December, 2010 Vol.9, No.4 Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n incorporating The Journal of Pipeline Integrity Great Southern Press Clarion Technical Publishers 4th Quarter, 2010 209 The Journal of Pipeline Engineering incorporating The Journal of Pipeline Integrity Volume 9, No 4 • Fourth Quarter, 2011 Sa no m t f ple or c di op st y rib ut io n Contents James Watt ................................................................................................................................................................213 Carbon dioxide transport infrastructure key learning and critical issues Saulat Lone, Dr Tim Cockerill , and Prof. Sandro Macchietto . ......................................................................... 223 The techno-economics of a phased approach to developing a UK carbon dioxide pipeline network Dr Brian N Leis, Dr James H Saunders, Ted B Clark, and Dr Xian-Kui Zhu .................................................... 235 Transporting anthropogenic CO2 in contrast to pipelines supporting early EOR Graeme G King and Satish Kumar ....................................................................................................................... 253 How to select wall thickness, steel toughness, and operating pressure for long CO2 pipelines Prof. Haroun Mahgrefteh, Solomon Brown, and Peng Zhang ........................................................................... 265 A dynamic boundary ductile-fracture-propagation model for CO2 pipelines Dr Robert Andrews, Dr Jane Haswell, and Russell Cooper ................................................................................ 275 Will fractures propagate in a leaking CO2 pipeline? Dr Tim Cockerill, Dr Naser Odeh, and Scott Laczay . ......................................................................................... 285 Greenhouse gas emissions from electricity generating CCS upstream and downstream transport processes ❖❖❖ Our cover photo shows construction work under way on Denbury Resources’ 24-in diameter, 512-km long Green Pipeline for both natural and man-made CO2. The new pipeline will be one of the first designed to transport anthropogenic CO2 in the Gulf Coast area of the US. Photograph courtesy of Denbury Resources Inc., www.denbury.com 210 The Journal of Pipeline Engineering T HE Journal of Pipeline Engineering (incorporating the Journal of Pipeline Integrity) is an independent, international, quarterly journal, devoted to the subject of promoting the science of pipeline engineering – and maintaining and improving pipeline integrity – for oil, gas, and products pipelines. The editorial content is original papers on all aspects of the subject. Papers sent to the Journal should not be submitted elsewhere while under editorial consideration. Authors wishing to submit papers should send them to the Editor, The Journal of Pipeline Engineering, PO Box 21, Beaconsfield, HP9 1NS, UK or to Clarion Technical Publishers, 3401 Louisiana, Suite 255, Houston, TX 77002, USA. Instructions for authors are available on request: please contact the Editor at the address given below. All contributions will be reviewed for technical content and general presentation. The Journal of Pipeline Engineering aims to publish papers of quality within six months of manuscript acceptance. Notes 4. Back issues: Single issues from current and past volumes are available for US$87.50 per copy. Sa no m t f ple or c di op st y rib ut io n 1. Disclaimer: While every effort is made to check the accuracy of the contributions published in The Journal of Pipeline Engineering, Great Southern Press Ltd and Clarion Technical Publishers do not accept responsibility for the views expressed which, although made in good faith, are those of the authors alone. 5. Publisher: The Journal of Pipeline Engineering is published by Great Southern Press Ltd (UK and Australia) and Clarion Technical Publishers (USA): 2. Copyright and photocopying: © 2010 Great Southern Press Ltd and Clarion Technical Publishers. All rights reserved. No part of this publication may be reproduced, stored or transmitted in any form or by any means without the prior permission in writing from the copyright holder. Authorization to photocopy items for internal and personal use is granted by the copyright holder for libraries and other users registered with their local reproduction rights organization. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising and promotional purposes, for creating new collective works, or for resale. Special requests should be addressed to Great Southern Press Ltd, PO Box 21, Beaconsfield HP9 1NS, UK, or to the editor. 3. Information for subscribers: The Journal of Pipeline Engineering (incorporating the Journal of Pipeline Integrity) is published four times each year. The subscription price for 2010 is US$350 per year (inc. airmail postage). Members of the Professional Institute of Pipeline Engineers can subscribe for the special rate of US$175/year (inc. airmail postage). Subscribers receive free on-line access to all issues of the Journal during the period of their subscription. v Great Southern Press, PO Box 21, Beaconsfield HP9 1NS, UK tel: +44 (0)1494 675139 fax: +44 (0)1494 670155 email: [email protected] web: www.j-pipe-eng.com www.pipelinesinternational.com Editor: John Tiratsoo email: [email protected] Clarion Technical Publishers, 3401 Louisiana, Suite 255, Houston TX 77002, USA tel: +1 713 521 5929 fax: +1 713 521 9255 web: www.clarion.org Associate publisher: BJ Lowe email: [email protected] 6. ISSN 1753 2116 v v www.j-pipe-eng.com is available for subscribers 4th Quarter, 2010 211 Editorial Guest editorial: CO2 transportation by pipeline – a special issue T operational in the next five years. Much of the research has therefore to be conducted in parallel with FEED studies and it is vital that there is rapid information exchange between academia and industry. Sa no m t f ple or c di op st y rib ut io n HIS SPECIAL ISSUE of the Journal of Pipeline Engineering is dedicated to the topic of carbon dioxide transportation by pipeline. The subject of pipeline transportation of CO2 is becoming increasingly important as carbon capture and storage (CCS) schemes world-wide are moving from the pilot to the demonstration phases. In recognition of the increased interest in CO2 transport, Newcastle University – in association with Tiratsoo Technical and Clarion Technical Conferences – organized the first Forum on the transportation of CO2 by pipeline in July this year. The aims of the Forum were to highlight the key issues relating to CO2 pipelines and bring together leading academics and industry experts to discuss current and future CCS ventures and the international research activities being undertaken to support these projects. In this issue, five papers from the technical sessions of the Forum are published, together with two further papers from international authors working in the field. The conference opened with a session of keynote and scene-setting papers including the paper published in this volume by James Watt on �Carbon dioxide transport infrastructure: key learning and critical issues’. In this paper, the world-wide operational experience in the large-scale transportation of high pressure CO2 is reviewed in order that the issues relevant to the development of networks of pipelines for CCS schemes can be understood. It is estimated that there are approximately 6000km of CO2 pipelines globally, predominantly in North America; the majority of these pipelines are transporting naturally sourced CO2 for the purpose of enhanced oil recovery (EOR). Whilst it is concluded that the knowledge gained from this experience is vital, there remain areas pertinent to the development of CCS transport networks which still need to be addressed. However, one of the key messages of this paper is that, due to the urgent requirement to reduce CO2 emissions, the industry is not in the position of being able to wait until this research is complete before designing, constructing, and operating pipeline networks. If CCS is to have an impact in the reduction of CO2 emissions, then plants have to be One of the areas highlighted in James Watt’s paper for further research is that of materials’ selection. This important topic is the subject of the paper given at the Forum by Dr Paul and co-authors, who conclude that the main issues for the pipeline material are corrosion in the CO2 process stream, resistance to brittle and ductile fracture propagation, and degradation of polymers in supercritical CO2. With respect to corrosion, it is recognized that typical carbon steel materials used for pipelines are not corrosive in pure, dry, supercritical CO2, and the paper highlights that there is considerable experience and research in this area relevant to the currently operating pipelines. However, one of the significant differences that will be encountered in pipelines transporting CO2 in CCS schemes is that the product stream will contain impurities in combinations not currently transported in pipelines for EOR. There is very little research work on the effect of these impurities in the event of a process upset which could allow free water to enter the pipeline. In terms of fracture propagation, Dr Paul and his co-authors conclude that although there is extensive experience with the specification of material properties to prevent long running ductile and brittle fractures from propagating in natural gas pipelines, this knowledge cannot be directly applied to the design and fracture control of dense phase CO2 pipelines. This view is shared by Leis and co-authors in their paper published here on �Transporting anthropogenic CO2 in contrast to pipelines supporting early EOR’. In particular, they conclude that the main approach used for specifying ductile fracture control requirements in natural gas pipelines, the Battelle Two-Curve Method, has not been validated for CO2 containing impurities from a capture plant, and neither has the equation of state which forms the basis of this model. One of the key conclusions from this paper is that the decades of experience with the transportation of CO2 212 The Journal of Pipeline Engineering for EOR should be considered carefully when applying this knowledge to the design of pipelines for CCS, particularly in the area of fracture control. Indeed it is highlighted that many of the early CO2 pipelines were retrofitted with crack arrestors to manage concerns with fracture arrest. In a complementary paper, King and Kumar illustrate the issues to be considered in the design of a high-pressure CO2 pipeline for ductile fracture arrest using the example of the propsoed CO2 Masdar pipeline in Abu Dhabi. In this paper it is demonstrated that ductile fracture arrest is possible in high-pressure CO2 pipelines using the pipe wall thickness alone without the need for crack arrestors. Finally, Dr Tim Cockerill and co-authors present a lifecycle analysis to determine the CO2 emissions which are generated by the pipeline transportation of CO2 and the impact that these might have on the CCS process. The analysis indicates that the greenhouse gas emissions from the transport phase are almost negligible and therefore the optimal location of power plant in the design of networks should be driven by the requirement to minimize fuel processing and transport rather than CO2 transportation. As you read through the papers in this issue, one of the recurring conclusions drawn by the authors is that there is a still a significant amount of research required in the area of CO2 pipeline transportation. In particular, the experimental database on which so much of the knowledge relating to natural gas pipelines is built has not been established for CO2. There is therefore an urgent need for researchers to work in collaboration in order that the required research can be completed within timescales which allow CO2 pipelines for CCS to be designed and operated safely and efficiently. Sa no m t f ple or c di op st y rib ut io n As well as concerns over ductile fracture propagation, as discussed in the previous papers, one of the issues with CO2 is that there is a significant temperature drop around the leak site as the escaping fluid expands, due to the Joule-Thompson effect. It is postulated that this drop in temperature could cause continuous brittle initiation of a crack in a CO2 pipeline as the steel is cooled below the ductile-brittle transition temperature. This effect is investigated in the paper by Andrews, Haswell, and Cooper �Will fractures propagate in a leaking CO2 pipeline?’. infrastructure is set up to which smaller CO2 sources could then be added. The analysis conducted indicates that there comes a point in the network development where addingin smaller sources, particularly those that are remote from clusters of CO2 sources, considerably increases the marginal costs and diminishes the returns. An important additional conclusion drawn by King and Kumar is that, when designing a pipeline or network, there are a number of potential options that could be selected and cost optimization techniques, in combination with the technical requirements, should be implemented to derive the final solution. This theme is taken up by Saulat Lone and his co-authors in their paper �The techno-economics of a phased approach to developing a UK CO2 pipeline network’. This paper investigates the establishment of CO2 networks in the UK using idealized scenarios in which a backbone Dr Julia Race Senior Lecturer in Pipeline Engineering School of Marine Science and Technology, Newcastle University, Newcastle-upon-Tyne, UK [email protected] 4th Quarter, 2010 213 Carbon dioxide transport infrastructure: key learning and critical issues by James Watt AMEC Power and Process, Europe, Darlington, UK C ARBON CAPTURE AND STORAGE is acknowledged as one of the key technologies in carbon dioxide abatement. Whilst not a permanent solution, it can enable the continued use of hydrocarbon based power generation and reduce emissions from industrial processes. This is critical in decarbonizing whilst renewable and cleaner energy sources come online to resolve the issues around energy security and security of supply. Sa no m t f ple or c di op st y rib ut io n Transportation of carbon dioxide in a CCS scheme is critically important but has not been well addressed. While increasing attention is paid to storage and capture technology, transportation issues still lag behind. In particular, storage assessments and research increases and demonstration projects drive capture technology development. Shipping is rising as a potential solution and more consideration is being given to the impact of clusters and networks. The amount of carbon dioxide pipelines is approximately 6000km globally, the majority of which are in North America. Compared to other pipeline distances for natural or hydrocarbon pipelines, this is a relatively small experience base. Systems that do exist are also different. The majority of pipelines are installed for the purposes of enhanced oil recovery, often using natural sources. There are anthropogenic sources, but not many. Whilst pipeline design is common practice, the concern – if any – is the fluid being transferred and the dynamics of the system. In such a new field set for rapid growth industry needs to understand and make use of what experience is available and transferable and, more importantly, identify the gaps. Maturity of transport systems When considering the use of pipelines it is important first to take stock of the existing facilities. Pipelines are good references, but the knowledge embedded in them comes with the supporting research, engineering, and learning that comes with each design. The majority of carbon dioxide pipelines are in the USA and Canada, along with substantial in-field pipework for EOR schemes. There are other pipelines, including 90km in Turkey, and pipelines in Algeria and Hungary, but there are relatively few outside North America. Only two projects have offshore pipelines: Snovhit and Sleipner in the North Sea. This paper was presented at the First International Forum on Transportation of CO2 by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical and Clarion Technical Conferences, and with the support of the University of Newcastle and the Carbon Capture and Storage Association. Author’s contact details tel: +44 (0)1740 646100 email: [email protected] In North America since 1972 carbon dioxide has been used for enhanced oil recovery from some man-made sources, but the majority of transported carbon dioxide is from naturally occurring gasfields along the mid-continental mountain ranges and Mississippi Basin, as shown in Fig.1 and Table 1; the gas is transmitted above critical pressure, Fig.2. The maturity of the systems is still limited, but the experience that is there has formed a nucleus upon which to build CCS. The pipelines have been developed at the required scale; for example, the Cortez pipeline delivers 19.3 million tonnes per year, roughly equivalent to one of Europe’s largest power station at Drax, UK. So the pipelines have been proved at scale in terms of flow. That only 6000km of pipelines are in place is often pointed to as a weakness, not providing an extensive knowledge base, when compared for instance to the 490,000km of gas pipelines or the 278,000km of hazardous liquid pipelines in the USA alone. Whilst specific knowledge may be limited to 2% or less of the hazardous liquids’ experience in the US, there is still a knowledge base to consider: over 600,000km [1] of pipeline designed to the same codes of practice. The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n 214 Fig.1. North American CO2 pipelines Location Capacity (Mt CO2/y) Length (km) Pressure (bar) Year Complete USA 19.3 808 186 1984 Sheep Mountain USA 9.5 660 132 Bravo USA 7.3 350 165 1984 Bravo Dome Canyon reef Carriers (SACROC) USA 5.2 225 175 1972 Gasification Val Verde USA 2.5 130 - 1998 Val Verde Gas Plants Turkey 1.1 90 170 1983 Dodan field USA & Canada 5 328 Up to 204 2000 Gasification Pipeline Cortez Bati Raman Weyburn Origin of CO2 McElmo Dome Sheep Mountain Table 1: Major carbon dioxide pipelines [26]. Within the US and Canada, regulatory frameworks that govern carbon dioxide pipelines have been developed and deployed for a number of years. The design of such pipelines is essentially uses the same standards as for any hazardous liquid pipeline such as ethylene, crude oil, or petroleum products. These codes are, for the US 49 CFR 195, Transportation of hazardous liquids by pipeline, and for Canada Z662-07, Oil and gas pipeline systems. For gaseous carbon dioxide, 49 CFR 192 applies rather than the liquidspecific 49 CFR 195 [2]. The approach within the US is that the code or regulation contains a number of established standards from the API, AMSE, ASTM and others. Those referenced in the regulation or quoted, such as ASME B31.4, are the required minimum and therefore become part of the regulation, in theory a hard standard on which to base design. The core standard is ASME B31.4 – the code for liquid pipelines – see Fig.3. However, evidence suggests that ASME B31.8 is also applied. This gas-specific code is used to evaluate the safety issues around a gas pipeline, applying these rules to carbon dioxide liquid lines as the fluid transitions to gas on release. The process of design for pipelines in hazardous liquid service is therefore robust and well understood. European experience Current European experience rests with Statoil at Sleipner and Snovhit, although there is some carbon dioxide experience on the Continent. Predominantly, current transportation is either by ship or road tanker. The expectation is that European CCS will evolve to target 215 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.2. Carbon dixoide phase diagram and North American pipeline operating envelope. Fig.3. Prescribed standards and codes under 49 CFR 195. Fig.4. ISO 13623 petroleum and natural gas industries – pipeline transportation systems. 216 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.5. BS PD 8010 steel pipelines on land. offshore storage sites, and due to the potential for CCS clusters, develop in a network format. In some estimates the infrastructure for European CCS is estimated at between 30,000 and 150,000km [3]. This brings new challenges that have not been addressed in the current experience base, both in terms of scale and the need for offshore pipelines. As in the US and Canada, European and international codes are also mature: ISO 13623 (BS EN 14161) and DNV OS-F101 are equivalent to the US codes, although neither specifically consider dense-phase or supercritical carbon dioxide. However, both are adequate codes, but should be supported by industry best practice. In the UK the HSE recommends that BS PD 8010 is used as a wider appreciation of good practice in pipeline design. In each case the standards recommend the same or similar sub-codes and standards, as shown in Figs 4 and 5. Requirements in CCS systems The infrastructure needs for CCS are different from those of simpler enhanced oil recovery schemes. In the existing schemes, the carbon dioxide is a by-product or natural emission that is used to drive crude oil out of the ground. The economic driver is therefore the demand for the oil, and when the demand is not there, the extraction of natural carbon dioxide need not occur. Future CCS schemes will not have the same relationship. Whilst the sensitivity of the storage is an issue, the requirements here of flow, pressure, and temperature set the downstream conditions. The emitter in CCS provides another set of upstream conditions in terms of flow rates, ramp rates, composition, temperature, and pressure, and each of these systems imposes conditions on the transport infrastructure. This is multiplied when considering networks. Transportation systems have little ability to respond to variances: essentially, even the flow rate is dictated by the power station or storage. The dynamics of the whole chain system, with drivers at both ends, is therefore more complex. There will be an economic driver here to maintain transportation to the storage and maintain a storage solution that is flexible and allows accommodation of the emitter’s operating regime. What has the US learned? The experience in the US has highlighted the critical issues that must be considered for carbon dioxide service [4, 5]. These include • process conditions • properties • operating conditions at entry and exit • flow calculation method • transient (surge) modelling • flow characteristics • typical carbon dioxide compositions • piping design • fracture propagation • blowdown assembly design • blowdown rate basis and calculation • line break controls • pig trap • depth of cover • routeing topography • safety and environmental • ambient/ground temperature • blowdown rate basis and calculation 4th Quarter, 2010 217 Fig.6. CO2 pipeline incidents 1986-2008. dispersion pattern frequency and position of block valves leak-detection systems line inventory For inspection there have been problems with dense-phase or supercritical pipeline inspection and cleaning using pigs. Simple scraper pigs used in cleaning pipelines need a lubricating fluid such as diesel, and they are adversely affected by the carbon dioxide which damages some non-metallic materials [6, 7]. For inspection, the use of intelligent pigs is routine; however, inspection experience in the US has highlighted a key problem. The carbon dioxide penetrates the non-metallic components and, as the tool is depressurized in the receiver, the systems are often subject to rapid gas decompression, destroying the unit. In fact Oosterkamp [8] indicates that it has been reported that by 2008 only two intelligent pig operations in North America have resulted in the expensive tool surviving. Sa no m t f ple or c di op st y rib ut io n • • • • • material selection • pipeline materials • carbon equivalent • hardness value • fracture strength • valve, fitting and trim types • seal, packing materials of construction • valve actuators • cleaning and strength testing • cleaning • hydrostatic testing/drying/dewatering • construction techniques • corrosion monitoring • external corrosion • fracture propagation • special construction and welding • stress relief • pipeline operation • refrigeration effects during start-up/blowdown • start-up/shutdown methodology • line pressuring • requirement for blowdown noise control • environmental considerations • operational problems • operational safety • measurement • custody transfer methods • moisture analysis The list is extensive and typical to the design of pipelines no matter the fluid, but has particular relevance here. Particular learning can be drawn from the experience in the areas of inspection, corrosion, material specification, operational safety, and thermodynamics. Corrosion of carbon dioxide/water systems has been studied extensively, not just for the carbon dioxide EOR industry but also for processes involving the production of ammonia, urea, steam reforming, and the sweetening of acid gas. The key example to be considered is the research and practical observation that was undertaken for the SACROC project1. The research into corrosion by Schremp and Roberson [9] tested a number of compositions against three common weld types. The samples were full-size X60 12-in and 16-in pipe sections that were exposed to a carbon dioxide mix consistent with real SACROC pipeline conditions. Tests were conducted at two design temperatures and pressures, and at chemical compositions including 600-800ppm of H2S and 800-1000ppm of water. As a result, it is now accepted that corrosion in carbon dioxide pipelines will not occur if the water content is kept less than 60% of the saturation value [10]. The original test used a water content 20 times higher than specified for the SACROC pipeline, and concluded that corrosion did not occur, even at this elevated water content. Operational experience showed that after 12 years the SACROC pipeline, with a 50-ppm water content limit, had a corrosion rate of 0.25-2.5μm/y [11, 12]. 1 The so-called SACROC (Scurry Area Canyon Reef Operators Committee) unit in West Texas near the town of Snyder, one of the US’ largest oil fields, initiated a 350-km long carbon dioxide injection pipeline network on 26 January, 1972. 218 The Journal of Pipeline Engineering Fig.7.Typical orifice meter. The operational problems most reported are safety incidents reported in the US to PHMSA for all pipelines. In the period from 1986 to March 2008 there were 42 [16] reported incidents to PHMSA, Fig.6. This relates to approximately 0.36 incidents/1000km/year, (assuming US pipeline distance average of 5000km over the period), compared to 2447 incidents on the US gas transmission network of 488,000km [17] or 0.22 incidents/1000km/year. In the period 1990-2001, the incident rate for natural gas was 0.17, while for hazardous liquids it was 0.82 [18], so the incident rate is still comparable and within the expected bounds for a hazardous liquid. Sa no m t f ple or c di op st y rib ut io n Care always has to be taken with regard to corrosion; however, it is clear from continued operation that carbon steel is acceptable as material for carbon dioxide pipelines. There have been incidents, however, with corrosion: SACROC reports some issues [13], notably a corrosion incident in a spur line from a main pipeline. The area had free-standing water remaining from a hydrostatic test, and so corrosion set in. This experience shows that care must be taken not only to ensure a water content limit in the entry specification, but also for the water introduced during testing, commissioning, and maintenance activities. In specifying water content the industry-accepted level are conservatively specified as between 288-480mg/m 3 [14]. In addition the presence of other additional �acid gases’ such as H2S, SOx, and NOx needs to be considered. In the case of CCS, the control of SOx and NOx compounds will to be subject to earlier restrictions in the capture process and are unlikely to present themselves in the pipeline in significant volume. Hydrogen sulphide may be an issue for some processes but should be kept below the 200-ppm limit from Dynamis2 [15], Table 2; this restriction and the water content specification should prevent corrosion occurring. Non-metallic components such as seals, valve seats, O-rings, and even greases have also shown to be affected by carbon dioxide. Petroleum-based seals can become saturated with the high-pressure fluid and rapidly decompress when the pressure is reduced or structurally weakened. Some greases are also known to become hard and no longer effective. Inorganic materials and greases are therefore more often recommended. Operational safety Operational problems are not necessarily reported in the public domain unless a release occurs, so public information refers to where an incident has occurred, and has not necessarily been prevented. In effect, only the failure is recorded, so this limits the learning from operational sources. 2 The DYNAMIS project is investigating viable routes to large-scale cost-effective hydrogen production with integrated CO2 management. The project is an element of the HYPOGEN initiative, and forms a part of the European Commission’s QuickStart Programme for the Initiative for Growth. HYPOGEN has the goal of providing Europe with a viable route to a hydrogen economy and includes, as an interim step, the construction of a large-scale test facility for the production of hydrogen and electricity from decarbonized fossil fuels with permanent CO2 storage. Another reference source for pipelines (although not specifically carbon dioxide) is CONCAWE. The data held by CONCAWE indicates a significant number of incidents relate to corrosion and third-party intervention. Hence, for carbon dioxide systems, drying is a critical aspect of design and operation in addition to the usual protection applied to pipelines against third-party incidents. There are two other operational issues to be considered. The first is valve operation: here it is recommended that all valves are slow opening, as this avoids damage to the valves and surging in the pipeline, and this is particularly important for blowdown valves. Where segments are above ground, thermal relief should be provided and valves must be capable of seating under high carbon dioxide pressures. Current practice is not to work on pressurized pipelines at all and, where necessary, sections and valves bodies are blown down before removal of a valve or other equipment item. The thermodynamics of the fluid are generally based on the determination of its properties by using equations of state. The correct selection of the equations is therefore critical. Experience has also shown that for key operations such as start-up and blowdown, the thermodynamic characteristics require much longer periods to avoid the very low temperatures that are possible with carbon dioxide. There are other key lessons to consider, including transient fluid effects, leak proving (for which the use of air or nitrogen is not sufficient), spacing of block valves, the high level of sensitivity to temperature and pressure, and the attendant effect on pipeline operations. All of these need to be carefully considered. 4th Quarter, 2010 Component 219 Post Combustion IGCC Oxyfuel Weyburn Dynamis CO2 >95% >95% N2/Ar <4% <4% (for noncondensable gases) 0.01 0.03 – 0.06% 4.1% <4% EOR 1001000ppm O2 Hydrocarbons 0 0.01% H2 0 0.8 – 2% H2O 0 H2S 0 Hg Sox Nox Glycol 0 0 <100ppm <500ppm <1450ppmv <200ppmv Sa no m t f ple or c di op st y rib ut io n CO 0.01-0.6% <5% Saline Formation <4% EOR <2% <4% for all noncondensable gases and Hydrocarbons 0 0.03-0.4% 0 <2000ppmv <0.01%, <100ppm <0.01%, <100ppm 0.5% <100 ppm <0.01%, <100ppm <0.01%, <100ppm 0.01% <100 ppm 0 Table 2. Examples of carbon dioxide stream composition. Major differences There are major technical and economic differences to be considered in CCS schemes. The demand-led EOR schemes in which the need for oil production places a requirement on the provider and the provision of carbon dioxide. The compressor operating regime is dictated by the production rate required and the geological configuration of the reservoir. In CCS schemes, the same geological factors dictate to the transport system the operating conditions and flow rates, but there is also the influence of the upstream technology. The power plant or industrial process does not fit the same operating profile as a geologic storage. Carbon dioxide from CCS-enabled plant must be accommodated, or the emitter will have to free vent, incurring penalties and making the idea of a CCS scheme redundant. This has an impact on the design of the infrastructure, particularly the downstream configuration at the storage site. Whilst the influences from the storage sites are common with EOR, the addition of deep saline formations to the mix adds another level of complexity and another series of unknowns. The upstream processes also vary and whilst not relevant for a bespoke source-to-sink design, they are for network considerations. There are a number of issues that need to be addressed in the system design, and key amongst these has to be the composition of the carbon dioxide stream entering any transport system. Two considerations have to be made here: firstly safety, and secondly technical. In terms of safety, the impact of contaminants needs to be considered alongside carbon dioxide. It is not enough to model the dispersion of a carbon dioxide stream, but also its constituent parts must be modelled. This has been done by the Dynamis [19] project, which recommended the specification in Table 2. Typically entry into a US scheme is similar; the Canyon Reef project advises [20] the following specification for carbon dioxide: • 95% mol carbon dioxide minimum • 0.489g/m3 (250ppm wt) water in the vapour phase, no free water • 1500 ppm (w/w) hydrogen sulphide • 1450 ppm (w/w) total sulphur • 4% mole nitrogen 220 The Journal of Pipeline Engineering Fig.8.Typical vortex meter and contaminant analysis. • 5% mole, < -28.9°C dewpoint for hydrocarbons • 10ppm (w/w) oxygen The Dynamis project considered not only the technical issues, deviation of properties, density-phase envelopes, but also the safety aspects. As a result the criteria for H2S, CO, SOx, and NOx are established on health and safety grounds. Sa no m t f ple or c di op st y rib ut io n In terms of design safety, the methods are almost global, although this is one of the major issues in the industry today. Whilst we can draw methods and experience from the US and Canadian pipelines, there is an underlying issue. It can be argued that the EOR-based systems are more tolerant, more conservative in their approach, and with lower population densities, and income-generating oil revenue from EOR can afford this. The example to consider here is the approach to dispersion modelling adopted in the US. In the example there is a single assumption that a percentage of the carbon dioxide immediately upon release forms a solid which falls to the ground [21]. The remaining vapour is then modelled using the EPA’s Aloha programme or the models DEGADIS or SLAB. The rate of flow from the pipeline is also simplified to a common assessment. This methodology is conservative and the approach in Europe is much more precautionary. Hence a more defined understanding is desired in understanding key risks, such as fracture propagation and dispersion modelling. It must be stressed that the US approach is not wrong, but the driver in CCS appears to be much more of a considered approach, more accurate and more economic. The dynamics of a CCS system become more complex when considering networks. Large networks of CCS infrastructure are already proposed in Scotland, as well as in the Humber, Mersey and Dee, Thames, Teesside, and Rotterdam regions. These clusters range from 20 million to 90 million tonnes per year of carbon dioxide, and link diverse emitters in industry and power generation. How these networks behave and cope in different operating scenarios is critical in terms of both the design of the system and also, more importantly, the economics. In terms of system dynamics and the interactions of all the processing elements, CCS has a series of different operating modes to consider at both ends. The preference for storages is to be a generally constant flow, whereas power station emitters are cyclic and diurnal, and here is little scope for change between the two. The driver here is economics and affects the capture plant as well: the emitter can vent if a storage closes or requires a reduce inlet rate or pressure, but there will be an associated cost penalty. The simpler systems in the US with their demand-led supply can only provide some of the operational dynamic models that CCS requires. It is likely that the storage will have to be flexible enough in terms of dispersed entry points or multiple storage options. Flow measurement in carbon dioxide is also acknowledged as a possible challenge. Previous experience shows that two measurements are made: contaminants and flow. Typically either orifice or vortex flowmeters (Figs 7 and 8) in pipeline systems are insulated to limit temperature-induced density changes, and all meters are fitted with flow computers. The major difference in CCS is that the metering scheme will have to be compliant with the EU ETS requirements, and the meters will have to be of a fiscal standard. Whether this is possible is the subject of continuing research [22], and careful consideration must be given to the requirements of EU ETS when metering flow. To gain funding for the provision of networks, building them to cope with future additions, or �right sizing’ as it has become known, mandates that the economics and tariffs that could be expected need to be clear. This is not the funding model elsewhere, and for investment to move forward in CCS, the issue needs to be clearly understood. The need for networks are fundamental: CCS may not succeed if the reliance is on multiple source-to-sink solutions. Cluster networks enable significant savings over the collected costs of A to B solutions. In these areas, whilst current learning can be applied, CCS is fundamentally different. The learning required here to give confidence to investors, industry, and other stakeholders cannot come from the current experience or knowledge alone. Some of the lessons industry needs to learn will only come from the first CCS schemes and a high rate of knowledge sharing. Issues and approaches There are a number of issues that need to be considered in transportation and pipelines. These issues do not prevent 4th Quarter, 2010 221 development, but their resolution would aid both technical and economic decisions. The technical challenges are already known, but an understanding of how to address them needs to be developed. Some good practice-based guidelines by the Energy Institute and DNV aim to resolve some of these issues, whilst research programmes aim to address gaps in the knowledge. Typically existing experience has either highlighted the issue, or it is specifically CCS-related. At the other end of the scale, the storage companies need to understand that the power utilities and emitters may dictate the system parameters, not the storage. The dynamics of these systems are different and the future operators and owners need to understand this. There is a cultural shift within corporations, business models, and individual sites that needs to be considered and addressed. When deployment of CCS comes, it must be received by an intelligent, informed, and correctly-resourced workforce with training complete. When considering resources, there needs to be an understanding of the market size and rate of deployment. Any evaluation of the CCS market is made with caution due to the number of dependencies involved. The IEA roadmap [25] provides one such analysis, stating the expectation that by 2020 we need 100 projects, by 2030 we need 850, and by 2050 we need 3400 projects, in order to meet the BLUE map scenario. Projects rely on three issues: the supply chain, engineers to design, and skilled technicians to construct. How many people does it take to design and build 100 projects in 10 years? The engineers in the marketplace now are needed in existing areas. New and emerging areas like CCS and biofuels, or even new nuclear construction, will draw on the same resource pool and supply chain. The CCS industry and academia must address these issues rapidly and build capacity now, ready for delivery to the market in a decade. Sa no m t f ple or c di op st y rib ut io n Flow-assurance guidelines for carbon dioxide, and an understanding of the fluid behaviour, needs to be clearer. Experience in the US [23] indicates that there is an issue with surge and transient pressure. Because the US experience is restricted to onshore pipelines, and with offshore pipelines anticipated for CCS, there is a concern. Few mitigation measures can be added to offshore pipelines, particularly those that terminate at subsea completions. In this discipline there are also concerns that, even at low water content in the fluid, carbon dioxide clathrates (hydrates) may form. The definition of clathrate formation behaviour, and this area of vapour-liquid equilibrium, clearly require further research. emitters, particularly power generators, face in the addition of a capture plant, pipeline, and storage are considerable. The current staff will have new processes to control and monitor, and chemical stocks to maintain and dispose of. Whilst these processes are common to the energy industry, they are not common to the power-generation companies. Importantly, the resolution of physical properties of possible fluid streams remains an issue. Equation-of-state selection is covered in multiple academic papers, but empirical data to support the assumptions and outputs of the predictive methods are needed. Whilst engineering design can be achieved using the predictions, the design margins applied may prove to be excessive. This is particularly of interest in phase-envelope predictions to protect systems against multi-phase flow. The data must cover the range of possible contaminants and process conditions for a pipeline system. Dispersion modelling is already acknowledged as an issue, and some experimental work has been done by BP [24] and Scottish and Southern Electricity for the now-cancelled DF1 project. Further modelling work is also planned by DNV in the Pipetrans research programme. The issues around dispersion are complex. Clear dispersion modelling and the behaviour of a depressuring pipeline are critical elements in both determining the major accident response and also in defining safe distances. The behaviour of carbon dioxide at the release point, the source terms, needs to be defined, and the computational models validated against it. Computational-fluid-dynamic models can be used instead of the simpler commercial programmes. However, the efficient and accurate modelling which is at an early stage for the purposes of safety cases and route definition is key to efficient, practical, and safe design. Human factor While the critical technological issues discussed here aid design, safety, and integrity of the pipeline, the human factor should not be ignored. Important in the learning from the US is that the schemes there are operated in an oil and gas production environment. The changes that the Perhaps the biggest potential issue for the industry to face is the public. There is a growing need to address public education and perception. Poor responses in Continental Europe and opposition in the USA are already showing public resistance to storage and pipelines. These issues already exist in all fields, and experience can be drawn from the UK’s position on gas storage. The reserve capacity in the UK gas supply is approximately 4% of the annual consumption, which compares poorly to that of France at 24% or Germany at 21%. The number of potential projects to rectify this is significant, but the majority of projects suffer in planning, face local opposition, or enter public enquiry, lengthening the development time and cost. Despite the urgent need for capacity, the poor educational message and engagement has prevented a number of projects proceeding in a timely fashion. This challenge now faces CCS: failure to engage with the public will have devastating effect on project development. There needs to be project-specific information, but also a wider education programme. Conclusions In conclusion, it is apparent that the knowledge gained from the current carbon dioxide pipelines will prove vital. It is 222 The Journal of Pipeline Engineering not the only the experience and solutions, but also those issues that have not been fully addressed. There are critical areas to address, including flow assurance, dispersion, properties, and engagement. These need to be resolved to enable deployment. However the body of evidence, experience, and proven design methodology, codes, and regulation all enable pipelines and infrastructure to be designed. While these parameters may prove to be conservative and therefore more costly, it does not prevent pipelines from moving forward. Nor can CCS afford to wait: design for the transport infrastructure is underway now. The industry must maximize the transfer of knowledge from carbon dioxide and hazardous liquid pipeline design, while parallel research must address the issues discussed here and generate resources and tools that can meet the challenge to bring to wide-scale deployment. References Sa no m t f ple or c di op st y rib ut io n 1. www.bts.gov/publications/national_transportation_statistics/ html/table_01_10.html 2. CFR – Code of Federal Regulation. 