JPE CO2 Transportation by Pipeline Special Issue

December, 2010 Vol.9, No.
Journal of
Pipeline Engineering
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
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
1. Disclaimer: While every effort is made to check the 4. Back issues: Single issues from current and past volumes
accuracy of the contributions published in The Journal of are available for US$87.50 per copy.
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 5. Publisher: The Journal of Pipeline Engineering is
those of the authors alone. published by Great Southern Press Ltd (UK and Australia)
and Clarion Technical Publishers (USA):
2. Copyright and photocopying: © 2010 Great Southern Great Southern Press, PO Box 21, Beaconsfield
Press Ltd and Clarion Technical Publishers. All rights HP9 1NS, UK
reserved. No part of this publication may be reproduced, tel: +44 (0)1494 675139
stored or transmitted in any form or by any means without fax: +44 (0)1494 670155
the prior permission in writing from the copyright holder. email: jtiratsoo@gs-press.com
Authorization to photocopy items for internal and personal web: www.j-pipe-eng.com
use is granted by the copyright holder for libraries and www.pipelinesinternational.com
other users registered with their local reproduction rights
organization. This consent does not extend to other kinds Editor: John Tiratsoo
of copying such as copying for general distribution, for email: jtiratsoo@gs-press.com
advertising and promotional purposes, for creating new
collective works, or for resale. Special requests should Clarion Technical Publishers, 3401 Louisiana,
be addressed to Great Southern Press Ltd, PO Box 21, Suite 255, Houston TX 77002, USA
Beaconsfield HP9 1NS, UK, or to the editor. tel: +1 713 521 5929
fax: +1 713 521 9255
web: www.clarion.org
3. Information for subscribers: The Journal of Pipeline
Engineering (incorporating the Journal of Pipeline Integrity) Associate publisher: BJ Lowe
is published four times each year. The subscription price for email: bjlowe@clarion.org
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). 6. ISSN 1753 2116
Subscribers receive free on-line access to all issues of the
Journal during the period of their subscription.
v v v
www.j-pipe-eng.com
is available for subscribers
Editorial
Guest editorial: CO2 transportation by pipeline – a special issue
T HIS SPECIAL ISSUE of the Journal of Pipeline

Engineering is dedicated to the topic of carbon
dioxide transportation by pipeline. The subject of pipeline
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
transportation of CO2 is becoming increasingly important academia and industry.
as carbon capture and storage (CCS) schemes world-wide
are moving from the pilot to the demonstration phases. One of the areas highlighted in James Watt’s paper for further
In recognition of the increased interest in CO2 transport, research is that of materials’ selection. This important topic
Newcastle University – in association with Tiratsoo Technical is the subject of the paper given at the Forum by Dr Paul and
and Clarion Technical Conferences – organized the first co-authors, who conclude that the main issues for the pipeline
Forum on the transportation of CO2 by pipeline in July material are corrosion in the CO2 process stream, resistance
this year. The aims of the Forum were to highlight the key to brittle and ductile fracture propagation, and degradation
issues relating to CO2 pipelines and bring together leading of polymers in supercritical CO2. With respect to corrosion,
academics and industry experts to discuss current and future it is recognized that typical carbon steel materials used for
CCS ventures and the international research activities being pipelines are not corrosive in pure, dry, supercritical CO2,
undertaken to support these projects. and the paper highlights that there is considerable experience
and research in this area relevant to the currently operating
In this issue, five papers from the technical sessions of the pipelines. However, one of the significant differences that
Forum are published, together with two further papers from will be encountered in pipelines transporting CO2 in CCS
international authors working in the field. The conference schemes is that the product stream will contain impurities
opened with a session of keynote and scene-setting papers in combinations not currently transported in pipelines for
including the paper published in this volume by James EOR. There is very little research work on the effect of these
Watt on ‘Carbon dioxide transport infrastructure: key impurities in the event of a process upset which could allow
learning and critical issues’. In this paper, the world-wide free water to enter the pipeline.
operational experience in the large-scale transportation
of high pressure CO2 is reviewed in order that the issues In terms of fracture propagation, Dr Paul and his co-authors
relevant to the development of networks of pipelines for conclude that although there is extensive experience with the
CCS schemes can be understood. It is estimated that there specification of material properties to prevent long running
are approximately 6000km of CO2 pipelines globally, ductile and brittle fractures from propagating in natural gas
predominantly in North America; the majority of these pipelines, this knowledge cannot be directly applied to the
pipelines are transporting naturally sourced CO2 for design and fracture control of dense phase CO2 pipelines.
the purpose of enhanced oil recovery (EOR). Whilst it is This view is shared by Leis and co-authors in their paper
concluded that the knowledge gained from this experience published here on ‘Transporting anthropogenic CO2 in
is vital, there remain areas pertinent to the development of contrast to pipelines supporting early EOR’. In particular,
CCS transport networks which still need to be addressed. they conclude that the main approach used for specifying
However, one of the key messages of this paper is that, due ductile fracture control requirements in natural gas pipelines,
to the urgent requirement to reduce CO2 emissions, the the Battelle Two-Curve Method, has not been validated
industry is not in the position of being able to wait until for CO2 containing impurities from a capture plant, and
this research is complete before designing, constructing, and neither has the equation of state which forms the basis of this
operating pipeline networks. If CCS is to have an impact model. One of the key conclusions from this paper is that
in the reduction of CO2 emissions, then plants have to be the decades of experience with the transportation of CO2
for EOR should be considered carefully when applying this infrastructure is set up to which smaller CO2 sources could
knowledge to the design of pipelines for CCS, particularly then be added. The analysis conducted indicates that there
in the area of fracture control. Indeed it is highlighted comes a point in the network development where adding-
that many of the early CO2 pipelines were retrofitted with in smaller sources, particularly those that are remote from
crack arrestors to manage concerns with fracture arrest. clusters of CO2 sources, considerably increases the marginal
In a complementary paper, King and Kumar illustrate the costs and diminishes the returns.
issues to be considered in the design of a high-pressure CO2
pipeline for ductile fracture arrest using the example of the Finally, Dr Tim Cockerill and co-authors present a lifecycle
propsoed CO2 Masdar pipeline in Abu Dhabi. In this paper analysis to determine the CO2 emissions which are generated
it is demonstrated that ductile fracture arrest is possible in by the pipeline transportation of CO2 and the impact that
high-pressure CO2 pipelines using the pipe wall thickness these might have on the CCS process. The analysis indicates
alone without the need for crack arrestors. that the greenhouse gas emissions from the transport phase
are almost negligible and therefore the optimal location of
As well as concerns over ductile fracture propagation, as power plant in the design of networks should be driven by
discussed in the previous papers, one of the issues with CO2 the requirement to minimize fuel processing and transport
is that there is a significant temperature drop around the leak rather than CO2 transportation.
site as the escaping fluid expands, due to the Joule-Thompson
effect. It is postulated that this drop in temperature could As you read through the papers in this issue, one of the
cause continuous brittle initiation of a crack in a CO2 recurring conclusions drawn by the authors is that there is a
pipeline as the steel is cooled below the ductile-brittle still a significant amount of research required in the area of
transition temperature. This effect is investigated in the CO2 pipeline transportation. In particular, the experimental
paper by Andrews, Haswell, and Cooper ‘Will fractures database on which so much of the knowledge relating to
propagate in a leaking CO2 pipeline?’. natural gas pipelines is built has not been established for
CO2. There is therefore an urgent need for researchers to
An important additional conclusion drawn by King and work in collaboration in order that the required research can
Kumar is that, when designing a pipeline or network, there be completed within timescales which allow CO2 pipelines
are a number of potential options that could be selected for CCS to be designed and operated safely and efficiently.
and cost optimization techniques, in combination with the
technical requirements, should be implemented to derive the Dr Julia Race
final solution. This theme is taken up by Saulat Lone and Senior Lecturer in Pipeline Engineering
his co-authors in their paper ‘The techno-economics of a School of Marine Science and Technology,
phased approach to developing a UK CO2 pipeline network’. Newcastle University,
This paper investigates the establishment of CO2 networks Newcastle-upon-Tyne, UK
in the UK using idealized scenarios in which a backbone j.m.race@ncl.ac.uk
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.
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 In North America since 1972 carbon dioxide has been used
for enhanced oil recovery from some man-made sources, but
When considering the use of pipelines it is important first the majority of transported carbon dioxide is from naturally
to take stock of the existing facilities. Pipelines are good occurring gasfields along the mid-continental mountain
references, but the knowledge embedded in them comes ranges and Mississippi Basin, as shown in Fig.1 and Table
with the supporting research, engineering, and learning 1; the gas is transmitted above critical pressure, Fig.2.
that comes with each design. The majority of carbon
dioxide pipelines are in the USA and Canada, along with The maturity of the systems is still limited, but the experience
substantial in-field pipework for EOR schemes. There are that is there has formed a nucleus upon which to build CCS.
other pipelines, including 90km in Turkey, and pipelines The pipelines have been developed at the required scale; for
in Algeria and Hungary, but there are relatively few outside example, the Cortez pipeline delivers 19.3 million tonnes
North America. Only two projects have offshore pipelines: per year, roughly equivalent to one of Europe’s largest power
Snovhit and Sleipner in the North Sea. 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
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 to the 490,000km of gas pipelines or the 278,000km of
and Clarion Technical Conferences, and with the support of the University of hazardous liquid pipelines in the USA alone. Whilst specific
Newcastle and the Carbon Capture and Storage Association.
knowledge may be limited to 2% or less of the hazardous
Author’s contact details liquids’ experience in the US, there is still a knowledge
tel: +44 (0)1740 646100
email: james.watt@amec.com
base to consider: over 600,000km [1] of pipeline designed
to the same codes of practice.
Fig.1. North American CO2 pipelines
Capacity Length Pressure Year

