. Similarly, the offshore structural industry has in the past carried out extensive investigations into local brittle zones in the HAZ but there has been no experience of service failures [15]. 0 20 40 60 80 100 120 140 160 60 70 80 90 100 110 120 Grade, ksi T o u g h n e s s
( J )
o r
l e n g t h
( m m ) Charpy toughness Crack length Figure 2 Variation of flow stress dependent crack length and Charpy toughness required to ensure flow stress dependence in a 36 inch, 12.7 mm wall pipeline operating at 80% SMYS 4 3.3 Mechanical damage resistance The resistance of high strength steels to mechanical damage has received little attention to date. There are two major concerns; resistance of the material to puncture and the integrity of damage remaining after a non-puncturing Incident. The puncture resistance of line pipe has been extensively evaluated by the European Pipeline Research Group (EPRG) [16]. Empirical relationships derived from full-scale testing indicate that the principal elements of puncture resistance are the pipe geometry (wall thickness) and the tensile strength. Hence, although no data are available for materials above Grade X80, puncture resistance is not expected to be an issue for higher strength pipelines. The integrity of remaining damage is potentially a greater concern. Damage can take the form of dents, gouges and combined dent-gouges; the last of these is the most critical type. Operators and regulators of pipelines constructed from high strength linepipe will require assurance that the material can resist external damage. Models for assessing dent-gouges are empirical, being derived from an extensive database of full-scale pressure tests [16, 17]. They include by inference the reductions in toughness due to the denting and rerounding process. As high strength materials tend to have comparatively lower uniform strain at failure [11, 18], the strain induced in damaging the pipe will be a larger proportion of the failure strain and can be expected to cause a greater reduction in toughness than would occur in a low grade material. Hence the failure pressure of damaged Grade X100 may be less than that predicted by models based on lower grade materials. This is clearly an area where further work is necessary to develop satisfactory models describing behaviour. 3.4 Girth weld defect tolerance The fracture control plan for a pipeline should include consideration of girth weld failure. It is expected that automated welding and automated ultrasonic testing will be utilised for the trunkline welds; this will be supported by a fitness-for-purpose approach to setting weld defect standards and inspection requirements. This approach is particularly needed in areas where strains above yield are predicted, for example in active seismic regions or zones of discontinuous permafrost. Under these conditions traditional workmanship acceptance levels are not usable and alternative methods are required to relate the acceptance criteria to the anticipated loading and the material properties. The guidelines produced by EPRG [19] are a good example of this type of approach. Tier 2 of these allows axial strains up to 0.5% and relates the defect length to the wall thickness; this level can be used provided a specified Charpy impact energy is reached. The published guidelines are currently limited to grades up to X70, but recent work [20] has extended them to X80, albeit with lower margins of safety. Extension beyond X80 has not yet been considered by EPRG, but results for X100 are beginning to become available. Again, this is an area where further work will be required to confirm satisfactory behaviour and produce general guidance. In the near term it is likely that project specific acceptance criteria will be needed. 3.5 Fittings A pipeline will require fittings such as bends, tees and valves, together with more complex components such as pig launchers and receivers. These will require sufficient toughness to ensure resistance to fracture initiation; there is no requirement for crack propagation control as generally the stresses are lower than in the linepipe to compensate for the lower strength levels usually used in fittings. The toughness levels can be set based on the requirements for linepipe, or using approaches in the pressure vessel design codes. Some care is needed when applying the pressure vessel codes to pipeline components, as the traditional pressure vessel materials do not approach X100 strength levels. For example the UK pressure vessel code PD 5500 [21] does not include any Carbon- Manganese steels with yield strengths approaching even X70. A further problem is that the pressure vessel codes generally limit the design allowables to a lower proportion of yield than in a pipeline where design factors of 80% are used. These problems can be offset by an increase in thickness, but this will increase the weight and fabrication costs and will make it more difficult to achieve adequate toughness. It may be necessary to move away from traditional pressure vessel steels to resolve these issues. 