3. Building the cost curves for CO2 storage: European sector. Report 2005/2 IEA GHG, UK. 4. M.Mohitpour, 2007. Pipeline design and construction: a practical approach. ASME Press, New York. 5. J.Barrie. Carbon dioxide pipelines: a preliminary review of design and risks. 7th Int. Conf. on Greenhouse Gas Technologies. http://uregina.ca/ghgt7/PDF/papers/ peer/126.pdf 6. M.Mohitpour, 2008. A generalized overview of requirements for the design, construction, and operation of new pipelines for CO2 sequestration. The Journal of Pipeline Engineering. 7. A.Oosterkamp and J.Ramsen, 2008. State-of-the-art overview of CO2 pipeline transport with relevance to offshore pipelines. Polytec, Norway. 8. Idem, ibid. 9. F.Schremp and G.Roberson, 1978. Effect of supercritical carbon dioxide on construction materials. Society of Petroleum Engineers. 10. G.Najera, 1986. Maintenance techniques proven on CO2 line. Oil and Gas Journal. 11. T.Gill, 1985. Canyon Reef Carriers Inc, CO2 pipeline description and 12 years of operation. ASME. 12. M.Seiersten, 2001. Material selection for transportation and disposal for CO2. Corrosion 2001. 13. L.Newton, 1977. Corrosion and operational problems, CO2 project, SACROC Unit. Society of Petroleum Engineers. 14. M.Mohitpour, 2007. Pipeline design & construction: a practical approach. ASME Press, New York. 15. de Visser et al., 2007. Towards hydrogen and electricity production with carbon dioxide capture and storage. Dynamis Consortium. 16. http://www.phmsa.dot.gov/pipeline 17. http://www.eia.doe.gov/pub/oil_gas/natural_gas/analysis_ publications/ngpipeline/ 18. J.Gale and J.Davison, 2004. Transmission of CO2 – safety and economic considerations. Energ,29. 19. de Visser et al., 2007. Towards hydrogen and electricity production with carbon dioxide capture and storage. Dynamis Consortium. 20. IPCC, 2005. Special report on carbon capture and storage. 21. www.energy.ca.gov/sitingcases/hydrogen_energy/documents/ applicant/revised_afc/Volume_II/Appendix%20E.pdf 22. A study of measurement issues for carbon capture and storage (CCS). Report 2009/54, April 2009, TUVNEL. 23. M.Mohitpour, 2007. Pipeline design & construction: a practical approach. ASME Press, New York. 24. http://archivos.labcontrol.cl/wcce8/offline/techsched/ manuscripts%5C8mnkk4.pdf 25. www.iea.org/papers/2009/CCS_Roadmap.pdf 26. IPCC, 2005. Special report on carbon capture and storage. 4th Quarter, 2010 223 The techno-economics of a phased approach to developing a UK carbon dioxide pipeline network by Saulat Lone1, Dr Tim Cockerill*2, and Prof. Sandro Macchietto3 1 Sui Northern Gas Pipelines Ltd, Pakistan 2 ICEPT, Imperial College London, UK 3 Department of Chemical Engineering, Imperial College London, UK A T Sa no m t f ple or c di op st y rib ut io n PHASED APPROACH to developing a CCS pipeline network would see an initial �backbone’ system constructed to collect carbon dioxide from the very largest sources. Once the backbone was in place, it might be possible to add a large number of smaller sources for relatively little additional cost.This paper analyses the techno-economics of a phased approach to rolling-out a comprehensive UK CO2 onshore pipeline network. We have developed a series of idealized scenarios where, initially, a new UK network is established to carry emissions from large-scale producers of carbon dioxide, defined here as more than 3Mtonnes per annum. In a second phase of development, medium-scale emitters are added to the network. A final third phase incorporates small producers with emissions in the range 0.5-1 Mtonnes per annum. For all scenarios, two different approaches to network construction have been compared, one using intermediate re-pressurization stations and one relying only on initial pressurization. Our results compare the construction and transportation costs of the different network configurations in each scenario, indicating the cost per tonne of CO2 transport. While there are some benefits offered to smaller sources by a phased approach, a rule of diminishing returns operates, with each tier experiencing an increase in marginal transport costs. The sensitivity of the costs to changes in the network configuration and design assumptions is investigated. HERE IS CURRENTLY interest in the technoeconomics of constructing a carbon-dioxide pipeline network for the UK to support the future deployment of CCS. It is likely that early development will focus on transporting CO2 from a small number of large carboncapture-fitted power stations to a limited number of offshore storage sites. As the number of carbon-capture-fitted point sources increases there may be some benefit from developing a common pipeline infrastructure. Once a �backbone’ network starts to emerge, an appealing possibility is that a large number of smaller CO2 producers could be connected to the network for relatively little cost. To investigate the feasibility of this idea, we have analysed the techno-economics of a phased approach to rolling-out a This paper was presented at the First International Forum on Transportation of CO2 by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical and Clarion Technical Conferences, and with the support of the University of Newcastle and the Carbon Capture and Storage Association. *Author’s contact details email: [email protected] comprehensive UK CO2 onshore pipeline network. We have developed a series of idealized scenarios where, initially, a new UK network is established to carry emissions from large-scale producers of carbon dioxide. In subsequent phases, smaller CO2 sources are progressively added to the growing network. The aims of this paper are to: • Develop an idealized, but representative, spatial data set detailing the locations of carbon dioxide sources and the quantities likely to be produced. The locations in the data set are disaggregated by size, such that they may be connected to facilitate study of a phased development of a carbon dioxide pipeline network. • Establish a method for modelling the phased development of a CO2 network. • Examine the technical requirements and performance of the network at each phase, and estimate the capital costs associated with network roll-out. Sa no m t f ple or c di op st y rib ut io n REGISTE R TODA Y Visit the Clarion website to register now www.cla rion.org Marriott Westchase Hotel Houston, Texas, USA Courses Conference Exhibition Now entering it 23rd year, the PPIM Conference is recognized as the foremost international forum for sharing and learning about best practices in lifetime maintenance and condition-monitoring technology for natural gas, crude oil and product pipelines. Plan to be there: www.clarion.org or call us at +1 713 521 5929 PLATINUM SPONSOR SILVER SPONSOR The international gathering of the global pigging industry! Conference Organizers 4th Quarter, 2010 225 Areas considered Approximate Year References North West UK 2006 [1] Scotland, Yorkshire 2008-2010 [2] East of England 2007 [3] Scottish Regional Carbon Capture and Storage (CCS) Study Scotland and Northern England 2009 [4] Yorkshire Forward Study Yorkshire and Humber 2008 [5] Europe Wide 2008 [6] Study Feasibility Study on the Transmission of CO2 CO2 Aquifer Storage Site Evaluation and Monitoring (CASSEM) North Sea Basin Task Force Study Ramp-up of large-scale CCS infrastructure in Europe Table 1. Selected studies of UK CCS pipeline infrastructure roll out. Brief UK literature overview For the tier of the largest producers, the GIS was used to identify corridors along which a backbone network of carbon dioxide pipelines could be constructed. The pipeline network was then designed using hydraulic analysis techniques. Finally a simple unit cost per in-km approach was used to estimate construction costs, assuming all pipelines are newly built. Sa no m t f ple or c di op st y rib ut io n Most work on the development of CO2 infrastructure for the UK has been performed by commercial consortia, with a focus on a regional development approaches. Figure 1 summarizes the areas considered by the first five major studies listed in Table 1. of a phased pipeline roll out, the sources were classified into three tiers according to their annual CO2 production. The literature suggests that links between regions have not been considered for developing a country-wide central CO2 transportation network. One drawback of regional approaches is the potential underutilization of storage capacity. Similarly, local CO2 transmission and storage capacity constraints could limit the extent to which CCS could be deployed within any region. This is particularly true in the North West of the country, where storage capacity is limited, but there are a large number of CO2 sources. Finally, a series of optimizations at local level is likely to deliver suboptimal solutions for the UK as a whole. Integration of regions with a central CO2 transmission system will help maximize utilization of both capture and storage potential. A similar approach was used to add all sources in each of the smaller production tiers to the backbone network. This allowed the implications of a phased approach to network development to be assessed, and in particular the marginal cost of adding smaller sources to be estimated. Calculations were carried out for networks with and without intermediate recompression stations. Identification of export terminals and sources Analytical approach Fig.1. Areas considered by UK pipeline studies. Figure 2 illustrates the methodology adopted for this study, which only considers the development of onshore pipelines connecting point carbon dioxide sources to a limited number of export terminals located on the coast. Further offshore pipelines will carry the CO2 to offshore storage locations, but the development of these components of the network are not examined here. In common with several other studies, it is assumed that onshore/offshore connections will only be permitted at existing pipeline terminals. Hence the first stage of the analysis was to identity existing UK pipeline terminals and their locations. Secondly, a spatial database of existing and prospective CO2 producers was developed within a geographical-information system (GIS) drawing on data from several sources. Due to the large number of potential sources, the database only included those above a threshold value. To facilitate study Identification of CO2 export terminals The UK’s existing oil and gas terminals and the nearest offshore oil and gas sedimentary basins with CO2 storage potential are summarized in Table 2. These CO2 export terminals would be equipped with compressor or pumping units for export of CO2 to the offshore storage. The storage potentials mentioned in the Table represent the realistic storage potential for each basin suggested by the British Geological Survey (BGS), which has been divided against each CO2 export terminal. At present, it is difficult to establish when individual fields or transport pipelines will become available due to commercial confidentiality of information. Therefore, for this study, it is assumed that UK’s offshore oil and gas fields will be available for CO2 storage when required. 226 The Journal of Pipeline Engineering Name of terminal St Fergus Gas terminal Nearest UK offshore Oil & Gas sedimentary basin CO2 Storage Capacity Northern & Central North Sea basin 1,346 Southern North Sea Basin 3,886 East Irish Sea Basin 1,043 Teesside Terminal Easington/Dimlington gas terminal Theddlethorpe gas terminal Bacton gas terminal Point of Ayr terminal Barrow-in-Furness gas terminal Table 2. UK existing onshore oil and gas terminals. Types of emitter Tier-0 3 million and above Coal & CCGT Power stations, Refineries, Steel industry Tier-1 1 million – 3 million CCGT & Oil Power stations, Refineries, Cement factories, CHP Tier-2 0.5 million – 1 million Sa no m t f ple or c di op st y rib ut io n CO2 Emissions Range Tonnes per annum Cement factories, CCGT Power stations, fertilizer, petrochemical complexes Table 3. Classification of emitters according to emissions. Design Approach Pipeline based Pipeline + Compression based Scenario name Captured CO2 volumes References S1A 156 Design “pipeline based” transmission network for Tier-0 emitters S2A 156 + 57 = 213 Design “pipeline based” transmission network for Tier-0 and Tier-1 emitters S3A 213 + 15 = 228 Design “pipeline based” transmission network for Tier-0, Tier-1 and Tier-2 emitters S1B 156 Design “pipeline + compression based” transmission network for Tier-0 and Tier-1 emitters S2B 156 + 57 = 213 Design “pipeline + compression based” transmission network for Tier-0 and Tier-1 emitters S3C 213 + 15 = 228 Design “pipeline + compression based” transmission network for Tier-0, Tier-1 and Tier-2 emitters Table 4. Definition of pipeline design scenarios. Selection of CO2 sources For this study, all current and planned (to 2015) industrial and power station CO2 emitting sources in the UK with CO2 emissions greater than 500,000t/a have been considered. In total, these sources account for 228Mt CO2 – approximately 50% of UK current CO2 emissions. The cut-off threshold is broadly consistent with the UK Government policy that any new combustion power station at or over 300MWe should be built �carbon-capture ready’ (CCR), as existing coal-fired power stations in UK with the threshold capacity would be expected to emit approximately 500,000t/a of CO2. The sources considered here therefore include larger power stations and large industrial sources. Non-power plant industrial sources of this scale comprise refineries, steel manufacturers, petrochemical complexes, fertilizers and cement factories. CO2 emissions data for the selection process were taken from a variety of sources, including the Environment Agency, published data under the EUETS programme, and the annual reports of each emitter. Figure 3 illustrates the approach. 227 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.2. Analytic approach used in this study. Division of emitters by CO2 emissions range Emitters above the threshold value have been categorized into three tiers as set out in Table 3. The division points were selected based on engineering judgment, with each tier distinguishing decreasing importance of the individual emitter for UK greenhouse gas emissions. • Tier-0 emitters: Tier-0 comprises emitters producing CO2 emissions of 3Mt/a and above, typically large coal-fired power stations with generation capacity from 1.5-4.0GWe. The largest emitter in this category is Drax power station, with the largest industrial emitter being Corus’ steel works. In total there are 20 existing power stations, five industrial installations, and five new power projects, totalling 156Mt/a (68% of the total UK emissions considered). and oil power stations with generation capacities in the range of 300–500MWe. Across the tier there are 13 existing power stations, 16 industrial installations, and three new power projects, totalling 15Mt/a (7% of the total UK emissions considered). Pipeline development scenarios The onshore CO2 pipeline transmission network for UK is designed in this study by considering following capacity phased approach: • Scenario 1 (S1): Evaluate the design and topology of a CO2 onshore �backbone’ transmission network to collect CO2 emissions from Tier-0 emitters only. • Tier-1 emitters: Tier-1 emitters are typically CCGT power stations and refineries, producing CO2 emissions in the range of 1–3Mt/a. The total inventory includes 19 existing power stations, 10 industrial installations, and eight new power projects, totalling 57Mt/a (25% of the total UK emissions considered). • Scenario 2 (S2): Re-evaluate the design of an onshore CO2 transmission network where Tier-1 emitters are added to the already laid transmission system for Tier-0 emitters. Determine the additional changes required in the transmission network to handle combined flows of Tier-0 and Tier-1 emitters to the CO2 export terminals. • Tier-2 emitters: Most Tier-2 emitters are industrial sources, typically cement and petrochemical plant, with emissions ranging from 0.5-1Mt/a. Power stations included in the Tier-2 band are predominantly CCGT • Scenario 3 (S3): Complete the design of the onshore CO 2 transmission system by considering the connection of Tier-2 emitters to the transmission system already in place for Tier-0 and Tier-1 emitters. 228 The Journal of Pipeline Engineering Fig.3. Approach to developing CO2 sources spatial database. flange of the CO2 compressor installed at export terminals to inject gas in offshore pipelines for subsequent storage in offshore oil and gas fields. Sa no m t f ple or c di op st y rib ut io n Two separate types of transmission models have been developed for each scenario, one based entirely on pipelines without re-pressurization (A) and the other using combinations of pipeline and intermediate compression stations (B). The full set of scenarios is summarized in Table 4. Data and assumptions Pipeline design assumptions Many factors have to be considered in the design of the new pipelines (assumed buried underground), including the properties and quantities of the fluid to be transmitted, underground conditions, and safety requirements. Some key data are set out in Table 5, with further discussion following. It is assumed here that CO2 transportation systems will be based on new onshore carbon steel pipelines, with a maximum diameter of 1,067mm. No re-use of existing infrastructure is accommodated within the study. The quantities of CO2 provided to the network by each source have been estimated by taking a carbon dioxide capture efficiency of 90% for power stations and 60% for industrial installations. The networks are designed for carbon dioxide in a supercritical state. Keeping in view the critical point of pure CO2, i.e. 74 bar, the minimum pressure at which CO2 would leave each source is taken to be 95 bar. This allows for a pipeline network pressure drop of 20 bar above the CO2 critical pressure. The operating pressure of the pipeline transmission network is specified at 100 bar due to the limitation of the operating pressures of the flanges and fittings at intermediate facilities. For intermediate booster stations, where used, it is assumed that recompression would be required when the pressure drops to 85 bar after which the pressure will be boosted again to 100 bar. It is assumed that arrival pressure at each CO2 export terminal would be 85 bar. This arrival pressure is the pressure at the upstream Another assumption is that CO2 streams from all the sources will be dehydrated to -5°C dewpoint, representing the temperature in the CO2 mixture at which water will start condensing. It is assumed that for Tier-0 and Tier-1 emitters, CO2 drying facilities would be installed at the premises as a part of the CO2 capture plant. The CO2 streams from Tier-2 emitters would be collected together at different locations for centralized drying. For deciding the -5°C water dew point of CO2 streams from each emitter, the available UK soil temperature data at 30cm depth for peak winter month (January) has been used. Since the majority of emitters are located in England, an average subsurface soil temperature of 3–5°C is taken for this study. Network topology A crucial factor in the design of any pipeline transmission network is the determination of pipeline route corridors. In this study, it is assumed that wherever feasible, the CO2 transmission network will follow the existing route corridors of onshore oil and gas pipelines in the country. However some new pipeline route corridors have been proposed for the emitters which are located away from existing oil and gas onshore pipeline infrastructure. Cost data Detailed construction costings for CO2 pipelines are difficult to obtain thanks to the limited number of CO2 pipelines operating worldwide. Instead, we assume that new carbon steel pipelines will be laid, and rely on approximate cost data for natural gas pipelines construction drawn from the IPCC special report [7] and the Oil and Gas Journal 2005-2008 [8]. In this study, the base unit pipeline total construction cost (including materials, labour, and all works) is taken to be US$ 30,000/in–km, including pipeline material costs, 4th Quarter, 2010 229 Fluid CO2 purity 100% Phase of CO2 Supercritical Critical Temperature 31°C Critical Pressure 74 bar Pipeline PN100 (100 bar nominal operating pressure) Standard used for pipeline fittings and equipment DIN2512 Pipeline Material A105-Carbon Steel Standard used for pipeline design criteria BS EN 14161 / BS EN 1594 Maximum allowable operating pressure of pipeline network 110 bar Pipeline internal design pressure 100 bar CO2 pressure leaving emitter’s premises 95 bar CO2 temperature leaving emitter’s premises 35 oC CO2 arrival pressure at export terminals 85 bar Minimum pipeline diameter 323.9 mm Maximum pipeline diameter 1,067 mm Onshore pipeline buried depth 1.2 - 1.8 m Sa no m t f ple or c di op st y rib ut io n Pressure rating of valves & fittings CO2 capture plants Efficiency of CO2 capture plant at power stations 90% Efficiency of CO2 capture plant at industrial installations 60% Table 5. Summary of pipeline design assumptions. labour costs, and costs related to intermediate facilities, but excluding additional compression. The impact of pipeline construction cost increases over this base level is later evaluated by performing sensitivity analysis. The cost of intermediate booster stations, where employed, has been estimated based on literature data. A review of several studies [9, 10, 11, 12] suggests that capital costs for installed compressor stations would range from $1500-4800/ kW. This wide variation can be attributed to differences in the year of installation, type of compression, geographical location, and installation of intermediate heating or cooling facilities. For the results reported here, a value for compression station costs towards the middle of the literature range, specifically $2500/kW, is adopted. This encompasses the cost of buildings, compressor units, prime mover, pipeline fittings, and drying equipment. Only the costs of intermediate booster stations are accounted for, with compression at sources and export terminals excluded from the calculations. However, for the booster stations, it is assumed that 50% of the compression requirement would be available as stand-by in order to make up in case of any emergency or during maintenance of main unit. All cost data have been converted to UK Pounds Sterling using 2008 exchange rates. Hydraulic modelling Introduction For designing the CO2 pipeline transmission system, a stateof-the-art hydraulic simulator, PipelineStudio version 3.0 has been used to simulate the transmission pipeline system for steady-state operation. For all scenarios, the pipeline network model has been constructed using: pipeline lengths and elevations from the GIS database, CO2 volumes from each emitter and location of CO2 export terminals, subsoil temperature data depending upon the region. Simulation methodology For the hydraulic simulation of the CO2 transmission network, the arrival pressure at each CO2 export terminal has been fixed at 85 bar whereas the captured CO2 volumes are fed at the source nodes. The pipeline diameters are then 230 The Journal of Pipeline Engineering Fig.4. CO2 transmission network for (A) Tier-0 emitters, (B) Tier 0+1 emitters, (C) Tier 0+1+2 emitters. Numbers 1,2,3… denotes CO2 export terminals as mentioned in modelling results table. where Sa no m t f ple or c di op st y rib ut io n optimized to keep the calculated pressures of each source within the pipeline operating limits and above the critical pressure of CO2. For estimating the thermodynamic properties of the CO2 mixtures, the Peng–Robinson equation of state has been used. While keeping the input flows constant, the pressures at each emitter are then back-calculated by the software by using the �general flow equation’ after Menon [11]: Q = gas flow rate, measured at standard conditions, ft3/day (SCFD) f = friction factor, dimensionless Pb = base pressure, psia Tb = base temperature, R (460 + F) P1 = upstream pressure, psia P2 = downstream pressure, psia G = gas gravity (air =1.00) Tf = average gas flowing temperature, R L = pipe segment length, mi Z = gas compressibility factor at the flowing temperature, dimensionless D = pipe inside diameter, in For compression-based scenarios, pipeline diameters of the transmission model have been optimized while considering intermediate booster stations. The intermediate booster stations are placed where the pipeline pressure drops to 85 bar after which the pressure is boosted again to 100 bar. The power requirement at each booster station are then calculated by the software using Eqn 2, again after Menon [11]: where HP = compressor horsepower γ = ratio of specific heats of gas, dimensionless Q = gas flow rate, MMSCFD T1 = suction temperature of gas, oR P1 = suction pressure of gas, psia P2 = discharge pressure of gas, psia Z1 = compressibility of gas at suction conditions, dimensionless Z2 = compressibility of gas at discharge conditions, dimensionless ηa = compressor adiabatic (isentropic) efficiency, decimal value Mass balance The pipeline transmission model drawn in the network has been verified by balancing the inputs and outputs of the transmission network. For steady-state calculations it is assumed that there will be no accumulation of CO2 volumes in the pipeline transmission network. Results Main network characteristics Table 6 summarizes the main technical results of the simulation and optimization study, and the network layouts are shown in Fig.4. It is clear that the pipeline-only scenarios (A) have pipeline diameters larger than the scenarios where re-compression is used (B). Without re-compression, larger 4th Quarter, 2010 231 Emission Sources: Tier-0 Scenario S1A Tier- 0 + 1 S1B S2A Tier- 0 + 1 + 2 S2B S3A S3B Basis: Power station CO2 capture Eff = 90% , Industrial CO2 capture Eff = 60% CO2 volume reaching export terminals Mtpa 34 34 71 58 75 75 2-Easington Gas Terminal Mtpa 46 46 55 55 56 56 3-Point of Ayr Terminal Mtpa 10 10 16 16 20 20 4-Theddlethorpe Gas Terminal Mtpa 44 44 40 53 42 42 5-Barrow-In-Furness Terminal Mtpa - - - - 1 1 6-Ireland Platform Mtpa - - 3 3 4 4 7-Teesside Gas Terminal Mtpa 19 19 25 25 27 27 8-St Fergus Gas Terminal Mtpa 3 3 3 3 3 3 Sub Total Mtpa 156 156 213 213 228 228 Pipeline Network Length Miles 883 883 1359 1359 1559 1559 Network Equivalent Diameter* mm 3,689 3,200 4,301 3,897 4,632 4,376 Sa no m t f ple or c di op st y rib ut io n 1- Bacton Gas Terminal Compression Requirements at compressor stations Comp-1 MW 2.0 6.6 6.5 Comp-2 MW 4.4 3.3 5.0 Comp-3 MW 5.3 3.5 4.8 Comp-4 MW 3.3 4.2 6.0 Sub Total MW 0 15.1 0 17.6 0 22.3 Table 6. Summary of main technical results for pipeline networks. * The term “network equivalent diameter” is used in this study only for indicative purpose in order to make an equal comparison of results of different scenarios. For each scenario it characterizes the diameter of the entire pipeline network as single diameter for ease of comparison of the results. pipeline diameters must be used to meet the pressure requirements at the CO2 export terminals. Transmission network capital costs A summary of total costs required to establish the CO2 transmission networks is presented in Table 7. All the pipeline-only scenarios (A) have higher up-front capital costs when compared to the total costs of the corresponding recompression based scenarios (B). The higher up-front costs of scenarios (A) are due to the larger pipeline diameters employed. Considering the overall CO2 network costs (Table 7 and Fig.5), the results suggest that systems relying on recompression offer more attractive techno-economics than those without. However, re-compression based networks will have increased operation and maintenance (O&M) costs in comparison to pipeline-only systems: these are not accounted Fig.5. Summary of CO2 costs for all base scenarios. for here, and more-detailed work is required to determine the optimum combination of compressor requirements and related O&M costs. 232 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.6. Marginal tier costs, defined as the incremental costs divided by the increase in quantity of carbon dioxide transported, for all scenarios. Fig.7. Range of transmission costs identified across all cases considered. For each set of scenarios, the cost of building a pipeline transmission network increases with the addition of smaller emitters. This trend is reflected in transportation costs, because the extra pipe length required increases more quickly than the additional CO2 provided by each tier. Adding all Tier-1 and 2 sources to the network would produce an increase in per tonne transportation capital costs of approximately 15% if no re-compression is used and 22% in a system that relies on re-compression. to a network with compression is 44% higher than that for the Tier-0 sources. The comparable marginal cost increase for a network without compression is 21%. Incorporating increasingly small sources demonstrates poorer returns, as further adding Tier-2 sources to networks with and without compression provides further marginal cost increases of 94% and 109% respectively. The value of adding such smaller sources to a network must be questioned, particularly if they are remote from natural clusters of carbon dioxide sources. Analysis Sensitivity analysis Marginal tier costs The marginal costs for each scenario have been calculated to determine the extra investment required for the addition of emissions of lower tiers by dividing the incremental capital costs of each scenario by the incremental increase in emissions. The results are shown in Fig.6. Marginal costs increase with the addition of emissions from the lower tiers. The marginal cost of adding Tier-1 sources For this study, sensitivities have been calculated to determine the impact of different parameters on CO2 transportations costs. Results of each sensitivity analysis have been compared with a base case. Change in pipeline network length This sensitivity analysis has been performed to determine the impact of 15% and 20% increases in pipeline length on CO2 transportation cost, perhaps due to the need to avoid 4th Quarter, 2010 233 Pipeline only based scenarios Basis: Power station CO2 capture Eff = 90% , Industrial CO2 capture Eff = 60% Capital Cost CO2 Transportation Cost Captured Emissions Pipeline Compression Total Pipeline Compression Total Mtpa M£ M£ M£ £ / Tonne £ / Tonne £ / Tonne S1A 156 952 0 952 6.12 0.00 6.12 S2A 213 1,375 0 1,375 6.46 0.00 6.46 S3A 228 1,608 0 1,608 7.05 0.00 7.05 Scenario Pipeline + compression based scenarios Basis: Power station CO2 capture Eff = 90% , Industrial CO2 capture Eff = 60% CO2 Transportation Cost Pipeline Compression Total Pipeline Compression Total Mtpa M£ M£ M£ £ / Tonne £ / Tonne £ / Tonne S1B 156 822 38 860 5.28 0.24 5.53 S2B 213 1,270 44 1,314 5.97 0.21 6.18 S3B Sa no m t f ple or c di op st y rib ut io n Capital Cost Captured Emissions 228 1,490 56 1,546 6.53 0.24 6.77 Scenario Table 7. Summary of CO2 transmission network costs for pipeline and compression-based scenarios. geographical obstacles not accounted for in the relatively high-level study. For the base-case scenario, the pipeline route was evaluated using a GIS database. It was assumed that pipeline diameters, CO2 emissions, and pipeline unit construction costs remained constant in all cases. However, in scenarios involving re-compression, compression powers were calculated again. Further hydraulic simulations determined the operating pressures of the transmission network and pressures required at sources. For a 15% increase in pipeline length, pressure at sources increased by an average of 2% as compared to the base-case pressures. Similarly for a 20% increase in pipeline length, the pressure at sources increased by an average of 3%. In all cases, the calculated pressures remained below the maximum allowable pipeline operating pressures as well as adequately above the CO2 critical pressure. Costs were then re-evaluated for the increased in pipeline lengths, unsurprisingly showing an increase in transmission costs in all cases (Fig.7). Change in CO2 capture plant efficiency This sensitivity analysis has been performed to determine the changes in CO2 transmission network resulting from changes in the CO2 removal efficiencies of the capture plants installed at each point source. The base-case assumption in this work is that the CO2 removal efficiency for power stations is 90% and 60% for the industrial installations. Any change in capture efficiencies will increase or decrease the volume of carbon dioxide flowing though the network, and hence the design and infrastructure requirements. Two perturbations of the base-case were examined, changing the capture efficiency at the power station sites to 85% and 95%. It was assumed that CO2 capture efficiency of the industrial units remained constant at the base-case value. Using the new flow volumes, new hydraulic simulations were carried out, assuming that pipeline lengths and route corridors will remain unchanged. Only changes in pipeline diameters and compression requirements have been re-calculated. In all cases the network pipeline diameters increased or decreased in accordance with the change in carbon dioxide flow volumes. For the compression-based cases, the compression energy requirements increased for the 95% capture case and decreased for 85% capture case. However for the 95% capture scenario, CO2 transportation cost decreased compared to the base-case whereas for 85% capture scenario, the unit cost increased. The decrease in CO2 transportation cost for 95% case scenarios is countered by increased capital cost due to using larger-diameter pipelines. Similarly for 85% scenario, pipeline diameters have been reduced due to a reduction in emissions volume. The change in CO2 transportation costs in all cases is comparatively small, at around 2%. Increase in pipeline construction costs This sensitivity analysis investigates the impact of increased pipeline unit construction costs, representing the effect 234 The Journal of Pipeline Engineering of increased steel prices, difficult construction terrain, or other project-management-related factors. As part of this sensitivity analysis, a sub-analysis has been conducted to check the impact of increased pipeline construction costs on the previously-calculated sensitivities to CO2 capture efficiencies and increased pipeline lengths. The results of the combined sensitivity analysis shown in Fig.7 indicate the range of CO2 transportation costs across all the cases considered. In all cases the transportation costs of compression-based scenarios are less than for non-compression scenarios. The average costs of the three scenarios – shown as small, light-coloured squares in Fig.7 – range from 8.24 £/tonne for scenario S1 to 9.2 £/tone for scenario S3; the cost variability within each scenario is rather larger, approximately 7-8 £/tonne. Conclusions These general trends are hardly influenced by changes in the major underlying assumptions and design choices, although of course the absolute value of the CO2 transportation cost does vary substantially.   