Pipeline Location Origin of CO2
(Mt CO2/y) (km) (bar) Complete
Cortez USA 19.3 808 186 1984 McElmo Dome
Sheep Mountain USA 9.5 660 132 Sheep Mountain
Bravo USA 7.3 350 165 1984 Bravo Dome
Canyon reef Carriers
USA 5.2 225 175 1972 Gasification
(SACROC)
Val Verde USA 2.5 130 - 1998 Val Verde Gas Plants
Bati Raman Turkey 1.1 90 170 1983 Dodan field
USA &
Weyburn 5 328 Up to 204 2000 Gasification
Canada
Table 1: Major carbon dioxide pipelines [26].
Within the US and Canada, regulatory frameworks that hard standard on which to base design. The core standard
govern carbon dioxide pipelines have been developed is ASME B31.4 – the code for liquid pipelines – see Fig.3.
and deployed for a number of years. The design of such However, evidence suggests that ASME B31.8 is also applied.
pipelines is essentially uses the same standards as for any This gas-specific code is used to evaluate the safety issues
hazardous liquid pipeline such as ethylene, crude oil, or around a gas pipeline, applying these rules to carbon dioxide
petroleum products. These codes are, for the US 49 CFR liquid lines as the fluid transitions to gas on release. The
195, Transportation of hazardous liquids by pipeline, and for process of design for pipelines in hazardous liquid service
Canada Z662-07, Oil and gas pipeline systems. For gaseous is therefore robust and well understood.
carbon dioxide, 49 CFR 192 applies rather than the liquid-
specific 49 CFR 195 [2]. European experience
The approach within the US is that the code or regulation Current European experience rests with Statoil at Sleipner
contains a number of established standards from the API, and Snovhit, although there is some carbon dioxide
AMSE, ASTM and others. Those referenced in the regulation experience on the Continent. Predominantly, current
or quoted, such as ASME B31.4, are the required minimum transportation is either by ship or road tanker. The
and therefore become part of the regulation, in theory a expectation is that European CCS will evolve to target
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.
Fig.5. BS PD 8010 steel pipelines

on land.
offshore storage sites, and due to the potential for CCS infrastructure. This is multiplied when considering networks.
clusters, develop in a network format. In some estimates the Transportation systems have little ability to respond to
infrastructure for European CCS is estimated at between variances: essentially, even the flow rate is dictated by
30,000 and 150,000km [3]. This brings new challenges that the power station or storage. The dynamics of the whole
have not been addressed in the current experience base, chain system, with drivers at both ends, is therefore more
both in terms of scale and the need for offshore pipelines. complex. There will be an economic driver here to maintain
transportation to the storage and maintain a storage solution
As in the US and Canada, European and international that is flexible and allows accommodation of the emitter’s
codes are also mature: ISO 13623 (BS EN 14161) and operating regime.
DNV OS-F101 are equivalent to the US codes, although
neither specifically consider dense-phase or supercritical What has the US learned?
carbon dioxide. However, both are adequate codes, but
should be supported by industry best practice. In the UK The experience in the US has highlighted the critical issues
the HSE recommends that BS PD 8010 is used as a wider that must be considered for carbon dioxide service [4, 5].
appreciation of good practice in pipeline design. In each These include
case the standards recommend the same or similar sub-codes
and standards, as shown in Figs 4 and 5. • process conditions
• properties
Requirements in CCS systems • operating conditions at entry and exit
• flow calculation method
The infrastructure needs for CCS are different from those • transient (surge) modelling
of simpler enhanced oil recovery schemes. In the existing • flow characteristics
schemes, the carbon dioxide is a by-product or natural • typical carbon dioxide compositions
emission that is used to drive crude oil out of the ground. • piping design
The economic driver is therefore the demand for the oil, • fracture propagation
and when the demand is not there, the extraction of natural • blowdown assembly design
carbon dioxide need not occur. • blowdown rate basis and calculation
• line break controls
Future CCS schemes will not have the same relationship. • pig trap
Whilst the sensitivity of the storage is an issue, the • depth of cover
requirements here of flow, pressure, and temperature set • routeing topography
the downstream conditions. The emitter in CCS provides
another set of upstream conditions in terms of flow rates, • safety and environmental
ramp rates, composition, temperature, and pressure, and • ambient/ground temperature
each of these systems imposes conditions on the transport • blowdown rate basis and calculation
Fig.6. CO2 pipeline incidents

1986-2008.
• dispersion pattern For inspection there have been problems with dense-phase
• frequency and position of block valves or supercritical pipeline inspection and cleaning using
• leak-detection systems pigs. Simple scraper pigs used in cleaning pipelines need a
• line inventory lubricating fluid such as diesel, and they are adversely affected
by the carbon dioxide which damages some non-metallic
• material selection materials [6, 7]. For inspection, the use of intelligent pigs
• pipeline materials is routine; however, inspection experience in the US has
• carbon equivalent highlighted a key problem. The carbon dioxide penetrates
• hardness value the non-metallic components and, as the tool is depressurized
• fracture strength in the receiver, the systems are often subject to rapid gas
• valve, fitting and trim types decompression, destroying the unit. In fact Oosterkamp [8]
• seal, packing materials of construction indicates that it has been reported that by 2008 only two
• valve actuators intelligent pig operations in North America have resulted
in the expensive tool surviving.
• cleaning and strength testing
• cleaning Corrosion of carbon dioxide/water systems has been studied
• hydrostatic testing/drying/dewatering extensively, not just for the carbon dioxide EOR industry
• construction techniques but also for processes involving the production of ammonia,
• corrosion monitoring urea, steam reforming, and the sweetening of acid gas. The
• external corrosion key example to be considered is the research and practical
• fracture propagation observation that was undertaken for the SACROC project1.
• special construction and welding
• stress relief The research into corrosion by Schremp and Roberson [9]
tested a number of compositions against three common
• pipeline operation weld types. The samples were full-size X60 12-in and 16-in
• refrigeration effects during start-up/blowdown pipe sections that were exposed to a carbon dioxide mix
• start-up/shutdown methodology consistent with real SACROC pipeline conditions. Tests were
• line pressuring conducted at two design temperatures and pressures, and at
• requirement for blowdown noise control chemical compositions including 600-800ppm of H2S and
• environmental considerations 800-1000ppm of water. As a result, it is now accepted that
• operational problems corrosion in carbon dioxide pipelines will not occur if the
• operational safety water content is kept less than 60% of the saturation value
• measurement [10]. The original test used a water content 20 times higher
• custody transfer methods than specified for the SACROC pipeline, and concluded
• moisture analysis that corrosion did not occur, even at this elevated water
content. Operational experience showed that after 12 years
The list is extensive and typical to the design of pipelines no the SACROC pipeline, with a 50-ppm water content limit,
matter the fluid, but has particular relevance here. Particular had a corrosion rate of 0.25-2.5μm/y [11, 12].
learning can be drawn from the experience in the areas of
inspection, corrosion, material specification, operational 1 The so-called SACROC (Scurry Area Canyon Reef Operators Committee) unit
safety, and thermodynamics. 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.
Fig.7.Typical orifice meter.
Care always has to be taken with regard to corrosion; however, The operational problems most reported are safety incidents
it is clear from continued operation that carbon steel is reported in the US to PHMSA for all pipelines. In the
acceptable as material for carbon dioxide pipelines. There period from 1986 to March 2008 there were 42 [16] reported
have been incidents, however, with corrosion: SACROC incidents to PHMSA, Fig.6. This relates to approximately
reports some issues [13], notably a corrosion incident in a 0.36 incidents/1000km/year, (assuming US pipeline
spur line from a main pipeline. The area had free-standing distance average of 5000km over the period), compared
water remaining from a hydrostatic test, and so corrosion set to 2447 incidents on the US gas transmission network of
in. This experience shows that care must be taken not only 488,000km [17] or 0.22 incidents/1000km/year. In the
to ensure a water content limit in the entry specification, but period 1990-2001, the incident rate for natural gas was 0.17,
also for the water introduced during testing, commissioning, while for hazardous liquids it was 0.82 [18], so the incident
and maintenance activities. rate is still comparable and within the expected bounds for
a hazardous liquid.
In specifying water content the industry-accepted level are
conservatively specified as between 288-480mg/m 3 [14]. Another reference source for pipelines (although not
In addition the presence of other additional ‘acid gases’ specifically carbon dioxide) is CONCAWE. The data held
such as H2S, SOx, and NOx needs to be considered. In by CONCAWE indicates a significant number of incidents
the case of CCS, the control of SOx and NOx compounds relate to corrosion and third-party intervention. Hence, for
will to be subject to earlier restrictions in the capture carbon dioxide systems, drying is a critical aspect of design
process and are unlikely to present themselves in the and operation in addition to the usual protection applied
pipeline in significant volume. Hydrogen sulphide may to pipelines against third-party incidents.
be an issue for some processes but should be kept below
the 200-ppm limit from Dynamis2 [15], Table 2; this There are two other operational issues to be considered.
restriction and the water content specification should The first is valve operation: here it is recommended that
prevent corrosion occurring. all valves are slow opening, as this avoids damage to the
valves and surging in the pipeline, and this is particularly
Non-metallic components such as seals, valve seats, O-rings, important for blowdown valves. Where segments are above
and even greases have also shown to be affected by carbon ground, thermal relief should be provided and valves
dioxide. Petroleum-based seals can become saturated with must be capable of seating under high carbon dioxide
the high-pressure fluid and rapidly decompress when the pressures. Current practice is not to work on pressurized
pressure is reduced or structurally weakened. Some greases pipelines at all and, where necessary, sections and valves
are also known to become hard and no longer effective. bodies are blown down before removal of a valve or other
Inorganic materials and greases are therefore more often equipment item.
recommended.
The thermodynamics of the fluid are generally based on
Operational safety the determination of its properties by using equations of
state. The correct selection of the equations is therefore
Operational problems are not necessarily reported in the critical. Experience has also shown that for key operations
public domain unless a release occurs, so public information such as start-up and blowdown, the thermodynamic
refers to where an incident has occurred, and has not characteristics require much longer periods to avoid the
necessarily been prevented. In effect, only the failure is very low temperatures that are possible with carbon dioxide.
recorded, so this limits the learning from operational sources.
There are other key lessons to consider, including transient
2 The DYNAMIS project is investigating viable routes to large-scale cost-effective
fluid effects, leak proving (for which the use of air or nitrogen
hydrogen production with integrated CO2 management. The project is an element
of the HYPOGEN initiative, and forms a part of the European Commission’s Quick- is not sufficient), spacing of block valves, the high level of
Start Programme for the Initiative for Growth. HYPOGEN has the goal of providing sensitivity to temperature and pressure, and the attendant
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
effect on pipeline operations. All of these need to be
electricity from decarbonized fossil fuels with permanent CO2 storage. carefully considered.
Post Com-
Component IGCC Oxyfuel Weyburn Dynamis
bustion
CO2 >95% >95%
<4% (for non-