3.6 Brittle fracture propagation resistance For conventional line pipe grades and thickness relatively little attention has been given in recent years to brittle fracture propagation as modern high quality line pipe can readily achieve the specified shear area requirements in Drop Weight Tear Tests (DWTT), and these have been shown to 5 be conservative compared to full scale pipe burst (West Jefferson) tests. Nevertheless it has been considered necessary to confirm the situation [22] for grade X100. Figure 3 presents the results of Charpy, DWTT and West Jefferson tests for a 1422mm diameter pipe demonstrating that the behaviour of X100 is similar to that of lower grade pipes and that the same criteria for brittle fracture resistance can be applied. 3.7 Ductile fracture arrest Extensive research over many years has been directed towards understanding and predicting the conditions for ductile fracture propagation and arrest [5, 6] and the issues concerning higher strength steels are similarly receiving considerable attention [22, 23]. The most widely used predictive method is the Battelle two curve approach [24], where the crack tip driving force at arrest is considered and then a relationship is obtained between the arrest stress, the crack velocity, the pipeline dimensions and material properties. This relation was calibrated against over 200 full-scale fracture arrest tests on line pipe up to X65 grade pressurized with air or lean natural gas. This curve defines the driving force required to cause the crack to propagate; the driving force available is derived from modelling the decompression behaviour of the escaping gas. During the initial research into ductile fracture arrest in the 1970s the full two curve model was cumbersome to use, and a simple formula was fitted to the results of applying the two curve model to a range of pipeline designs typical of the period. Other research groups also developed predictive formulae with varying degrees of empiricism [6]. These equations have been used to set toughness specification levels in various codes, but current projects and proposed designs are now well outside the range of data on which the equations were originally calibrated. To examine the applicability of these approaches to X100 line pipe, a programme of full-scale fracture arrest tests has been completed recently by Advantica, on behalf of an international group of sponsors. 914mm diameter grade X100 pipes were obtained from three leading Japanese suppliers, NKK, Sumitomo and Kawasaki, and were used to undertake two full-scale tests, pressurised using lean natural gas, at Advanticas Spadeadam test site. The testing conditions are summarised in Table 1; the test facility and procedures are described more fully elsewhere [25]. Figure 4 shows the test section after the first test. The results from these tests, the first ever conducted on X100 pipe pressurised using natural gas, are compared with other test results on high strength large diameter pipes, including two air-pressurised X100 tests [22] in Figure 5, using the Battelle two-curve approach as the basis for comparison. These first results show that the full-scale fracture arrest behaviour of Grade X100 appears to be consistent with the data and trends developed for other high strength (X80) steels, taking into account the amount of scatter shown in current and previous results. Figure 5 shows that the factors relating predicted and actual behaviour were 1.3-1.5 (propagate) and 1.5 1.9 (arrest) for the individual pipes tested, compared to around 1.4 for other data. However it can also be seen that one of Figure 3 DWTT and Charpy transition curves and West Jefferson test results for 1422 mm diameter, 19 mm wall thickness X100 pipe. Data from [22]. 6 the air-pressurised tests propagated at a factor of 1.7. This degree of scatter makes it difficult to determine a factor that can be applied consistently and used for other designs. There is clearly a need for more understanding of the fracture behaviour of X100, to eliminate some of the sources of uncertainty.
Test 1 Test 2 Nominal diameter 914 mm (36 inch) 914 mm (36 inch) Nominal thickness 13 mm (1/2 inch) 15 mm Test pressure 136 bar 180 bar Test design factor 69 % SMYS 79 % SMYS Backfill Soil, 1 metre Soil, 1 metre Test temperature 8.5 C 15.0 C C1 96.45 Gas Composition (average for Tests 1 and 2) C2 3.26 C3 0.10 C4+ 0.02 Nitrogen 0.17 CO2 0.00 Table 1 Test conditions for X100 fracture propagation tests