References 1. GASTEC at CRE Ltd (GaC), 2006. Feasibility study on the transmission of CO2. GaC Report 3484, October [Available at www.gastecuk.com/case-studies-detail.php?id=3 ]. 2. G.Pickup, 2009. CASSEM Overview, November, www.geos. ed.ac.uk/ccs/UKCCSC/Pickup.pdf [Consulted May 2010]. 3. UK DBERR. Development of a CO2 transport and storage network in the North Sea, [Available at www.nsbtf.org/documents/file42476.pdf ]. 4. Scottish Centre for Carbon Storage & Scottish Government, 2009. Opportunities for CO2 storage around Scotland. [www. geos.ed.ac.uk/sccs/regional-study/CO2-JointStudy-Full.pdf]. 5. Yorkshire Forward, 2008. A carbon capture and storage network for Yorkshire and Humber. [Available at: www.yorkshireforward.com/sites/default/files/ documents/Yorkshire%20%20Humber%20Carbon%20 Capture%20%20Storage%20Network.pdf]. 6. J.Kjärstad and F.Johnsson, 2008. Ramp-up of large-scale CCS infrastructure in Europe. Int. J. of Greenhouse Gas Control, 2, 4, pp417-438, October. 7.E.S.Menon, 2005. Gas pipelines hydraulics. CRC Press, May. 8. Ramgen Compressors, 2010. Ramgen’s low-cost, highefficiency CO2 compressor technology. www.ramgen.com/ apps_comp_unique.html [Consulted May 2010]. 9. D.Simbeck and E.Chang, 2002. Hydrogen supply cost estimate for hydrogen pathways scoping analysis. US National Renewable Energy Laboratory (NREL) Report/SR-540-32525, November. 10. US Department of Energy, 2007. Conceptual engineering/ socioeconomic impact study – Alaska spur pipeline : Appendix 3-5: compressor cost estimate. Report on DOE-NTL Contract No DE-AM26-05NT42653, January. [Available at www. jpo.doi.gov/SPCO/DOE%20Spurline%20Documents/ Appendix%203-5%20Compressor%20Cost%20Estimate.pdf] 11. B.Metz, O.Davidson, H.de Coninck, M.Loos, and L.Meyer, (Eds), 2005. IPCC special report on carbon dioxide capture and storage, IPCC. 12. Oil and Gas Journal (various issues, 2005-2008). PennWell Petroleum Group. Sa no m t f ple or c di op st y rib ut io n Approximately 50% of UK industrial and energy CO2 emissions are produced by emitters that generate more than 500,000Mt/a of CO2. Providing a CO2 pipeline transport network for each of these sources has the potential, assuming they are also retro-fitted with carbon capture equipment, to facilitate a major reduction in UK CO2 emissions. without re-compression. The CO2 transport cost per tonne are overall smaller, as the expense of the compression stations is outweighed by the reduced cost of smaller-diameter pipes. Our simplified estimates take no account of operations and maintenance costs, meaning this difference is certainly within the range of uncertainty for the results. More-detailed work is required to determine the optimum combination of compressor requirements and related O&M costs. The conceptual design and techno-economics of a phased approach to rolling out such a network have been investigated, subject to a number of simplifying assumptions. If a �backbone’ network connecting the very largest sources is constructed first, smaller sources could be later added to the network with a relatively small impact on the transportation cost per tonne of carbon dioxide. Typically, adding all Tier-1 and-2 sources to the network would produce an increase in per unit transportation costs of approximately 15% if no re-compression is used and 22% in a system that relies on re-compression. Considering the marginal cost of making these additions, however, tells a different story, as increasingly large expense is required to add sources of rapidly decreasing size. The marginal per tonne cost of adding Tier-1 sources to a network with compression is 44% higher than that for the Tier-0 sources. The comparable marginal cost increase for a network without compression is 21%. Incorporating increasingly small sources demonstrates diminishing returns, as further adding Tier-2 sources to networks with and without compression provides further marginal per tonne cost increases of 94% and 109% respectively. The value of adding such smaller sources to a CCS network must be questioned, particularly if they are remote from natural clusters of carbon dioxide sources. The relatively simple cost analysis carried out for this work suggests that a system relying on re-compression perhaps offers a 10% capital cost (about £100m) advantage over one 4th Quarter, 2010 235 Transporting anthropogenic CO2 in contrast to pipelines supporting early EOR by Dr Brian N Leis*, Dr James H Saunders, Ted B Clark, and Dr Xian-Kui Zhu Energy Systems and Carbon Management, Battelle Columbus Laboratory, Columbus, OH, USA T Sa no m t f ple or c di op st y rib ut io n HIS PAPER EXAMINES aspects in quantifying arrest-toughness requirements to control running fracture in CO2 pipelines. Four key risk and safety discriminators were used to contrast transport of nearly pure CO2 for EOR to that of CO2 for CCS applications, including: the retrofitting of early EOR pipelines to provide for fracture control; differing impurities for CCS service that can increase required arrest toughness as compared to EOR applications; differing transported volume and routeing that lead to increased risk exposure for CCS pipelines; and technological uncertainties in assessing fracture-control requirements that develop due to impurity effects. Against this background key elements such as the equation of state and critical assumptions are evaluated as the basis for establishing practical direction for fracture control. Finally, the historic design space for many CO2 pipelines supporting EOR is contrasted to that for pipelines in CCS service. Whereas some widely recognized reports indicate that technology to design CO2 pipelines is mature, some significant gaps were identified for CCS applications. The results indicated that arrest toughness is a very strong function of the minimum CO2 level, with an order of magnitude swing in the arrest toughness required for a 10% swing in minimum CO2 content as compared to pure CO2, with subtle differences in the constituents present being a major driver. On-line monitoring of injected streams was suggested to help manage the related risk. Finally, the often-used Battelle two-curve model adapted to CO2 applications – while validated in regard to near-pure CO2 applications and cases involving rich (dense-phase) natural gas – remains unvalidated in application to typical CCS product streams. Such was also the case for many supporting elements like the equation of state, with an expanded empirical database being key to ensuring viable fracture-arrest predictions. T HE LITERATURE ON schemes to capture and store anthropogenic carbon dioxide (CO2) reflects the increase in concern for the effects of greenhouse gas (GHG) and the need for GHG management. The extent of concern is evident when �GHG + greenhouse gases’ is entered into a web browser, which recently pointed to 15,300,000 hits. Tracking the early history of GHG management leads to sites associated with the intergovernmental panel on climate change (IPCC, whose formation traces to the late 1980s), and other agencies such as the International Energy Agency (IEA), and its focus on GHG (IEAGHG). Early schemes to capture and store anthropogenic CO2 were based on its injection into the earth at sites local to its generation, which precludes the need for transport, and potentially underlies the commonly used acronym CCS 1 (carbon capture and storage). Over time it became evident that, in many cases, local CO2 storage sites might not provide adequate long-term storage integrity, resulting in CO2 seeping back to the surface. This could have consequences such as litigation, and/or other issues, like verification monitoring or economic drivers, which are beyond the present scope. As a result, transport became a necessary aspect of CCS, such that moving anthropogenic CO2 to sites or schemes better suited to its retention or mitigation is now a consideration in this process. The need for transport in conjunction with CCS implies that CCST might be the appropriate acronym, in lieu of CCS. However, searching CCST (as �CCST + carbon capture’) does not quickly lead to transport, but rather points to terms like CCS technology or training, possibly implying that transport (aside from economic considerations) does not pose concerns in parallel to the other elements of CCS. *Author’s contact details tel: +1 614 424 4421 email: [email protected] 1. With CCS defined in reference to managing GHG implies that re-injecting naturally sourced CO2 in support of EOR is not CCS – but simply a commercial operation that returns the CO2 below ground. In the same vein, EOR supported by CO2 from GHG is CCS, as it helps manage GHG emissions. Are you up to speed? 0 0 0 2 0 1 0 Training courses – 2011 TRAINING FEB 2011 January 25–28 Subsea Production Systems Engineering (Aberdeen) January 31– February 4 Subsea Pipeline Engineering Course (Amsterdam) January 31– February 4 In-line Inspection of Pipelines (Amsterdam) February 1– February 3 Defect Assessment in Pipelines (Amsterdam) February 14–15 Defect Assessment in Pipelines (Houston) February 14–15 DOT Pipeline Safety Regulations – Overview and Guidelines for Compliance (Houston) February 14–15 Pigging & In–line Inspection (Houston) February 14–15 Pipeline Repair Methods / In–Service Welding (Houston) February 14–15 Introduction to Excavation Inspection & Applied NDE for Pipeline Integrity Assessment (Houston) February 14–15 Performing Pipeline Rehabilitation (Houston) February 14–15 Stress Corrosion Cracking in Pipelines (Houston) February 14–15 Advanced Pipeline Risk Management (Houston) March 17–18 Microbiological Corrosion in Pipelines (Houston) March 21–25 Onshore Pipeline Engineering (Houston) March 30–31 Unpiggable pipeline solutions forum (Houston) April 25–29 Deepwater Riser Engineering Course (Houston) April 25–29 Subsea Pipeline Engineering Course (Houston) Sa no m t f ple or c di op st y rib ut io n JAN 2011 2011 MAR 2011 APR 2011 Working with a faculty of some 38 leading industry experts, Clarion and Tiratsoo Technical are privileged to provide some of the best available industry based technical training courses for those working in the oil and gas pipeline industry, both onshore and offshore. Complete syllabus and registration details for each course are available at: www.clarion.org 4th Quarter, 2010 237 Examination of some major governmental reports tends to underscore the view that the transport of anthropogenic CO2 regardless of its end use, is straightforward. For example, in 2005 the IPCC [1] stated that “many analysts consider CO2 pipeline technology to be mature.” Likewise, a US Congressional report in 2007[2] stated “pipeline transport of CO2 operates as a mature market technology,” a view that remained unchanged when the report was updated a year later to address jurisdictional issues. A quick check of the facts indicates that both statements are accurate in reference to transporting naturally occurring (relatively pure) CO2, in volumes needed to support enhanced oil recovery (EOR), as dome-sourced CO2 has been transported for decades. This opens to the question: why would this or other papers be written concerning CO2 transport? Sa no m t f ple or c di op st y rib ut io n This question is simply answered in terms of three important safety and risk discriminators that emerge if pipelines transporting anthropogenic CO2 service are contrasted to those in service moving relatively pure CO2 in support of EOR. These discriminators include: • the retrofitting of some early CO2 pipelines to provide for fracture control; • the role and significance of trace impurities; and • the volume transported that drives the use of largerdiameter pipelines, which tend to run longer distances with some traversing high-consequence areas. Fig.1. Photo after a running ductile fracture along a full-scale test section (circa 1970s). While not recognized in the IPCC and US Congressional reporting circa 2005 to 2007, these discriminators are significant, as elaborated later in this paper. After demonstrating the significance of the three safety and risk discriminators, this paper presents the technological background to ensure pipeline fracture control, considering important aspects like the equation of state (EoS) and the assumptions embedded in the analyses of fracture-arrest requirements to offset fracture concerns. The role of impurities is illustrated in regard to arrest requirements, and select design scenarios are considered as the basis for evaluating the practical implications of both fracture control and hydraulics in the light of various platforms available to quantify required arrest toughness. This discussion considers a range of fluid properties determined relative to past and proposed anthropogenic service conditions, along with are a range of potential pipeline designs. Finally, by reference to the results of such analysis, the historic design space for many CO2 pipelines is contrasted to the design space for pipelines in anthropogenic service. The paper closes with some important conclusions regarding safety and risk. What is running fracture, when is it an issue, and why? While consideration of and concern for running fracture are second nature to specialists in this technology, many in Fig.2. Retrofit fracture arrestors (courtesy of Clock Spring): (a top) successful fracture arrest on a CO2 pipeline; (b - bottom) installation of one retrofit arrestor scheme. 238 The Journal of Pipeline Engineering the design community seem unaware of the details and/or the safety implications, while others seem unaware that the decompression of a supercritical CO2 pipeline can involve multi-phase response that cannot be represented by the broadly available analysis for single-phase gas behaviour. Such scenarios are known first-hand to the authors within the last two years in the context of CO2 pipeline design, so such considerations are neither hypothetical nor are they relegated to the distant past. In addition to considerations such as loss of service and/or the cost to replace potentially significant lengths of pipeline, the consequences of running fracture for some transported products and pipeline locations require the certainty that, if initiated, such fractures would be quickly arrested. Considering the cost of lost service, product make-up for uninterrupted delivery contracts, and pipeline replacement, the lost value can easily amount to millions of dollars per mile through which the fracture propagates, depending on the location, product, and length of the propagation. However, far greater losses can accrue due to litigation consequent to such a failure. It follows that there is a need for rational schemes to assess and quantify fracture propagation and arrest in any scenario where a running ductile fracture is plausible. Models to quantify fracture propagation and arrest will be presented following consideration of the key differences between pipelines supporting EOR versus those involved in CCS service. Sa no m t f ple or c di op st y rib ut io n Fracture propagation occurs following the unlikely event of fracture initiation that leads to a rupture, because the stored energy in certain transported fluids, such as supercritical CO2, is sufficient in a high-pressure transmission pipeline to sustain the unstable axial extension of that rupture. Where the decompression front that develops as expansion waves propagate back into the pipeline runs axially at a speed greater than the fracture speed, the pipeline depressurizes faster than the fracture propagates, leading to fracture arrest when the pressure decays to a level that no longer will support axial extension. In contrast, so long as the crack-tip sustains pressure above that level, unstable axial propagation continues. Figure 1 (from Battelle’s archives) shows the outcome of propagating fracture following a fullscale test to quantify this behaviour, which is illustrative of ductile propagation that has caused failure on pipelines while in revenue-service. of the propagating crack. This opening is evident in Fig.1, where the upper quadrants of the pipe have opened. The size of the flaps reflects the energy associated with the passage of the crack and the extent of the energy stored prior to its passing. Flap formation and the extent of the opening develop longitudinal and circumferential stresses ahead of the crack. These stresses cause thinning of the wall thickness and induce significant ovality in the pipe’s cross-section. Figure 1 shows the effects of the longitudinal yielding ahead of propagation and adjacent to the crack, which is evident in the wavy response apparent on either side of the crack over the upper half of the pipeline. Some fracture-arrestor concepts rely on constraint of flap formation, on the presumption that the stresses due to the flap inertia and the related dynamics contribute to the crack’s advance. Other running ductile fracture-arrestor concepts act to reduce the wall stress, or dissipate the fracture energy, or provide enhanced fracture resistance. In addition, arrestors can act to �ring-off’ the cracking, although violent ring-off should be avoided due to the chance of fracture re-initiation. Factors that can contribute to arrest include: • dissipation of energy through the inherent toughness of the linepipe steel and its strain-hardening characteristics; • reduction in local stress, due to increased wall thickness at a road crossing or comparable condition; and • the effects of external factors that act to restrain crack advance or lower the net load carried by the pipecwall, such as due to a fracture arrestor. Early in-service failures involving propagating or running fractures occurred via �brittle’ fracture, which showed limited dissipation in terms of either deformation or crack-tip response. As changes to the pipe steels, coupled with appropriate steel specifications, managed concern for running brittle fracture, it became apparent that a running �ductile’ fracture was also possible. In contrast to the brittle scenario, a running ductile fracture exhibits significant plastic deformation and involves cracking mechanisms associated with locally ductile stretching, leading to void nucleation, growth, and coalescence. As fractures propagate axially along a pipe, the fracture tends to open in the wake of the crack, creating what have been termed �flaps’. For a running brittle fracture, the extent of flap-opening can be very limited, with some distribution pipe materials like polyethylene remaining tight in the wake of the crack due to residual stresses induced in pipe manufacture. In contrast, with a running ductile fracture the flaps open and can fully flatten the pipe in the wake Early pipelines supporting EOR in contrast to CCS applications This section elaborates on: • the need to retrofit some of the early CO2 pipelines to provide for fracture control; • the role and significance of impurities (including trace levels in some cases); and • the volume transported that motivates use of larger-diameter pipelines running longer distances and traversing high-consequence areas to identify differences in the safety and risk aspects for domesourced pipelines that support EOR versus those designed for CCS applications. 2. Some other practical concerns are provided for in design, such as hydraulics, compression/pumping, repair, start-up/shutdown, in addition to fracture control. Some of these impact efficiency and cost-drivers, while others impact safety and risk, with the focus here being fracture control because of its safety and risk implications. 4th Quarter, 2010 239 The focus here is aspects in which the design basis for such lines supporting EOR will underestimate what must be specified to ensure safe design in CCS applications, underscoring the fallacy of CO2 pipeline design as a mature technology. The need to retrofit fracture arrestors on CO2 pipelines Role of impurities The second key discriminator is the presence of impurities and their impact on the fluid’s properties, which becomes clear in contrasting results for near-pure naturally occurring �dome’ CO2 to that for some EOR service, and the flue gas anthropogenic CO2 mix emitted in fossil power generation4,5. While knowledge of and concern for this aspect also traces to the 1980s [17], it too became topical circa 2005 [14], and since has been emphasized in the broadly available literature [18-20], particularly in regard to CO2 compression and pipeline transport. Sa no m t f ple or c di op st y rib ut io n The view that pipeline transport of CO2 is a mature technology can be considered flawed unless there is a mature basis to quantify fracture-arrest requirements for such applications2. This view reflects the observation that some early-design CO2 pipelines supporting EOR had to be retrofitted to manage concern for running fracture3 [3, 4], a need that was realized after the first of the CO2 pipelines went into service. Likewise, it reflects the observation that while regulations covering CO2 pipeline design require that fracture arrest be considered since 1988 [7], it is only in 2010 that a recommended practice to address this aspect became available [8]. Finally, this view follows from the observation that the guidance available in Reference 8 is performance-based rather than prescriptive, with the tools needed to implement that guidance are still in development or in are use only by specialists, and generally lack full-scale proof of their utility/applicability, except for pure CO2. Accordingly, the design basis for CO2 pipelines is well short of mature, and use of the design basis of pipelines built to move nearly pure naturally occurring CO2 can be misleading when employed where the transported product is typical anthropogenic CO2. Concern for running fracture has been recognized for decades for any pipeline transporting compressed gases or supercritical fluids, with the latter being the common state used for CO2 transport due to related efficiencies. Yet the threat it poses is not evident in what was termed mature technology. The need to address fracture control also was emphasized in regard to CO2 pipelines in textbooks on pipeline design [12], which were published well before the IPCC and US Congressional reporting. This need has been emphasized by some regulators [13], and included in some pipeline regulations for decades [7], with the need for fracture control becoming topical again, circa 2005 [14, 15] into 2007 [16]. As becomes clear later, confidence in future design cannot be gained from such testing, nor can the design process for the pressure boundary of such pipelines – however mature – be taken as the benchmark for other pipelines, unless all relevant design parameters are comparable. Several resources that trace to the 1980s [9-11] provide perspective for the running-fracture concerns in CO2 pipelines designed then to support EOR. While such resources underlay this retrofitting, such is not evident in the post-2000 IPCC and US Congressional reporting. Reference 9 couples two similar papers that reflect on the unique traits of supercritical CO2, while Reference 10 outlines the threat posed by long-running ductile fracture, as does Reference 11, so the concern for complexity well beyond mature technology was in print and readily accessible for decades. In regard to retrofitting as discussed in Reference 3, a series of full-scale fracture-arrest tests was conducted to assess the utility of wrap-on fracture arrestors for such applications, but not broadly published. Figure 2a shows a view of a successful arrest (in a full-scale test) from that work. It is apparent from this photograph that while the fracture ran up to and below the arrestor, the pressure was reduced and the pipe restrained sufficiently to limit its propagation beyond the arrestor. Figure 2b illustrates the installation of such arrestors, although this is not specific to a CO2 pipeline application. References 9to 11 are notable in regard to the knowledgebase concerning the threat posed running fracture in CO2 pipelines circa the 1980s, with other resources also available then. 3. Pipelines already in EOR service were retrofitted after the concern for running fracture became apparent. Others for which pipe had been ordered, or at various stages beyond that into construction, addressed the concern for running fracture by applying fracture arrestors during construction [5, 6]. It is apparent from the work reported circa 2005 [14] that the saturation pressure changes significantly due to the presence of modest impurity levels, and it can be inferred that the critical temperature will show a comparable dependence. While such outcomes are specific to the EoS that was used (the Peng-Robinson [21] (P-R) EOS was adopted for that work), nevertheless the concern for the role of impurities was clear then, which given the impact of saturation pressure on fracture propagation and arrest must be addressed in CO2 pipeline design. Recent work indicates the P-R EoS can significantly underestimate the saturation pressure [22] in contrast to other EoSs, and the density dependence of fluid on pressure, depending on the initial conditions. In turn this means that fracture control of pipelines in CCS service based on the P-R EoS could be less conservative than required if referenced to the inappropriate mix of CO2 and impurities, and their relative percentages. This becomes an even more acute issue if their design basis was benchmarked to outcomes for nearly pure CO2 (i.e., typical EOR cases). 4. Depending on the fossil generation plant (coal versus gas-fired) and the combustion process (pre versus post-combustion versus oxy-fuel) the impurities present, their absolute or relative levels, and their effects differ significantly – for example, see Sass et al. [14]. 5. Continuing work indicates that nominal impurity levels reported for a given process can vary significantly from what is considered nominal due to fuel and process variations. 240 The Journal of Pipeline Engineering Fig.3.The BTCM. Transport volume, distances, and routeing also be addressed. This particularly true given that some early-use EOR pipelines required retrofitting. Sa no m t f ple or c di op st y rib ut io n The third key discriminator involves the volume of CO2 to be transported, the distance moved, and the locations traversed. While the above two aspects affect differences in the threat posed by running fractures, parameters including transported volume affect pipe size and pressure, which like routeing and distance, impact public exposure and the consequences associated with the unlikely occurrence of fracture initiation. While there is no simple avenue to determine the eventual volume, it has been stated [1] that “pipelines can be expected to play a significant role in the required transportation infrastructure.” This indicates it is likely that many CO2 pipelines will be built for CCS-related service, which will develop in addition to the existing network of more than 30 pipelines supporting EOR6 [23]. Reference 23 designates three types of CO2 pipelines in EOR service circa 2007, with the discriminator being the transported product specification. What was termed Type I covered special, single-use pipelines with case-by-case (rather open) specifications, with Type II being multiple source / user lines and �strict’ specifications (this type was noted as typical of most of the North American network), and Type III represented hybrid lines, with relaxed but �controlled’ CO2 composition. These three scenarios differ somewhat from the compositions moved by pipelines supporting early EOR – say prior to 1985. Of the six or so CO2 pipelines operating circa 1985, about half moved nearly pure dome-sourced CO2, while the others moved CO2 with small amounts of hydrocarbons whose levels depended on the source field and its variability. If history is to serve as a benchmark for current practices, the role of impurities must be considered. If that same history is to be a benchmark for CCS applications – where CCS involves anthropogenic sources given its GHG roots – then differences in impurities and their levels must 6. Of the 30 pipelines reported in EOR service, about one-third transport CO2 whose source is gas plants or other processing facilities that results in CO2 separated by man from a process stream that originated naturally – as such it is not anthropogenic as a consequence of CCS, nor is it naturally occurring in the form it is transported. The sizes of the two CO2 pipeline segments commissioned in the US in support of EOR prior to 1980 involved a diameter of 16in (406mm) or less, with the pipe wall made of Grade X65 (448MPa) or lower. The sizes and grades of the CO2 pipelines built in the interval thereafter, but before CCS transport became a major consideration, had diameters that ranged up to 30in, which were typically built in Grade X70 or below. Pipelines supporting EOR tended to run through quite remote areas, with source sites selected, in general, as close as possible to the reservoir being worked to minimize the length of the pipeline. For this reason, well over half of these pipelines have a length less than 100miles (160km), with almost all having lengths less than twice that distance. In general, the diameter of such lines is also small in comparison to natural-gas transmission pipelines, with well over half having a diameter of 12in (305mm) or less, with almost all having diameters less than twice that size. In contrast, to the authors’ knowledge, the diameter of lines currently now built in the US for CCS service run up to 24in7 (610mm), and are made in Grades up to X80 (551MPa). Trunklines are under consideration for anthropogenic CO2 that will run from the north to the south of the US, leading to distances the order of five to ten times that just cited. Currently constructed pipelines for CCS service, such as the Green Pipeline, run through high-consequence areas (HCAs) that pose a public concern given the asphyxiative properties of gaseous CO2, and its terrain-tracking transport as a dense (heavier than air) vapour for onshore pipelines, but are much less an issue in an offshore context [25]. Its properties as a supercritical fluid lead to unique concerns, which increase nonlinearly with diameter given the 7. As indicated above there are some quite large-diameter segments supporting EOR, notably the Cortez pipeline [5]. 4th Quarter, 2010 241 illustrates this graphical scheme circa the 1970s. The curve labelled �fracture’ reflects the relationship between fracture speed and pressure as a function of toughness, which was plotted by trial and error until the fracture speed for a given toughness was tangent to the curve labelled �gas’ that reflects speed of decompression in the wake of the expansion waves propagating back into the pipeline. Arrest is ensured by use pipe steel that is specified with that or greater toughness. As becomes is evident later, arrest occurs rapidly once the stress at the crack-tip drops below its critical value. Each of the above three considerations points to the need to address running fracture in regard to CO2 pipelines in CCS-related service. Likewise, it was evident that differences between pipelines designed to support EOR versus transport mixtures of CO2 and the trace impurities common to CCS service that precludes use of existing pipelines as design benchmarks. On this basis, the next section considers approaches to quantify fracture propagation and arrest in CCS applications. As noted above, such considerations are not new – they have been with us since the 1970s for natural gas transmission [26], the 1980s in reference to dome-sourced near-pure CO2 pipelines [4-6], and more recently in regard to CO2 pipelines in CCS service [15, 18-20]. Consequently, what follows is a brief introduction and review, which suffices as background to assess the viability of the assumptions and the viability of the model and its predictions. By the late 1970s analytic approaches [27] and semi-analytical schemes [28, 29] appeared, as did empiricism in the form of curve-fits to full-scale test data [30]. The mid-1980s saw continued attempts to analytically capture this phenomenon [31-33], including a comprehensive energy-balance formulation, and a large-scale numerical formulation that made use of crack-tip opening angle (CTOA) as its measure of cracking resistance. After significant work to refine the metric for fracture resistance, the formulation solidified [34], but its blind application to predict arrest for the Alliance Pipeline full-scale tests led to values of CTOA the order of 25º for arrest [35], well beyond the moderate toughness levels actually required. This gave rise to redefinition of CTOA and some reformulation of the model, which resulted in a more correct outcome the order of 11-12º [36]. Subsequently, definition of CTOA was further updated [37], and then revised again in the late 1990s along with additional �tweaks’ [32], which contributed to its improved case-specific predictability. While this work continues, primarily in Italy [34-38] and the UK [39], and holds much promise, day to day arrest-toughness predictions still make use of the BTCM, except as modified to address higher-toughness steels [40, 41]. Subsequent work has addressed other issues, but relies on the same concepts [42, 43]. In this context there are three approaches to make predictions for CO2 applications: Sa no m t f ple or c di op st y rib ut io n functional dependence of running fracture on diameter [26]. As such, the requirements for fracture control of pipelines transporting nearly pure CO2 in the volumes associated with EOR are much less demanding than for the typically largerdiameter systems anticipated for service transporting CO2 containing impurities typically associated with CCS. Tables 2 and 3 in Reference 14 identify possible trace impurities by source type and summarize their potential effects on capture, compression, pipeline transmission, and injection, and so are useful assessing related safety implications. Modelling to quantify fracture propagation and arrest Technology to quantify running fracture spans from raw empiricism into numerical formulations which, while potentially elegant in concept, still embed empirical calibration(s). While the rapid evolution of fracturemechanics’ theory and technology development that began in the 1960s and has continued since provides the foundation, running ductile fracture is a complex phenomenon. Running ductile fracture couples fluid and solid mechanics with fracture mechanics, gas dynamics, and thermodynamics. The coupled nonlinearities of these disciplines involve soilstructure interaction for buried onshore pipelines, and its parallel offshore, where the fact that pipelines can operate at significant depths adds further complexity. Work through the mid 1970s led to what has been referred to as the Battelle Two-Curve Model [25] (BTCM), which reflects the significant efforts of Maxey [26]. This formulation capitalized on the basic fracture concepts, and coupled that with gas dynamics, and thermodynamics to characterize driving force and the inherent fracture resistance. The BTCM is referred to as such because it quantifies the driving force reflecting the gas’ decompression speed versus pressureinduced wall stress response and the fracture speed versus pressure (stress) response through use of two curves, whose iterative solution for computational reasons in the 1970s was done by plotting their speed versus pressure trends. Figure 3 • the BTCM as is [16, 44]; • adaptations of GASDECOM with other elements of the BTCM largely intact [45, 46]; and • use of the concepts that underlie the BTCM with a return to first principles, as needed [22]. As for dense-phase (rich) natural gas, the expansions waves leading to decompression for CO2 propagate through a one or two/multi-phase medium. The gas dynamics’ aspects of the BTCM were formulated under the assumption that the flow is a homogeneous isentropic process. The model was formulated using the Benedict-Webb-Rubin EoS as modified by Starling (BWRS) [47], which was packaged as software that has become known as GASDECOM. Because this formulation was developed for typical natural gas producing fields, its scope included binary interactions for a range of NGLs, as well as for CO2. Use of this technology in applications to quite rich gases decades later showed it to be quite robust, well beyond expectations [41]. The basic GASDECOM model has seen only modest changes since, which focused on the solver and related algorithms early in 2000. 