N2/Ar <4% condensable
gases)
0.03 –
0.01 4.1% <4%
0.06% <4% for all non-
EOR 100-
O2 condensable
1000ppm
gases and
Hydrocarbons
Saline
Formation <4%
Hydrocarbons 0 0.01% <5%
EOR <2%
H2 0 0.8 – 2% 0
H2O 0 <100ppm <500ppm
H2S 0 0.01-0.6% 0 <1450ppmv <200ppmv

CO 0 0.03-0.4% 0 <2000ppmv
Hg 0
<0.01%, <0.01%,
Sox 0.5% <100 ppm
<100ppm <100ppm
<0.01%, <0.01%,
Nox 0.01% <100 ppm
<100ppm <100ppm
Glycol
Table 2. Examples of carbon dioxide stream composition.
Major differences The upstream processes also vary and whilst not relevant
for a bespoke source-to-sink design, they are for network
There are major technical and economic differences to be considerations. There are a number of issues that need to
considered in CCS schemes. The demand-led EOR schemes be addressed in the system design, and key amongst these
in which the need for oil production places a requirement has to be the composition of the carbon dioxide stream
on the provider and the provision of carbon dioxide. The entering any transport system. Two considerations have
compressor operating regime is dictated by the production to be made here: firstly safety, and secondly technical. In
rate required and the geological configuration of the terms of safety, the impact of contaminants needs to be
reservoir. In CCS schemes, the same geological factors considered alongside carbon dioxide. It is not enough
dictate to the transport system the operating conditions and to model the dispersion of a carbon dioxide stream, but
flow rates, but there is also the influence of the upstream also its constituent parts must be modelled. This has been
technology. The power plant or industrial process does not done by the Dynamis [19] project, which recommended
fit the same operating profile as a geologic storage. Carbon the specification in Table 2.
dioxide from CCS-enabled plant must be accommodated,
or the emitter will have to free vent, incurring penalties and Typically entry into a US scheme is similar; the Canyon
making the idea of a CCS scheme redundant. Reef project advises [20] the following specification for
carbon dioxide:
This has an impact on the design of the infrastructure, • 95% mol carbon dioxide minimum
particularly the downstream configuration at the storage site. • 0.489g/m3 (250ppm wt) water in the vapour phase,
Whilst the influences from the storage sites are common no free water
with EOR, the addition of deep saline formations to the • 1500 ppm (w/w) hydrogen sulphide
mix adds another level of complexity and another series • 1450 ppm (w/w) total sulphur
of unknowns. • 4% mole nitrogen
Fig.8.Typical vortex meter and