The limitations of approaches such as Battelle two-curve model, and the need for empirical adjustments to describe the behaviour of tough, high strength steels have focused attention on the limitations of the Charpy test as a measure of fracture resistance. Improvements can be made in several ways; firstly by ensuring that the test machine energy capacity is significantly greater than the fracture energy; secondly by separating the energy of propagation from the energy absorbed during initiation and bending. Instrumented Drop Weight Tear Tests have also been used extensively to measure propagation resistance in terms of fracture energy per unit area of crack extension. However none of these approaches has yet demonstrated the potential for significant improvement in predictive capability compared to the Battelle two-curve approach. Alternatively, deformation-based approaches to fracture arrest have been explored. Ductile fracture propagation in full scale pipe tests is associated with a significant concurrent plastic strain field, with an axial strain extending up to 2.5 pipe diameters ahead of the crack tip [6]. Consideration of these strain fields led to the concept of a complex plastic strain zone associated with a constant crack tip opening angle (CTOA) corresponding to the crack driving force during steady state propagation. It was also found by analysing DWTT specimens that a similar geometrical parameter (CTOA c ) could be identified to characterise the fracture propagation resistance of the material. Hence in full-scale Figure 4 Fracture propagation test on 36 inch X100 linepipe, showing pipe body arrest at far end of image. 7 fracture tests, propagation or arrest will be determined by whether the CTOA is greater or less than CTOA c . While the CTOA approach is beginning to show considerable promise, there are two key features of the fracture process that add significantly to the complexity of the analysis, particularly for high strength steels. Firstly anistropy of the material, in particular the axial and longitudinal differences in yield strength and initial work hardening rate, have a major influence on the shape, size and energy absorption characteristics of the dynamic biaxial plastic zone at the crack tip, and hence on both Figure 5 Fracture arrest predictions for X100 and other high strength linepipe Figure 6 CTOA measurements on X80 linepipe, showing consistency of measurement from crack flanks and grids marked at 1 mm intervals from the crack. Data from Shterenlikht et al [27]. 8 CTOA and CTOAc. Secondly, the nature of the fracture surface (the extent of flat and shear surfaces, the magnitude of any separations) can influence the effective CTOAc and render it geometry- dependent. In an attempt to address this latter issue, Advantica have recently undertaken a programme of work [26, 27] in collaboration with Sheffield University (UK) to measure CTOAc directly in a specimen with a much lower degree of bending and a longer amount of crack growth. In addition, the mode of loading in this specimen allows a transition to the fully shear crack propagation which is evident in full-scale pipe tests but not in the DWTT or Charpy geometries. Early results from this work, on X80 pipe, are encouraging; the CTOA reduced to a constant steady state value (Figure 6) as the fracture mode changed from flat to fully shear. This steady state value was maintained over a distance of about seven times the test section thickness. These latest developments may eventually result in improved understanding and definition of the critical parameters influencing crack propagation and arrest. At present however CTOA does not offer any improvement in predictive capability compared to the empirically adjusted Battelle two-curve approach. A further requirement for the fracture control plan is the predicted rupture length, the number of pipe joints through which the crack will propagate before it arrests. By taking account of the scatter in the full-scale test data, EPRG [28] developed a probabilistic approach to determine the probability of different rupture lengths occurring. The outcome is illustrated in Figure 7 for a typical X80 pipe design; the analysis incorporates the toughness distribution of the pipe actually supplied and estimates the proportion of pipe joints that can arrest a crack. This approach has the benefit of being compatible with the trend towards reliability-based design and operation of pipelines, but it does require knowledge of the pipe toughness distribution. Until X100 material enters full scale production it will not be known if the same level of consistency can be obtained as in lower grade pipe. It also assumes that cracks arrest within one pipe joint if the conditions for arrest are met. A dynamic variant of the Battelle two-curve approach has been developed by the Japanese High-Pressure Line Pipe Committee [29], enabling predictions of arrest over distances of several pipe joints. At present this approach has only been validated for a limited data set, but if extended and used in conjunction with the EPRG probabilistic approach it would offer the potential for demonstrating the likelihood of arrest in situations where conventional analyses produce marginal results. For several of the new pipeline projects and conceptual designs it is unlikely that the pipe body toughness available from leading suppliers will be sufficient to guarantee a high enough probability of crack arrest. In this situation the alternative is to use crack arrestors [30]. While some pipelines incorporating arrestors have been implemented, arrestor design methods are not well established or validated experimentally. Arrestors are more