242 The Journal of Pipeline Engineering in a pipeline that suddenly suffers a guillotine break. At the time of the break, the flow is zero at the exit plane and a sonic wave propagates upstream at the local speed of sound. The moving expansion wave converts fluid at rest into fluid in motion. As the exit plane velocity increases from zero, another expansion wave propagates into the fluid, which is now moving at a speed, denoted u, at a slightly reduced pressure and temperature. In reality, this process is continuous with expansion waves propagating upstream, each moving a little slower than the preceding one, with the exit plane velocity continuously increasing until the exit plane velocity equals the expansion wave speed and the flow is considered choked. The exit plane speed can be described by the following [51]: (1) where ρ is the density, a is the local speed of sound, and t is the time after the rupture. The speed of sound is a property of the fluid and can be found in the literature for a wide variety of fluids. For pure CO2, the speed of sound is available from a number of studies, with the results based upon the Span and Wagner [49] EoS, which is widely considered as highly accurate. Sa no m t f ple or c di op st y rib ut io n While adaptation of the scheme developed for methanedominated mixes was found necessary in Battelle’s 1980s work involving nearly pure CO2 [10, 48], others have used GASDECOM for gases dominated by CO2 at least in their initial work [16, 44]. As Battelle continued to experience issues in adapting the source-code for GASDECOM to the CO2-impurity mixes typical of CCS, we have since adopted alternative EoSs, depending on the application, and returned to a first-principles approach. The Span and Wagner EoS [49], which was developed and calibrated specifically for pure CO2 is used for this benchmark scenario. Both the P-R EoS [21] and the GERG EoS [50] were initially considered when impurities effects were involved, however, an adaptation of the GERG EoS is now preferred, with care still taken to assess the practical viability of the outcomes. The speed of decompression is then inferred from gas dynamics referenced to the speed of sound as a function of composition and density in the manner due to Liepmann and Rosko [51]. The fracture velocity is determined as a function of pressure and toughness using variations of historically proven practices, with the required arrest toughness being that needed to slow the speed of propagation to equal or less than that for decompression. Key assumptions and their implications Fundamental to predicting fracture arrest is the assumption that arrest occurs when the wall stress falls below some critical level due to decompression, because the rate of propagation has slowed such that the pressure-induced wall stress decreases. The historic view with the BTCM is that arrest ensues within a diameter or so once toughness suffices to cause tangency between the gas and fracture curves in Fig.3. A second key assumption is that the decompression response can be calculated independent of the fracture response, with these being brought together through an empirically determined function that couples their response, that was termed the backfill coefficient. Yet another key assumption is that the flow during decompression is a one-dimensional, homogeneous, isentropic process that develops in response to expansion waves that propagate back into the pipeline, the implication of which is that it is an isentropic, reversible process, with relative motion of liquid and vapour neglected. Finally, it is assumed that the speed of the expansion waves can be determined relative to the instantaneous local density of the fluid. These assumptions are considered next, more or less in the reverse order they have been cited. Speed of expansion waves and flow response Expansion waves of single-component, single-phase fluids have a long history of study, are generally considered well understood and thus provide a solid basis for understanding the more complex case of expansion-wave propagation in CO2 mixtures with impurities and phase change. Such analysis is formulated in regard to a fluid at high pressure In single-phase fluids, the process is considered isentropic and the density path can be found by tracing an isentropic path on a thermodynamic state diagram from the initial pressure to a final pressure. Under this assumption, and using the necessary data, Equn 1 can be readily integrated and the exit velocity obtained. The focus here is the expansion wave, which propagates at the local speed of sound in a frame moving with the exit velocity. In a frame at rest, the propagation velocity is then a + u (where the direction is positive in the upstream direction, so u < 0, a > 0). With this, the propagation velocity can be tracked and compared to the crack propagation velocity in the manner of the BTCM. For a mixture of CO2 and impurities as dictated by the product stream, the situation is more complex. First, the properties must be determined by an appropriate EoS. As noted above, these range from relatively simple expressions, such as the P-R EoS, to more-complex and realistic expressions that are grounded in thermodynamics and fit with extensive data sets. One of the most accurate sets is based upon the GERG database, as modified by Lemmon [52] of the US National Institute of Standards and Technology. These expressions are widely used for hydrocarbon mixtures, but do not have extensive data for mixtures with high fractions of CO2 and impurities, indicating that there is a need for additional data to support CO2 mixture studies. As the pressure drops in the pipeline, the fluid may undergo a change in phase. Figure 4a illustrates an isentropic expansion that starts at 2160psia (149bar) and 90°F (32°C) for a CO2 mixture specific to one client. Such plots develop for pure 243 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.4. Aspects of the thermodynamic state and expansion waves for CO2 pipelines: (a - top) thermodynamic state; (b - left) traits of process timeline. CO2, through mixtures of CO2 and impurities – the key differences being the locations of the phase boundaries and associated properties. The y-axis in this figure is pressure on a logarithmic scale, while the x-axis is density. The light layered trends are isotherms, while the heavy curved trend that runs from the upper right down to the lower left across the layered isotherms is the isentrope from the initial condition. The region above the critical point at the top of the two-phase dome comprises supercritical fluid, while the liquid phase is to the right side of the figure, and the vapour phase is to the left side, with the region within the dome being a mix of these two subcritical phases. For a supercritical CO2 pipeline, there is a spectrum of initial temperatures and pressures ranging from the inlet to the exit of the pipeline, so there can be a spectrum of responses depending on position along the pipeline considered in regard to the guillotine rupture, and its local pressure and temperature. For the particular initial conditions considered in Fig.4a, the CO2 starts as a supercritical fluid, passes into the liquid region, and then encounters the liquid-vapour saturation boundary, and remains therein through the remainder of the process. 244 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.5. Speed of sound of CO2 and CO2 mixtures referenced to three EoSs. In reference to Fig.4b, the x-axis is density as it is in Fig.4a, whereas the y-axis is speed in regard to each of the three curves shown in the figure. Each of these curves represent a process that begins at 70°F (21°C) and 2200psia (152bar), which in this figure originates from the upper right corner of the figure, as the local fluid density corresponds to that prior to rupture. The upper curve is the speed of sound, relative to the moving liquid. As time passes, the local density decreases as the fluid increases in velocity. As evident, the sonic velocity is strongly nonlinear with density initially, with this dependence diminishing with time. The sonic velocity is a property of the liquid and vapour determined from the equation of state, with multi-phase sonic speeds determined from the method outlined in Wallis [53]. The lower curve represents the outflow fluid velocity from Fig.6. Pressure versus time for a location 12ins (~30cm) upstream of the exit plane. Equn 1, starting from rest (highest density) and increasing to choked flow. The middle curve is the wave speed in a stationary reference frame and is simply the vector addition of the upper and lower curves. The wave speed starts at the sonic speed (highest density) and then decreases as the fluid speed increases, finally reaching zero, which is the choked flow condition. Each of Figs 4a and 4b are unique to the initial conditions, to the fluid mix in terms of the fraction of CO2, and to the mix of impurities and their relative levels. As becomes apparent next, they are also dependent on the EoS embedded in the process description, so a compendium of such behaviour as a function of circumstances along a pipeline, much less for differing product streams is implausible. 4th Quarter, 2010 245 Fig.7. Effect of impurities on decompression wave velocity. flow is a strong but common assumption; among others it assumes equality of liquid and vapour velocities, which at best is true only for low bubble volume fractions. For the results in Fig.5, the speed of sound of the liquid and vapour phases has been estimated in the context of REFPROP [52], while the speed of sound in the multi-phase region has been estimated using an expression from Wallis [53]. Sa no m t f ple or c di op st y rib ut io n Figure 5 further illustrates trends in the speed of sound for CO2 as a function of density and fluid composition, to better quantify the range of responses that can occur depending on pressure and composition, and also the manner these are related via the EoS that underlies the analysis. While speed was the dependent parameter in regard to density in Fig.4b, in Fig.5 the sonic speed is a function of pressure that is shown on the x-axis. It is clear from this figure that for any trend considered, the sonic speed is high when the transported fluid exists as a liquid, and then drops sharply as the isentrope crosses the liquid-vapour phase boundary, reaching a much lower value as a gas-liquid mixture, and even lower as a gas. This response is central to controlling running fracture, because this behaviour underlies changes that lead to decompression local to the rupture plane, such that arrest can occur, with sufficient toughness supplied such that the crack propagation speed is equal or less than the decompression speed. Consequently, understanding the speed of sound as a function of conditions is important to the designer, and especially understanding the boundary of the phase change regions for CO2 mixtures with impurities. Figure 5 provides insight into the role of the EoS that underlies the locations of the phase boundaries and how these move as a function of impurities, which taken together dictate the breakpoints in the trends shown in this figure. The speed of sound is shown for three fluids: pure CO2, a 95% CO2 and impurities mixture, and a 98% CO2 and impurities mixture. The mixes in both cases represent client estimates of what might be transported in their pipeline, which might be considered proprietary and so are not elaborated. Results are presented for these mixtures characterized by the GERG EoS and the P-R EoS, both of which were introduced previously along with the Span and Wagner EoS for pure CO2. These results reflect the assumption of homogenous one-dimensional flow in the two-phase region: that is, the properties are a weighted average based upon the liquid and vapour fractions. Homogeneous The expansion process for the mixtures in Fig.5 has been assumed isentropic – which is reasonable for the singlephase part of the expansion but is not generally valid for the multi-phase part. Essentially then it is the transition region between the liquid and vapour speed of sound that is open to question. However, it should be noted that the phase transition itself may not necessarily occur at the phase boundary, due to nucleation delays. Realizing that the underlying formulation can be used to quantify pressure response within the pipeline, it is instructive to consider that response and the rate of that process in contrast to crack speeds approaching arrest. Consider Fig.6 in this regard, which shows the pressure response 12ins (30cm) upstream of the exit (rupture) plane as a function of time. This result is generated for pure CO2, for isentropic expansion beginning at 2200psia (152bar) and 70°F (21°C). It is apparent from this figure that there is a strong drop in pressure until the liquid-vapour phase boundary is reached, and that the response time is the order of milliseconds. While this outcome reflects the same assumptions as noted above, it is apparent that when such events occur they do so at very high speeds. Because the timeframe is short the occurrence of subtleties arising due to three-dimensional versus one-dimensional response are indicated to have only modest influence on where along the pipeline – relative to its diameter – the decompression velocity is found to match the propagation speed, which in the BTCM framework defines arrest. As such, the minimum 246 The Journal of Pipeline Engineering Fig.8. Isentropic processes plausible in design vary significantly. practical applications involving rich (dense-phase) gases [41]. Furthermore, the lack of equality between liquid and vapour velocities generally means that these flows are not isentropic, since the relative phase motion results in interphase drag, heat transfer, and mixing. In addition, viscous and inertial effects will become important as the flow speed increases, which are not included here. Sa no m t f ple or c di op st y rib ut io n arrest toughness that results from this analysis is a reasonable indicator of the toughness needed for fracture control. Figure 7 develops in the context of analyses for the conditions that underlie Fig.5, presenting a plot of pressure shown on the y-axis and the decompression wave speed on the x-axis in regard to the pure CO2 and the 95% CO2 plus impurities cases. Comparing these trends shows a practically quite significant shift in response due to the presence of impurities. As can be seen, this shift for the impurities mix is associated with a higher pressure – which occurs due to the increased saturation pressure, and the shifting phase boundary in the pressure-density space. At pressures above the phase boundary, the decompression speed is slower for CO2 with impurities compared to pure CO2 at the same pressure. At pressures below the phase boundary, these trends slow as expansion occurs within the two-phase boundary and the temperature drops. Because expansions waves form continuously and there is feedback in the pressure-volumetime space consistent with the EoS, the decompression velocity gradually decreases, with that decrease generally stronger for the mixtures than for pure CO2. Some results indicate that a plateau develops once the phase boundary is crossed regardless of the EoS used, which is evident (for example) in References 45 and 46. Clearly this is in conflict with the view shown in Fig.7, but as yet the reason for such has not been identified. In this regard the need for data to support the reliable prediction of phase boundaries, equations of state, and the transport properties for these CO2 mixtures, is emphasized, as such data are generally lacking. In addition, the multi-phase flow regime is uncertain. The homogenous assumption has not been verified for these fluids; essentially it assumes that the liquid and vapour phases move at the same velocity and these velocities will depart significantly as the bubble volume increases. However, we note that the homogenous flow assumption has been successfully applied in many Clearly a compendium of pressure-volume outcomes would be useful from a design perspective, but is virtually impossible to generate because of the range of product compositions and initial conditions. That view follows from our experience in working with the product streams anticipated to be carried under conditions akin to common-carrier / commodity-type service, which indicates the range of possible mixes could be very large. It also follows from the observation that the results of analyses as those in Figs 5 and 7 are highly sensitive to both the impurities present and to subtle differences in this mix or the relative percentage of a given constituent. As such, the focus here is on cause-effect aspects, and the assumptions involved, and their implications, in lieu of attempts to trend such outcomes. The last point to make in this context involves multi-phase effects, which are as important to hydraulics and other design aspects (which are not considered herein) as they are to running fractures. Suffice it to note that multi-phase wave propagation has been extensively studied for supercritical water-blowdown events related to a loss of coolant accident scenario in the nuclear research community. While CO2 mixture flows are more complex, the existing related knowledge could provide a strong base for understanding such flow response. Discussion and design implications Prior sections have identified the dependence of the outcomes on the initial conditions used in the analysis (pressure and temperature), the scope of the impurities, and 4th Quarter, 2010 247 the assumptions made. The need for data to assess/establish the viability of the predictive schemes also exists, as illustrated, for example, by the apparent disparity in a plateau forming beyond the phase boundary in other work [45, 46] versus the steady decay in the pressure-velocity response evident in Fig.7. These are considered next in regard to decompression and arrest-toughness prediction, and then illustrated in regard to historic CO2 pipeline designs in contrast to the spectrum of potential EOR-CCS applications. Viability of decompression and arrest toughness predictions The design space for historic EOR versus plausible CCS applications Perhaps the most instructive contrast between the outcome of the design approach that underlies the early CO2 pipelines and has continued since – and is still considered by some to be mature technology [1, 2] – is to compare the required arrest toughness for a range of designs for pipelines transporting CO2 across a range of pipeline capacities and CO2 quality specifications. To simplify this comparison, the outcomes are presented in a normalized format – which avoids making this comparison specific to details such as inlet pressure and temperature. Suffice it to say that parameters typical of supercritical CO2 transport have been considered in regard to inlet pressure and temperature. To manage the scope of this comparison and focus the outcomes, a simple binary CO2 product stream is considered, with the second constituent being one common to both EOR and CCS product streams whose effect on arrest toughness is neither worst-case nor trivial. The CO2 content ranges from pure, which is well above what is cited as the usual Permian Basin EOR minimum CO2 content of 95% [54], down to 90%, which is equally below that minimum. Finally, to avoid the nonlinear influence of diameter on the outcome of this analysis, a constant diameter pipeline is considered. Sa no m t f ple or c di op st y rib ut io n As identified above, a valid concern exists in regard to the EoS used, which in turn impacts the saturation pressure and temperature, and the phase boundaries. It also impacts aspects such as compression / pumping decisions, and hydraulics issues that control the recompression distance for longer CO2 pipeline systems. Finally, such concerns impact predictions of the decompression behaviour of CO2 and CO2 mixtures, and carry through to the viability of the concepts that underlie the BTCM and its reliance on the Charpy V-notch (CVN) energy as the toughness metric. These are considered next in regard to decompression and arrest toughness prediction, and then illustrated in regard to historic CO2 pipeline designs in contrast to the spectrum of potential EOR-CCS applications. diagram it is readily apparent that substantially different behaviour can develop over the length of a given pipeline. This is evident, for example, in Fig.8, which illustrates the thermodynamic state of a 95% CO2 plus impurities mixture for isentropic processes that start at 40°F up to 160°F (4 to 71°C), all at 2160psia (104bar). Depending upon the initial temperature of the fluid, the expansion process can descend on either the liquid or vapour side of the critical point, or possibly through the critical point. Reality for a pipeline is that the pressure changes along its length, which also plays into the need for the designer to consider the full range of conditions that might develop across the full range of possible compositions. On this basis there is need to identify worst-case scenarios for purposes of decompression assessment and arrest-toughness prediction. Developing a viable database to address these concerns with scope adequate to address the range of potential EOR-CCS applications is a bit like a Christmas list where there is no Santa Claus. While work is planned or continuing to address these gaps, the scope is limited, so the outcomes tend to be marginal in contrast to what might be needed given the scope of concerns involved. More critical is the observation that some EoS and/or predictive schemes are comparable under certain circumstances but quite different for others, which complicates defining the empirical basis to discriminate between them. This inconsistent disparity is perhaps the biggest concern, as empirical understanding in regard to: • the EoS (saturation pressure and temperature, and the phase boundaries); • Multi-phase flow (compression/pumping decisions, hydraulics issues, recompression distance, decompression behaviour); and • the viability of the concepts that underlie the BTCM for CO2 applications could require a significant testing matrix. Design implications for CO2 pipelines An important consideration to the pipeline designer is the temperature, pressure, and composition of the CO2 mixture across the range of potential rupture locations. Given the random nature of threats such as third-party damage, fracture initiation leading to possible running fracture could occur most anywhere along the length of some pipelines. By following the isentropic processes on the state The reference pipeline used to normalize the outcomes in this comparison is an EOR-service pipeline that for the sake of simplicity transports pure CO2. To keep the normalization benchmark representative, this benchmark is sized slightly larger than �average’ relative to the outcomes in earlier discussion in the section titled Transport volume, distances, and routeing. On that basis, the benchmark has been chosen as a 20-in (508-mm) diameter pipeline made of Grade X65 (448MPa), whose wall was sized using a design factor of 0.72 for operation at a given inlet pressure and temperature8. To address the apparent shift from the earlier EOR designs through what could emerge if CCS service 8. While this benchmark design has been chosen to represent typical onshore CO2 pipelines, its dimensions and pipe grade are comparable to one of the early-design EOR pipelines that moved dome-sourced CO2, parts of which were retrofitted with fracture arrestors in the 1980s. 248 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.9. Design space for CO2 pipelines in terms of arrest toughness and capacity. Fig.10. Distribution of CVN energies for a major order of large-diameter linepipe. dominates CO2 pipelines, the need to increase pipeline capacity (transported volume) is inferred by considering a range of linepipe grades (diameter is fixed as noted above) that permit relatively higher stresses. The resulting wall thickness for a given grade is determined by design factor, pressure, and diameter, which are constant for all scenarios. Based on earlier discussion, the range of grades in addition to the X65 benchmark includes X70 (482MPa) and X80 (551MPa), which in current use in CCS service. Figure 9 presents the normalized results of this comparison in terms of required arrest toughness shown on the y-axis as a function of pipeline capacity shown on the x-axis. The normalized outcomes are trended as a function the impurity level, which as noted above involves the variation of one constituent in a binary fluid, which has been selected from among the candidate impurities because it is less demanding than most from a fracture-control perspective for usual EOR scenarios, and much less demanding than the constituents in typical CCS-related streams. Because the benchmark pipeline has been used to normalize the outcomes, that result is evident in the lower left corner of this figure at (1,1), which is emphasized by the solid circular symbol. Pipelines that require lower arrest toughness and are relatively smaller than the benchmark exist, as do others that are somewhat larger and require somewhat higher toughness, the scope of which is typical of many pipelines supporting EOR. This group of designs tends to fall within the dashed box that is roughly centred at the benchmark pipeline. As can be seen from the x-axis labels, only a 50% relative increase in capacity has been inferred for this analysis, which is less than some might anticipate. The scale along the y-axis runs up to a value of 10, with the worst-case considered in these analyses regarding capacity and impurities leading to a required arrest toughness that 4th Quarter, 2010 249 approaches ten-times the toughness required for arrest in the benchmark pipeline. Archival records at Battelle located for one of the EOR pipelines that had arrestors installed back in the 1980s indicates toughness levels (hereafter CVN full-size equivalent energy at service temperature) in excess of 60ft-lb (81J) in the sections where arrestors were installed. While this historic design scenario is comparable to the benchmark pipeline used to normalize the results shown in Fig.9, this actual toughness level is slightly larger than the required arrest toughness for the benchmark, which was the order of 40ft-lb (54J). This paper has examined aspects determining arresttoughness requirements to control running fracture in CO2 pipelines, and addressed the validity of the view that the design of these systems deploys mature technology, as has been suggested in some widely distributed reports on CCS as the means to manage GHG.[1, 2] While this might be inferred from decades of experience in transporting CO2 to support EOR, this view is open to question when moving large volumes of anthropogenic CO2 from four perspectives: • some of the early CO2 pipelines were retrofitted to provide for fracture control; • trace impurities in CCS and EOR product streams can significantly increase the arrest toughness required for fracture control; • the volume transported for CCS and its routeing implies risks that can be much greater than for many EOR applications; and • the technology used to quantify fracture control requirements is neither mature nor is the empirical database supporting it complete relative to current expectations. Sa no m t f ple or c di op st y rib ut io n It can be seen from Fig.9 that the presence of impurities drives the required arrest toughness much more so than pipeline capacity. This is evident in regard to the benchmark capacity (x-axis value equal to one), where a ten-fold increase in toughness is needed to deal with impurity levels the order of 10%, as compared to the modest increase in toughness due to capacity. Realizing that the minimum purity is often set at 95%, the decision to establish this level as compared to a lower level helps to offset concern for running fracture. While there is a desire to control the minimum CO2 content (apparently due to miscibility for the benefit of EOR, not safety), there is a clear need to monitor and/or control the product stream in regard to this parameter given its strong impact on required arrest toughness. Summary and conclusions Equally, there is a need to better understand sources of variability in toughness for linepipe, as such variability can diminish the all-heat average (AHA) toughness and opens the door to longer propagation, which undermines specification of toughness as the control for running fracture. The data in Fig.10 help to illustrate this point, where the results show the distribution of toughnesses for a recent large order of transmission linepipe. The toughness coordinates in this figure run from 50 to 400ft-lb, equally 68 to 542J. Whereas the AHA for this large pipe order exceeds 200ft-lb (271J), the lowest toughness measured was 56ft-lb (76J), with a significant fraction of the heats with toughness less than one standard deviation below the mean. A tighter population, or a population skewed to above the mean, would greatly improve fracture control, and could reduce the AHA – that might simplify making specifications where high AHA levels are needed. Central to fracture control is a fracture control plan, which should be developed during the FEED (front-end engineering and design), and thereafter serves as the basis to specify steel for the pipeline. Such planning offsets the eventual need to retrofit arrestors – which opens to costs and maintenance concerns that otherwise can be avoided. Without such planning, where the inherent toughness is found to be inadequate after the line is commissioned and the operational parameters are fixed, requires the use of retrofit arrestors. Their use and placement is motivated by risk assessment and other considerations, just as was done in the 1980s, which is about where this paper began. While not evaluated directly in regard to theses perspectives, some have considered running fracture a greater risk for CO2 pipelines than for hydrocarbon pipelines [55]. This paper has considered the technological background to ensure pipeline fracture control, considering important aspects like the EoS and the assumptions embedded in the analyses of fracture-arrest requirements to offset fracture concerns. That technology was used to illustrate the role of trace impurities after which the select design scenarios have been considered as the basis for evaluating practical implications in regard to both fracture control and related aspects. This discussion has considered a range of fluid properties for naturally occurring and anthropogenic CO2 sources, and the service conditions anticipated for a range of pipeline designs. Finally, the historic design space for many CO2 pipelines supporting EOR has been contrasted to that for pipelines in CCS service. Important conclusions that can be drawn from this work follow: • the minimum CO2 level can cause significant swings in the arrest toughness, such that toughness requirements for pipelines in use for dome-sourced CO2 underestimate that for CCS service by a factor of two – possibly more in contrast to relatively pure CO2; • minor out-of-specification swings below the usual 95% minimum can further increase the required arrest toughness – with a decrease in the minimum below that level of 2.5% increasing the required arrest toughness about two-fold; Held under the Patronage of His Excellency Dr. Abdul Hussain bin Ali Mirza, Minister of Oil & Gas Affairs and Chairman of the National Oil & Gas Authority, Kingdom of Bahrain PLATINUM SPONSOR 15–17 May 2011, Bahrain GULF CONVENTION CENTRE, BAHRAIN Sa no m t f ple or c di op st y rib ut io n ORGANIZERS Join leaders in the international pipeline industry as they converge for the Best Practice in Pipeline Operations and Integrity Management Conference and Exhibition in Bahrain. CONFERENCE EXHIBITION Six technical streams covering a wide range of subjects will run over the two and a half day event (and presented by industry leaders). • Planning, design, construction and materials • Operations and maintenance • Asset integrity management • Inspection and cathodic protection • Repair and rehabilitation • Automation and control • Leak detection Paper abstracts are now being accepted. A comprehensive exhibition will be part of the event, allowing companies from around the world to showcase their products and services. Contact us today to book your space. NETWORKING Throughout the event there will be ample opportunities to network with participants to further your business relationships. Meet with industry leaders from around the world. vent. ndmark e pen ons will o ti a tr is g e R this la u attend o y e r u s e 11 – mak 0 in early 2 www.pipelineconf.com 4th Quarter, 2010 251 • minor differences in the relative proportions of a trace impurity can have a significant influence on the required arrest toughness, with subtle differences in the constituents present also a major driver – highly volatile constituents can complicate the analysis and significantly drive fracture arrest requirements; • on-line monitoring of the transported stream and/ or the injected streams on a trunkline should be considered to help manage risk; • the BTCM adapted to CO2 applications has been validated in regard to near-pure CO2 applications and cases involving rich (dense-phase) natural gas: however, like most other aspects involved in fracture control, like the EoS, it remains unvalidated in applications to typical CCS product streams; and • an expanded empirical database is essential to ensure viable fracture arrest predictions. Sa no m t f ple or c di op st y rib ut io n Acknowledgements Useful discussions with Dr Bruce Sass of Battelle’s Energy Systems and Carbon Management product line are gratefully acknowledged, as is support from Battelle’s Science and Technology fund in modelling the EoS and the expansionwave response. References by regulating saturation arrest pressures. Oil and Gas Journal, pp44-46. 11. A.B.Rothwell, 1988. Fracture control in natural gas and CO2 pipelines. In: Microalloyed HSLA Steels. ASM International, pp95-108. 12. M.Mohitpour et al., 2003. Pipeline design and construction: a practical approach. 2nd edition, ASME Press, (Note: the 3rd edition is now available but is not used here, to emphasize the timeline of this 2nd edition). 13. anon., 2008. Interim guidance on conveying CO2 in pipelines in connection with carbon capture, storage and sequestration projects. UK-HSE. 14. B.Sass, B.Monzyk, S.Ricci, A.Gupta, B.Hindin, and N.Gupta, 2005. Impact of SOx and NOx in flue gas on CO2 separation, compression, and pipeline transmission. Carbon Dioxide Capture for Storage in Deep Geologic Formations, Vol.2, D.C.Thomas and S.M.Benson (Eds.), Elsevier. 15. J.Barrie, K.Brown, P.R.Hatcher, and H.U.Schellhase, 2005. Carbon dioxide pipelines: a preliminary review of design and risks. Proc. 7th Int. Conf. on Greenhouse Gas Control Technologies. 16. A.Cosham and R.Eiber, 2007. Fracture control in carbon dioxide pipelines. J. Pipeline Engineering, 6, 3, pp150–158. 17. C.B.Farris, 183. Unusual design factors for supercritical CO2 pipelines. Energy Prog. 3, 3. 18. P.N.Seevam, J.M.Race, M.J.Downie, and P.Hopkins, 2008. Transporting the next generation of CO2 for carbon, capture and storage: the impact of impurities on supercritical CO2 pipelines. Proc.7th Int. Pipeline Conf., IPC2008-64063, Calgary, October. 19. A.Cosham and R.J.Eiber, 2008. Fracture control in carbon dioxide pipelines: the effect of impurities. Idem, IPC2008-64346. 20. H.Li and J.Yan, 2009. Impacts of equations of state (EOS) and impurities on the volume calculations of CO2 mixtures in the applications of CO2 capture and storage (CCS) processes. J. Applied Energy, Elsevier. 21. D.-Y.Peng and D.B.Robinson, 1976. A new two-constant equation of state. Ind. Eng. Chem., Fundam., 15, 59–64. 22. B.N.Leis, J.H.Saunders, and E.B.Clark, 2010. Issues in quantifying the expansion wave response in CO2 pipelines. 9th Annual Conf. on Carbon Capture and Sequestration, DOE/NETL, Pittsburgh. 23. L.S.Meltzer, 2007. CO2 transport – building on the current framework to meet demands of widely deployed commercial scale CCS systems. 6th Annual Conf. on Carbon Capture and Sequestration, DOE/NETL, Pittsburgh. 24. www.statoil.com/en/ouroperations/explorationprod/ncs/ snoehvit/pages/default.aspx. 25. B.N.Leis and R.J.Eiber, 2010. Fracture control technology for transmission pipelines. PRCI Catalog L51846, 2010: updates Eiber, R. J., Bubenik, T. A. and Maxey, W. A., Fracture control technology for natural gas pipelines, NG-18 Report No. 208, Pipeline Research Council International, Project PR-3-9113, Battelle, 1993. 26. W.A.Maxey, 1974. Fracture, initiation, propagation, and arrest. 5th Symposium on Line Pipe Research, American Gas Association. 27. H.C.van Elst, 1974. Criteria for steady state crack extension in gas pipelines. Int. Conf. on Prospects of Fracture Mechanics, Netherlands, pp299–318, June 24-28. 28. P.A.McGuire, S.G.Sampath, C.Popelar, and M.F.Kanninen, 1978. A theoretical model for crack propagation and arrest in 1. Carbon dioxide capture and storage, IPCC Special Report, 2005. B.Metz, O.Davidson, H.de Coninck, M.Loos, and L.Meyer, (Editors). 2. P.W.Parformak and P.Folger, 2007. Carbon dioxide (CO2) pipelines for carbon sequestration: emerging policy issues. CRS Report for Congress, 2007: Updated in 2008 as A.Vann and P.W.Parfomak, �CRS report for Congress: regulation of carbon dioxide (CO2) sequestration pipelines: jurisdictional issues,’ which notes pipeline safety in a US jurisdictional perspective, but does not otherwise consider the topic. 3. D.I.Marsili and G.R.Stevick, 1990. Reducing the risk of ductile fracture on the Canyon Reef Carriers CO2 pipeline. SPE 65th Annual Technical Conference, New Orleans, LA, Proceedings 311-20, September. 4. J.Watts, 1983. Sheep Mountain CO2 pipeline to boost West Texas production. Pipeline & Gas J. Vol/Issue 210: 6, May 1. 5. W.R.Quarles, 1983. Willbros nears completion on Cortez CO2 trunkline. Pipe Line Industry, Aug. 6. C.Horner, 1985. Choctaw carbon dioxide line laid in Mississippi. Pipeline & Underground Utilities Construction, 40, 4, pp4-6, April. 7. anon. §195.111 Transportation of hazardous liquids by pipeline. US CFR Part 195. 8. anon., 2010. Design and operation of CO2 pipelines. Recommended Practice, DNV-RP-J202, April. 9. G.G.King, 1981. Design of carbon dioxide pipelines. EnergySources Technology Conference and Exhibition, Houston, January: see also G.G.King, Design considerations for carbon dioxide pipelines. Pipe Line Industry, 1981, pp125–132. 10. W.A.Maxey, 1986. Long shear fractures in CO2 lines controlled 252 The Journal of Pipeline Engineering Eng. Frac. Mech., 71, pp1997-2013, 2004. 40. B.N.Leis, R.J.Eiber, L.E.Carlson, and A.Gilroy-Scott, 1998. Relationship between apparent Charpy V-Notch toughness and the corresponding dynamic crack-propagation resistance. Int. Pipeline Conf., ASME, Calgary, pp723-732: see also Leis, B. N., Relationship between apparent Charpy V-Notch toughness and the corresponding dynamic crack-propagation resistance, 1997, Battelle Report to R J. Eiber, Consultant, Inc. Exhibit B-82, Proceeding GH 3-97, National Energy Board of Canada, 1997-1998. 41.R.J.Eiber, B.N.Leis, L.E.Carlson, N.Horner, and A.GilroyScott, 1999. Full-scale tests confirm pipe toughness. Oil & Gas Journal, Nov.8, pp48-54. 42. D.Rudland, D.-J.Shim, H.Xu, D.Rider, P.Mincer, D.Shoemaker, and G.Wilkowski, 2007. First major improvements to the two curve ductile fracture model. DOT-PHMSA No. DTRS5603-T-0007, May. 43. B.N.Leis, and T.P.Forte, 2007. New approach to assess running fracture in transmission pipelines. DOT/PHMSA DTRS5605-T-0003, February. 44. anon., 2010. Material requirements for CO2 line pipe. JFE Steel Corporation / Marubeni-Itochu Steel Inc Presentation, Battelle, February. 45. A.Cosham, 2009. CO2: it's a gas, Jim, but not as we know it. 5th Pipeline Technology Conference, Ostend, Belgium, October. 46. A.Cosham, R.J.Eiber, and E.B.Clark, 2010. GASDECOM: carbon dioxide and other components. 8th Int. Pipeline Conf., IPC2010-31572, October. 47. K.E.Starling, 1973. Fluid thermodynamic properties for light petroleum systems. Gulf Publishing Co., Houston: see also Hopke, S. W. and Lin, C. J., Applications of the BWRS equation to natural gas systems., paper presented at the 75th National AIChE Meeting, Denver, March 1974. 48. W.A.Maxey, 1983. Gas expansion studies. AGA NG-18 Report 133. 49. R.Span, and W.Wagner, 1996. A new equation of state for carbon dioxide covering the fluid region from the triple point temperature to 1100 K at pressures up to 800 MPa. J. Phys. Chem. Ref. Data, 25, 6. 50. O.Kunz, R.Klimeck, W.Wagner, and M.Jaeschke, 2007. The GERG-2004 wide-range equation of state for natural gases and other mixtures. GERG Technical Monograph 15. Fortschr.-Ber. VDI, VDI-Verlag, Düsseldorf. 51. H.W.Liepmann and A.Roshko, 1957. Elements of gas dynamics. John Wiley and Sons. 52. E.W.Lemmon, M.L.Huber, and M.O.McLinden, 2010. NIST Standard Reference Database 23: Reference Fluid Thermodynamic and Transport Properties-REFPROP,” Version 9.0, National Institute of Standards and Technology, Standard Reference Data Program, Gaithersburg. 53. G.B.Wallis, 1969. One-dimensional two-phase flow. McGrawHill. 54. anon., 2009. Table 2 in Guidelines and regulations for oxy-fuel carbon dioxide capture, transport and storage. IEA Oxy-Fuel Working Group. 55.J.Barrie, K.Brown, P.R.Hatcher, and H.U.Schellhase, 2004. Carbon dioxide pipelines: a preliminary review of design and risks. 7th Int. Conf. in GHG Control Tech., Vancouver. Sa no m t f ple or c di op st y rib ut io n pressurized pipelines. American Gas Association, Catalogue No. L00033, November. 29. W.A.Poynton, 1974. A theoretical analysis of shear fracture propagation in backfilled gas pipelines. Crack Propagation in Pipelines, Paper No. 14, published by the Institute of Gas Engineers, Newcastle upon Tyne, UK, March. 30.G.D.Fearnehough and D.G.Jones, 1980. Toughness specification for shear fracture arrest in pipelines. Int. Conf. On Analytical and Experimental Fracture Mechanics, Rome, June 23-27. See also Bonomo, F. et al., Survey and tentative revision of ductile fracture arrest criteria in pipelines for gas transmission, Ibid. See also Vogt, G. H., et al., Toughness for crack arrest in gas pipelines. EPRG Report, 3R International, 22, 1983, pp 98-105, and others. 