contaminant analysis.
• 5% mole, < -28.9°C dewpoint for hydrocarbons The dynamics of a CCS system become more complex when
• 10ppm (w/w) oxygen considering networks. Large networks of CCS infrastructure
are already proposed in Scotland, as well as in the Humber,
The Dynamis project considered not only the technical Mersey and Dee, Thames, Teesside, and Rotterdam regions.
issues, deviation of properties, density-phase envelopes, These clusters range from 20 million to 90 million tonnes
but also the safety aspects. As a result the criteria for per year of carbon dioxide, and link diverse emitters in
H2S, CO, SOx, and NOx are established on health and industry and power generation. How these networks behave
safety grounds. and cope in different operating scenarios is critical in terms
of both the design of the system and also, more importantly,
In terms of design safety, the methods are almost global, the economics.
although this is one of the major issues in the industry
today. Whilst we can draw methods and experience from Flow measurement in carbon dioxide is also acknowledged
the US and Canadian pipelines, there is an underlying as a possible challenge. Previous experience shows that two
issue. It can be argued that the EOR-based systems are measurements are made: contaminants and flow. Typically
more tolerant, more conservative in their approach, and either orifice or vortex flowmeters (Figs 7 and 8) in pipeline
with lower population densities, and income-generating oil systems are insulated to limit temperature-induced density
revenue from EOR can afford this. The example to consider changes, and all meters are fitted with flow computers. The
here is the approach to dispersion modelling adopted in major difference in CCS is that the metering scheme will
the US. In the example there is a single assumption that a have to be compliant with the EU ETS requirements, and
percentage of the carbon dioxide immediately upon release the meters will have to be of a fiscal standard. Whether this
forms a solid which falls to the ground [21]. The remaining is possible is the subject of continuing research [22], and
vapour is then modelled using the EPA’s Aloha programme careful consideration must be given to the requirements of
or the models DEGADIS or SLAB. The rate of flow from EU ETS when metering flow.
the pipeline is also simplified to a common assessment.
This methodology is conservative and the approach in To gain funding for the provision of networks, building them
Europe is much more precautionary. Hence a more defined to cope with future additions, or ‘right sizing’ as it has become
understanding is desired in understanding key risks, such as known, mandates that the economics and tariffs that could
fracture propagation and dispersion modelling. It must be be expected need to be clear. This is not the funding model
stressed that the US approach is not wrong, but the driver elsewhere, and for investment to move forward in CCS, the
in CCS appears to be much more of a considered approach, issue needs to be clearly understood. The need for networks
more accurate and more economic. are fundamental: CCS may not succeed if the reliance is on
multiple source-to-sink solutions. Cluster networks enable
In terms of system dynamics and the interactions of all significant savings over the collected costs of A to B solutions.
the processing elements, CCS has a series of different
operating modes to consider at both ends. The preference In these areas, whilst current learning can be applied, CCS is
for storages is to be a generally constant flow, whereas fundamentally different. The learning required here to give
power station emitters are cyclic and diurnal, and here is confidence to investors, industry, and other stakeholders
little scope for change between the two. The driver here is cannot come from the current experience or knowledge
economics and affects the capture plant as well: the emitter alone. Some of the lessons industry needs to learn will
can vent if a storage closes or requires a reduce inlet rate or only come from the first CCS schemes and a high rate of
pressure, but there will be an associated cost penalty. The knowledge sharing.
simpler systems in the US with their demand-led supply
can only provide some of the operational dynamic models Issues and approaches
that CCS requires. It is likely that the storage will have to
be flexible enough in terms of dispersed entry points or There are a number of issues that need to be considered in
multiple storage options. transportation and pipelines. These issues do not prevent
development, but their resolution would aid both technical emitters, particularly power generators, face in the addition
and economic decisions. The technical challenges are already of a capture plant, pipeline, and storage are considerable.
known, but an understanding of how to address them needs The current staff will have new processes to control and
to be developed. Some good practice-based guidelines by monitor, and chemical stocks to maintain and dispose of.
the Energy Institute and DNV aim to resolve some of these Whilst these processes are common to the energy industry,
issues, whilst research programmes aim to address gaps in they are not common to the power-generation companies.
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
Flow-assurance guidelines for carbon dioxide, and an the system parameters, not the storage. The dynamics of these
understanding of the fluid behaviour, needs to be clearer. systems are different and the future operators and owners
Experience in the US [23] indicates that there is an need to understand this. There is a cultural shift within
issue with surge and transient pressure. Because the US corporations, business models, and individual sites that
experience is restricted to onshore pipelines, and with needs to be considered and addressed. When deployment of
offshore pipelines anticipated for CCS, there is a concern. CCS comes, it must be received by an intelligent, informed,
Few mitigation measures can be added to offshore pipelines, and correctly-resourced workforce with training complete.
particularly those that terminate at subsea completions.
In this discipline there are also concerns that, even at When considering resources, there needs to be an
low water content in the fluid, carbon dioxide clathrates understanding of the market size and rate of deployment.
(hydrates) may form. The definition of clathrate formation Any evaluation of the CCS market is made with caution due
behaviour, and this area of vapour-liquid equilibrium, to the number of dependencies involved. The IEA roadmap
clearly require further research. [25] provides one such analysis, stating the expectation that
by 2020 we need 100 projects, by 2030 we need 850, and
Importantly, the resolution of physical properties of possible by 2050 we need 3400 projects, in order to meet the BLUE
fluid streams remains an issue. Equation-of-state selection map scenario. Projects rely on three issues: the supply chain,
is covered in multiple academic papers, but empirical data engineers to design, and skilled technicians to construct. How
to support the assumptions and outputs of the predictive many people does it take to design and build 100 projects in
methods are needed. Whilst engineering design can be 10 years? The engineers in the marketplace now are needed
achieved using the predictions, the design margins applied in existing areas. New and emerging areas like CCS and
may prove to be excessive. This is particularly of interest biofuels, or even new nuclear construction, will draw on
in phase-envelope predictions to protect systems against the same resource pool and supply chain. The CCS industry
multi-phase flow. The data must cover the range of possible and academia must address these issues rapidly and build
contaminants and process conditions for a pipeline system. capacity now, ready for delivery to the market in a decade.
Dispersion modelling is already acknowledged as an issue, Perhaps the biggest potential issue for the industry to face
and some experimental work has been done by BP [24] and is the public. There is a growing need to address public
Scottish and Southern Electricity for the now-cancelled DF1 education and perception. Poor responses in Continental
project. Further modelling work is also planned by DNV Europe and opposition in the USA are already showing
in the Pipetrans research programme. The issues around public resistance to storage and pipelines. These issues
dispersion are complex. Clear dispersion modelling and the already exist in all fields, and experience can be drawn
behaviour of a depressuring pipeline are critical elements from the UK’s position on gas storage. The reserve capacity
in both determining the major accident response and in the UK gas supply is approximately 4% of the annual
also in defining safe distances. The behaviour of carbon consumption, which compares poorly to that of France at
dioxide at the release point, the source terms, needs to be 24% or Germany at 21%. The number of potential projects
defined, and the computational models validated against it. to rectify this is significant, but the majority of projects suffer
Computational-fluid-dynamic models can be used instead in planning, face local opposition, or enter public enquiry,
of the simpler commercial programmes. However, the lengthening the development time and cost. Despite the
efficient and accurate modelling which is at an early stage urgent need for capacity, the poor educational message and
for the purposes of safety cases and route definition is key engagement has prevented a number of projects proceeding
to efficient, practical, and safe design. in a timely fashion. This challenge now faces CCS: failure
to engage with the public will have devastating effect on
Human factor project development. There needs to be project-specific
information, but also a wider education programme.
While the critical technological issues discussed here aid
design, safety, and integrity of the pipeline, the human Conclusions
factor should not be ignored. Important in the learning
from the US is that the schemes there are operated in an In conclusion, it is apparent that the knowledge gained from
oil and gas production environment. The changes that the the current carbon dioxide pipelines will prove vital. It is
not the only the experience and solutions, but also those 8. Idem, ibid.
issues that have not been fully addressed. There are critical 9. F.Schremp and G.Roberson, 1978. Effect of supercritical
areas to address, including flow assurance, dispersion, carbon dioxide on construction materials. Society of Petroleum
properties, and engagement. These need to be resolved to Engineers.
enable deployment. 10. G.Najera, 1986. Maintenance techniques proven on CO2 line.
Oil and Gas Journal.
However the body of evidence, experience, and proven 11. T.Gill, 1985. Canyon Reef Carriers Inc, CO2 pipeline
design methodology, codes, and regulation all enable description and 12 years of operation. ASME.
pipelines and infrastructure to be designed. While these 12. M.Seiersten, 2001. Material selection for transportation and
parameters may prove to be conservative and therefore disposal for CO2. Corrosion 2001.
more costly, it does not prevent pipelines from moving 13. L.Newton, 1977. Corrosion and operational problems, CO2
forward. Nor can CCS afford to wait: design for the project, SACROC Unit. Society of Petroleum Engineers.
transport infrastructure is underway now. The industry 14. M.Mohitpour, 2007. Pipeline design & construction: a practical
must maximize the transfer of knowledge from carbon approach. ASME Press, New York.
dioxide and hazardous liquid pipeline design, while parallel 15. de Visser et al., 2007. Towards hydrogen and electricity
research must address the issues discussed here and generate production with carbon dioxide capture and storage. Dynamis
resources and tools that can meet the challenge to bring Consortium.
to wide-scale deployment. 16. http://www.phmsa.dot.gov/pipeline
17. http://www.eia.doe.gov/pub/oil_gas/natural_gas/analysis_
References publications/ngpipeline/
18. J.Gale and J.Davison, 2004. Transmission of CO2 – safety and
1. www.bts.gov/publications/national_transportation_statistics/ economic considerations. Energ,29.
html/table_01_10.html 19. de Visser et al., 2007. Towards hydrogen and electricity
2. CFR – Code of Federal Regulation. production with carbon dioxide capture and storage. Dynamis
3. Building the cost curves for CO2 storage: European sector. Consortium.
Report 2005/2 IEA GHG, UK. 20. IPCC, 2005. Special report on carbon capture and storage.
4. M.Mohitpour, 2007. Pipeline design and construction: a 21. www.energy.ca.gov/sitingcases/hydrogen_energy/documents/
practical approach. ASME Press, New York. applicant/revised_afc/Volume_II/Appendix%20E.pdf
5. J.Barrie. Carbon dioxide pipelines: a preliminary review 22. A study of measurement issues for carbon capture and storage
of design and risks. 7th Int. Conf. on Greenhouse Gas (CCS). Report 2009/54, April 2009, TUVNEL.
Technologies. http://uregina.ca/ghgt7/PDF/papers/ 23. M.Mohitpour, 2007. Pipeline design & construction: a practical
peer/126.pdf approach. ASME Press, New York.
6. M.Mohitpour, 2008. A generalized overview of requirements 24. http://archivos.labcontrol.cl/wcce8/offline/techsched/
for the design, construction, and operation of new pipelines manuscripts%5C8mnkk4.pdf
for CO2 sequestration. The Journal of Pipeline Engineering. 25. www.iea.org/papers/2009/CCS_Roadmap.pdf
7. A.Oosterkamp and J.Ramsen, 2008. State-of-the-art overview 26. IPCC, 2005. Special report on carbon capture and storage.
of CO2 pipeline transport with relevance to offshore pipelines.
Polytec, Norway.
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 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.
T HERE IS CURRENTLY interest in the techno-

economics of constructing a carbon-dioxide pipeline
network for the UK to support the future deployment
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
of CCS. It is likely that early development will focus on producers of carbon dioxide. In subsequent phases, smaller
transporting CO2 from a small number of large carbon- CO2 sources are progressively added to the growing network.
capture-fitted power stations to a limited number of offshore
storage sites. As the number of carbon-capture-fitted point The aims of this paper are to:
sources increases there may be some benefit from developing
a common pipeline infrastructure. Once a ‘backbone’ • Develop an idealized, but representative, spatial
network starts to emerge, an appealing possibility is that a data set detailing the locations of carbon dioxide
large number of smaller CO2 producers could be connected sources and the quantities likely to be produced.
to the network for relatively little cost. The locations in the data set are disaggregated by
size, such that they may be connected to facilitate
To investigate the feasibility of this idea, we have analysed study of a phased development of a carbon dioxide
the techno-economics of a phased approach to rolling-out a pipeline network.
• Establish a method for modelling the phased
by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical development of a CO2 network.
and Clarion Technical Conferences, and with the support of the University of
Newcastle and the Carbon Capture and Storage Association.
• Examine the technical requirements and performance
*Author’s contact details of the network at each phase, and estimate the capital
email: t.cockerill@imperial.ac.uk
costs associated with network roll-out.
REGISTE
R TODA
Visit the Y
Clarion
website
to
register
now
www.cla
rion.org
Marriott Westchase Hotel