likely to be accepted in remote areas with low risk to public safety, and the spacing is likely to be based on the economic balance between arrestor installation costs and long rupture repair costs, including penalties for supply interruption. Figure 7 Prediction of ductile rupture lengths for a typical X80 pipeline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Rupture Length, pipes 10 -7 10 -6 10 -5 10 -4 10 -3 10 -2 10 -1 10 0 P r o b a b i l i t y 9 However arrestors may actually increase the likelihood of a fire at the arrestor location and this must be taken in account in any risk assessment. If arrestors are used for new highly-rated designs utilising high strength steels, an improved understanding of the dynamics of crack arrest will be essential to develop and validate the design methodology. 4 OVERVIEW AND CONCLUDING REMARKS In the 50 years since the first full-scale fracture arrest tests were conducted by the Battelle Memorial Institute, the understanding of the issues concerning the initiation, propagation and arrest of brittle and ductile fractures in pipelines has received considerable attention. This in turn has delivered substantial benefits in terms of the enviable safety record of gas transmission pipeline networks worldwide. When pipelines are designed and constructed in accordance with modern best practice the likelihood of failure is extremely low, especially in those areas where public safety or environmental sensitivity is an issue. This favourable situation should not however be an indication that relaxation of the standards is either possible or acceptable. Indeed, as the new designs for highly rated pipelines continue to extend the performance envelope, it has become more important to replace enlightened empiricism and engineering judgement, however sound, with comprehensive analytical understanding of all aspects of design, material performance, construction quality and operational requirements. The application of high strength steels is a classic example of the need to proceed with caution, adopting an openly reviewed/challenged multidisciplinary approach at all stages. The major new pipeline projects are of key importance to the whole energy industry and a high-profile failure would have very wide repercussions. Hence it is important to pay very close attention to every detail of Fracture Control in high strength pipelines. This paper has reviewed and highlighted many of the advances that have been made, not least through collaborative cross-industry efforts combining the latest analytical methods with carefully targeted full-scale testing. While many of the issues associated with fracture control can be addressed through already established approaches, areas where further work is still necessary include resistance to mechanical damage, girth weld defect tolerance and ductile fracture arrest. Advantica is proud to have been involved in developing solutions to Fracture Control issues for a significant portion of this 50 year period and we continue to be at the forefront of both experimental and analytical efforts to enable the use of higher strength steels for next-generation pipelines. 5 ACKNOWLEDGEMENTS The authors wish to thank the sponsors and supporters of the Joint Industry Project on Fracture Control in High Strength Steel Pipelines (Alliance Pipeline, BP Exploration, BG Group, TransCanada PipeLines, Kawasaki, NKK and Sumitomo) for agreeing to the publication of the full- scale fracture arrest tests on X00 line pipe. The authors also wish to acknowledge input and discussions with many present and former colleagues in Advantica and elsewhere, particularly in the European Pipeline Research Group (EPRG) and the Pipeline Research Council International (PRCI). Finally they wish to thank Advantica for permission to publish this paper. REFERENCES 1. Sanderson, N., Ohm, R. K. and Jacobs, M. (1999). Study of X-100 line pipe costs points to potential savings. Oil & Gas Journal, March 15, 97, 54-57. 2. Donati, E. and Ercolani, D. (2000). High pressure long distance pipelines: technical-economic analysis. IGU World Gas Conference, Nice, June 2000. 3. Glover, A. (2002). Application of Grade 550 (X80) and Grade 690 (X100) in Arctic climates. Proceedings, Application and Evaluation of High Grade Linepipes in Hostile Environments, Yokohama, Japan, 7-8 November 2002. Edited Denys, R. and Toyoda, M. 33-52 4. Gray, J. M. (2001). Niobium bearing steels in pipeline projects. Presented at Niobium 2001, to be published by ASM. 5. Eiber, R.J., Carlson, L. and Leis, B. (2000). Fracture control requirements for gas transmission pipelines. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys, R. Amsterdam: Elsevier Scientific, 437- 53. 10 6. Rothwell, A. B. (2000). Fracture propagation control for gas pipelines past, present and future. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys R Amsterdam: Elsevier Scientific, 387-405. 7. Kiefner, J.F., Maxey, W.A., Eiber, R. J. and Duffy, A. R. (1973) Failure stress levels of flaws in pressurized cylinders. Progress in flaw growth and fracture toughness testing ASTM STP 536. Philadelphia: American Society for Testing and Materials 461-81. 