31. F.Abbassian, 1985. Long-running ductile fracture of high pressure gas pipelines. Dissertation for PhD, University of Cambridge, November. 32. L.B.Fruend and D.M.Parks, 1980. Analytical interpretation of running ductile fracture experiments in gas-pressurized linepipe. Crack Arrest Methodology and Applications, ASTM STP 711, G. T. Hahn and M. F. Kanninen, Eds., pp359-378. 33. G.Buzzichelli, F.Nicolazzo, G.Demofonti, G.Re, S.Venzi, M.F.Kanninen, J.W.Cardinal, E.Z.Polch, T.B.Morrow, S.T.Green, and C.H.Popelar, 1987. Second annual report on the development of a ductile pipe fracture model. Southwest Research Institute report to PRCI, AGA/PRC Contract Nos. PR 182-527 and PR 15-527, May 1987. See also Kanninen, M. F., O’Donoghue, P. E., Cardinal, J. W., Leung, C. P., Morrow, T. B., Green, S. T., Popelar, C. F., Buzzichelli, G., Demofonti, G., Rizzi, L., and Venzi, S., Dynamic fracture mechanics analysis and experimentation for the arrest of ductile fracture propagation in gas transmission pipelines, Pipeline Technology Conference, Belgium, October. 34. G.Demofonti and I.Hadley, 1992. Review of fracture parameters for laboratory measurement of resistance to ductile crack propagation in line pipe steels. CANMET Pipeline Conference, Calgary. 35. S WRI, 1997. Written private communication to Von Rosenburg, E. L., March. 36. S WRI, 1997. Written private communication to Von Rosenburg, E. L., May. 37. G.Demofonti, S.Venzi, and M.Kanninen, 1995. Step by step procedure for the two specimen CTOA Test. 9th EPRG/PRCI Symposium, pp18-1 through 18-10. 38. G.Berardo, P.Salvini, G.Mannucci, and G.Demofonti, 2000. On longitudinal propagation of a ductile fracture in a buried gas pipeline: numerical and experimental analysis. Proc. 2000 Int. Pipeline Conf., 1, New York, ASME, pp287 –294. See also Salvini P., Fonzo A., and Mannucci G., Identification of CTOA and fracture process parameters by drop weight test and finite element simulation, Engineering Fracture Mechanics, 70, 3-4, 553-566, 2003. See also Mannucci, G., Buzzichelli, G., Salvini, P., Eiber, R. and Carlson, L., Ductile fracture arrest assessment in a gas transmission pipeline using CTOA. 3rd Int. Pipeline Conf., Calgary, Alberta, Canada, 1, pp315-320. 39. S.H.Hashimi, et al., 2004. A specimen for studying the resistance to ductile crack propagation in pipes. 5th Int. Pipeline Conf., IPC04-0610, Calgary, 2004: see also Shterenlikht, A., Hashemi, S. H., Howard, I. C., Yates, J. R. and Andrews, R. M., A specimen for studying the resistance to ductile crack propagation in pipes, 4th Quarter, 2010 253 How to select wall thickness, steel toughness, and operating pressure for long CO2 pipelines by Graeme G King*1 and Satish Kumar2 1 Tensor Engineering Ltd, Calgary, Alberta, Canada 2 Masdar Carbon, Abu Dhabi, UAE M ASDAR IS PLANNING to capture CO2 from power plants, smelters, steel works, industrial facilities, and oil and gas processing plants in Abu Dhabi in a phased series of projects. Captured CO2 will be transported in a new national CO2 pipeline network with a nominal capacity of 20 x 106 t/a to oil reservoirs where it will be injected for reservoir management and sequestration. I Sa no m t f ple or c di op st y rib ut io n The design of the wall thickness,pipe toughness,and operating pressure of the network considered fundamental thermodynamic properties of CO2, code requirements, toughness needed to control long ductile fractures, and cost optimization to resolve contention between the different technical requirements and arrive at a safe and economical pipeline design.The work selected a design pressure of 24.5MPa, well above the critical point for CO2 and much higher than is normally seen in conventional oil and gas pipelines. Despite its high operating pressure, the proposed network will be one of the safest pipeline systems in the world today. n June, 2007, Masdar announced the Abu Dhabi Carbon Capture and Storage (CCS) project. Front-end engineering design (FEED) for the first phase of the project started in November, 2008, and has now been completed: Fig.1 shows the proposed route, which will form part of the basic infrastructure that will significantly reduce greenhouse gas emissions in the UAE from 2020 onwards. Captured CO2 will be transported in a new national CO2 pipeline network to onshore oil reservoirs throughout Abu Dhabi where it will be injected for reservoir management and sequestration. Masdar is working closely with the Abu Dhabi National Oil Co (ADNOC) and the Abu Dhabi Co for Onshore Oil Operations (ADCO). The project will have a threefold benefit: it will reduce greenhouse gas emissions in the UAE, make CO2 available for enhanced oil recovery (EOR), and free-up natural gas that is currently being injected to maintain pressure in some of the fields. Three different and sometimes conflicting sets of requirements are needed to select wall thickness and toughness of CO2 pipelines. The first set of requirements limits stresses in the pipe wall under steady and transient *Author’s contact details tel: +1 403 398 3858 email: [email protected] operating pressures and temperatures, and is embodied in the governing pipeline codes and regulations. The first set of requirements is used to select the minimum allowable wall thickness of the pipe. The second set of requirements is based on the need to prevent longitudinal ductile fractures and has been developed by the gas pipeline industry using empirical data obtained from full-scale pipe-burst tests with natural gas. This set of requirements is used to establish pipe toughness and, if the required toughness is unachievable, it dictates the minimum required wall thickness or the use of crack arrestors. The final set of requirements is related to the optimization of total project owning and operating costs. Cost optimization provides a rational methodology for choosing between the apparently conflicting results of the first two sets of requirements, and leads to the selection of a pipeline that is both safe and economic. Nomenclature A = area beneath Charpy notch, m2 CS = velocity of sound in CO2, m/s CV = Charpy notch toughness, J D = pipe outside diameter, m E = modulus of elasticity, Pa EN = normalized toughness parameter, (-) F = hoop stress design factor, (-) The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n 254 Fig.1. Route map for Abu Dhabi CO2 pipeline network. FJ = weld joint factor (ASME B31.4 Table 402.4.3) Pd = decompressed pressure at phase boundary, Pa SMYS = specified minimum yield strength T1 = temperature at time of installation, °C T2 = maximum operating temperature, °C t = pipe wall thickness, m α = linear coefficient of thermal expansion, 1/°C ∆v = change in velocity of the fluid, m/s ∆P = surge pressure, Pa ν = Poisson’s ratio for steel, (-) ρ = density of the fluid, kg/m3 σd = hoop stress due to internal pressure Pd, Pa σf = pipe steel flow stress, Pa σh = hoop stress, Pa σL = longitudinal compressive stress, Pa Process design CO2 can be transported most efficiently over long distances in the dense phase. The Cortez pipeline, for example, transmits CO2 from Wyoming to Texas and operates at pressures up to 18MPa (2600psi), and the Weyburn-Souris pipeline in Montana and Saskatchewan has a maximum operating pressure of almost 21MPa (3000psi). These high operating pressures are due in part to the need for a high minimum pressure to maintain single-phase operation, and in part to the optimum frictional pressure loss along CO2 pipelines. As a result the optimum operating pressure can be higher than 21 MPa. For any specific project, the optimum operating pressure is influenced by the composition of the CO2 mixture it carries and specific details of cost and economic models used for the project. The dense phase was originally defined as a single phase separating the gas and liquid phases immediately above the two-phase region, where fluid properties transition between those of a gas and a liquid without any change of phase, and where fluid properties can be distinctly different from those of either a gas or a liquid [1]. Figure 2 shows the dense phase region on a pressure-enthalpy chart developed using the BWRS equation of state for a design mixture of 95% pure CO2. Specific heat is one of the properties of dense-phase fluids that is different from either a gas or a liquid. Figure 2 shows specific heat for gas, dense, and liquid phases (single-phase region): it shows that the specific heat of CO2 in both gas and liquid phases is less than 2.5kJ/kg-K but in the dense phase the specific heat is higher than 2.5kJ/kg-K and can reach values as high as 10kJ/kg-K. Another unusual property of dense-phase fluids is their volumetric sensitivity to changes in temperature. For example, the density of CO2 at 10MPa and 40°C is 400kg/ m3 (see Point 1 on Fig.2) and the density of carbon dioxide at 15°C and 10MPa is 800kg/m3 (see Point 2 on Fig.2). If the fluid followed the gas law, this 25°C change in temperature from 40°C (313K) to 15°C (288K) would cause the density to change only 8% from 500 to 540kg/m3. But dense-phase CO2 is an order of magnitude more sensitive to changes 4th Quarter, 2010 255 Location Dwellings per Mile General Description Design Factor Class 1 <10 Sparsely populated wasteland, wilderness, grazing and farmland 0.72 Class 2 10 to 46 Intermediate fringe and areas around cities and towns 0.60 Class 3 >46 Suburban residential and industrial areas 0.50 Class 4 >46 Multi-story urban areas with high traffic and buried utilities 0.40 Tier-2 0.5 million – 1 million Cement factories, CCGT Power stations, fertilizer, petrochemical complexes Table 1. Basic design factor (F). in temperature, and in this particular case a temperature change of only 25°C causes the density to change more than 100% from 400 to 800kg/m3. parameters subject to all the relevant technical restraints imposed by the requirements of hydraulic performance, governing codes, and good engineering practice. The UAE has long summers with air temperatures rising to about 48°C between May and September, and short moderate winters between December and March with air temperatures rarely dropping below 6°C. The maximum design temperature of CO2 from aerial coolers after compression has therefore been set at 55°C. The ground temperature at pipeline depth fluctuates between 13°C in late winter and 38°C in late summer. Allowable stress Sa no m t f ple or c di op st y rib ut io n Requirements for CO2 pipelines are included in ASME B31.4 Pipeline transportation systems for liquid hydrocarbons and other liquids [2], in which paragraph 402.3. uses a design factor (F) of 0.72 to limit the allowable hoop stress (σh) under steady flow conditions: The volumetric sensitivity of dense-phase CO2 to changes in temperature means that during summer when the ground temperature at pipeline depth is close to 40°C, the density of carbon dioxide at the inlet to intermediate booster stations and down-hole injection facilities would be approximately 400kg/m3, but at low flow during winter when the ground temperature at pipeline depth is around 15°C, the density of CO2 at pump suction would be approximately 800kg/ m3. Wide swings in density between summer and winter, and between high and low flows, can create challenges to the smooth operation of the pipeline unless the magnitude of the swings is identified early in the design and solutions are found and implemented. If centrifugal equipment is used to pump the dense-phase fluid, wide swings in fluid density cause proportionally wide swings in differential pressure between suction and discharge. Speed control can be used to manage the swings by allowing the equipment to be run faster during summer when density is low and slower in winter when density is high. Variablefrequency and variable-speed drives (VFDs and VSDs) are now widely used throughout the pipeline industry, so that once variations in density have been properly evaluated it is not difficult to size and configure variable-speed centrifugal equipment to handle the full range of operating flows, pressures, and temperatures throughout the year. The overall goal of pipeline design work therefore is to configure the proposed system to handle a nominal flow of 20 × 106t/a of CO2 with a maximum design temperature of 55°C, ground temperatures between of 17°C and 38°C, and a maximum design pressure of around 21MPa, and to optimize σh ≤ 0.72 FJ SMYS (1) Metallurgical investigations conducted as part of the optimization work looked at a range of steel strengths and selected API 5L X65 (L450) linepipe as the most suitable for the project. Table 402.4.3 of ASME B31.4 sets the joint factor (FJ) in Equn 1 for pipe made from this steel equal to unity so that the allowable hoop stress for the pipeline network under steady flow conditions is 72% SMYS. In addition, paragraphs 402.3.2(c) and 419.6.4(b) of ASME B31.4 limit the combined stress of buried pipelines: σh + σL ≤ 0.90 SMYS (2) In Equn 2, σL is the net longitudinal compressive stress due to the combined effects of temperature rise and fluid pressure computed from: σL = Eα(T2 – T1) – νσh (3) Paragraph 419.6.4(b) of ASME B31.4 makes it clear that bending stresses only need to be included in combined stress calculations when designing above-grade portions of restrained lines and do not need to be considered when designing buried portions providing, of course, that the pipeline is built using proven pipeline construction practices, good engineering, and other more-specific rules included in Chapter V of ASME B31.4. If a design factor of 0.72 is used, Equns 2 and 3 allow a maximum temperature differential of 160°C between construction and operating conditions for buried pipelines before extra wall thickness is required to keep the combined 256 The Journal of Pipeline Engineering Fig.2. Pressure-enthalpy diagram for 95% pure CO2 showing specific heat in liquid, dense, and gas phases. density over the normal range of operating conditions is approximately 800kg/m3. Sa no m t f ple or c di op st y rib ut io n stress below 0.90 SMYS. The maximum temperature differential between construction and operation for the proposed pipeline is less than 50°C, and the combined stress limitations of ASME B31.4 therefore do not require additional wall thickness. In order to enhance safety, the design factor of 0.72 has been modified depending on its location by additional restrictions of ASME B31.8 Gas transmission and distribution piping systems [3]. Table 1 summarizes the design factors for the four location classes specified in Table 841.114A of ASME B31.8. These additional requirements over and above the requirements of ASME B31.4 increase wall thickness of the CO2 pipeline network in populated areas. Surge pressures Surge pressures are produced by changes in velocity of the moving fluid that result from shutting-down pumps or pump stations, closing valves, or otherwise blocking the flow. Internal viscous effects and inelastic properties of the backfill and pipe wall attenuate surge pressure waves as they move away from the point of origin. Paragraph 402.2.4 of ASME B31.4 requires pipeline designers to make surge calculations and to provide adequate controls and protective equipment to prevent pressure rise due to surges and other variations from normal operations from exceeding the internal steady state design pressure anywhere in the pipeline system and equipment by more than 10%. The maximum surge pressure (∆P) can be precisely calculated from the fundamental equation: ∆P = ρ CS ∆v (4) Figure 3 shows the density of dense-phase CO2 over a broad range of operating pressures and temperatures. Typical fluid Figure 4 shows the velocity of sound in CO2, the typical value of which over the normal range of operating conditions is approximately 500m/s. Both density and velocity of sound in CO2 are slightly lower than for oil, so that the magnitude of surge pressure waves in CO2 pipelines can be expected to be less than in oil pipelines. Figure 5 shows the unit surge pressure defined as the change in pressure for a change in flow velocity of 1m/s. The typical value for unit surge pressure over the normal range of operating conditions is approximately 0.4MPa/(m/s). Surge pressure is proportional to change in flow velocity, and maximum surge pressure therefore depends on the maximum flow velocity. The optimum flow for CO2 systems is less than 4m/s and therefore, for cost efficiency, it is best to design and build CO2 systems so that they operate at velocities less than 4m/s. The maximum surge pressure in a properly designed CO2 pipeline, if a valve suddenly closes or a pump station suddenly stops working, is therefore less than 1.6MPa (that is, 4m/s x 0.4MPa/(m/s)). Since the design pressure of the proposed pipeline network is 24.5MPa, the maximum increase in pressure due to sudden valve closure or station outage is no more than 6.5% above the design pressure. Because the maximum surge pressure is less than allowed by the code (that is, less than 10% above the design pressure), no increase in wall thickness is required. The next section will comment on the effect of surge pressure on ductile fracture arrest in CO2 pipelines. Ductile fractures Paragraph 402.5.1 of ASME B31.4 requires the designer of CO2 pipeline systems to consider the possibility of ductile 4th Quarter, 2010 257 Fig.3. Density of CO2 with highlighting showing the normal range of operating conditions. Decompression and crack velocity Sa no m t f ple or c di op st y rib ut io n fractures and provide reasonable protection to limit their occurrence and length throughout the pipeline, with special consideration at crossings and other appropriate locations. More specifically, Paragraph 402.5.3 requires the designer to minimize ductile fracture propagation by the selection of pipe steel with appropriate fracture toughness and/or by the installation of suitable fracture arrestors. Design considerations include pipe diameter, wall thickness, fracture toughness, yield strength, operating pressure, operating temperature, and the decompression characteristics of CO2 with its associated impurities. The possibility that very long ductile fractures can occur in CO2 pipelines was first identified in the late 1970s [4] almost 10 years after the first CO2 transmission pipelines had been built in North America. Crack arrestors were subsequently retrofitted to existing CO2 pipelines to help reduce risk. Since that time new CO2 pipelines have been built with crack arrestors. Thicker walls and tougher steels can be used instead of crack arrestors effectively to control long ductile fractures. If a longitudinal subcritical crack in the pipe wall grows during operation, it can initiate a longitudinal tear in the pipe wall. If the pipe wall is not strong enough or tough enough to resist the force of the decompressed pressure pushing against the unrestrained walls of the pipe on either side of the crack, a ductile fracture will form and run along the pipe. The solution is to increase either wall thickness or fracture toughness of the pipe steel, or both. Increasing wall thickness helps by reducing the stress at the tip of the fracture, and increasing toughness helps by enabling the steel to absorb more energy when it tears. Alternatively, crack arrestors can be installed to limit the length of the fracture. In the unlikely event that a dense-phase CO2 pipeline bursts and a ductile fracture starts to run, the sudden loss of gas causes the CO2 to decompress isentropically into the two-phase region. The decompression is rapid and highly turbulent, and the gas and liquid components do not have time to separate. The two phases behave like an homogenous single phase with a very low sonic velocity (less than 100m/s). The low sonic velocity causes a sustained pressure at the moving tip of the fracture equal to the pressure at which the decompression crosses the phase boundary into the two-phase region. This high sustained pressure acts on the unrestrained flaps of the fractured pipe just behind the moving fracture tip, and concentrates stresses at the tip of the fracture large enough to tear the steel wall and drive the fracture along the length of the pipe. Pipe steels need to be both strong and tough to resist the forces and prevent the crack from propagating. In order to calculate the pressure at the tip of the moving fracture it is necessary to understand how CO2 decompresses and the effect of phase behaviour. The actual composition of CO2 in the proposed pipeline is important in this regard. CO2 from identified sources is relatively pure but during design it is not possible to identify all the future sources of CO2. Therefore a worst-case estimate of composition from present and future sources needs to be used during design to protect the pipeline system against long ductile fractures over its full operating life. Of the possible impurities, hydrogen has the largest effect on the phase boundary and decompression behaviour of CO2. Nitrogen has a lesser but still large effect while 258 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.4.Velocity of sound in CO2 with highlighting showing the normal range of conditions. Fig.5. Unit pressure surge for a change in flow velocity of 1m/s. other possible impurities have relatively minor effects. The effect of hydrogen on the properties of CO2 has been determined accurately by experimental and theoretical work of Prausnitz and Gunn [5] and others [6, 7, 8]. Comparison of experimental results with predictions of equations of state using parameters developed by Prausnitz and Gunn [5] have shown that the PR equation of state is sufficiently accurate for design work over a wide range of pressures and temperatures. and that the BWRS equation of state was sufficiently accurate for temperatures greater than -20°C. After studying a range of possible compositions from a variety of sources that might need to be transported in the proposed network, a design-case mixture of 95% CO2 with 1% hydrogen and 4% nitrogen by volume was chosen as the basis for design work to prevent long ductile fractures over the operating life of the project. Figures 6, 7, and 8 illustrate how the pipe toughness that is needed to arrest ductile fractures is found using the Battelle two-curve method [9]. Each figure shows two sets of curves: the first is a set of decompression pressure-wave velocity curves for the design-case mixture for a range of temperatures from 15 to 55°C (59 to 131°F) when the pipeline decompresses following a rupture. The decompression pressure wave curves define the speed that each pressure level propagates back along the pipe from the initial fracture site as the pipeline decompresses. Figures 6, 7, and 8 are for sudden decompression from operating pressures of 17.5, 21, and 24.5MPa (2538, 3046, and 3553 psi) respectively, and were calculated using the BWRS equation of state. The second set of curves in Figs 6, 7, and 8 is a set of Battelle fracture-velocity J-curves for a range of toughness values. For illustrative purposes the J-curves were calculated for NPS 18 (18-in diameter) Grade L450 (X65) pipe with wall thickness for the three different operating pressures calculated using a design factor of 72%. The fracture-velocity J-curves define the relationship between the pressure driving the fracture along the pipe and the velocity of the fracture. The toughness 4th Quarter, 2010 259 Sa no m t f ple or c di op st y rib ut io n Fig.6. Decompression pressure wave velocity curves for design-case CO2 decompressing from 17.5MPa intersecting toughness J-curves for NPS18 L450 pipe with 12.4-mm wall. required to arrest a ductile fracture is found where the two sets of curves are tangential (just touching). Figures 6, 7 and 8 show that higher operating pressures require lower-toughness pipe steel. Figure 6 for decompression from 17.5MPa shows a required toughness of 120J; Fig.7 for decompression from 21MPa shows a toughness of 60J; and Fig.8, for decompression from 24.5MPa, shows a toughness requirement of 40J. An interesting observation from these figures is that, when the initial operating temperature is kept constant, the sustained pressure plateau is slightly lower for higher initial operating pressures. This means that to prevent long ductile fractures in CO2 pipelines, unlike gas pipelines, they need to be designed to handle low operating pressures rather than the maximum allowable operating pressure. If CO2 pipelines are designed to prevent ductile fractures at low operating pressures they are generally safer from the point of view of controlling ductile fractures when they are operated at higher pressures, up to the maximum hoop stress allowed by the code. Another conclusion that can be drawn from this counterintuitive result is that increases in operating pressure due to surge waves do not need to be considered when selecting pipe toughness to arrest ductile fractures. This counterintuitive behaviour of dense-phase pipelines is distinctly different from gas pipelines where the sustained pressure plateau increases markedly as operating pressure increases. Figure 9 illustrates the reason for the difference, and shows decompression paths from four different operating points on a pressure-enthalpy diagram developed using the BWRS equation of state for a mixture of 95% pure CO2. Decompression paths from Points 1 and 2 in the dense phase intersect the bubble-point line, whereas decompression paths from Points 3 and 4 in the gas phase intersect the dew-point line. Because bubble-point and dewpoint lines have different slopes, dense-phase decompression from high-pressure Point 1 intersects the phase boundary at a lower pressure than decompression from low-pressure Point 2. On the other hand, gas-phase decompression from high-pressure Point 3 intersects the phase boundary at a higher pressure than decompression from low-pressure Point 4. This difference explains why it is easy to control ductile fractures in dense-phase pipelines by increasing the design pressure, but not so easy in gas-phase pipelines. Toughness calculations The decompression path from 24.5MPa and 40°C for design-case CO2 is shown in Fig.8. It shows a sustained pressure plateau extending between 60m/s and 330m/s at a pressure of 8.4MPa. The toughness requirement where the decompression path just touches the fracture velocity curve is determined from the low-velocity end of the pressure plateau. at a velocity of 60m/s. At velocities as low as 60m/s, the fracture-velocity J-curve is horizontal; it can be seen from Fig.8 that if the decompressed pressure is kept constant at 8.4MPa there is no practical difference between the toughness required for a fracture speed of 60m/s and the toughness required for a fracture speed of 0m/s. For CO2 pipelines, this observation allows a simplification, with no loss of accuracy for steel toughness up to approximately 200J, by setting fracture velocity equal to zero and fracture arrest pressure equal to the highest pressure at which the decompression enters the two-phase region [4]. 260 The Journal of Pipeline Engineering Wall @ 175 bar & 72% Wall @ 210 bar & 72% Wall @ 245 bar & 72% Diameter Steel Grade Thickness Toughness Thickness Toughness Thickness Toughness inches MPa mm J mm J mm J 8.625 450 5.9 ∞(1) 7.1 35.6 8.3 22.7 10.75 450 7.4 ∞(1) 8.8 44.4 10.3 28.3 12.75 450 8.7 ∞(1) 10.5 52.6 12.2 33.6 16 450 11.0 ∞(1) 13.2 66.0 15.4 42.1 20 450 13.7 ∞(1) 16.5 82.5 19.2 52.7 24 450 16.5 ∞(1) 19.8 99.0 23.0 63.2 Table 2.Wall thickness to satisfy the requirements of ASME B31.4 and notch toughness to prevent long ductile fractures in pipelines carrying the design-case mixture. Note (1): use crack arrestor or increase wall thickness when the decompressed stress ratio (σd/σf) exceeds 0.28 (5) where There are approximations in the development of the Battelle fracture-arrest equations and uncertainties in their application to CO2 pipelines that need to be considered. The underlying theory was originally calibrated using results from full-scale pipe-burst tests to advance the design of pipelines carrying natural gas. The theory has been extended without additional experimental validation to CO2 pipelines operating at much higher pressures than natural gas pipelines. Full-scale burst tests with proposed pipe, proposed operating pressure, and proposed pipelinequality CO2, are recommended to validate the applicability of the theory to new CO2 pipelines. Sa no m t f ple or c di op st y rib ut io n For many CO2 pipelines the fracture arrest pressure will be equal to the cricondenbar of the mixture selected for the design work. If the fracture velocity is set equal to zero and the arrest pressure is set equal to the highest pressure at which the decompression path intersects the phase boundary (Pd), the Battelle fracture-arrest equations [9], after rearranging terms, become: (6) (7) As the ratio of hoop stress at the fracture tip to flow stress (σd/σf) approaches 0.30, the normalized toughness (EN) in Equn 6 approaches infinity and becomes highly sensitive to small errors in the estimation of decompressed pressure and flow stress. When the ratio exceeds 0.28, wall thickness rather than toughness should be increased to control ductile fractures, or crack arrestors should be used. As a result, the practical limit of applicability of Equns 5, 6, and 7 is given by: (8) Equations 5 to 8 can be used to calculate the toughness (CV) required for the arrest of ductile fractures using wall thickness (t), pipe diameter (D), decompressed plateau pressure (Pd), and flow stress (σf). Alternatively, the equations can be solved iteratively to find the wall thickness (t) needed to arrest ductile fractures for any given toughness (CV). Future variations in product quality have been accounted for in this work by selecting an appropriate design-case mixture but, until full-scale burst tests are conducted to validate the Battelle fracture-arrest equations for CO2 pipelines, an additional margin of safety is required in order to account for uncertainties in the theory itself and its extension to dense-phase CO2. This has been done by adding 0.4MPa to the arrest pressure (cricondenbar). The cricondenbar of the design-case mixture is 8.4MPa, so that the arrest pressure selected for this project for use in Equn 7 was therefore 8.8MPa. Table 2 shows the minimum wall thickness required to satisfy the hoop-stress requirements of ASME B31.4 with a design factor of 0.72 as well as the Charpy V-notch toughness required to prevent long ductile fractures using fracture arrest Equns 5, 6, and 7 when the pipeline is transporting the design-case mixture. Crack arrestors are indicated in Table 2 for cases where the decompressed stress ratio (σd/ σf) exceeds 0.28. Optimization Table 2 defines three long CO2 transmission pipelines with design pressures of 17.5, 21.0, and 24.5MPa using pipe sizes from NPS 8 (8.625ih) to NPS 24 (24in). The three different systems can be characterized as follows: 4th Quarter, 2010 261 Sa no m t f ple or c di op st y rib ut io n Fig.7. Decompression pressure wave velocity curves for design-case CO2 decompressing from 21MPa intersecting toughness J-curves for NPS18 L450 pipe with 14.9-mm wall. Fig.8. Decompression pressure wave velocity curves for design-case CO2 decompressing from 24.5MPa intersecting toughness J-curves for NPS18 L450 pipe with 17.3-mm wall. • pipelines with design pressure of 17.5MPa and the thinnest wall of the three alternatives using crack arrestors instead of toughness to arrest ductile fractures; • pipelines with design pressure of 21.0MPa using relatively high toughness steel to arrest ductile fractures; • pipelines with design pressure of 24.5MPa and the thickest wall of the three alternatives using relatively low-toughness steel to arrest ductile fractures. Optimization on a cost basis can be used to choose between these alternatives. To make a fair comparison of systems with different maximum design pressures, compressor facilities for all the systems were designed to take CO2 from source at the minimum pipeline operating pressure and deliver it to injection facilities at the same minimum operating pressure. Initial and intermediate compressor stations were located as required to boost pressure from the minimum operating pressure to the system-design pressure. The pipe costs were based on current prices for API 5L L450 PSL2 pipe and do not include a premium for highertoughness steels. It was assumed that crack arrestors could be installed at no additional cost so that the lower-pressure systems would not be unfairly penalized. Figure 10 shows 262 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.9. Decompression paths for 95% pure CO2 showing why burst pressure inside gas pipelines is higher when initial pressure is higher, but in densephase pipelines it is lower when initial pressure is higher. Fig.10. Comparative unit cost of transportation curves for dense phase CO2 pipelines showing optimum capacities for each pipe size. comparative unit cost of transportation curves based on total lifetime owning and operating costs for the three systems: it can be seen that the lowest unit transportation cost (lowest lifetime owning and operating cost) is obtained with a design pressure of 24.5MPa for all pipe sizes. This means it is more cost-effective for long CO2 transmission pipelines to use pipe with thick walls, high operating pressures, and low toughness steels, rather than thin walls, low operating pressures, and high-toughness steels or crack arrestors. A further increase in operating pressure is not expected to further reduce cost of transportation because 24.5MPa is close to the pressure rating of 1500 Class flanges at the maximum pipeline operating temperature, and the high cost of heavier flanges and fittings (2500 Class) would put a step in the optimization curves. Conclusions ASME B31.4 sets out the equations required for calculating minimum wall thickness of CO2 pipelines for any given operating pressure and the Battelle fracture arrest Equns 5 to 7 in combination with Equn 8 can be used to calculate toughness to arrest ductile fractures, as well as the need for extra wall thickness or crack arrestors. Finally, cost optimization provides a rational methodology for choosing between the different alternatives produced by the first two sets of requirements, leading to the selection of a pipeline system that is both safe and economic. For this project a design pressure of 24.5MPa was selected as the most costeffective alternative resulting in the adoption of thick walls and low-toughness steels for the arrest of ductile fractures. Although it is more cost effective to use the pipe wall rather 4th Quarter, 2010 263 than crack arrestors to arrest ductile fractures, the main benefit is the increase in safety that comes from limiting the length of ductile fractures to much shorter lengths than is practical with crack arrestors. A secondary benefit of using the pipe wall rather than crack arrestors to control ductile fractures is that the pipeline is built with a thicker wall making it more resistant to third-party damage. References Sa no m t f ple or c di op st y rib ut io n 1. D.L.Katz and G.G.King, 1973. Dense phase transmission of natural gas. Energy Processing Canada, November and December. 2. American Society of Mechanical Engineers, 2006. ASME B31.42006 Pipeline transportation systems for liquid hydrocarbons and other liquids. 3. Idem, 2007. ASME B31.8-2007 Gas transmission and distribution piping systems. G.G.King, 1981. Design considerations for CO2 pipe line. Pipe Line Industry, November, pp125-132. 4. J.M.Prausnitz and R.D.Gunn, 1958. Volumetric properties of nonpolar gaseous mixtures. A.I.Ch.E Journal, 4, 4, December, pp430-435. 5. C.Yokoyama, K.Arai, S.Saito, and H.Mori, 1988. Bubble-point pressures of the H2-CO-CO2 system. Fluid Phase Equilibria, 39, 101-110. 6. J.O.Spano, C.K.Heck, and P.L.Brick, 1968. Liquid-vapor equilibria of the hydrogen-CO2 system. J. Chem. Eng. Data, 13, 168-171. 7. C.Y.Tsang and W.B.Streett, 1981. Phase equilibria in the H2/ CO2 system at temperatures from 220 to 290 K and pressures to 172 MPa. Chem. Eng. Sci., 36, 993-1000. 8. R.J.Eiber, T.A.Bubenik, and W.A.Maxey, 1993. Final report on fracture control technology for natural gas pipelines. Project PR-3-9113, NG-18 Report No. 208, Pipeline Research Committee, American Gas Association, December. UPSF This two-day forum will address capabilities and guidance concerning tools for corrosion and mechanical damage inspection of �unpiggable’ oil, gas and hazardous liquids pipelines. The focus will be on existing, new, and developing in-line tools as well as those in research and development. Sa no m t f ple or c di op st y rib ut io n Advances and best practices in guided-wave technologies will also comprise a significant part of the program. Live-line access techniques, combined with low-flow restriction robotic and wire-line-powered internal inspection tools, will also be a special focus, especially for pipeline segments that cannot be taken out of service when these tools suggest a need for further integrity inspections. Contact us today for information on registration or about sponsorship & exhibiting options. Program Chairman Dr. Keith Leewis, P-PIC Program advisory Committee Mark Andraka, PECO Energy Drew Hevle, El Paso Corp. Richard Kania, TransCanada Pipelines Garry Matocha, Spectra Energy Bryan Melan, Marathon Oil Co. Andrew Pulsifer, CenterPoint Energy Albert Van Roodselaar, Chevron Energy Technology houston marriott Westchase hotel ConferenCe organizers B.J. Lowe, Clarion Technical Conferences John Tiratsoo, Tiratsoo Technical Call +1 713 521 5929 or visit www.clarion.org 4th Quarter, 2010 265 A dynamic boundary ductilefracture-propagation model for CO2 pipelines by Prof. Haroun Mahgrefteh*, Solomon Brown, and Peng Zhang Department of Chemical Engineering, University College London, UK T T Sa no m t f ple or c di op st y rib ut io n HE DEVELOPMENT and testing of a dynamic boundary ductile-fracture-propagation model for pressurized CO2 pipelines is presented. The model accounts for all the important fluid-structure interactions governing the fracture process.These include expansion-wave propagation, real fluid behaviour, pipe/wall fiction, and heat transfer, as well as the rapidly diminishing dynamic loading effects as the crack tip opens. The resistance to crack-tip propagation is determined based on the drop-weight tear test energy approach. The performance of the fracture model is tested by comparison of its predictions of the crackpropagation velocity versus crack length against real data. The latter include the High-Strength Line Pipe Committee, ECSC X100 and Alliance full-scale burst tests conducted for pipes containing either air or rich gas mixtures. In all cases good agreement is obtained between the model predictions and the real data.The validated model is used to test the propensity of a hypothetical but realistic pressurised CO2 pipeline to ductile fracture propagation failure. The simulations indicate the remarkably significant role of the starting line temperature on fracture propagation in CO2 pipelines. HOUSANDS OF KILOMETRES of pressurized pipelines are used to transport large amounts of hydrocarbons across the world. Although this method of transportation is generally considered to be safe, pipeline failures do occur with some leading to catastrophic consequences (see for example Refs 1, 2). In the US alone, despite having one of the most stringent safety requirements across the globe, over 202 pipeline incidents were reported during 2005 – 2009 [3]. These resulted in an estimated $2 billion of damage leading to 69 deaths and 254 serious injuries. Propagating factures are considered as by far the most catastrophic type of pipeline failure. Such failures involve the rapid axial splitting or tearing of the pipeline, sometimes running over distances of several hundred meters resulting in massive loss of inventory in a very short time. Deservedly, understanding and modelling of the mechanisms responsible for such type of failure has led to a large number of studies (see for example Refs 4, 5). Such interest has intensified recently [6-8] given the prospect of using pressurised pipelines for transporting captured CO2 from fossil plants for subsequent storage. This paper is based on one presented at the First International Forum on Transportation of CO2 by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical and Clarion Technical Conferences, and with the support of the University of Newcastle and the Carbon Capture and Storage Association. *Author’s contact details tel: +44 (0)20 7679 3835 email: [email protected] Given that CO2 at concentration of >10% v/v is likely to be instantly fatal [9], the rupture of a CO2 pipeline near a populated area can lead to catastrophic consequences. Fractures can initiate from defects introduced into the pipe by outside forces such as mechanical damage, soil movement, corrosion, material defects, or adverse operating conditions. Fractures propagate when the stresses acting on the defect overcome the fracture initiation tolerance of the pipe, reaching a critical size based on the pipeline material properties and operating condition. As such it is highly desirable to design pipelines such that when a defect reaches a critical size and fails, the result is a leak rather than a long running facture. The above requires a two-tiered design approach involving: • providing sufficient fracture initiation resistance, mainly via specifying the required pipe toughness, wall thickness and operating conditions • ensuring sufficient fracture propagation resistance such that if a running fracture occurs its length is limited to a short distance Notwithstanding cost implications, fracture initiation can be largely controlled a priori by specifying the required fracture initiation toughness, minimum wall thickness and the maximum stresses acting upon the defect. 