Houston, Texas, USA
Courses Now entering it 23rd year, the PPIM
Conference is recognized as the
Plan to be there: www.clarion.org
or call us at +1 713 521 5929
foremost international forum for
sharing and learning about best
Conference practices in lifetime maintenance PLATINUM SPONSOR SILVER SPONSOR
and condition-monitoring technology
for natural gas, crude oil and product
Exhibition pipelines.
The international gathering of the global pigging industry!
Conference Organizers
Approximate
Study Areas considered References
Year
Feasibility Study on the Transmission of CO2 North West UK 2006 [1]
CO2 Aquifer Storage Site Evaluation and Monitoring
Scotland, Yorkshire 2008-2010 [2]
(CASSEM)
North Sea Basin Task Force Study East of England 2007 [3]
Scottish Regional Carbon Capture and Storage Scotland and Northern
2009 [4]
(CCS) Study England
Yorkshire Forward Study Yorkshire and Humber 2008 [5]
Ramp-up of large-scale CCS infrastructure in Europe Europe Wide 2008 [6]
Table 1. Selected studies of UK CCS pipeline infrastructure roll out.
Brief UK literature overview

of a phased pipeline roll out, the sources were classified
into three tiers according to their annual CO2 production.
Most work on the development of CO2 infrastructure for
the UK has been performed by commercial consortia, with For the tier of the largest producers, the GIS was used to
a focus on a regional development approaches. Figure 1 identify corridors along which a backbone network of carbon
summarizes the areas considered by the first five major dioxide pipelines could be constructed. The pipeline network
studies listed in Table 1. was then designed using hydraulic analysis techniques. Finally
a simple unit cost per in-km approach was used to estimate
The literature suggests that links between regions have construction costs, assuming all pipelines are newly built.
not been considered for developing a country-wide central
CO2 transportation network. One drawback of regional A similar approach was used to add all sources in each of
approaches is the potential underutilization of storage the smaller production tiers to the backbone network. This
capacity. Similarly, local CO2 transmission and storage allowed the implications of a phased approach to network
capacity constraints could limit the extent to which CCS development to be assessed, and in particular the marginal
could be deployed within any region. This is particularly true cost of adding smaller sources to be estimated. Calculations
in the North West of the country, where storage capacity is were carried out for networks with and without intermediate
limited, but there are a large number of CO2 sources. Finally, recompression stations.
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 Identification of export terminals
maximize utilization of both capture and storage potential.
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 Identification of CO2 export terminals
connecting point carbon dioxide sources to a limited number
of export terminals located on the coast. Further offshore The UK’s existing oil and gas terminals and the nearest
pipelines will carry the CO2 to offshore storage locations, offshore oil and gas sedimentary basins with CO2 storage
but the development of these components of the network are potential are summarized in Table 2. These CO2 export
not examined here. In common with several other studies, terminals would be equipped with compressor or pumping
it is assumed that onshore/offshore connections will only units for export of CO2 to the offshore storage. The storage
be permitted at existing pipeline terminals. Hence the first potentials mentioned in the Table represent the realistic
stage of the analysis was to identity existing UK pipeline storage potential for each basin suggested by the British
terminals and their locations. Geological Survey (BGS), which has been divided against
each CO2 export terminal. At present, it is difficult to
Secondly, a spatial database of existing and prospective CO2 establish when individual fields or transport pipelines will
producers was developed within a geographical-information become available due to commercial confidentiality of
system (GIS) drawing on data from several sources. Due to information. Therefore, for this study, it is assumed that
the large number of potential sources, the database only UK’s offshore oil and gas fields will be available for CO2
included those above a threshold value. To facilitate study storage when required.
Nearest UK offshore Oil & Gas

Name of terminal CO2 Storage Capacity
sedimentary basin
St Fergus Gas terminal Northern & Central North Sea basin 1,346
Teesside Terminal
Easington/Dimlington gas terminal
Southern North Sea Basin 3,886
Theddlethorpe gas terminal
Bacton gas terminal

Point of Ayr terminal
East Irish Sea Basin 1,043
Barrow-in-Furness gas terminal
Table 2. UK existing onshore oil and gas terminals.
CO2 Emissions Range

Types of emitter
Tonnes per annum
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 Cement factories, CCGT Power stations, fertilizer, petrochemical complexes
Table 3. Classification of emitters according to emissions.
Captured CO2
Design Approach Scenario name References
volumes
Design “pipeline based” transmission network for
S1A 156
Tier-0 emitters
Pipeline based S2A 156 + 57 = 213
Tier-0 and Tier-1 emitters
S3A 213 + 15 = 228
Tier-0, Tier-1 and Tier-2 emitters
Design “pipeline + compression based” transmis-
S1B 156
Pipeline sion network for Tier-0 and Tier-1 emitters
+ Design “pipeline + compression based” transmis-
Compression based S2B 156 + 57 = 213
sion network for Tier-0 and Tier-1 emitters
Design “pipeline + compression based” transmis-
S3C 213 + 15 = 228
sion network for Tier-0, Tier-1 and Tier-2 emitters
Table 4. Definition of pipeline design scenarios.
Selection of CO2 sources

(CCR), as existing coal-fired power stations in UK with the
threshold capacity would be expected to emit approximately
For this study, all current and planned (to 2015) industrial 500,000t/a of CO2. The sources considered here therefore
and power station CO2 emitting sources in the UK with CO2 include larger power stations and large industrial sources.
emissions greater than 500,000t/a have been considered. In Non-power plant industrial sources of this scale comprise
total, these sources account for 228Mt CO2 – approximately refineries, steel manufacturers, petrochemical complexes,
50% of UK current CO2 emissions. fertilizers and cement factories. CO2 emissions data for
the selection process were taken from a variety of sources,
The cut-off threshold is broadly consistent with the UK including the Environment Agency, published data under
Government policy that any new combustion power station the EUETS programme, and the annual reports of each
at or over 300MWe should be built ‘carbon-capture ready’ emitter. Figure 3 illustrates the approach.
Fig.2. Analytic approach used in

this study.
Division of emitters by CO2 emissions range and oil power stations with generation capacities in
the range of 300–500MWe. Across the tier there are
Emitters above the threshold value have been categorized 13 existing power stations, 16 industrial installations,
into three tiers as set out in Table 3. The division points and three new power projects, totalling 15Mt/a (7%
were selected based on engineering judgment, with each of the total UK emissions considered).
tier distinguishing decreasing importance of the individual
emitter for UK greenhouse gas emissions. Pipeline development scenarios
• Tier-0 emitters: Tier-0 comprises emitters producing
CO2 emissions of 3Mt/a and above, typically large The onshore CO2 pipeline transmission network for UK
coal-fired power stations with generation capacity is designed in this study by considering following capacity
from 1.5-4.0GWe. The largest emitter in this category phased approach:
is Drax power station, with the largest industrial
emitter being Corus’ steel works. In total there are 20 • Scenario 1 (S1): Evaluate the design and topology
existing power stations, five industrial installations, of a CO2 onshore ‘backbone’ transmission network
and five new power projects, totalling 156Mt/a (68% to collect CO2 emissions from Tier-0 emitters only.
of the total UK emissions considered).
• Scenario 2 (S2): Re-evaluate the design of an onshore
• Tier-1 emitters: Tier-1 emitters are typically CCGT CO2 transmission network where Tier-1 emitters are
power stations and refineries, producing CO2 added to the already laid transmission system for
emissions in the range of 1–3Mt/a. The total inventory Tier-0 emitters. Determine the additional changes
includes 19 existing power stations, 10 industrial required in the transmission network to handle
installations, and eight new power projects, totalling combined flows of Tier-0 and Tier-1 emitters to the
57Mt/a (25% of the total UK emissions considered). CO2 export terminals.
• Tier-2 emitters: Most Tier-2 emitters are industrial • Scenario 3 (S3): Complete the design of the onshore
sources, typically cement and petrochemical plant, CO 2 transmission system by considering the
with emissions ranging from 0.5-1Mt/a. Power stations connection of Tier-2 emitters to the transmission
included in the Tier-2 band are predominantly CCGT system already in place for Tier-0 and Tier-1 emitters.
Fig.3. Approach to developing CO2