8. Shannon, R. W. E. (1974). The failure behaviour of linepipe defects. International Journal of Pressure Vessels and Piping, 2: 243-55. 9. Mannucci, G. and Harris, D. (2002). Fracture properties of X100 gas pipeline steels. Final report on ECSC Agreement 7210-PR/058. Brussels: ECSC. 10. Demofonti, G., Mannucci, G., Harris, D., Barsanti, L. and Hillenbrand, H-G. (2002). Fracture behaviour of X100 gas pipeline by full scale tests. Proceedings, Application and Evaluation of High Grade Linepipes in Hostile Environments, Yokohama, Japan, 7-8 November 2002. Edited Denys, R. and Toyoda, M. 245-261 11. Fu, B., Millwood, N. A., Vu, D. Q., Matthews, S. and Jacobs, M. (2000). Failure behaviour of ultra high strength line pipe. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys R Amsterdam: Elsevier Scientific 497-507. 12. BSI. (2000). Guide to methods for assessing the acceptability of flaws in metallic structures BS 7910:1999 Incorporating Amendment 1, April 2000. London: British Standards Institution. 13. Ohm, R. K., Martin, J. T. and Orzessek, K. M. (2000). Characterization of ultra high strength line pipe. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys R. Amsterdam: Elsevier Scientific 483-96. 14. Martin, J. T., Koers, R. W. J. and Ohm, R. K. (2000). Seam weld defect tolerance in ultra high strength line pipe. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys R. Amsterdam: Elsevier Scientific 521-30. 15. Batte, A. D. and Kirkwood, P R. (1998). Developments in the weldability and toughness of steels for offshore structures. Proceedings, Microalloying 85, Chicago, ASM, 175-188. 16. Bood R.J., Galli M.R., Marewski U., Roovers P., Steiner M. and Zarea M. (1999). EPRG methods for assessing the tolerance and resistance of pipelines to external damage Part 1. 3R International 38, 739-744; Part 2: 3R International 38, 806-811. 17. Maxey, W. A. (1986). Outside force defect behaviour. Seventh symposium on line pipe research. Houston, October 14-16 1986. Arlington, Virginia: Pipeline Research Committee Paper 14. 18. Takeuchi, I., Fujino, J., Yamamoto, A. and Okaguchi, S. (2002). The prospect of high grade steel pipe for gas pipelines. Proceedings, Application and Evaluation of High Grade Linepipes in Hostile Environments, Yokohama, Japan, 7-8 November 2002. Edited Denys, R. and Toyoda, M. 185-203. 19. Knauf, G. and Hopkins, P. (1996). The EPRG guidelines on the assessment of defects in transmission pipeline girth welds. 3R International 35(10/11):620-624. 20. Knauf, G. (2002). EPRG: Crack arrest and girth weld acceptance criteria for high pressure gas transmission pipelines. Proceedings, Application and Evaluation of High Grade Linepipes in Hostile Environments, Yokohama, Japan, 7-8 November 2002. Edited Denys, R. and Toyoda, M. 475-500. 21. BSI. (2000). Unfired fusion welded pressure vessels. PD 5500:2000 London: British Standards Institution. 22. Demofonti, G., Mannucci, G., Harris, D., Barsanti, L. and Hillenbrand, H-G. (2002). Fracture behaviour of X100 gas pipeline by full scale tests. Proceedings, Application and Evaluation of High Grade Linepipes in Hostile Environments, Yokohama, Japan, 7-8 November 2002. Edited Denys, R. and Toyoda, M. 245-261. 23. Batte, A. D. (2001). Developments in high-strength steel for pipelines. Pipeline World 10(12):16. 24. Maxey, W. A. (1974). Fracture Initiation, Propagation and Arrest. Fifth Symposium on Line Pipe Research, Houston, November 20-22 1974. AGA Catalogue L30174. Arlington, Virginia: Pipeline Research Committee Paper J. 25. Johnson, D. M., Horner, N., Carlson, L. and Eiber, R. J. (2000). Full scale validation of the fracture control of a pipeline designed to transport rich natural gas. Pipeline Technology Proceedings of the third international pipeline technology conference Volume I, Brugge, Belgium, May 21-24 2000. Edited: Denys, R. Amsterdam: Elsevier Scientific 191-209. 26. Andrews, R. M., Shterenlikht, A., Howard, I. C. and Yates, J. R. (2002). Measurement and modelling of the crack tip opening angle in a pipeline steel. International Pipeline Conference 2002. Edited: Ellwood, J. R. New York: ASME. 11 27. Shterenlikht, A., Hashemi, S. H., Howard, I. C., Yates, J. R. and Andrews, R. M. (2003). A specimen for studying the resistance to ductile crack propagation in pipes. Submitted to Engineering Fracture Mechanics. 28. Dawson, J. and Pistone, V. (1998). Probabilistic evaluation of the safety embodied in the EPRG recommendations for shear fracture arrest toughness. 3R International 37(10/11):728-33. 29. Makino, H., Kubo, T., Shiwaku, T., Endo, S., Inoue, T., Kawaguchi, Y. et al. (2001). Prediction for crack propagation and arrest of shear fracture in ultra high pressure natural gas pipelines. ISIJ International 41(4):381-8. 30. Venton, P. B. and Dietsch, A. E. (1997). Design of crack arrestors. Proceedings, International Seminar on Fracture Control in Gas Pipelines. Edited: Rothwell, A. B. Lidcombe New South Wales: Welding Technology Institute of Australia. Paper 9.