266 The Journal of Pipeline Engineering Parameter HLP ECSC Alliance Rich Gas (see table 2) Air Rich Gas (see table 2) 1.182 1.182 1.4223 0.8856 0.0183 0.0183 0.0183 0.0191 0.0142 Initial pressure (bara) 116 116 104 126 120.2 Initial temperature (oC) 12 6 -5 20 23.9 Ambient pressure (bara) 1.01 1.01 1.01 1.01 1.01 Ambient temperature (oC) 20 20 20 20 20 Pipe length (m) 35 35 35 35 100 Tensile stress (MPa) 505 505 505 807 505 Yield stress (MPa) 482 482 482 728 482 Pipe grade X70 X70 X70 X100 X70 B1 C2 Inventory Air Air Internal diameter (m) 1.182 Pipe thickness (m) Sa no m t f ple or c di op st y rib ut io n A1 Table 1. Pipeline characteristics and prevailing conditions used for the full-scale burst tests. However, controlling fracture propagation once a leak has formed is more complex, presenting a unique set of challenges. As well as the fracture toughness of the steel and the backfill conditions, the fracture-propagation velocity and arrest length depend on the depressurization rate, the thermal stresses, and the minimum pipe wall temperature relative to its ductileto-brittle transition temperature. To model the above and hence develop methodologies for overcoming such a type of failure, we need to understand the nature of the processes taking place once a fracture has been initiated. The onset of a leak in the pressurized pipeline results in a series of expansion waves that propagate from the rupture plane towards the intact end of the pipeline at the speed of sound [10]. As the main driving force for crack propagation is the crack tip pressure [11], the precise tracking of the expansion waves, and their effect on the pressure profile along the pipeline, is essential for the proper modelling of fracture propagation. Fig.1. Schematic representation of the experimental setup used in the HLP full-scale pipe burst tests [20]. Additionally, given the significant drop in the speed of sound and hence the depressurization rate during the transition from the gaseous to the two-phase region [12], such analysis must also account for real fluid behaviour through the use of an appropriate equation of state. Also, non-isentropic effects such as the fluid/pipe wall friction and heat transfer must be accounted for as these also directly affect the depressurization rate. Finally the temperature drop as a result of the Joule-Thomson expansion cooling [13] of the fluid within the pipeline during discharge can be significant. In the case of CO2, depending on the starting conditions, such temperatures can reach as low as -70°C resulting in very significant localised cooling of the pipe wall in contact with the escaping fluid. The minimum pipe wall temperature reached relative to its ductile to brittle transition temperature will dictate whether the pipeline will fail in the ductile or brittle fracture manner. The modelling of brittle fractures in 267 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.2.Variation of crack velocity with crack length for test A1 south-running crack. Inventory: air, initial pressure = 116bara, initial temperature = 12°C. Curve A: experimental data [20]. Curve B. DBFM prediction. pressurized pipelines has been presented in the authors’ previous publication [13]. Ductile fractures are the focus of our attention in this work. The so called Battelle Two-Curve (BTC) approach by Maxey [5] was the first used to express the criterion for the propagation of a ductile fracture in terms of the relation between the fluid decompression-wave velocity and the crack-propagation velocity. If the fluid decompressionwave velocity is larger than the crack velocity, the crack tip stress will decrease, eventually dropping below the arrest stress and causing the crack to arrest. Conversely, if the decompression-wave velocity remains smaller than the crack velocity, the crack tip pressure will remain constant resulting in indefinite propagation. Several studies have since been conducted for modelling ductile fractures based on the BTC approach (see for example Refs14, 15). Some employ sophisticated finiteelement methods for simulating material deformation but use over-simplistic transient fluid flow models for predicting the rupture plane pressure and hence the crack driving force (see for example Refs 16, 17). Others, on the other hand, although accounting for the transient depressurization profile within the pipeline, do not deal with the impact of pipe wall heat transfer and friction effects on the fluid decompression behaviour (see for example Refs 18, 19). Additionally a reliable decompression model must also incorporate a suitable equation of state. This is especially important in the case of CO2 pipelines given the unique depressurization thermodynamic trajectory of CO2 [8]. Crucially none of the studies reviewed simulate the dynamic interaction between the rapidly changing crack tip opening area and the pressure loading as the crack propagates. In this paper, we report the development and validation of a rigorous dynamic boundary ductile-fracture-propagation model which takes into account all of the important transient fluid/structure interactions governing the fracture process. The performance of the model in terms of predicting the crack-propagation velocity and arrest length is tested by comparison against real data. These full-scale burst tests conducted by the High-Strength Line Pipe Committee[20], ECSC X100 [21] and Alliance[22] for pipes containing either air or rich gas mixtures. The validated model is used to test the propensity of a hypothetical but realistic pressurized CO2 pipeline to ductile fracture propagation failure, paying particulate attention to the impact of the starting line temperature. The latter investigation was prompted by the findings of Cosham and Eiber [7] indicating the significance of the starting temperature on the CO2 depressurization trajectory relative to its phase transition boundary. 268 The Journal of Pipeline Engineering Theory Sa no m t f ple or c di op st y rib ut io n Fig.3.Variation of crack velocity with crack length for test B1 south-running crack. Inventory: air, initial pressure = 116bara, initial temperature = 6°C. Curve A: experimental data [20]. Curve B: DBFM prediction. The full background theory of the fluid flow model employed in this study to predict the fluid decompression velocity and the crack tip pressure for a given opening area is given elsewhere [23-26], and hence only a brief account of its main features is given here. Based on the homogeneous flow assumption, in the case of unsteady, one-dimensional flow the mass, momentum and energy conservation equations are respectively are given by: where ρ, u, P and h are the density, velocity, pressure and specific enthalpy of the homogeneous fluid as function of time, t, and space, x; qh is the heat transferred through the pipe wall to the fluid and βy is the friction force term given by: where, fw is the Fanning friction factor and D the pipeline diameter. Also, where θ is the angle of inclination of the pipeline to the horizontal. Equations 1-3 are quasi-linear and must be solved numerically. In this study, the Method of Characteristics (MOC) [27] is used as the numerical solution method, as opposed to other numerical techniques such as finite-element [28, 29] and finite-difference methods [30-32] as both have difficulty in handling the choking condition at the rupture plane. The MOC handles the choked flow intrinsically via the Mach line characteristics. Moreover, MOC is considered to be more accurate than the finite-difference method as it is based on the characteristics of wave propagation. Hence, numerical diffusion associated with a finite-difference approximation of partial derivatives is reduced. The key step in the BTC method is the derivation of two sets of curves: one set describing the crack velocity, and the other the velocity of fluid decompression wave. The resistance to crack propagation is indicated by the Charpy V-Notch (CVN) 269 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.4.Variation of crack velocity with crack length for test C2 south-running crack. Inventory: rich gas (Table 2), initial pressure = 104bara, initial temperature = -5°C. Curve A: experimental data [20]. Curve B: DBFM prediction. energy [5]. However, in the full-scale pipe bust tests conducted by the High-Strength Line Pipe Committee (HLP) [33], the BTC theory is used in conjunction with the drop-weight tear test (DWTT) energy, as this is shown to provide a more accurate indication of the pipeline resistance to fracture. Consequently this is the model applied in this work. The two-curve model for the crack propagation velocity, vc, and crack arrest pressure, Pa, are respectively given by [33]: Component HLP C2 Alliance Test 1 CH4 89.57 80.665 C2H6 4.7 15.409 C3H8 3.47 3.090 iC4H10 0.24 0.232 nC4H10 0.56 0.527 iC5H12 0.106 0.021 nC5H12 0.075 0.014 nC6H14 0.033 0.003 nC7H16 0.017 0 nC8H18 0.008 0 nC9H20 0.001 0 N2 0.5 0.039 CO2 0.72 0 Table 2. Rich gas feed compositions. where σflow, Dp, and Ap are respectively the flow stress (the mean value of the tensile and yield stresses), pre-cracked DWTT energy and ligament area of a pre-cracked DWTT specimen. On the other hand Pt and tw are the crack tip pressure and pipe wall thickness respectively. The crack tip pressure Pt is taken to be the choked pressure at the pipeline release plane. For brevity, full details of the coupling of the fracture and the fluid decompression models will be reported in a separate The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n 270 Fig.5.Variation of crack velocity with crack length for test ECSC X100 south-running crack. Inventory: air, initial pressure = 126bara, initial temperature = 20°C. Curve A: experimental data [33]. Curve B: DBFM prediction. study. The calculation algorithm automatically corrects for the effective pipeline length and hence the crack tip pressure as crack opens. The required fluid decompression velocity and the crack tip pressure are determined from the numerical solution of the conservation equations (1-3) using our CFD computational package, PipeTech [34]. The Peng-Robinson [35] equation of state (PR EoS) is used for the prediction of the pertinent fluid phase equilibrium data for both air and rich gas mixtures. In the case of CO2, the Modified Peng-Robinson [36] EoS is used. As compared to the PR EoS, this equation has been shown to produce better predictions of the phase equilibrium data during the most part of the depressurization process [37]. Results and discussion Validation The following shows the results relating to the validation of the dynamic boundary ductile-fracture model presented above, hereby referred to as DBFM, by comparison of its predictions against the following published experimental data: • HLP full-scale burst test [20] • ECSC X100 pipe full-scale burst test [21] • Alliance full-scale burst tests [22] Table 1 shows the pertinent conditions relating to each test. Table 2 on the other hand shows the rich gas feed compositions for HLP C2 and Alliance tests. The full-burst-test pipelines used comprised several sections of differing toughness for which the corresponding DWTT energy may be calculated. In all simulations, the pipe wall roughness and heat-transfer coefficient are taken as 0.05 mm and 5 W/(m2 K) respectively. The latter correspond to the uninsulated pipeline exposed to still air in all simulations. An equidistant grid system comprising 100 nodal points is employed for the fluid dynamic simulations using PipeTech. The corresponding discretisation time element is determined using 90% of the Courant, Friedrichs and Lewy value [38]. The HLP full-scale experiments involved three series of burst tests, referred to as test series A, B, and C using X70 API grade pipelines containing air and a rich gas mixture. Pipeline fracture was initiated using an explosive charge. Figure 1 shows a schematic representation of the pipe setup. Figures 2 to 4 show the variation of the crack velocity with crack length for the south-running A1, B1 and C2 tests respectively for the HLP full-scale experiments. Curves A show the measured crack length; Curves B, on the other hand are the simulation predictions. In all cases, the 271 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.6.Variation of crack velocity with crack length for test Alliance Test 1. Inventory: rich gas (Table 2), initial pressure = 120.2bara, initial temperature = 23.9°C. Curve A: experimental data [22]. Curve B: DBFM prediction. corresponding Charpy Energy, Cv, for each pipe section is given in the figures. Figures 5 and 6 show the corresponding data for ECSC X100 [21] and Alliance full-scale burst tests [22], respectively. Returning to Figs 2-6, as it may be observed, the crack velocity significantly decreases with increase in crack length. This is due to the significant rapid decrease in the crack tip pressure as the pipeline depressurizes. As an example, such behaviour expressed in terms of the variation of the cark tip pressure with time is shown in Fig.7 for the HLP A1 south-running crack. Also as expected, the crack velocity decreases as the crack propagates into the pipeline section with the higher toughness, eventually coming to rest in all cases. As expected, the data in Fig.4 show the smallest crack length as compared to the other tests due to the combination of the much higher fracture toughness pipe material employed together with the lowest initial pressure. The initial rapid increase in the crack velocity observed in many of the test data is due to the finite time taken for the initial notch to fully develop into an open flap following detonation. This time domain is ignored in the present simulations. Returning to the simulation data (curves A), given the experimental uncertainties, it is clear that in all cases the DBFM predictions produce reasonably good agreement with the test data. CO2 pipeline ductile fracture investigation The following describes the results of the application of the validated DBFM to the rupture of a hypothetical CO2 pipeline. To ensure practical relevance, the respective pipeline internal diameter and wall thickness of 590.7mm and 9.45mm are employed in the simulations. Cosham and Eiber [7] suggest that such dimensions are the most likely for CO2 pipelines to be employed in CCS. The same authors also propose that a Cv of 50 J would be sufficient to arrest a fracture for typical operating conditions of 100barg and 10oC. The same pipeline operating conditions are chosen in the proceeding simulations. Figure 8, curve A, shows the predicted variation of the crack velocity versus crack length based on the above conditions for the CO2 pipeline. For the sake of comparison, the analogous data for methane (curve B) and natural gas (85% methane-15% ethane, curve C) inventories are also presented. The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n 272 Fig.7. Simulated variation of crack tip pressure versus time for test HLP A1 south-running crack. As may be observed following an initial drop, the natural gas pipeline (curve C) exhibits a relatively constant high-velocity crack which propagates through almost the entire length of the pipeline before coming to rest at approximately 89m. Similar trends in behaviour but of a smaller magnitude is observed in the case of the methane pipeline (curve B) where the fracture comes to rest at a distance of approximately 18m. Of the three cases examined, the CO2 pipeline (curve A) offers the best resistance to ductile fracture. Here the fracture almost instantaneously comes to rest at a distance of only 6m. Impact of line temperature Based on an analysis of the CO2 depressurization trajectory, Cosham and Eiber [7] postulate that the initial temperature of the CO2 pipeline may have a significant impact on its resistance to ductile fracture failure. Figure 9 shows impact of the line temperature on the variation of fracture velocity versus fracture length for the CO2 pipeline at four selected temperatures of 30°C (curve A), 20°C (curve B), 10°C (curve C), and 0°C (curve D). As it may be observed, in the range 0-20°C, an increase in temperature results in a relatively modest increase in the fracture velocity and fracture arrest length. The data at 30°C (curve A) is an exception to this rule. Remarkably only a 10°C rise in the line temperature results in a fast-running propagating fracture which covers the entire length of the pipeline. To explain the above, Fig.10 shows the variation of the crack tip pressure with temperature for the starting line temperatures of 10, 20, and 30°C relative to the CO2 saturation curve. The latter is generated using the Span and Wagner [39] equation of state for CO2. The calculated crack arrest pressure of 43bara is also indicated in the same figure. In the case of the highest line temperature of 30°C (curve A), soon after pipeline failure, the dense-phase CO2 inventory crosses the saturation curve at the maximum pressure of ca. 60bara, some 17bara higher than the crack arrest pressure of 43bara, thus resulting in a propagating fracture. The fracture comes to rest once the crack tip pressure is equal to the crack arrest pressure. By the time this occurs in the case of the 30°C pipeline, the crack will have already propagated through the entire length of the pipe. Conclusion The development and validation of a dynamic boundary ductile-fracture-propagation model for pressurized pipelines was presented. The model, based on the coupling of a semi-empirical fracture model with the transient real fluid 273 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.8.Variation of crack velocity with crack length for a 100-m long pipe at 100barg and 10°C containing various inventories. flow simulator, PipeTech, takes into account all of the important fluid/structure interactions governing the fracture propagation and arrest process. A particularly important feature is accounting for the change in the effective pipeline length as the pipeline unzips and its impact on the crack tip pressure. Following its successful validation against real pipe burst data reported for air and rich-gas mixtures, the model is used to test the propensity of a hypothetical but realistic pressurized CO2 pipeline to ductile-fracture-propagation failure. Such investigations are particularly timely given the real prospects of using CO2 pipelines as part of the carbon capture and sequestration (CCS) chain. propagation. A relatively modest increase in the line temperature from 20 to 30°C resulted in a running fracture which propagated through the entire length of the 100m pipe. Analysis of the data revealed that upon crack initiation, the pipeline inventory rapidly transforms from the dense phase into the saturated state, thereafter following a relatively prolonged depressurization trajectory along the saturation curve. The crack will propagate for as long as the crack tip pressure remains higher than the crack arrest pressure. In the case of the CO2 pipeline at 30°C this cross over will not happen before the fracture has propagated through the entire length of the pipe. Obviously such phenomenon will have significant practical implications when transporting CO2 at different ambient temperatures as part of the CCS. Simulations conducted using a 100-m long pipeline containing methane, natural gas or CO2 at 100barg and 10°C revealed that whereas for the natural gas and methane inventories the fracture propagated through most of the pipe length, in the case CO2 pipe, the crack length was limited to only a short distance. The study assumes that the adopted rather simplistic but nevertheless effective drop-weight tear test (DWTT) energy approach validated for air and rich gas mixtures is also applicable to CO2. This assumption is justified given the fact that the only fluid parameter introduced into the DWTT approach is the crack tip pressure. The change in the temperature of CO2 however was found to have a remarkable impact on the resistance to fracture Furthermore the fluid-flow model employed is based on the plausible homogenous flow assumption in which the The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n 274 Fig.9.Variation of crack velocity with crack length for the 100-m long, 100barg CO2 pipe at different starting line temperatures. constituent fluid phases remain at thermal and mechanical equilibrium during the fracture-propagation process. Addressing the non-equilibrium phase behaviour, together with the impact of the typical impurities present within the CO2 stream for the various capture technologies on the ductile fracture behaviour of CO2 pipelines, is currently being investigated by the authors. References 1. A.Bartenev, 1996. Statistical analysis of accidents on the Middle Asia-Centre gas pipelines. J. Haz. Mat., 46, 1, pp57-69. 2. G.Papadakis, 1999. Major hazard pipelines: a comparative study of onshore transmission accidents. J. Los. Prev. Proc. Ind., 12, 1, pp91-107. 3. PHMSA, 2010. Pipeline incidents and mileage reports. Retrieved from primis.phmsa.dot.gov/comm/reports/safety/PSI.html. 4. B.N.Leis, X.Zhu, and T.Forte, 2005. Modelling running fracture in pipelines: past, present and plausible future directions. Proc. 11th Int. Conf. Frac., Italy. 5. W.A.Maxey, 1974. Fracture initiation, propagation and arrest. In: Proc. 5th Symposium in Line Pressure Research. Houston. 6. H.Mahgerefteh, G.Denton, and Y.Rykov, 2008. CO2 pipeline rupture. IChemE Symposium Series: HAZARDS XX Process Safety and Environmental Protection, 154, pp869 - 879. Manchester: IChemE. 7. A.Cosham and R.J.Eiber, 2008. Fracture propagation in CO2 pipelines. J. Pipe. Eng., 7, pp115-124. 8. M.Bilio, S.Brown, M.Fairwheather, and H.Mahgerefteh, 2009. CO2 pipelines material and safety considerations. IChemE Symposium Series: HAZARDS XXI Process Safety and Environmental Protection, 155, pp423–429. Manchester: IChemE. 9. H.Kruse and M.Tekiela, 1996. Calculating the consequences of a CO2-pipeline rupture. En. Conv. Man., 37, 95, pp1013-1018. 10. D.J.Picard and P.R.Bishnoi, 1988. The importance of realfluid behaviour and non-isentropic effects in modeling decompression characteristics of pipeline fluids for application in ductile fracture propagation analysis. Cana. J Chem. Eng., 66, 1, pp3-12. 11. G.Fearnehough, 1974. Fracture propagation control in gas pipelines: a survey of relevant studies. Int. J. Pres. Ves. Piping, 2, 4, pp257-282. 12. Idem., 1987. Calculation of the thermodynamic sound velocity in two-phase multicomponent fluids. Int. J. Multip. Flow, 13, 3, pp295–308. 13. H.Mahgerefteh and O.Atti, 2006. Modelling low-temperatureinduced failure of pressurized pipelines. AIChE J., 52, 3, pp1248-1256. 14. B.N.Leis, 1997. Relationship between apparent (total Charpy V-Notch toughness) and the corresponding dynamic crack- 275 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.10.The variation of crack tip pressure with temperature for the 100m long, 100-barg CO2 pipe at different starting line temperatures. propagation resistance. Proc. Int. Pipe. Conf., Calgary: ASME. 15. H.Makino, I.Takeuchi, and R.Higuchi, 2008. Fracture propagation and arrest in high-pressure gas transmission pipeline by ultra strength line pipes. In: 7th Int. Pipeline Conf.. Calgary. 16. P.E.O’Donoghue, M.F.Kanninen, C.P.Leung, G.Demofonti, and S.Venzi, 1997. The development and validation of a dynamic fracture propagation model for gas transmission pipelines. Int. J. of Pres. Ves. Piping, 70, 1, pp11-25. 17. Z.Zhuang, 1999. Analysis of dynamic fracture mechanisms in gas pipelines. Eng. Frac. Mech., 64, 3, pp271-289. 18. H.Makino, T.Sugie, H.Watanabe, T.Kubo, T.Shiwaku, S.Endo, et al., 2001. Natural gas decompression behaviour in high pressure pipelines. ISIJ Int., 41, 4, pp389-395. 19. A.Terenzi, 2005. Influence of real-fluid properties in modelling decompression wave interacting with ductile fracture propagation. Oil Gas Sci. Tech., 60, 4, pp711–719. 20. T.Inoue, H.Makino, S.Endo, T.Kubo, and T.Matsumoto, 2003. Simulation method for shear fracture propagation in natural gas transmission pipelines. In: International Offshore and Polar Engineering Conf., 5, pp121-128. Honolulu. 21. I.Takeuchi, H.Makino, S.Okaguchi, N.Takahashi, and A.Yamamoto, 2006. Crack arrestability of high-pressure gas pipelines by X100 or X120. In: 23rd World Gas Conference. Amsterdam. 22. D.M.Johnson, N.Horner, L.Carlson, and R.J.Eiber, 2000. Full scale validation of the fracture control of a pipeline designed to transport rich natural gas. Pipeline Technology, 1, 331. 23. H.Mahgerefteh, O.Atti, and G.Denton, 2007. An interpolation technique for rapid CFD simulation of turbulent two-phase flows. Pro. Safe. Env. 24. H.Mahgerefteh, A.Oke, and Y.Rykov, 2006. Efficient numerical solution for highly transient flows. Chem. Eng. Sci., 61, 15, pp5049-5056. 25. H.Mahgerefteh, P.Saha, and I.G.Economou, 1999. Fast numerical simulation for full bore rupture of pressurized pipelines. AIChE J., 45, 6, pp1191–1201. 26. A.Oke, H.Mahgerefteh, I.Economou, and Y.Rykov, 2003. A transient outflow model for pipeline puncture. Chem. Eng. Sci., 58, pp4591-4694. 27. M.J.Zucrow and J.D.Hoffman, 1975. Gas dynamics. Wiley, New York. 28. C.Bisgaard and H.Sørensen, 1987. A finite element method for transient compressible flow in pipelines. Int. J. Numer. Methods Fluids, 7, pp291-303. 29. E.Lang, 1991. Gas flow in pipelines following a rupture computed by a spectral method. J. App. Math. Phy., 42, March. 30. K.Bendiksen, D.Maines, and R.Moe, 1991. The dynamic twofluid model OLGA: theory and application. SPE Production, 6, 6, pp171–180. Society of Petroleum Engineers. 276 The Journal of Pipeline Engineering 36. D.Wu and S.Chen, 1997. A modified Peng-Robinson equation of state. Chem. Eng. Comm., 156, 1, pp215-225. 37. H.Mahgerefteh, G.Denton, and Y.Rykov, 2008. A hybrid multiphase flow model. AIChE J., 54, 9, pp2261–2268. 38. H.Mahgerefteh, Y.Rykov, and G.Denton, 2009. Courant, Friedrichs and Lewy (CFL) impact on numerical convergence of highly transient flows. Chem. Eng. Sci., 64, 23, pp4969-4975. 39. R.Span and W.Wagner, 1996. A new equation of state for carbon dioxide covering the fluid region from the triple-point temperature to 1100 K at pressures up to 800 MPa. J. Phys. Chem. Ref. Data, 25, 6. Sa no m t f ple or c di op st y rib ut io n 31. J.R.Chen, S.M.Richardson, and G.Saville, 1995. Modelling of two-phase blowdown from pipelines – 1: a hyperbolic model based on variational principles. Chem. Eng. Sci., 50, 4, pp695–713. 32. idem, 1995. Modelling of two-phase blowdown from pipelines – 2: a simplified numerical method for multi-component mixtures.Ibid., 50, 13, pp2173-2187. 33. H.Makino, I.Takeuchi, M.Tsukamoto, and Y.Kawaguchi, 2001. Study on the propagating shear fracture in high strength line pipes by partial-gas burst test. ISIJ Int., 41, 7, pp788-794. 34. PipeTech. Pipeline rupture simulation software: www. pipetechsoftware.com. 35. D.Peng and D.B.Robinson, 1976. A new two-constant equation of state. Ind. Eng. Chem. Fund., 15, 1, pp59-64. 4th Quarter, 2010 277 Will fractures propagate in a leaking CO2 pipeline? by Dr Robert Andrews*1, Dr Jane Haswell2, and Russell Cooper3 1 BMT Fleet Technology, Loughborough, UK 2 Pipeline Integrity Engineers, Newcastle upon Tyne, UK 3 National Grid Gas Transmission, Warwick, UK A Sa no m t f ple or c di op st y rib ut io n HYPOTHETICAL CONCERN has been raised that leaks in a CO2 pipeline could escalate to a propagating fracture.This is due to the potentially large temperature drop associated with the expansion of either gaseous or dense-phase CO2 to ambient conditions. It is suggested this local cooling would lower the pipe wall temperature to an extent that a brittle fracture would initiate followed by a transition to a propagating fracture. Although such a mechanism could theoretically occur in natural gas pipelines, there is increased concern for CO2 transport because of the different thermodynamic behaviour of the contents, particularly for dense-phase transport. This paper critically reviews the literature associated with this postulated failure mechanism and other studies on the cooling of cracks and holes by escaping fluid. It is concluded that pipelines constructed to modern standards are not at risk. Limited crack extension may occur when the leak is through a �tight’ crack in a material of low toughness. However, the crack will arrest as it enters warmer material remote from the leak. Escalation to a propagating fracture can be controlled using methods which are widely used and understood in the pipeline industry. Introduction Background With the current drive towards reducing emissions of carbon dioxide (CO2) gas to the atmosphere, pipeline operators and designers are investigating options for the transport of bulk quantities of CO2. This will involve moving CO2 from sources such as power stations equipped with carbon capture to storage sites such as aquifers or depleted oil or gas fields. This will require either the construction of new pipelines, or the re-use of existing pipelines originally constructed to carry natural gas or other fluids. A concern has been raised that leaks in a CO2 pipeline could escalate to a propagating fracture. This is due to the potentially large temperature drop associated with the expansion of either gaseous or dense-phase CO2 to ambient conditions. It is suggested this cooling would lower the pipe wall temperature to an extent that a brittle fracture This paper was presented at the First International Forum on Transportation of CO2 by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical and Clarion Technical Conferences, and with the support of the University of Newcastle and the Carbon Capture and Storage Association. *Author’s contact details tel: +44 (0)1509 621814 email: [email protected] would occur followed by a propagating fracture. Although such a mechanism could theoretically occur in natural gas pipelines, there is increased concern for CO2 pipelines because of the different thermodynamic behaviour of the contents, particularly for dense-phase transport. There is no public domain evidence that such a failure mechanism has occurred either in natural gas pipelines in service or in laboratory experiments. The postulated failure mechanism involves a complex interaction of the thermo-fluid mechanics of a leaking pipeline, heat transfer driven by cold fluid escaping through either a crack or a hole, crack initiation and propagation, and crack arrest. These issues are discussed separately, after a summary of the postulated failure mechanism in the next section. The issues considered are grouped as fluid flow and heat transfer at a leak, fracture initiation, immediate arrest after initiation, and full-bore fracture propagation, and these are considered successively in the subsequent sections. This arrangement of the material does involve some repetition, but it was considered the best approach to ensure that all issues are covered. A general discussion and conclusions then follow. It should be noted that this paper has been prepared in the context of UK pipeline design practices, although most of the analysis and the conclusions should be appropriate for other locations. 278 The Journal of Pipeline Engineering Terminology The following terminology has been used to distinguish between different possible types of defect. It is important that these distinctions are maintained, as there can easily be confusion between the different possible types of loss of containment in the pipe wall. • Leak – a break in the pipe wall through which fluid can escape. This is used as a generic term. A leak is stable at the current length, applied stress and metal temperature. • Crack – a sharp-ended feature which can be analysed by fracture mechanics methods which characterize the crack tip stresses as a singularity. • Hole – a rounded feature, typically (but not necessarily) circular. Fracture mechanics’ methods are not applicable to holes. • Although not completely clear in the published papers, it appears to be assumed in [2] that once the condition in Eqn (1) is satisfied, a propagating fracture occurs immediately. In [2] this is described as “a secondary more catastrophic running brittle fracture”; in [1] both ductile and brittle propagating fractures are described. No experimental evidence is given to support this mechanism. In [3] it is suggested that this mechanism was the cause of the gas pipeline failure at Ghislenghien, Belgium, in July 2004. At the time of writing this incident is still the subject of litigation and the official Belgian government enquiry has not issued a report, so it is not clear if this speculation is correct. Sa no m t f ple or c di op st y rib ut io n • Rupture – a break in the pipeline which has an opening equivalent to at least the full-bore area. A rupture may remain stable or may escalate to a propagating fracture. • were Kc is described as “the critical fracture toughness below which a fracture propagates”, Y is a “shape factor depending on the crack length and geometry”, a is the crack half length, and σ is “the sum of the pressure and thermal stresses”. • Propagating fracture – a fracture moving continuously along the pipeline at a velocity of around 100 m/s or higher. Postulated failure due to local cooling at a leak The failure mechanism postulated in [1, 2] can be summarized as follows: • Fluid (usually a gas but possibly a dense-phase fluid) escapes through a leak in the pipeline wall, expanding and cooling. • The low temperature fluid cools the pipe wall at the leak. • Cooling a carbon steel reduces its toughness as the material becomes brittle at lower temperatures. Additionally, the temperature differential between cold material at the leak and warmer areas remote from the leak induces thermal stresses. • The combined effect of reduced toughness and thermal stress initiates a fracture. The condition for fracture is set in reference [2] by a pure linear elastic fracture mechanics (LEFM) approach, with fracture assumed to occur when: (1) Heat transfer and fluid flow at a leak As noted above, heat transfer between the leaking fluid and the pipe wall is essential if the material around the leak is to be cooled and possibly reach the lower shelf of the ductilebrittle transition curve. This section considers these issues. Heat transfer and flow at a leak in a pipeline The model in [2] assumes that flow both within the pipeline and through the leak in the wall is turbulent and that heat transfer is by forced convection. The leak is considered to be a 5-mm diameter hole, with a crack extending from it for 50mm. It is not clear if leakage is assumed through only the hole, or through both hole and crack. Heat transfer coefficients are estimated using coefficients from the literature. Although not clearly stated, it appears that the flow through the leak is treated in the same way as flow within the pipe. The model also does not allow for heat recovery from the surrounding soil as [2] refers to natural and forced convective heat transfer “to ambient”. Whilst in the immediate vicinity of the leak the pipe may be exposed, even here it will be surrounded by escaped gas rather than the ambient. Remote from the leak the pipeline will be buried and heat transfer would be expected to be by conduction from the soil, not by a convective mechanism. An analysis has been carried out for leaking ethylene pipelines in [4]. This analysed flow and heat transfer at circular holes in the wall of ethylene pipelines, with contents in both the gas and dense phases. In most cases the flow through the hole was choked and the bulk of the fluid expansion actually occurred in a shock wave outside the pipe wall. 4th Quarter, 2010 279 Sa no m t f ple or c di op st y rib ut io n The fluid flowing through the hole was thus at a similar temperature to the bulk fluid and so there was little local cooling of the material around the hole. The exterior of the pipe was surrounded by a cloud of cooled gas, but the heat transfer from this gas to the pipe wall was relatively low. Other points from this study were: Fig.1. Predicted local wall temperatures from [4] during blowdown of a supercritical ethylene pipeline through a 50-mm diameter hole at either the top or bottom of the pipeline. greatest cooling occurs when the pipeline is shut-in so that the bulk fluid pressure and temperature fall. For the more likely case of a small undetected leak, where flow continues, the temperatures return to ambient over shorter distances as the pipe wall is heated by the warm flowing bulk contents. • Steady-state pipe metal temperatures were reached within 1 minute, rather than taking up to 10 minutes (the timescale predicted in [2]) to reach a steady state. • Figure 1 shows predicted wall temperatures from [4] as a function of pressure during blowdown of a supercritical ethylene pipeline. The lowest predicted temperature is -58°C but this occurs at only approximately 25% of the initial pressure, which would considerably reduce the driving force for fracture initiation. • For supercritical conditions, Fig.1 shows that the location of the hole affects the temperatures and behaviour once the pressure falls to the boiling point and two-phase flow is established. This is because – with a hole at the bottom of the pipe – relatively warm liquid is forced through the hole; with a hole at the top, colder gas escapes from the hole. • The model in [2] appears to assume the pipe is above ground, as there are references to convective heat transfer. In contrast, Saville [4] explicitly models a buried pipeline and includes heat recovery from the soil. Crater formation at the leak due to displacement of the soil by the escaping fluid was also considered. Although there are differences in the treatment of the local fluid flow and heat transfer, both models show that the metal temperature recovers to the bulk fluid temperature over relatively short distances. Saville’s model predicts a more rapid recovery as there is less local cooling. Both models also predict that the Experimental studies on leakage from full-scale pipelines are rare, as they are difficult and expensive to perform. One relevant study was carried out by SZMF [5] on a 1067-mm diameter, 33.5-mm wall thickness grade L555 (X80) vessel simulating a storage vessel for compressed natural gas (CNG) transport applications. Leakage of lean natural gas through a fatigue crack produced a lowest measured temperature (on the outer surface near the crack tip) of -70°C, in agreement with predictions made by an unknown method, and these predictions are reproduced in Fig.2. They show a significant through-wall temperature gradient, with the crack at inner surface of the vessel at around -10°C. The cooling is greater than [2] predicts for natural gas, albeit starting from a higher pressure of 180 bar. This result suggests that leakage through cracks may be more severe than leakage through holes. Flow through cracks Workers at Sheffield University have studied the flow of fluids through narrow cracks, rather than through a hole [6-8]. Whilst they did not study heat transfer effects, and the pressure differentials were small, they did identify an effect of surface roughness on the flow regime. At very small openings, comparable to the surface roughness, the flow was essentially laminar and “followed” the local surface roughness. With increasing separation of the crack faces, the flow could move between the peaks of the surface and transitioned to turbulent flow at higher velocities. This work shows that there may be complex effects of the surface and crack opening; the heat transfer between the fluid and the crack would be affected by the flow regime and the fluid velocity. 