sources spatial database.
Two separate types of transmission models have been flange of the CO2 compressor installed at export terminals
developed for each scenario, one based entirely on to inject gas in offshore pipelines for subsequent storage in
pipelines without re-pressurization (A) and the other using offshore oil and gas fields.
combinations of pipeline and intermediate compression
stations (B). The full set of scenarios is summarized in Table 4. Another assumption is that CO2 streams from all the
sources will be dehydrated to -5°C dewpoint, representing
Data and assumptions 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
Pipeline design assumptions part of the CO2 capture plant. The CO2 streams from Tier-2
emitters would be collected together at different locations
Many factors have to be considered in the design of the for centralized drying. For deciding the -5°C water dew
new pipelines (assumed buried underground), including point of CO2 streams from each emitter, the available UK
the properties and quantities of the fluid to be transmitted, soil temperature data at 30cm depth for peak winter month
underground conditions, and safety requirements. Some key (January) has been used. Since the majority of emitters are
data are set out in Table 5, with further discussion following. located in England, an average subsurface soil temperature
of 3–5°C is taken for this study.
It is assumed here that CO2 transportation systems will
be based on new onshore carbon steel pipelines, with a Network topology
maximum diameter of 1,067mm. No re-use of existing
infrastructure is accommodated within the study. The A crucial factor in the design of any pipeline transmission
quantities of CO2 provided to the network by each source network is the determination of pipeline route corridors.
have been estimated by taking a carbon dioxide capture In this study, it is assumed that wherever feasible, the CO2
efficiency of 90% for power stations and 60% for industrial transmission network will follow the existing route corridors
installations. of onshore oil and gas pipelines in the country. However
some new pipeline route corridors have been proposed for
The networks are designed for carbon dioxide in a the emitters which are located away from existing oil and
supercritical state. Keeping in view the critical point of gas onshore pipeline infrastructure.
pure CO2, i.e. 74 bar, the minimum pressure at which CO2
would leave each source is taken to be 95 bar. This allows Cost data
for a pipeline network pressure drop of 20 bar above the
CO2 critical pressure. The operating pressure of the pipeline Detailed construction costings for CO2 pipelines are difficult
transmission network is specified at 100 bar due to the to obtain thanks to the limited number of CO2 pipelines
limitation of the operating pressures of the flanges and operating worldwide. Instead, we assume that new carbon
fittings at intermediate facilities. For intermediate booster steel pipelines will be laid, and rely on approximate cost data
stations, where used, it is assumed that recompression would for natural gas pipelines construction drawn from the IPCC
be required when the pressure drops to 85 bar after which special report [7] and the Oil and Gas Journal 2005-2008
the pressure will be boosted again to 100 bar. It is assumed [8]. In this study, the base unit pipeline total construction
that arrival pressure at each CO2 export terminal would be cost (including materials, labour, and all works) is taken to
85 bar. This arrival pressure is the pressure at the upstream be US$ 30,000/in–km, including pipeline material costs,
Fluid
CO2 purity 100%
Phase of CO2 Supercritical
Critical Temperature 31°C
Critical Pressure 74 bar
Pipeline
Pressure rating of valves & fittings 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 net-
110 bar
work
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
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 All cost data have been converted to UK Pounds Sterling
excluding additional compression. The impact of pipeline using 2008 exchange rates.
construction cost increases over this base level is later
evaluated by performing sensitivity analysis. Hydraulic modelling
The cost of intermediate booster stations, where employed, Introduction
has been estimated based on literature data. A review of
several studies [9, 10, 11, 12] suggests that capital costs for For designing the CO2 pipeline transmission system, a state-
installed compressor stations would range from $1500-4800/ of-the-art hydraulic simulator, PipelineStudio version 3.0
kW. This wide variation can be attributed to differences in has been used to simulate the transmission pipeline system
the year of installation, type of compression, geographical for steady-state operation. For all scenarios, the pipeline
location, and installation of intermediate heating or network model has been constructed using:
cooling facilities. For the results reported here, a value
for compression station costs towards the middle of the pipeline lengths and elevations from the GIS database,
literature range, specifically $2500/kW, is adopted. This CO2 volumes from each emitter and location of CO2
encompasses the cost of buildings, compressor units, prime export terminals,
mover, pipeline fittings, and drying equipment. subsoil temperature data depending upon the region.
Only the costs of intermediate booster stations are accounted Simulation methodology
for, with compression at sources and export terminals excluded
from the calculations. However, for the booster stations, it is For the hydraulic simulation of the CO2 transmission
assumed that 50% of the compression requirement would network, the arrival pressure at each CO2 export terminal
be available as stand-by in order to make up in case of any has been fixed at 85 bar whereas the captured CO2 volumes
emergency or during maintenance of main unit. are fed at the source nodes. The pipeline diameters are then
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.
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 where
mixtures, the Peng–Robinson equation of state has been
used. While keeping the input flows constant, the pressures HP = compressor horsepower
at each emitter are then back-calculated by the software by γ = ratio of specific heats of gas, dimensionless
using the ‘general flow equation’ after Menon [11]: 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
where Z2 = compressibility of gas at discharge conditions,
dimensionless
Q = gas flow rate, measured at standard conditions, ηa = compressor adiabatic (isentropic) efficiency,
ft3/day (SCFD) decimal value
f = friction factor, dimensionless
Pb = base pressure, psia Mass balance
Tb = base temperature, R (460 + F)
P1 = upstream pressure, psia The pipeline transmission model drawn in the network
P2 = downstream pressure, psia has been verified by balancing the inputs and outputs of
G = gas gravity (air =1.00) the transmission network. For steady-state calculations it is
Tf = average gas flowing temperature, R assumed that there will be no accumulation of CO2 volumes
L = pipe segment length, mi in the pipeline transmission network.
Z = gas compressibility factor at the flowing
temperature, dimensionless Results
D = pipe inside diameter, in
For compression-based scenarios, pipeline diameters of the Main network characteristics

transmission model have been optimized while considering
intermediate booster stations. The intermediate booster Table 6 summarizes the main technical results of the
stations are placed where the pipeline pressure drops to 85 simulation and optimization study, and the network layouts
bar after which the pressure is boosted again to 100 bar. The are shown in Fig.4. It is clear that the pipeline-only scenarios
power requirement at each booster station are then calculated (A) have pipeline diameters larger than the scenarios where
by the software using Eqn 2, again after Menon [11]: re-compression is used (B). Without re-compression, larger
Emission Sources: Tier-0 Tier- 0 + 1 Tier- 0 + 1 + 2

Scenario S1A S1B S2A S2B S3A S3B
Basis: Power station CO2 capture Eff = 90% , Industrial CO2 capture Eff = 60%
CO2 volume reaching export terminals
1- Bacton Gas Terminal 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
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 re-
compression 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 re- Fig.5. Summary of CO2 costs for all base scenarios.
compression offer more attractive techno-economics than
those without. However, re-compression based networks will for here, and more-detailed work is required to determine
have increased operation and maintenance (O&M) costs in the optimum combination of compressor requirements and
comparison to pipeline-only systems: these are not accounted related O&M costs.
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 to a network with compression is 44% higher than that for
transmission network increases with the addition of smaller the Tier-0 sources. The comparable marginal cost increase
emitters. This trend is reflected in transportation costs, for a network without compression is 21%. Incorporating
because the extra pipe length required increases more quickly increasingly small sources demonstrates poorer returns, as
than the additional CO2 provided by each tier. Adding further adding Tier-2 sources to networks with and without
all Tier-1 and 2 sources to the network would produce compression provides further marginal cost increases of 94%
an increase in per tonne transportation capital costs of and 109% respectively. The value of adding such smaller
approximately 15% if no re-compression is used and 22% sources to a network must be questioned, particularly if they
in a system that relies on re-compression. are remote from natural clusters of carbon dioxide sources.
Analysis Sensitivity analysis
For this study, sensitivities have been calculated to determine

Marginal tier costs the impact of different parameters on CO2 transportations
costs. Results of each sensitivity analysis have been compared
The marginal costs for each scenario have been calculated with a base case.
to determine the extra investment required for the addition
of emissions of lower tiers by dividing the incremental Change in pipeline network length
capital costs of each scenario by the incremental increase
in emissions. The results are shown in Fig.6. This sensitivity analysis has been performed to determine
Marginal costs increase with the addition of emissions from the impact of 15% and 20% increases in pipeline length on
the lower tiers. The marginal cost of adding Tier-1 sources CO2 transportation cost, perhaps due to the need to avoid
Pipeline only based scenarios

Captured Capital Cost CO2 Transportation Cost

Scenario 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
Pipeline + compression based scenarios
Captured Capital Cost CO2 Transportation Cost

Scenario Emissions 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 228 1,490 56 1,546 6.53 0.24 6.77
Table 7. Summary of CO2 transmission network costs for pipeline and compression-based scenarios.
geographical obstacles not accounted for in the relatively volume of carbon dioxide flowing though the network, and
high-level study. For the base-case scenario, the pipeline hence the design and infrastructure requirements.
route was evaluated using a GIS database. It was assumed
that pipeline diameters, CO2 emissions, and pipeline unit Two perturbations of the base-case were examined, changing
construction costs remained constant in all cases. However, the capture efficiency at the power station sites to 85% and
in scenarios involving re-compression, compression powers 95%. It was assumed that CO2 capture efficiency of the
were calculated again. industrial units remained constant at the base-case value.
Using the new flow volumes, new hydraulic simulations were
Further hydraulic simulations determined the operating carried out, assuming that pipeline lengths and route corridors
pressures of the transmission network and pressures required will remain unchanged. Only changes in pipeline diameters
at sources. For a 15% increase in pipeline length, pressure and compression requirements have been re-calculated.
at sources increased by an average of 2% as compared to the
base-case pressures. Similarly for a 20% increase in pipeline In all cases the network pipeline diameters increased
length, the pressure at sources increased by an average of or decreased in accordance with the change in carbon
3%. In all cases, the calculated pressures remained below dioxide flow volumes. For the compression-based cases,
the maximum allowable pipeline operating pressures as well the compression energy requirements increased for the
as adequately above the CO2 critical pressure. 95% capture case and decreased for 85% capture case.
However for the 95% capture scenario, CO2 transportation
Costs were then re-evaluated for the increased in pipeline cost decreased compared to the base-case whereas for 85%
lengths, unsurprisingly showing an increase in transmission capture scenario, the unit cost increased. The decrease in CO2
costs in all cases (Fig.7). transportation cost for 95% case scenarios is countered by
increased capital cost due to using larger-diameter pipelines.
Change in CO2 capture plant efficiency Similarly for 85% scenario, pipeline diameters have been
reduced due to a reduction in emissions volume. The change
This sensitivity analysis has been performed to determine in CO2 transportation costs in all cases is comparatively
the changes in CO2 transmission network resulting from small, at around 2%.
changes in the CO2 removal efficiencies of the capture plants
installed at each point source. The base-case assumption Increase in pipeline construction costs
in this work is that the CO2 removal efficiency for power
stations is 90% and 60% for the industrial installations. Any This sensitivity analysis investigates the impact of increased
change in capture efficiencies will increase or decrease the pipeline unit construction costs, representing the effect
of increased steel prices, difficult construction terrain, or without re-compression. The CO2 transport cost per tonne
other project-management-related factors. As part of this are overall smaller, as the expense of the compression stations
sensitivity analysis, a sub-analysis has been conducted to is outweighed by the reduced cost of smaller-diameter pipes.
check the impact of increased pipeline construction costs Our simplified estimates take no account of operations
on the previously-calculated sensitivities to CO2 capture and maintenance costs, meaning this difference is certainly
efficiencies and increased pipeline lengths. within the range of uncertainty for the results. More-detailed
work is required to determine the optimum combination of
The results of the combined sensitivity analysis shown in compressor requirements and related O&M costs.
Fig.7 indicate the range of CO2 transportation costs across
all the cases considered. In all cases the transportation These general trends are hardly influenced by changes in the
costs of compression-based scenarios are less than for major underlying assumptions and design choices, although
non-compression scenarios. The average costs of the three of course the absolute value of the CO2 transportation cost
scenarios – shown as small, light-coloured squares in Fig.7 does vary substantially.
– range from 8.24 £/tonne for scenario S1 to 9.2 £/tone
for scenario S3; the cost variability within each scenario is References
rather larger, approximately 7-8 £/tonne.
1. GASTEC at CRE Ltd (GaC), 2006. Feasibility study on the
Conclusions transmission of CO2. GaC Report 3484, October [Available
at www.gastecuk.com/case-studies-detail.php?id=3 ].
Approximately 50% of UK industrial and energy CO2 2. G.Pickup, 2009. CASSEM Overview, November, www.geos.
emissions are produced by emitters that generate more than ed.ac.uk/ccs/UKCCSC/Pickup.pdf [Consulted May 2010].
500,000Mt/a of CO2. Providing a CO2 pipeline transport 3. UK DBERR. Development of a CO2 transport and storage
network for each of these sources has the potential, assuming network in the North Sea,
they are also retro-fitted with carbon capture equipment, to [Available at www.nsbtf.org/documents/file42476.pdf ].
facilitate a major reduction in UK CO2 emissions. 4. Scottish Centre for Carbon Storage & Scottish Government,
2009. Opportunities for CO2 storage around Scotland. [www.
The conceptual design and techno-economics of a phased geos.ed.ac.uk/sccs/regional-study/CO2-JointStudy-Full.pdf].
approach to rolling out such a network have been investigated, 5. Yorkshire Forward, 2008. A carbon capture and storage network
subject to a number of simplifying assumptions. If a for Yorkshire and Humber.
‘backbone’ network connecting the very largest sources is [Available at: www.yorkshireforward.com/sites/default/files/
constructed first, smaller sources could be later added to the documents/Yorkshire%20%20Humber%20Carbon%20
network with a relatively small impact on the transportation Capture%20%20Storage%20Network.pdf].
cost per tonne of carbon dioxide. Typically, adding all Tier-1 6. J.Kjärstad and F.Johnsson, 2008. Ramp-up of large-scale CCS
and-2 sources to the network would produce an increase infrastructure in Europe. Int. J. of Greenhouse Gas Control, 2,
in per unit transportation costs of approximately 15% if 4, pp417-438, October.
no re-compression is used and 22% in a system that relies 7. E.S.Menon, 2005. Gas pipelines hydraulics. CRC Press, May.
on re-compression. 8. Ramgen Compressors, 2010. Ramgen’s low-cost, high-
efficiency CO2 compressor technology. www.ramgen.com/
Considering the marginal cost of making these additions, apps_comp_unique.html [Consulted May 2010].
however, tells a different story, as increasingly large expense 9. D.Simbeck and E.Chang, 2002. Hydrogen supply cost estimate
is required to add sources of rapidly decreasing size. The for hydrogen pathways scoping analysis. US National Renewable
marginal per tonne cost of adding Tier-1 sources to a Energy Laboratory (NREL) Report/SR-540-32525, November.
network with compression is 44% higher than that for 10. US Department of Energy, 2007. Conceptual engineering/
the Tier-0 sources. The comparable marginal cost increase socioeconomic impact study – Alaska spur pipeline : Appendix
for a network without compression is 21%. Incorporating 3-5: compressor cost estimate. Report on DOE-NTL Contract
increasingly small sources demonstrates diminishing returns, No DE-AM26-05NT42653, January. [Available at www.
as further adding Tier-2 sources to networks with and jpo.doi.gov/SPCO/DOE%20Spurline%20Documents/
without compression provides further marginal per tonne Appendix%203-5%20Compressor%20Cost%20Estimate.pdf]
cost increases of 94% and 109% respectively. The value of 11. B.Metz, O.Davidson, H.de Coninck, M.Loos, and L.Meyer,
adding such smaller sources to a CCS network must be (Eds), 2005. IPCC special report on carbon dioxide capture
questioned, particularly if they are remote from natural and storage, IPCC.
clusters of carbon dioxide sources. 12. Oil and Gas Journal (various issues, 2005-2008). PennWell
Petroleum Group.
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
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 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)
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
and the need for GHG management. The extent of concern seeping back to the surface. This could have consequences such
is evident when ‘GHG + greenhouse gases’ is entered into as litigation, and/or other issues, like verification monitoring
a web browser, which recently pointed to 15,300,000 hits. or economic drivers, which are beyond the present scope. As
Tracking the early history of GHG management leads to a result, transport became a necessary aspect of CCS, such
sites associated with the intergovernmental panel on climate that moving anthropogenic CO2 to sites or schemes better
change (IPCC, whose formation traces to the late 1980s), suited to its retention or mitigation is now a consideration
and other agencies such as the International Energy Agency in this process. The need for transport in conjunction with
(IEA), and its focus on GHG (IEAGHG). CCS implies that CCST might be the appropriate acronym, in
lieu of CCS. However, searching CCST (as ‘CCST + carbon
Early schemes to capture and store anthropogenic CO2 capture’) does not quickly lead to transport, but rather points
were based on its injection into the earth at sites local to to terms like CCS technology or training, possibly implying
its generation, which precludes the need for transport, that transport (aside from economic considerations) does
and potentially underlies the commonly used acronym not pose concerns in parallel to the other elements of CCS.
1. With CCS defined in reference to managing GHG implies that re-injecting

*Author’s contact details naturally sourced CO2 in support of EOR is not CCS – but simply a commercial
tel: +1 614 424 4421 operation that returns the CO2 below ground. In the same vein, EOR supported
email: leis@battelle.org 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
JAN 2011 January 25–28 Subsea Production Systems Engineering (Aberdeen)
January 31–
Subsea Pipeline Engineering Course (Amsterdam)
February 4
January 31–
In-line Inspection of Pipelines (Amsterdam)
February 4
FEB 2011 February 1–

Defect Assessment in Pipelines (Amsterdam)
February 3
February 14–15 Defect Assessment in Pipelines (Houston)
DOT Pipeline Safety Regulations – Overview and Guidelines for Compliance
February 14–15
(Houston)
February 14–15 Pigging & In–line Inspection (Houston)
February 14–15 Pipeline Repair Methods / In–Service Welding (Houston)
Introduction to Excavation Inspection & Applied NDE for Pipeline Integrity
February 14–15
2011
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)
MAR 2011 March 17–18 Microbiological Corrosion in Pipelines (Houston)

March 21–25 Onshore Pipeline Engineering (Houston)
March 30–31 Unpiggable pipeline solutions forum (Houston)
APR 2011 April 25–29 Deepwater Riser Engineering Course (Houston)