280 The Journal of Pipeline Engineering the leak, and so assuming a uniform stress will overestimate the stress-intensity factor. The analysis also incorrectly treats the thermal stress as a primary stress, rather than secondary. Fig.2. Predicted temperatures at a through-wall fatigue crack due to leakage of natural gas. Figure 3 of [5], Reepmeyer, Lothe,Valsgaard, Erdelen-Peppler and Knauf: Full-scale gas leak test at a large-diameter X-80 DSAW pipe. ASME International Pipelines Conference 2006, IPC2006 10005. Copyright ASME 2006, used with permission. Nuclear industry studies The model in [2] assumes the presence of a sharp crack in all cases. Work on the failure of volumetric corrosion defects in low-toughness linepipe [12, 13] has shown a substantial influence of defect geometry on the failure behaviour. The most significant factor appears to be the local defect geometry, in particular the acuity of the defect. The tests were carried out using material that was on the lower shelf of the Charpy transition curve at room temperature, but the behaviour of blunt machined defects simulating volumetric corrosion at quasi-static strain rates was predicted by conventional ductile-failure models. The trend of the results is shown in Fig.3, which shows the transition temperature as a function of the stress-concentration factor associated with the defect. The stress-concentration factor increases as the defect becomes sharper, showing that the effective transition temperature of a rounded defect can be well below the Charpy transition temperature. Thus, for leakage through holes rather than cracks, it is unlikely that fracture initiation would occur, even in low-toughness material. Sa no m t f ple or c di op st y rib ut io n The nuclear industry has studied the flow of fluids through cracks as part of �leak before break’ (LBB) arguments, which aim to show that a through-wall crack will leak before breakage or rupture occurs. A part of these arguments is to show that a leak can be detected before further sub-critical crack growth leads to rupture; this requires an estimate of leakage rates, and much of this work is summarized in [9]. However, the nuclear LBB studies have concentrated on isothermal leakage, presumably because the containment is effectively at constant temperature. This is not relevant to the current issue for CO2 pipelines where heat transfer to the containment (the pipe) from the fluid expanding in a non-isothermal manner is the key issue. The fracture analysis should, in any case, use more modern methods such as the failure-assessment diagram approach of [10] or [11]. These can take account of thermal-stress gradients if these are significant. However, it is true that a carbon steel pipeline will show transition behaviour and, depending on the material properties, local cooling may reduce the toughness. Fracture initiation This Section considers issues associated with the analysis and prediction of fracture initiation from a locally cooled area of a pipeline. Fracture analysis methods The model in [2] assumes LEFM applies, so that the initiation of fracture is controlled by the stress-intensity factor; the paper also assumes a step transition between ductile and brittle behaviour, although this is an oversimplification of the true behaviour of a pipeline steel. Ignoring the beneficial effects of warm pre-stressing, discussed below, the behaviour of the pipeline should reflect the low constraint of a thin-walled structure loaded in tension. Thus, even on the lower shelf, the behaviour is likely to be better than would be expected from simple LEFM considerations. Fracture initiation should be predicted using an elastic-plastic measure of toughness such as the crack-tip-opening displacement (CTOD) or the J-integral. There are other concerns with the fracture-analysis method in [2], as the authors appear to assume that the thermal and primary stresses can simply be added to give a higher uniform hoop tensile stress when calculating the stressintensity factor. The thermal stress will decay remote from Crack initiation due to local cooling Fearnehough [14] carried out experiments using liquid nitrogen vapour to cool steel plates containing a centre crack of length 600mm and subject to remote axial tension stresses in the range 15 to 124N/mm2. The aim was to simulate the effects of spilling LNG onto the outer containment of a storage tank. The material tested was a grade 43A structural steel, which is approximately equivalent in strength to a Grade L290 (X42) pipeline steel. The reported 30-J Charpy impact temperature was -15°C; as the material thickness was 6.3mm, this would have passed the typical UK gas transmission pipeline requirement of 27J at 0°C in a 2/3 specimen. Local cooling to around 80°C below the bulk metal temperature was required to initiate cracks, which then arrested. It was argued that the fracture was initiated by the additional thermal stress generated by the local cooling and the applied remote stress level was not a factor. It should be noted, however, that the stresses used in this study were relatively low by the standards of many pipelines. The experimental study on a Grade L555 vessel [5] discussed above shows that fracture initiation will not 4th Quarter, 2010 281 Sa no m t f ple or c di op st y rib ut io n Fig.3. Predicted variation of the transition temperature in a brittle linepipe for blunt corrosion defects as a function of the elastic stress concentration factor.Trend of results from [12]. Fig.4. British Gas relation between temperature and stress level for brittle crack arrest, showing operating point for case study. necessarily occur at a sharp crack, even when the material is subject to significant local cooling. The fatigue crack in the vessel did not propagate and remained stable throughout the experiment. This was a modern hightoughness material, with an upper-shelf Charpy energy exceeding 400J at -10°C, so it is possible that it was not on the lower shelf at –70°C. An arrested crack with a larger opening would be closer to a hole than a crack. If, as the work considered above suggests, there are differences in the fluid flow and heat transfer of holes and cracks, a larger defect would behave in a manner similar to a hole and there would be less cooling. This would reduce the likelihood of subsequent re-initiation after an arrest. Immediate arrest after initiation The study by Fearnehough [14] shows that cracks initiated by local cooling will arrest when they run into warmer material with a higher toughness. The amount of crack extension, or equivalently the position of the crack arrest, was considered to be a function of both stress level and temperature, as the propagation distance was greater in the specimens subjected to higher remote tension. This observation that arrest was a function of stress and temperature is consistent with pipeline experience with propagating brittle fractures [15]. This section considers crack arrest after initiation from a locally cooled leak. This is important, as the postulated failure mechanism described above does not clearly allow for this. Whilst the enlargement of a small leak to a bigger leak, or a full-bore rupture, is not desirable, such an escalation is preferable to the formation of a propagating fracture. 282 The Journal of Pipeline Engineering The West Jefferson test was developed for brittle crack arrest tests on pipeline steels and is an example of brittle crack arrest occurring under a constant applied stress. In this test a through-wall crack is initiated in a pressurized vessel (so the stress ahead of the crack is constant at the hoop stress) and propagates along the length of the pipe. The fracture appearance and the temperature at the point of crack arrest correlate well with full-scale test results and the results of the Battelle drop-weight tear test (DWTT). The West Jefferson methodology has been used for many years and is not dependent on the pipe contents – in fact typically the tests fill the pipe section with 95% of water and pressurize the remaining space with air or nitrogen to reduce the stored energy in the vessel. Fracture propagation and arrest It is possible that cooling of the remaining contents during blowdown following a full-bore rupture could lower the temperature of the pipe wall below the DWTT transition temperature. However, this cooling would also be accompanied by a reduction in pressure which would reduce the hoop stress. Again, the relationship between arrest temperature and hoop stress shown in Fig.4 could be used if the pressure–temperature trajectory is known from simulations of the depressurization. An alternative approach to predicting brittle fracture arrest in a blowdown would be that proposed by Battelle [17], which relates the arrest behaviour to the Charpy transition curve. Sa no m t f ple or c di op st y rib ut io n The phenomena of both brittle and ductile fracture propagation in pipelines are well understood and can be analysed by accepted methods. The two types of propagating fracture are considered separately below. It should also be noted that propagating fractures can originate from pipeline damage, often due to mechanical interference or corrosion, that does not involve the postulated cooling and fracture mechanism. Pipeline designers already design against propagating fractures for this reason. It may be necessary to modify the design methods to take account of the specific decompression behaviour of CO2, but these changes will be required for a CO2 pipeline in any case. Thus it is considered that, provided a pipeline is operated above its DWTT transition temperature, a propagating brittle fracture should not escalate from a leak, even when operating at stress levels up to 72% SMYS. If an old pipeline has a high DWTT transition temperature, it could be operated at lower stress levels using the relationship between arrest temperature and hoop stress developed by Fearnehough [15] and included in Edition 2 of IGE/TD/1 [16]. This relation is shown in Fig.4. Propagating brittle fractures These fractures are characterized by little plastic deformation and are driven by the elastic stress in the pipe wall. The fractures travel at axial velocities above the acoustic velocity in the fluid. As a result, the crack tip is continually propagating into a region where the pipeline contents are undisturbed. These cracks cannot be affected by rapid cooling of escaping fluid as they are moving faster than the fluid can depressurize and cool. The steel ahead of the crack will remain essentially at the ambient temperature prior to a leak occurring. Thus a propagating brittle crack in a CO2 pipeline will behave in the same manner as a propagating brittle crack in a natural gas pipeline. The standard fracture-control approach of ensuring that the pipe steel is operating above its DWTT transition temperature will prevent propagating brittle fractures. A DWTT requirement has been routinely specified for gas transmission pipelines in the UK since the introduction of Gas Council specifications for steel linepipe in the late 1960s, and most existing UK gas transmission pipelines will satisfy this requirement. Pipelines pre-dating this with a high DWTT transition temperature would require special consideration, but the approach would be the same as has been used for natural gas pipelines since the publication of Edition 2 of IGE/TD/1 [16]. For new build, any competent pipe supplier should be able to produce pipe meeting the usual DWTT requirement of 85% shear area for design temperatures around 0°C. Propagating ductile fractures The phenomenon of propagating ductile fractures in gas pipelines has been known for over 30 years, and has been extensively studied. Rothwell gives a good overview of these studies [18]; most recent work has concentrated on testing very high strength linepipe with yield strength over 690N/ mm2 [19] and rich gas mixtures [20]. Propagating ductile fractures typically run at axial velocities in the range 200300m/s, although in some full-scale tests velocities down to 100m/s have been measured. These velocities are less than the acoustic velocity in the fluid, and the crack is driven by the pressure of the escaping fluid acting on the �flaps’ developing behind the crack tip as the pipe cracks. At these velocities there will be negligible heat transfer to or from the crack tip area, and so the cooling effects described in Section 2 will not occur. The steps required to control propagating ductile fractures in a gas pipeline are also well understood, so that formal fracture control plans are now explicitly required in some pipeline design codes such as the Australian code AS 2885 [21]. UK codes such as TD/1 [22] and PD 8010 [23] do not have such an equivalent explicit requirement, but the material toughness requirements in the codes will achieve a high level of resistance to propagating fractures. The special requirements for CO2 pipelines discussed in [1] are well known in the pipeline industry. Maxey identified them in [24] and by 1990 crack arrestors had been installed on the Canyon Reef Carriers CO2 [25] pipeline in west Texas because of concerns over ductile crack propagation. Whilst some of the effects of rapid decompression of dense-phase CO2 are counter-intuitive, they can be predicted [26]. The 4th Quarter, 2010 283 effects of impurities in the gas have also been considered [27], but again the fundamental approach to fracture control remains unchanged. The greatest uncertainty appears to be in predicting the decompression behaviour of the contents. If necessary, experimental confirmation of predictions by shock tube testing may be required. It is accepted that the control of propagating ductile fractures in a CO2 pipeline may be more difficult than in a typical UK onshore natural gas pipeline, particularly for a dense-phase pipeline. However, the methodology for control is known. For a new build, these problems should be manageable with correct specification of the material properties. If an existing pipeline is being converted to transmit CO2, then an assessment of the toughness will be required to show that propagating fractures can be controlled. General discussion The available work suggests there may be differences between the cooling experienced at a hole and that at a �tight’ fatigue crack. If [4] is correct and the flow through a hole is choked so that the bulk of the expansion and cooling takes place outside the pipe wall, there will be little cooling at a leak through a hole. If this is the case, the likelihood of the postulated fracture mechanism occurring in practice reduces. In contrast, the test with a fatigue crack reported in [5] suggests that expansion through a tight crack may generate a substantial temperature drop in the material. The work discussed in above also suggests that there are differences in flow behaviour for tight cracks which would affect heat transfer. This difference requires further investigation. Fracture initiation The available evidence shows that if a fracture should initiate from a locally cooled crack, it will arrest as it extends into warmer material. Fearnehough’s tests [14] conclusively demonstrated this effect, and crack arrest is included in the R6 procedure [10]. Generally it is assumed that crack arrest can occur when either the growing crack has a decreasing stress intensity factor or when the crack grows into an area of increasing fracture toughness. In the present case, the fracture toughness will increase as the crack is growing into warmer material. A longer crack would have an increased stress-intensity factor simply due to the greater crack length, but would have grown away from the high stresses associated with the locally cooled area. Sa no m t f ple or c di op st y rib ut io n Heat transfer It was noted above that there were more-realistic fractureassessment methods than the simple LEFM approach used in [2]. A further consideration is the effect of local constraint. It is now widely accepted that pipelines are a �low-constraint’ structure, as they are thin and defects are loaded predominantly in tension. This has the effect of increasing the effective toughness above that measured in a standard fracture-toughness test. Constraint-based arguments have been used to assess the behaviour of longitudinal seam-weld defects [29] and pipeline girth welds [30]. These methods can be used to assess the behaviour of through-wall cracks in a pipeline subject to local cooling. Methods of including constraint effects in fracture assessments are given in Section III.7 of R6 Rev 4 [10]. The available evidence for fracture initiation due to local cooling is conflicting. Fearnehough’s tests [14] using liquid nitrogen vapour showed that fracture can be initiated at a locally cooled crack, and there are other cases known of failures occurring when components are cooled under load: for example the failure of a heat exchanger in the Longford incident [28] was attributed to a combination of thermal shock and local embrittlement of cold material. However, the test in [5] showed that fracture will not necessarily occur under these conditions, and tests on simulated corrosion defects discussed above have shown that the defect acuity appears to be a major factor. A combination of a substantial temperature drop, a sharp crack, and a material with a high transition temperature is required to initiate a fracture due to local cooling of a leaking pipeline. Quantifying these effects will require understanding of the temperatures generated at a leak, together with fracture analyses. It is possible that a crack may grow due to the postulated local embrittlement mechanism and reach a length exceeding the stable through-wall crack length for the pipeline. This would cause the leak to escalate to a rupture. The critical through-wall crack length can be calculated using the standard methods for an axial crack in a pipeline. Fracture propagation As noted above, it is considered that accepted techniques can be used to assess fracture propagation in CO2 pipelines, and indeed these methods will be required to control propagating fractures originating from other forms of damage such as external interference. It is understood that work is being carried out elsewhere to address issues specific to the control of propagating fractures in CO2 pipelines, in particular the rapid decompression behaviour of CO2. Warm pre-stressing Warm pre-stressing is a phenomenon where the low temperature fracture toughness of a material is improved by a prior loading at a higher temperature. Various explanations exist for the effect, such as crack-tip blunting at the higher temperature, the generation of beneficial residual stresses at a crack tip by the prior loading, or metallurgical effects such as the de-cohesion of brittle second phase particles during the high temperature loading. Guidance on the application of warm pre-stressing arguments is given in Annex O of BS 7910 [11] and Section III.10 of R6 Rev 4 [10]. Rio Pipeline FP1 2011 Conference & Exposition September 20-22 Sa no m t f ple or c di op st y rib ut io n 11/10 Rio de Janeiro • Brazil Call for Papers: Submission Deadline December 17, 2010 Participation Information: Phone.: (+55 21) 2112-9000 Fax: (+55 21) 2220-1596 e-mail: [email protected] www.riopipeline.com.br Organization / Realization 4th Quarter, 2010 285 For a defect created in an operating pipeline, without unloading and re-loading, crack tip residual-stress effects cannot be significant. Thus the operating mechanisms are likely to be crack-tip blunting and, possibly, metallurgical effects. There is ample evidence that brittle and transition fracture toughnesses are strongly influenced by notch acuity: all the fracture-toughness testing standards require a fatiguesharpened crack, because this gives lower-bound results. Limits are set on the maximum stress intensity factor Kmax when fatigue pre-cracking fracture-toughness specimens to ensure the crack is not artificially blunted before the main fracture test. The stress acting on a crack before cooling can be considered to be analogous to a fatigue pre-cracking cycle with a high Kmax, leading to crack-tip blunting and an elevation of the fracture toughness. It can be argued that the initial load due to internal pressure before the onset of any cooling at a leak is the warm pre-stress, and the additional local thermal stress generated by cooling is an additional load, giving a �load – cool – fracture’ case. The applied stress-intensity factor will be increasing during the cooling due to thermal stresses adding to the driving force due to pressure loading (which will remain constant if the leak is small and undetected). As a result the simplified warm pre-stressing argument, that failure is avoided if the stress-intensity factor is monotonically falling during cooling, will not hold, and a more-detailed analysis is required. Case study Crack initiation To assess the possibility of fracture occurring at a sharp crack under these conditions, the fracture resistance of the material at -17°C must be estimated. Unless test data exist, it is only possible to estimate the effects of cooling from the linepipe specification acceptance criteria. Such an approach should be conservative, as the actual material properties should be better than the specification minima. LX/1 compliant material would have achieved 27J Charpy energy at 0°C in a 2/3 size specimen, and would have 75% shear area in the DWTT at 0°C. Unfortunately this information cannot be used directly to provide the toughness at a different temperature. The approach developed by Battelle for brittle fracture arrest predictions [17] can be used, as follows. Sa no m t f ple or c di op st y rib ut io n In the authors’ opinion it is not appropriate to claim credit for the commissioning hydrotest as a prior warm load if warm pre-stressing arguments are used. This is because most defects will have been introduced after the hydrotest; for example mechanical damage will have produced a gouge or dent-gouge defect at some time after commissioning. Hence the hydrotest will have stressed defect-free material, and there can have been no effect of the hydrotest on a crack-tip region which did not exist at the time of the test. of 0.56°C per bar, the maximum cooling associated with a leak would be 21°C. If the operating temperature is +4°C, the metal temperature could reach as low as -17°C at a leak, if perfect heat transfer occurred and there is no heat recovery from the bulk contents or surrounding soil. This section applies the arguments developed in this paper to a case where an existing pipeline is being re-used to transmit CO2. Such re-use is environmentally beneficial as it avoids the carbon costs associated with the construction of a new pipeline. Figure 6 of [17] can be used to estimate that the Charpy transition temperature in a 2/3 size specimen would be 18°F (10°C) below the DWTT transition temperature, which is known to be 0°C or lower for LX/1 material. Hence the estimated Charpy transition temperature is -10°C, and so when fully cooled, the material is operating at 7°C below the transition temperature. Using Fig.5 of the Battelle report, the shear area in a 2/3 size Charpy specimen tested at 7°C below the transition temperature is 72%. At this shear area Fig.7 then shows that the impact energy is 75% of the maximum upper-shelf energy. Thus the minimum Charpy toughness of the material at a fully cooled leak would be predicted to be 0.75 x 27 = 20J. In practice, the material upper-shelf energy would be above the specification minimum, and it is likely that the transition temperature would be lower than that estimated, so the 20-J estimate is conservative. The effect of reducing the impact energy from 27J to 20J is to reduce the critical through-wall crack length calculated using the NG-18 toughness-dependent model [31] by about 10%, from 350mm to 320mm. Given the conservative assumptions in the analysis, and the likely benefits of warm pre-stressing, it is judged that this difference is not significant and that brittle fracture initiation is unlikely to occur for this specific case. Basic data The pipeline is an existing gas transmission pipeline being reused to transmit gas phase CO2 at 38bar. This pipeline is a 36-in (914mm) nominal diameter line with a nominal wall thickness of 12.7mm, and was constructed from pipe complying with the former British Gas LX/1 specification. Thus the material would be expected to have a DWTT transition temperature of 0°C. The hoop stress under the proposed operating conditions is 137N/mm2. For an X60 material this is equivalent to 33% SMYS based on the nominal wall thickness. Assuming Joule-Thomson cooling The change in toughness predicted by the analysis in the previous paragraph is a function of the temperature change and, to a limited extent, the wall thickness. The pipe wall thickness affects the shift between the Charpy transition temperature and the assumed DWTT transition temperature of 0°C for LX/1 compliant material. Smaller shifts will occur for thinner pipe; other cases will require specific evaluations. For modern linepipe, the degree of cooling estimated in this case would not be expected to be a problem, as the material would be expected to still be on the upper shelf at temperatures around -20°C. 286 The Journal of Pipeline Engineering It should also be noted that in Fearnehough’s tests on steel plates [14] the highest stress was 124N/mm2, comparable to the nominal stress of 137N/mm2 in this case. A cooling of over 100°C was required to initiate fracture in this test. The cracks in these tests were 600mm long, greater than the tolerable through-wall crack length for this pipeline. Crack propagation Using the British Gas relation between hoop stress and temperature shift below the DWTT transition temperature, shown in Fig.4, it can be seen that the operating point at a hoop stress of 33% SMYS and a temperature shift of 17°C below the DWTT transition temperature is just above the 95% probability of arrest line. Thus it is considered that even if a long length of pipeline were cooled to a low metal temperature, a propagating brittle fracture is not a credible event under the proposed operating conditions. The issues regarding a postulated failure mechanism for leaking CO2 pipelines have been reviewed. The main conclusions are: • It is considered possible that under a restricted range of circumstances a leak in a CO2 pipeline may enlarge due to local cooling causing a brittle fracture. Such enlargement is only likely to occur for a sharp crack in a material with a high ductile–brittle transition temperature in a pipeline operating at a high pressure. • The main uncertainty in assessing if such enlargement could occur is in the understanding of fluid flow, heat transfer, and cooling associated with the leak. This will be strongly influenced by the leak geometry. • Industry-standard fracture-assessment methods can be used to determine if fracture will occur at a leak. These methods can take account of elastic– plastic fracture and constraint effects to eliminate unnecessary conservatism. Sa no m t f ple or c di op st y rib ut io n Further work Concluding remarks The greatest unknown in assessing whether it is possible for a leak to escalate to a rupture due to the postulated local cooling mechanism is the amount of cooling generated at the leak. Once the cooling is quantified, given knowledge of the material’s toughness transition curve, fracture-mechanics’ methods can be used to assess if brittle fracture will occur. The published work suggests that the amount of cooling is strongly influenced by the geometry of the leak. Work is required to address this issue. It is suggested that the heat transfer and local cooling issue will require both numerical and experimental work. Whilst numerical studies can give an indication of trends and the important parameters, it is considered that experimental verification is required. An initial study using computational-fluid-dynamics’ (CFD) methods could be used to estimate leakage rates and heat-transfer coefficients; this would be combined with finite-element heat-transfer calculations for the steel pipe wall and stress analysis to determine the thermal stresses. The analyses should consider a range of hole sizes and crack geometries, up to a limiting length of the critical through-wall crack length for the geometry and material. Fracture analyses would be carried out using the results from this study. The results of the numerical work would be used to design an experimental programme. This is likely to require tests on a tough material, where crack growth is not expected, and a low-toughness material with a high ductile–brittle transition temperature where brittle crack extension is predicted. The tests should also include a range of defect acuity, from a round hole to a tight fatigue crack. Instrumentation of the tests should include temperature, strain, and crack growth. • The beneficial effects of warm pre-stressing should be quantified and included in future assessments. • For the specific case study of an existing pipeline being re-used to carry gas phase CO2 at 38bar, it is considered that crack extension by the postulated cooling and embrittlement mechanism will not occur. Acknowledgements This study was funded by National Grid Gas Transmission, which is gratefully acknowledged. References 1. M.Bilio, S.Brown, M.Fairweather, and H.Mahgerefteh, 20009. CO2 pipelines material and safety considerations. In: Hazards XXI - Process Safety and Environmental Protection in a Changing World. IChemE, Manchester, pp423-429. 2. H.Mahgerefteh and O.Atti, 2006. Modelling low-temperatureinduced failure of pressurized pipelines. AIChE Journal, 52, 3, 1248-1256. 3. Ibid., 2006. An analysis of the gas pipeline explosion at Ghislenghien, Belgium. American Institute of Chemical Engineers, Orlando, FL, USA. 4. G.Saville, S.M.Richardson, and P.Barker, 2004. Leakage in ethylene pipelines. Process Safety and Environmental Protection, 82, 1, 61-68. 5. O.Reepmeyer, P.Lothe, S.Valsgard, M.Erdelen-Peppler, and G.Knauf, 2006. Full scale gas leak test at a large diameter X-80 DSAW pipe. Paper IPC06-10005, vol. 3, Part A, American Society of Mechanical Engineers, Calgary, AB, Canada, pp1-8. 6. N.M.Bagshaw, S.B.M.Beck, and J.R.Yates, 2000. Identification of fluid flow regimes in narrow cracks. Proc. Institution of 4th Quarter, 2010 287 20. D.M.Johnson, N.Horner, L.Carlson, and R.J.Eiber, 2000. Full scale validation of the fracture control of a pipeline designed to transport rich natural gas. In: Proc 3rd Int. Pipeline Technology Conference, I, Bruges, Belgium, Elsevier Scientific, Amsterdam, pp191-209. 21. Standards Australia, 2007. Pipelines - gas and liquid petroleum, Part 1: design and construction. AS 2885.1: Standards Australia, Homebush, New South Wales. 22. IGE, 2008. Steel pipelines and associated installations for high pressure gas transmission. IGE/TD/1 Edition 5. Institution of Gas Engineers and Managers, Loughborough. 23. BSI, 2004. Code of practice for pipelines - Part 1: Steel pipelines on land PD 8010-1. British Standards Institution, London. 24. W.A.Maxey 1986. Long shear fractures in CO2 lines controlled by regulating saturation, arrest pressures. Oil and Gas Journal, 84, 31, 44-46. 25. D.L.Marsili and G.R.Stevick, 1990. Reducing the risk of ductile fracture on the Canyon Reef Carriers CO2 pipeline. In: Society of Petroleum Engineers Annual Conference, New Orleans, Vol. 3. Society of Petroleum Engineers, Richardson, Texas, pp311-319. 26. A.Cosham, 2009. “It’s a gas, Jim, but not as we know it”. Pipeline Technology 2009. R.Denys, editor, Ostend, Belgium, Scientific Surveys Ltd, Beaconsfield, UK. 27. P.N.Seevam, J.M.Race, M.J.Downie, and P.Hopkins, 2008. Transporting the next generation of CO2 for carbon capture and storage: the impact of impurities on supercritical CO2 pipelines. IPC2008-64063. Int. Pipeline Conference Calgary. ASME, New York, pp1-13. 28. D.M.Dawson and B.J.Brooks, 1999. The Esso Longford gas plant accident report of the Longford Royal Commission. Melbourne. Government Printer for the State of Victoria. 29. R.M.Andrews, G.C.Morgan, and W.J.Beattie, 2004. The significance of low toughness areas in the seam weld of linepipe. IPC04-0422. Proc. International Pipeline Conference. ASME, New York, pp1-10. 30. Y.Y.Wang and D.J.Horsley, 2003. Tensile strain limits of pipelines. Paper 35. 14th Joint Technical Meeting on Linepipe Research, Berlin, May. European Pipeline Research Group, Duisburg, Germany, pp1-14. 31. J.F.Kiefner, W.A.Maxey, R.J.Eiber, and A.R.Duffy, 1973. Failure stress levels of flaws in pressurized cylinders. Progress in flaw growth and fracture toughness testing ASTM STP 536. American Society for Testing and Materials, Philadelphia, pp461-481. Sa no m t f ple or c di op st y rib ut io n Mechanical Engineers, Part C. Journal of Mechanical Engineering Science, 214, 8, 1099-1106. 7. S.B.M.Beck, N.M.Bagshaw, and J.R.Yates, 2005. Explicit equations for leak rates through narrow cracks. Int.J. Pressure Vessels and Piping, 82, 7, 565-570. 8. L.V.Clarke, H.Bainbridge, S.B.M.Beck, and J.R.Yates, 1997. Measurement of fluid flow rates through cracks. Idem, 71, 1, 71-75. 9. J.P.Taggart and P.J.Budden, 2008. Leak before break: studies in support of new R6 guidance on leak rate evaluation. J.Pressure Vessel Technology, Transactions of ASME. 130, 1, 01140210114026. 10. Anon., 2001. Assessment of the integrity of structures containing defects. R6 Revision 4. Barnwood. British Energy Generation. 11. BSI, 2005. Guide to methods for assessing the acceptability of flaws in metallic structures BS 7910:2005 incorporating Amendment 1, September 2007. British Standards Institution, London. 12. R.M.Andrews, M. Martin, and V.Chauhan, 2006. Assessment of corrosion defects in old low toughness pipelines. IPC0610140. Proc.International Pipeline Conference, ASME, New York, pp1-14. 13. G.Wilkowski, D.Rudland, D.Rider, P.Mincer, and W.Sloterdijk, 2006. When old line pipes initiate fracture in a ductile manner. IPC06-10326. Proc. International Pipeline Conference. Calgary, Canada September, ASME, New York, pp1-12. 14. G.D.Fearnehough, 1985. Progressive crack extension due to local cooling of a crack in LNG storage tank material. Int.J. Pressure Vessels and Piping, 19, 283-292. 15. G.D.Fearnehough, D.W.Jude, and R.T.Weiner, 1971. The arrest of brittle fracture in pipelines. In: Practical application of fracture mechanics to pressure vessel technology. Institution of Mechanical Engineers, London, pp156-162. 16. IGE, 1984. Recommendations on transmission and distribution practice: steel pipelines for high pressure gat Transmission. IGE/TD/1 Complete Edition 2. Institution of Gas Engineers, London. 17. W.A.Maxey, J.F.Kiefner, and R.J.Eiber, 1983. Brittle fracture arrest in gas pipelines. NG-18 Report 135 PRCI Catalogue L51436. In: PRCI Report, Washington. Pipeline Research Committee. 18. A.B.Rothwell, 2007. Evolution and current status of approaches to fracture control design for gas pipelines. Paper 15. In: 16th Joint Technical Meeting on Linepipe Research, Canberra, April. Australian Pipeline Industry Association, pp1-15. 19. R.M.Andrews, N.A.Millwood, A.D.Batte, and B.J.Lowesmith, 2004. The fracture arrest behaviour of 914mm diameter X100 grade steel linepipes. Paper 0596. Proc. International Pipeline Conference, ASME, New York, pp1-9. Sa no m t f ple or c di op st y rib ut io n PIPE The global organization for oil and gas pipeline engineers • Recognizing your skills and status • Promoting the highest engineering standards • Providing a professional network SIGN UP TODAY! www.pipeinst.org 4th Quarter, 2010 289 Greenhouse gas emissions from electricity generating CCS upstream and downstream transport processes by Dr Tim Cockerill*1, Dr Naser Odeh2, and Scott Laczay1 1 ICEPT, Imperial College London, UK 2 AEA PLC (formerly at University of Reading), UK H Sa no m t f ple or c di op st y rib ut io n EADLINE FIGURES suggest CCS technology will capture 90% or more of the CO2 produced by a power plant. While this may be true at the stack, on a full lifecycle basis the �greenhouse gas’ (GHG) savings offered are more modest thanks to significant resource consumption in upstream and downstream processes. Our analysis suggests that lifecycle GHG emissions can be reduced to approximately 170gCO2/ kWh for an integrated gasification combined cycle (IGCC) plant with 90% capture efficiency. This still represents around an 80% saving compared to conventional coal plant, but is considerably higher than the better-performing renewables such as wind that produces only 10-30gCO2/kWh in good locations This paper examines the origin and importance of upstream and downstream CCS GHG emissions, in particular identifying those associated with transport processes. Sensitivity studies investigate which major characteristics of a CCS system are likely to have an important impact on transport GHG emissions. The scope for combining biofuels with CCS in order to improve lifecycle performance is considered. In principle BioCCS could produce a system with overall negative atmospheric GHG emissions. However that potential is constrained by emissions arising from the production and transportation of biofuels. Finally some general conclusions for design approaches for CCS systems aimed at minimizing system GHG emissions are drawn. Some key areas of uncertainty are also identified for further work. H EADLINE FIGURES suggest CCS technology will capture 90% or more of the CO2 produced by a power plant. While this may be true at the stack, on a full lifecycle basis the GHG savings offered are likely to be more modest thanks to significant resource consumption in upstream and downstream processes. This paper summarizes results from a series of lifecycle analysis investigations of hypothetical fossilfuel-based electricity-generating CCS plant, emphasizing the role, albeit relatively small, that transport systems play in contributing to the overall emissions. The discussion encompasses both downstream transport systems (i.e. for carbon dioxide) and upstream systems (i.e. for fuel and consumable materials). Much of this discussion draws on This paper was presented at the First International Forum on Transportation of CO2 by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical and Clarion Technical Conferences, and with the support of the University of Newcastle and the Carbon Capture and Storage Association. *Author’s contact details email: [email protected] analysis substantially reported in two publications by two of the authors [1], [2]. The performance of the fossil-fuel-based CCS systems is compared to that of several electricity generating renewable energy technologies. Subsequently the potential offered by combining biomass and CCS technologies, with the ultimate objective of producing net carbon dioxide capture from the atmosphere, is examined. The results from this section of the paper are necessarily indicative as there are many uncertainties about the GHG impacts of biomass production and combustion, let alone the complexities introduced by combining biomass combustion with carboncapture technologies. In consequence, considerable care must be exercised in quantitatively comparing the various results presented throughout the paper. Nevertheless, the qualitative trends are clear enough to allow some useful conclusions to be drawn. FP2 Sa no m t f ple or c di op st y rib ut io n The new online information service that unlocks the secrets of the global pipeline industry Pipelines International Premium is the international oil and gas pipeline industry’s foremost in-depth source of information, comprising a digest of high-quality papers covering the latest technology and reviews of the pipeline industry worldwide, and a comprehensive project database. It is comprised of: Pipelines International Digest which provides a monthly update of papers covering all areas of the industry – from key projects, and engineering and construction issues, to environmental, regulatory, legal and financial issues. Pipelines International Projects which allows subscribers to access a searchable database of completed and current projects. Subscribe or get a free 14 day trial now at www.pipelinesinternational.com/premium 291 Sa no m t f ple or c di op st y rib ut io n 4th Quarter, 2010 Fig.1. Summary of system boundaries for lifecycle analysis. Overview of the lifecycle analysis approach Lifecycle analysis (LCA) is a standardized method for evaluating the environmental impacts of a given process or different competing processes. Greenhouse gas emissions, other air and water emissions, resource consumption, and energy use are evaluated using energy and material balances. The evaluation procedure covers all sub-processes within the lifecycle of the system, starting from raw material production and ending with product and waste disposal. By evaluating the environmental impacts of different systems, recommendations can be made to reduce possible effects. The work reported here makes use of LCA in studying the impacts (with emphasis on GHG emissions) of fossil-fuel power generation with and without CCS. The main objective is to evaluate the actual reduction in GHG emissions that can be realized by CCS in various configurations. Broadly speaking, each of the systems studied consists of fuel production, its transportation to the power plant, power plant construction, power plant operation, and any processes related to power/capture plant operation. For all CCS technologies, the analysis also includes the capture plant construction and operation in addition to CO2 transport and storage. System specification System boundaries Figure 1 summarises the extent of the system considered by the lifecycle analysis results presented here, with further detail of individual elements in Table 1. The key element of course is the power plant itself, and our calculations include direct emissions from combustion, plant internal energy consumption, energy used in operating an monoethanolamine-based capture system, and energy for plant maintenance activities. Note that the impacts of electricity transmission beyond the power station are not included, and thus the results quoted here are for electricity produced rather than delivered. Upstream process direct emissions include those arising from fuel production activities, encompassing the energy used to operate the machinery and transport systems required. Also included within upstream processes is the production of other consumable materials such as limestone, ammonia, and monoethanolamine (MEA). For gas-cycle systems, leakage from pipelines is accounted for. Indirect upstream emissions take account of equipment manufacturing, recognizing that production facilities are not dedicated to servicing the power plant and the associated emissions should be shared across 292 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.2. Breakdown of lifecycle GHG atmospheric emissions from power plant technologies. Category Coal-based systems Natural gas-based systems • Power plant construction • Capture plant construction • CO2 transport pipeline • • • • Fuel combustion Direct CO2 emissions Direct CO2 emissions Fuel production Mining (Equipment manufacture, Mining Gas extraction (platform construction, activities, Methane leakage, Coal cleaning, gas sweetening and flaring, methane Land recovery for surface mining) leakage) Other material production • • • • • Limestone/ammonia production SCR catalyst production Water treatment and distribution MEA production NaOH/activated C production • • • • • • • • • Boiler/ESP/Gasifier ash FGD waste SCR catalyst waste MEA re-claimer waste • SCR catalyst waste • MEA re-claimer waste Construction Waste disposal Transport • Coal transport • Local by rail • International by ship • Limestone transport by truck • Chemical transport by rail • Waste transport by truck • CO2 compression and injection Power plant Gas pipeline Capture plant CO2 transport pipeline Ammonia production SCR catalyst production Water treatment and distribution MEA production NaOH/activated C production • Gas transport • gas compression • onshore processing • Methane leakage • Chemical transport by rail • Waste transport by truck • CO2 compression and injection Table 1. Details of the sub-processes included within the lifecycle analysis described in the paper. 4th Quarter, 2010 293 all uses. Downstream processes consider waste transport and disposal in relatively nearby locations. As with the upstream analysis where non-dedicated facilities are used, emissions are attributed appropriately across all uses. As the figure shows, emissions arising from the construction phases across all the supply chain elements are included in the calculations, taking account of : • materials • material production processes • material transportation by truck over an average distance of 50km • on-site energy consumption, comprising 80% diesel and 20% electricity taken from the UK grid. Power plant decommissioning is accounted for, but decommissioning of upstream and downstream equipment is not included. Overall results A breakdown of the results is shown in Fig.2 and Table 7. Unsurprisingly, the supercritical plant without CCS produces by far the largest atmospheric GHG emissions per unit of electricity output, dominated by the direct emissions from combustion. All the CCS-fitted plant produce considerably less atmospheric emissions, with the coal-fired IGGC plant generating the smallest quantity of emissions per unit (kWh) of electricity produced. Emissions associated with transport processes Downstream In all the CCS cases, emissions associated with CO2 transport are very small, representing between 1% and 1.8% of the total per unit of electricity produced. This of course only applies for the system specification set out above. The component of these emissions arising from powering any recompression stations could change if energy sources with differing carbon intensities were employed: the present study has assumed they are powered by electricity generated with the UK grid average carbon intensity. Sa no m t f ple or c di op st y rib ut io n External factors As well as the system boundary, another important influence on the assessment is the location of the plant and the resources consumed in construction and operation. The analysis below has been developed on the basis that the plant is located in North-East England, and that fuel and other materials are sourced relatively locally so far as possible; Table 2 provides more comprehensive details. Life cycle emissions for fossilfuelled CCS plant Power plant types of only 37.5% are achieved due to commissioning and decommissioning activities respectively. Further outline details of the power plant are contained in Tables 3-6, but for full information reference should be made to the authors’ previous work [1, 2]. With longer or less-secure pipelines, the impact of CO2 transport could increase considerably, as investigated below. Other downstream processes include ash and other solid waste disposal, which also make only a very small contribution to overall emissions. Four types of fossil-fuel plant are considered here: Upstream • a supercritical pulverized (SuperPC) coal-fired plant with selective catalytic reduction (SCR), electrostatic precipitation (ESP) and flue-gas desulphurization (FGD) pipe-end clean-up technologies; • a similar supercritical coal-fired plant, but fitted additionally with an MEA-based CO2 capture unit having 90% CO2 capture efficiency; • a natural gas fired combined-cycle (NGCC) plant fitted with similar MEA-based capture unit; • a coal-fired integrated gasification combined-cycle (IGCC) plant fitted with Selexol-based carbon dioxide capture, again having 90% capture efficiency. The calculations assume that the plant has a rated capacity of 500MWe, an operating lifetime of 30 years, and take three years to construct. Load factors are taken as 75% except in the first and last year of operation, where factors Upstream emissions are included within the �operation’ components of Fig.2, which also includes the very small downstream operational emissions associates with ash and waste disposal. The upstream calculations assume coal is produced from a nearby UK deep mine and subsequently transported to the power plant by rail. This represents something of an idealized best case, as the limited number of UK mines means transport will in general be over longer distances. Limestone is also UK sourced and transported by truck, with other consumables such as solvents transported by rail. Natural gas is assumed to be sourced from the UK North Sea, carried via pipelines with the specification set out in Table 10. For the base coal-CCS configurations considered here, in general just less than 50% of upstream emissions arise from mining, with a similar contribution coming from the production and transport of all other consumables. Coal transport accounts for approximately 1.5% of upstream 294 The Journal of Pipeline Engineering Category Coal-based power plants Gas-based power plants Power Plant Location Teesside Mine location Surface mine: Maiden’s Hall Extension, Northumberland Deep mine: Killingley Colliery, North Yorkshire – Limestone Quarry North Yorkshire: 50 km from power plant – Ammonia production Billingham, Durham, 20 km from power plant Concrete manufacturer Leeds, 100 km from power plant Steel manufacturers Teesside Gas field – Southern North Sea On-shore gas processing – Hartlepool Gas pipeline – Offshore: 100 km, on-shore: 50 km Teesside within 50 km from power plant 0 CO2 pipeline 50 km on-shore, 150 km offshore 0.039 Bunter Sandstone-Southern North Sea, Closure 0 CO2 storage Sa no m t f ple or c di op st y rib ut io n CO2 on-shore collection point Table 2. Location of power plant and key material inputs. emissions, though as will be seen later, this low value is largely a reflection of very optimistic assumptions. For the natural gas CCS system considered, the majority of the calculated upstream GHG emissions arise because of escapes in the natural gas supply system, though it should be noted that we have assumed comparatively high leakage rate of 1%. Smaller contributions are distributed over the production and transport of other consumables. Across all the results for CCS plant, it is clear that the operational GHG emissions are much large than those from CO2 transport. Indeed, under the assumptions set out here, emissions from pipeline CO2 transport are almost negligible compared to direct emissions, operational emissions, and emissions associated with capture. In the carbon capture transport and storage (CCTS) system, the transport element appears to have a tiny impact on GHG emissions. Sensitivity study Figure 3 illustrates how sensitive GHG emissions from each of the CCS plant are to system changes, with an emphasis on transport processes. It is immediately clear that the details of the CO2 transport system have relatively little effect, as increasing the pipeline network length by 100 km raises lifecycle emissions by between 0.05% (PC+CCS) and 0.08% (IGCC+CCS). The details of the other downstream processes, and in particular plant ash waste disposal, also seem unimportant. Upstream transport processes are much more important. Importing coal from Russia, rather than relying on local production, has a severe influence on the emissions for the PC and IGCC plant. Much of this is due to emissions from the transportation processes, though it has also been assumed here that a poorer quality coal is delivered. Similarly, supplying the NGCC+CCS plant from a gas network with two percentage points greater leakage increases lifecycle GHG emissions by about one-third. Also of great importance is the effectiveness of CO2 capture. A 5% reduction in the overall proportion of CO2 captured unsurprisingly gives a substantial increase in GHG emissions in all cases. This result remains qualitatively true irrespective of whether the reduction is due to less-effective capture equipment, or increased leakage from a pipeline transport system. The key conclusion is that, in terms of GHG emissions of a CCS system, most of the details of the downstream processes are relatively unimportant. This presents a stark contrast to upstream transport processes, which our results suggest have a much larger impact on emissions. If a design objective is to minimize lifecycle emissions, CCS systems should in general be situated to promote ease and effectiveness of fuel supply, and with little regard to the implications for the CO2 pipeline transport network. However, one important factor impacting the lifecycle performance is the total CO2 captured, and the downstream transport system has the potential to 4th Quarter, 2010 295 Sa no m t f ple or c di op st y rib ut io n Fig.3. Sensitivity of CCS power plant lifecycle GHG emissions to system changes.The first columns show the percentage increase in emissions for PC+CCS and IGCC+CCS plant if all coal Is imported from Russia, rather than locally sourced.The second column shows the impact for NGCC+CCS plant if methane leakage from the supply network increases by two percentage points.The third column represents the impact of recycling 50% of ash and FGD waste as construction materials.The fourth column illustrates the result of lengthening the CO2 transmission network by 100km.The final column shows the effect of decreasing the CCS capture efficiency by 5% (assuming all other plant parameters remain the same). influence this with respect to its resistance to leaks. While the overall configuration of the CO2 transport system has little impact on GHG emissions, it is vitally important that transport is as secure as possible. Comparison with other lowcarbon energy sources For comparison purposes, Table 8 shows ranges of values for GHG emissions from other low-carbon electricityproduction systems taken from the literature. In general it is clear that, despite producing much lower carbon emissions than conventional fossil-fuelled plant, CCS cannot produce electricity that is as low carbon as most renewables. It should be noted, though, that the dividing line is rather fuzzy and dependent on the location and system boundary of the renewable-energy technology. Solar PV in locations with poor resources can result in emissions per kWh higher than those calculated here for IGCC systems. Equally, most of the renewable-energy systems’ assessments do not take account of the impact of intermittency on the lifecycle emissions. Including electricity storage facilities within the system boundary, for example, can dramatically worsen the environmental performance. This raises a number of complex issues that are beyond the scope of this paper, but it should be kept in mind that fossil-fuel-based CCS systems have the potential to offer supply controllability and a geographical independence that certain renewables find difficult to match without additional facilities. One possible way of further reducing the emissions of CCS systems is by combining them with biomass fuels. The potential of this technology is considered in the remainder of this paper. Lifecycle analysis of BioCCS Objectives CCS with biofuel firing (BioCCS) offers the attractive potential of producing a net removal of carbon dioxide from the atmosphere, since the carbon dioxide released by biomass combustion was originally absorbed from the atmosphere in photosynthesis. A further high-level study has examined the lifecycle implications of BioCCS drawing on the outputs of the UKCCSC supported study, supplemented with data from the literature as described in detail by Laczay [3]. Both pure-biofuel and coal co-firing cases have been examined, with the pure-biofuel cases considering both miscanthus1 and RC willow as fuels. For the co-firing case, only miscanthus was analysed. Approach and assumptions The study considers a circulating fluidized bed (CFB) power plant comparable to the 550MWth / 240MWe facility operated by Alholmens Kraft in Pietarssaari, Finland [4]. This plant operates at a typical thermal efficiency of 38%, but it was assumed that a 90% effective CO2 capture system would reduce the power output by 25%, giving an overall conversion efficiency for the BioCCS system of 28.5%. 1 Miscanthus is a tall perennial grass that has been evaluated in Europe in recent years as a new bioenergy crop. It is sometimes confused with elephant grass (Pennisetum purpureum) and has been called both �elephant grass’ and �E-grass’. 296 The Journal of Pipeline Engineering Parameter Value Parameter Ambient temperature, ºC 15 Ambient pressure, kPa 101 Steam cycle heating rate, MJ/kWh 7.4 Gasifier temperature, ºC Excess air, % 20 Gasifier pressure, MPa Temperature of flue gas exiting boiler, ºC 370 Load factor, % 75 Steam input to gasifier, mol H2O / mol C Life time, years 30 ID fan efficiency, % 85 Table 3. Key parameters for the supercritical-PC plant. Parameter Number of gas turbines Value 2 Excess air, % 180 NOx emissions rate, ppm 10 15.7 Compressor efficiency, % 70 Pressure loss across combustor, kPa 28 Temperature into turbine, ºC 1330 Turbine isentropic efficiency, % 85 Mechanical and generator efficiencies, % 98 Table 4. Key parameters for the NGCC plant. Due to the complexities associated with biomass lifecycle analysis, this work used a simplified approach. In particular, emissions associated with power plant and CO2 pipeline construction processes have been neglected. As the latter are relatively small this assumption is unlikely to have a significant influence on the results. The former are more likely to have an impact on the detail of the calculations, but not the qualitative conclusions. Emissions associated with the up-keep of the carbon capture system, for example solvent replacement, have also been neglected. It should also be kept in mind that the study takes no account of the whole system indirect impacts of wider biomass use, such as induced land use change (ILUC) which, it is argued, could have a devastating effect on the lifecycle sustainability of certain biofuels. There has been much recent debate about the GHG emissions that should be associated with biofuel production, typified by the Searching-Wang debate (see for example [5]) with the Gallagher Review providing an excellent reference [6]. The calculations reported here account for only the emissions that arise directly from biomass cultivation, processing and harvesting operations. The Biomass Environmental Assessment Tool (BEAT2) [7, 8] was used to calculate the energy yield and combustion products in all cases, with the following parameters: GE oxygenblown 1250 6 0.45 Carbon loss, % 1 Oxidant pressure (at outlet of ASU), MPa 4 Oxidant composition, %O2 : % Ar : % N2 95 : 4: 1 Particulate removal efficiency from syngas, % 50 COS to H2S conversion efficiency, % 98 H2S removal efficiency, % 98 COS removal efficiency, % 40 CO to CO2 conversion efficiency, % 95 Sulphur recovery efficiency, % 95 Steam added to shift reactor, mol H2O/ mol CO converted 1 Sa no m t f ple or c di op st y rib ut io n Air compressor ratio Type of gasifier Value Table 5. Key parameters for the IGCC plant with Selexol capture.Table 5. Key parameters for the IGCC plant with Selexol capture. • All power plant operate with an annual load factor of 90% for miscanthus yields are 18 wet tonnes per hectare per year, having 30% moisture content. Once harvested the feedstock is naturally dried in storage for 40 days reducing the moisture content to 10%. 60kg of nitrogen fertilizer is used when establishing each hectare, and a production cycle lasts 15 years after which the plantation must be cleared and re-established. • For willow yields are 14 wet tones with 50% moisture content, with the feedstock dried to 10% moisture content. • Energy crops are transported by truck 100km from the plantation to the storage/processing site, with a further 100-km journey to the power plant. • Losses of 7% occur, 11% during storage and 3% during transport of energy crops. • Coal is imported from South America, USA, Australia and South Africa Results Table 9 shows the calculated lifecycle GHG emissions for the pure-biomass based CFB power plant with CCS. Both cases 4th Quarter, 2010 297 Fig.4. Lifecycle comparison of fossil CCS plant with other low carbon energy systems.  Parameter CO2 removal efficiency, % Sa no m t f ple or c di op st y rib ut io n show strongly negative net GHG emissions, with the 90% capture rate more than compensating for the emissions that do reach the atmosphere. The net emissions are sufficiently negative that the simplifications outlined earlier are very unlikely to change the qualitative conclusion More detail is shown in Fig.6, which compares the origins of the emissions reaching the atmosphere. The upstream processes for miscanthus and SRC willow show slight differences in emissions. Miscanthus has more emissions associated with cultivation/harvest compared to willow. This is due to differences in the planting and harvesting processes of the two energy crops. Miscanthus also has much higher transport emissions than SRC willow because it is bailed rather than chipped. Chips are more densely transported, and as a result, miscanthus transport emissions are nearly twice that of chipped SRC willow. Results for co-firing with coal are shown in Fig.7, where the vertical axis represents net atmospheric GHG emissions per kWh of electricity produced relative to a supercritical coal power plant without a CO2 capture unit. Unsurprisingly, net emissions reduce almost in direct proportion to the proportion by energy value of biomass in the fuel mix. A useful observation is that miscanthus-based BioCCS appears to become GHG neutral for a co-firing level of approximately 20%. Higher proportions of biomass produce net capture from the atmosphere, though some care is necessary in interpreting the values in the light of the simplifications outlined earlier. Discussion Emission minimization strategies Presumably a key objective in the design of any electricityproducing CCS system is to generate power with the lowest achievable GHG emissions per unit. To reach this objective, Value 90 SO2 removal efficiency in capture plant, % 99 SO2 removal efficiency in FGD, % 98 SO3 removal efficiency in capture plant, % 99 HCl removal efficiency in FGD, % 95 NO2 removal efficiency in capture plant, % 25 Ash removal efficiency in FGD, % 50 MEA concentration, %w/w 30 Lean CO2 loading, mol CO2/mol MEA 0.2 Blower efficiency, % 75 Pressure across blower, kPa 15 Sorbent pump efficiency, % 75 Pressure across pump, kPa 200 Compressor efficiency, % 80 CO2 outlet pressure, MPa 13.5 Table 6. Key parameters for MEA-based capture process. the results in this paper suggest that there is some value in adopting an integrated approach to the design of the whole system, as decisions made in one part of the CCTS chain can have implications for the GHG emission of another. Minimizing overall emissions requires that such interactions are fully accounted for. In the cases considered here, downstream pipeline-based CO2 transport does not have a substantial influence on overall GHG emissions, and thus can largely be designed independently from the rest of the system in this regard. 298 The Journal of Pipeline Engineering Source of GHG Emissions (gCO2e/kWh) Plant Type Total Emissions Construction Direct Operation CO2 Capture CO2 Transport Super-Crit Coal 2 788 91 0 0 881 S-C Coal + CCS 3 107 124 22 3 258 NGCC + CCS 3 42 118 25 2 190 IGCC + CCS 3 90 73 1 3 170 Table 7. Summary of lifecycle GHG for representative power plant. Range of GHG emissions (gCO2/kWh) References Hydro 3-33 [11], [12], [13], [14] Geothermal 15-23 [13], [14] Solar PV 39-217 [15], [14], [16], [17], [18] Highly location dependent. Some higher values included battery storage Solar thermal (to electricity) 30-120 [19] Parabolic trough, centralized receiver & parabolic dish Wind 9.7-29.5 [20],[14] Wind with pumped hydro storage 20 [21] Wind with compressed air storage 109 [21] 6-24.2 [12], [22], [14] Nuclear fission Comments Sa no m t f ple or c di op st y rib ut io n Technology Higher values generally for offshore Table 8. Representative values for GHG emissions from several low-carbon electricity production technologies. Note that the large ranges arise partially from incompatible assumptions between the studies considered.  Fuel Miscanthus Willow 73 57 CO2 from biomass combustion 1291 1449 Combustion CO2 captured 1162 1304 Combustion CO2 to atmosphere (B) 129 145 Net combustion CO2 emissions (C) -1162 -1304 Other power plant GHG emissions (D) 19 15 Direct emissions to atmosphere (A+B+D) 221 217 -1070 -1232 Upstream Process GHG Emissions (A) NET GHG EMISSIONS (A+C+D) Table 9. Indicative lifecycle GHG emissions for CFB biofuel to electricity plant with a 90% capture efficiency carbon dioxide capture plant, operating on two fuels.The overall conversion efficiency to electricity is taken to be 28.5%. All emissions are stated in gCO2e/kWh(e). 4th Quarter, 2010 299 Fig.5. Biomass with CCS system summary.  Pipeline CO2 transport can influence system GHG emissions via leakage, and hence minimizing escapes should be a primary design objective. Thickness, mm On-shore Off-shore On-shore Off-shore 75 100 7.8 9 Table 10. Diameter and wall thickness for natural gas pipeline. Sa no m t f ple or c di op st y rib ut io n For countries of scales similar to the UK, pipeline length does not have a significant effect on overall GHG emissions. Hence the CO2 transport distance should not play major role in CCS plant site selection. Upstream transport processes, notably fuel transport, have a much stronger impact on emissions and should have an influence on site selection. Clearly any whole-system GHG-minimization strategy should focus on simplifying fuel rather than CO2 processing and transport, so long as any risk of CO2 leakage is avoided. This is particularly true with biomassbased CCS systems. Diameter, cm Areas of uncertainty While undertaking this work we have identified a number of areas where limited technical understanding constrains the usefulness of LCA approaches for the analysis and optimization of future CCS systems. The two most important, in the opinion of the authors, are discussed in this section. Operational effects Most LCA analyses assume that the systems they study operate under steady-state, full-load conditions, and this is true of almost all CCS studies. Where some account is taken of variable loading, typically, analysts use only a load factor approach to account for periods of non-generation. Where plant will be used predominantly to supply base load, this is a reasonable assumption, especially as construction makes a relatively small contribution to overall emissions even for CCS power plant. The vagaries of the electricity market mean that base-load operation is unlikely for all CCS systems, in practice. As a result, some plant might be subject to substantial numbers of cold-start and shut-down procedures, as operators try to optimize their financial return. Operators may also wish to run plant at part load. From a lifecycle GHG-emissions perspective, non-steady-state and part-load operations are likely to exhibit much poorer efficiency than steady-state full-load operation. In consequence they have the potential to substantially increase the overall GHG emissions of a CCS system, particularly if they are frequent events. New CCS plant will most likely be initially conceived for base-load operation, and thus it could be argued that their lifecycle performance will not be impacted by non-steadystate operation. Over their lifetime, though, there will be substantial changes in energy and electricity markets, and as they age, CCS plant are likely be moved towards a peaking role as is common with existing old fossil-fuel plant [9]. Moreover, expected increasing penetration of intermittent renewables will push even CCS fossil plant towards operating regimes that are more variable than those experienced by existing fossil plant. Evaluating the impact of transient operation on GHG performance is hindered by poor understanding of both CCS plant and carbon dioxide transport systems under such conditions. Further work is required in these areas in order to fully evaluate the �real world’ performance of future CCS systems. Impact of biomass combustion products on CCS efficiency While the results in this paper suggest that biomass-to-power combined with CCS has the potential to produce negative lifecycle GHG emissions, it is important to keep in mind that the underlying calculations assumed there was no detrimental interaction between the biomass combustion products and both the capture system together with the CO2 transport system. In general this would seem a reasonable assumption, it being widely accepted that co-firing reduces the emission of pollutant elements (including sulphur, nitrogen, and mercury) in comparison to pure coal. However, biomass 300 The Journal of Pipeline Engineering Sa no m t f ple or c di op st y rib ut io n Fig.6. Breakdown of contributions to direct GHG atmospheric emissions (i.e. A+B+D in Table 10) for the biomass with CCS system described in the text, operating on two fuels Fig.7. Relative impact of increasing levels of co-firing with miscanthus on lifecycle GHG emissions for a supercritical coal plant, with an without a carbon dioxide capture unit with a 90% capture efficiency. The vertical axis shows the emissions compared to a representative coal super-critical plant without CCS. co-firing can yield increased concentration of hydrochloric acid in flue gases [10]. Again there is scant data available in the literature regarding the effect of biomass combustion products on CO2 capture and transport processes. Further technical data are required before the LCA can be taken forward. Conclusions Transport emissions are a relatively small component of the GHG emissions from CCS systems, though the quantities vary considerably with the assumptions underlying the lifecycle analysis. As a general rule, downstream emissions associated with pipeline CO2 transport are almost negligible, certainly with respect to the construction of short pipelines. Operational emissions depend on the source of the energy used to power recompressions stations. However any CO2 leakage from the pipeline system to the atmosphere has the potential to dramatically increase the impact of downstream transport processes. Upstream transport emissions, predominantly for fuel, are more important, typically representing at least 2% of all GHG emissions. For biofuels, upstream emissions are considerably more important. This is true even without taking account of the current uncertainty surrounding the whole-system sustainability of biofuels, as typified by the Searchinger-Wang debate. The results have implications for reducing lifecycle GHG emissions from CCS plant by optimizing plant location. In Journal of Pipeline Engineering Editorial Board - 2010 There are two key areas of uncertainty that have had relatively scant coverage in the literature and require further work. Firstly, most lifecycle studies of CCS systems make substantial simplifications with respect to the operational regime of the plants under consideration. Secondly, the impact that combustion of biomass fuels might have on the performance of carbon dioxide capture systems does not appear to have been extensively considered in LCA studies of CCS systems. Such an extension can be readily included within the lifecycle methodology in principle, but there is a scarcity of data to support such work. Sa no m t f ple or c di op st y rib ut io n Obiechina Akpachiogu, Cost Engineering Coordinator, Addax Petroleum Development Nigeria, Lagos, Nigeria Mohd Nazmi Ali Napiah, Pipeline Engineer, Petronas Gas, Segamat, Malaysia Dr Michael Beller, NDT Systems & Services AG, Stutensee, Germany Jorge Bonnetto, Operations Vice President, TGS, Buenos Aires, Argentina Mauricio Chequer, Tuboscope Pipeline Services, Mexico City, Mexico Dr Andrew Cosham, Atkins Boreas, Newcastle upon Tyne, UK Prof. Rudi Denys, Universiteit Gent – Laboratory Soete, Gent, Belgium Leigh Fletcher, MIAB Technology Pty Ltd, Bright, Australia Roger Gomez Boland, Sub-Gerente Control, Transierra SA, Santa Cruz de la Sierra, Bolivia Daniel Hamburger, Pipeline Maintenance Manager, El Paso Eastern Pipelines, Birmingham, AL, USA Prof. Phil Hopkins, Executive Director, Penspen Ltd, Newcastle upon Tyne, UK Michael Istre, Engineering Supervisor, Project Consulting Services, Houston, TX, USA Dr Shawn Kenny, Memorial University of Newfoundland – Faculty of Engineering and Applied Science, St John’s, Canada Dr Gerhard Knauf, Salzgitter Mannesmann Forschung GmbH, Duisburg, Germany Lino Moreira, General Manager – Development and Technology Innovation, Petrobras Transporte SA, Rio de Janeiro, Brazil Prof. Andrew Palmer, Dept of Civil Engineering – National University of Singapore, Singapore Prof. Dimitri Pavlou, Professor of Mechanical Engineering, Technological Institute of Halkida , Halkida, Greece Dr Julia Race, School of Marine Sciences – University of Newcastle, Newcastle upon Tyne, UK Dr John Smart, John Smart & Associates, Houston, TX, USA Jan Spiekhout, Kema Gas Consulting & Services, Groningen, Netherlands Dr Nobuhisa Suzuki, JFE R&D Corporation, Kawasaki, Japan Prof. Sviatoslav Timashev, Russian Academy of Sciences – Science & Engineering Centre, Ekaterinburg, Russia Patrick Vieth, Senior Vice President – Integrity & Materials, CC Technologies, Dublin, OH, USA Dr Joe Zhou, Technology Leader, TransCanada PipeLines Ltd, Calgary, Canada Dr Xian-Kui Zhu, Senior Research Scientist, Battelle Pipeline Technology Center, Columbus, OH, USA particular, the lifecycle GHG impacts are much more sensitive to fuel transport processes than downstream carbon dioxide transport. From the GHG perspective, it is suggested that optimal plant location strategies for the UK should focus on minimizing fuel processing and transport, and not be overly concerned about CO2 transportation distance. This is especially true for biomass-derived fuels, even if they are sourced from the UK. ❖❖❖ Acknowledgements This paper draws on CCS systems’ analysis projects undertaken by the authors over the last eight years. Major support has been provided by the Tyndall Centre, and the UK Natural Environment Research Council (NERC) within the scope of the UK Carbon Capture and Storage Consortium. References 1. N.A.Odeh and T.T.Cockerill, 2008. Life cycle analysis of UK coal fired power plants. Energy Conversion and Management, 49, 212–220. doi:10.1016/j.enconman.2007.06.014. 2. Ibid., 2008. Lifecycle GHG assessment of fossil fuel power plants with carbon capture and storage. Energy Policy, 36, 367–380. 3. S.A.Laczay, 2009. A comparative analysis of the economics and GHG emissions of fossil fuel, co-fired and dedicated biomass electricity generation systems with carbon capture and storage in the UK. Master’s thesis, Centre for Environmental Policy, Imperial College London, September. 4. Alholmens Kraft. The history. Website [accessed 15 July 2009] www.alholmenskraft.com/en/history/index.htm. 5. T.Searchinger, R.Heimlich, R.A.Houghton, F.Dong, A.Elobeid, J.Fabiosa, S.Tokgoz, D.Hayes, and T.-H.Yu, 2008. Use of US croplands for biofuels increases greenhouse gases through emissions from land use change. Science, 319, 1238–1240. 6. EGallagher, ed., 2008. The Gallagher Review of the indirect effects of biofuels production. Renewable Fuels Agency. Available at www.renewablefuelsagency.gov.uk/reportsandpublications/reviewoftheindirecteffectsofbiofuels. 7. AEA. Biomass Environmental Assessment Tool Version 2 User Guide., 2008. Available at http://www.biomassenergycentre. org.uk/pls/portal/url/ITEM/5A8E649760A5B873E04014A C08044B4D. 8. AEA, 2007. The biomass environmental assessment tool (BEAT2). www.biomassenergycentre.org.uk/portal/page?_ pageid=74,153193&_dad=portal&_schema=PORTAL. 9. DECC, 2009. Towards carbon capture and storage: Government response to consultation. Technical report, UK Department of Energy and Climate Change. 10. M.F.G.Cremers, 2009. Technical status of biomass co-firing. Technical Report IEA Bioenergy Task 32 Deliverable 4, International Energy Agency, August. 11. L.Gagnon and J.Van de Vate, 1997. Greenhouse gas emissions from hydropower: The state of research in 1996. Energy Policy, 25, 1, 7–13. 12. L.Gagnon, C.Belanger, and Y.Ychiyama, 2001. Life-cycle assessment of electricity generation options: The status of research in year 2001. Idem, 30. 13. Y.Uchiyama, 1996. Life cycle analysis of electricity generation and supply systems. In Proceedings of a symposium on electricity, health and the environment: Comparative assessment in support of decision making, pp279–291, Vienna, Austria, October. International Atomic Energy Agency. 14. H.Hondo, 2005. Life cycle GHG emission analysis of power generation systems: Japanese case. Energy, 30, 2042–2056. 15. P.Denholm and G.Kulcinski, 2003. Net energy balance and greenhouse gas emissions from renewable energy storage system. Technical Report 223-1, Energy Centre of Wisconsin. Available at http://fti.neep.wisc.edu/pdf/fdm1261.pdf. 16. J.L.R.Proops, P.W.Gay, S.Speck, and T.Schroder, 1996. The lifetime pollution implications of various types of electricity generation. Energy Policy, 24, 3, 229–237. 17. R.Kannan, K.C.Leong, R.Osman, and H.K.Ho, 2007. Life cycle energy, emissions and cost inventory of power generation technologies in Singapore. Renewable and Sustainable Energy Reviews, 11, 702–715. 18. H.Schaefer and G.Hagedorn, 1992. Hidden energy and correlated environmental characteristics of PV power generation. Renewable Energy, 2, 2, 159–166. 19. C.Dey and M.Lenzen, 2000. Greenhouse gas analysis of electricity generation systems. In Proceedings of the ANZSES Solar 2000 Conference, pages 658–668, Griffith University, Queensland, Australia, November. 20 L.Schleisner, 2000. Life cycle assessment of a wind farm and related externalities. Renewable Energy, 20, 279–288. 21. P.Denholm and G.L.Kulcinski, 2004. Life cycle energy requirements and greenhouse gas emissions from large scale energy storage systems. Energy Conversion and Management, 45, 2153–2172. 22. R.Dones, T.Heck, M.F.Emmenegger, and N.Jungbluth, 2005. Life cycle inventories for the nuclear and natural gas energy systems, and examples of uncertainty analysis: EcoInvent: Energy Supply. Int.J. of Life Cycle Assessment, 10, 1, 10–23.
© Copyright 2024 Paperzz