April 25–29 Subsea Pipeline Engineering Course (Houston)
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
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?
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; Fig.1. Photo after a running ductile fracture along a full-scale
• the role and significance of trace impurities; and test section (circa 1970s).
• the volume transported that drives the use of larger-
diameter pipelines, which tend to run longer distances
with some traversing high-consequence areas.
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?
Fig.2. Retrofit fracture arrestors (courtesy of Clock Spring): (a -
While consideration of and concern for running fracture top) successful fracture arrest on a CO2 pipeline; (b - bottom)
are second nature to specialists in this technology, many in installation of one retrofit arrestor scheme.
the design community seem unaware of the details and/or of the propagating crack. This opening is evident in Fig.1,
the safety implications, while others seem unaware that the where the upper quadrants of the pipe have opened. The size
decompression of a supercritical CO2 pipeline can involve of the flaps reflects the energy associated with the passage
multi-phase response that cannot be represented by the of the crack and the extent of the energy stored prior to
broadly available analysis for single-phase gas behaviour. its passing. Flap formation and the extent of the opening
Such scenarios are known first-hand to the authors within develop longitudinal and circumferential stresses ahead of
the last two years in the context of CO2 pipeline design, so the crack. These stresses cause thinning of the wall thickness
such considerations are neither hypothetical nor are they and induce significant ovality in the pipe’s cross-section.
relegated to the distant past. Figure 1 shows the effects of the longitudinal yielding ahead
of propagation and adjacent to the crack, which is evident
Fracture propagation occurs following the unlikely event of in the wavy response apparent on either side of the crack
fracture initiation that leads to a rupture, because the stored over the upper half of the pipeline. Some fracture-arrestor
energy in certain transported fluids, such as supercritical concepts rely on constraint of flap formation, on the
CO2, is sufficient in a high-pressure transmission pipeline to presumption that the stresses due to the flap inertia and the
sustain the unstable axial extension of that rupture. Where related dynamics contribute to the crack’s advance. Other
the decompression front that develops as expansion waves running ductile fracture-arrestor concepts act to reduce
propagate back into the pipeline runs axially at a speed the wall stress, or dissipate the fracture energy, or provide
greater than the fracture speed, the pipeline depressurizes enhanced fracture resistance. In addition, arrestors can act
faster than the fracture propagates, leading to fracture to ‘ring-off’ the cracking, although violent ring-off should be
arrest when the pressure decays to a level that no longer avoided due to the chance of fracture re-initiation.
will support axial extension. In contrast, so long as the
crack-tip sustains pressure above that level, unstable axial In addition to considerations such as loss of service and/or
propagation continues. Figure 1 (from Battelle’s archives) the cost to replace potentially significant lengths of pipeline,
shows the outcome of propagating fracture following a full- the consequences of running fracture for some transported
scale test to quantify this behaviour, which is illustrative of products and pipeline locations require the certainty that, if
ductile propagation that has caused failure on pipelines initiated, such fractures would be quickly arrested. Considering
while in revenue-service. the cost of lost service, product make-up for uninterrupted
delivery contracts, and pipeline replacement, the lost value
Factors that can contribute to arrest include: can easily amount to millions of dollars per mile through
which the fracture propagates, depending on the location,
• dissipation of energy through the inherent toughness product, and length of the propagation. However, far greater
of the linepipe steel and its strain-hardening losses can accrue due to litigation consequent to such a
characteristics; failure. It follows that there is a need for rational schemes
• reduction in local stress, due to increased wall thickness to assess and quantify fracture propagation and arrest in any
at a road crossing or comparable condition; and scenario where a running ductile fracture is plausible. Models
• the effects of external factors that act to restrain to quantify fracture propagation and arrest will be presented
crack advance or lower the net load carried by the following consideration of the key differences between pipelines
pipecwall, such as due to a fracture arrestor. supporting EOR versus those involved in CCS service.
Early in-service failures involving propagating or running Early pipelines supporting EOR in
fractures occurred via ‘brittle’ fracture, which showed
limited dissipation in terms of either deformation or contrast to CCS applications
crack-tip response. As changes to the pipe steels, coupled
with appropriate steel specifications, managed concern for This section elaborates on:
running brittle fracture, it became apparent that a running
‘ductile’ fracture was also possible. In contrast to the brittle • the need to retrofit some of the early CO2 pipelines
scenario, a running ductile fracture exhibits significant plastic to provide for fracture control;
deformation and involves cracking mechanisms associated • the role and significance of impurities (including
with locally ductile stretching, leading to void nucleation, trace levels in some cases); and
growth, and coalescence. • the volume transported that motivates use of
larger-diameter pipelines running longer distances
As fractures propagate axially along a pipe, the fracture tends and traversing high-consequence areas to identify
to open in the wake of the crack, creating what have been differences in the safety and risk aspects for dome-
termed ‘flaps’. For a running brittle fracture, the extent of sourced pipelines that support EOR versus those
flap-opening can be very limited, with some distribution designed for CCS applications.
pipe materials like polyethylene remaining tight in the
wake of the crack due to residual stresses induced in pipe 2. Some other practical concerns are provided for in design, such as hydraulics,
compression/pumping, repair, start-up/shutdown, in addition to fracture control.
manufacture. In contrast, with a running ductile fracture Some of these impact efficiency and cost-drivers, while others impact safety and risk,
the flaps open and can fully flatten the pipe in the wake with the focus here being fracture control because of its safety and risk implications.
The focus here is aspects in which the design basis for such lines Concern for running fracture has been recognized for
supporting EOR will underestimate what must be specified decades for any pipeline transporting compressed gases or
to ensure safe design in CCS applications, underscoring the supercritical fluids, with the latter being the common state
fallacy of CO2 pipeline design as a mature technology. used for CO2 transport due to related efficiencies. Yet the
threat it poses is not evident in what was termed mature
The need to retrofit fracture arrestors on CO2 technology. The need to address fracture control also was
pipelines emphasized in regard to CO2 pipelines in textbooks on
pipeline design [12], which were published well before the
The view that pipeline transport of CO2 is a mature IPCC and US Congressional reporting. This need has
technology can be considered flawed unless there is a been emphasized by some regulators [13], and included in
mature basis to quantify fracture-arrest requirements for some pipeline regulations for decades [7], with the need
such applications2. This view reflects the observation that for fracture control becoming topical again, circa 2005
some early-design CO2 pipelines supporting EOR had to be [14, 15] into 2007 [16]. As becomes clear later, confidence
retrofitted to manage concern for running fracture3 [3, 4], in future design cannot be gained from such testing, nor
a need that was realized after the first of the CO2 pipelines can the design process for the pressure boundary of such
went into service. Likewise, it reflects the observation that pipelines – however mature – be taken as the benchmark
while regulations covering CO2 pipeline design require for other pipelines, unless all relevant design parameters
that fracture arrest be considered since 1988 [7], it is only are comparable.
in 2010 that a recommended practice to address this aspect
became available [8]. Finally, this view follows from the Role of impurities
observation that the guidance available in Reference 8 is
performance-based rather than prescriptive, with the tools The second key discriminator is the presence of impurities
needed to implement that guidance are still in development and their impact on the fluid’s properties, which becomes
or in are use only by specialists, and generally lack full-scale clear in contrasting results for near-pure naturally occurring
proof of their utility/applicability, except for pure CO2. ‘dome’ CO2 to that for some EOR service, and the flue gas
Accordingly, the design basis for CO2 pipelines is well anthropogenic CO2 mix emitted in fossil power generation4,5.
short of mature, and use of the design basis of pipelines While knowledge of and concern for this aspect also traces
built to move nearly pure naturally occurring CO2 can be to the 1980s [17], it too became topical circa 2005 [14], and
misleading when employed where the transported product since has been emphasized in the broadly available literature
is typical anthropogenic CO2. [18-20], particularly in regard to CO2 compression and
pipeline transport.
Several resources that trace to the 1980s [9-11] provide
perspective for the running-fracture concerns in CO2 It is apparent from the work reported circa 2005 [14] that
pipelines designed then to support EOR. While such the saturation pressure changes significantly due to the
resources underlay this retrofitting, such is not evident in the presence of modest impurity levels, and it can be inferred
post-2000 IPCC and US Congressional reporting. Reference that the critical temperature will show a comparable
9 couples two similar papers that reflect on the unique traits dependence. While such outcomes are specific to the
of supercritical CO2, while Reference 10 outlines the threat EoS that was used (the Peng-Robinson [21] (P-R) EOS was
posed by long-running ductile fracture, as does Reference adopted for that work), nevertheless the concern for the
11, so the concern for complexity well beyond mature role of impurities was clear then, which given the impact
technology was in print and readily accessible for decades. of saturation pressure on fracture propagation and arrest
must be addressed in CO2 pipeline design. Recent work
In regard to retrofitting as discussed in Reference 3, a series indicates the P-R EoS can significantly underestimate the
of full-scale fracture-arrest tests was conducted to assess the saturation pressure [22] in contrast to other EoSs, and
utility of wrap-on fracture arrestors for such applications, but the density dependence of fluid on pressure, depending
not broadly published. Figure 2a shows a view of a successful on the initial conditions. In turn this means that fracture
arrest (in a full-scale test) from that work. It is apparent from control of pipelines in CCS service based on the P-R EoS
this photograph that while the fracture ran up to and below could be less conservative than required if referenced to
the arrestor, the pressure was reduced and the pipe restrained the inappropriate mix of CO2 and impurities, and their
sufficiently to limit its propagation beyond the arrestor. Figure relative percentages. This becomes an even more acute
2b illustrates the installation of such arrestors, although this issue if their design basis was benchmarked to outcomes
is not specific to a CO2 pipeline application. References 9to for nearly pure CO2 (i.e., typical EOR cases).
11 are notable in regard to the knowledgebase concerning
the threat posed running fracture in CO2 pipelines circa the 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
1980s, with other resources also available then. absolute or relative levels, and their effects differ significantly – for example, see
Sass et al. [14].
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 5. Continuing work indicates that nominal impurity levels reported for a given
stages beyond that into construction, addressed the concern for running fracture by process can vary significantly from what is considered nominal due to fuel and
applying fracture arrestors during construction [5, 6]. process variations.
Fig.3.The BTCM.
Transport volume, distances, and routeing

also be addressed. This particularly true given that some
The third key discriminator involves the volume of CO2 early-use EOR pipelines required retrofitting.
to be transported, the distance moved, and the locations
traversed. While the above two aspects affect differences in The sizes of the two CO2 pipeline segments commissioned
the threat posed by running fractures, parameters including in the US in support of EOR prior to 1980 involved a
transported volume affect pipe size and pressure, which diameter of 16in (406mm) or less, with the pipe wall made
like routeing and distance, impact public exposure and of Grade X65 (448MPa) or lower. The sizes and grades
the consequences associated with the unlikely occurrence of the CO2 pipelines built in the interval thereafter, but
of fracture initiation. While there is no simple avenue to before CCS transport became a major consideration, had
determine the eventual volume, it has been stated [1] that diameters that ranged up to 30in, which were typically
“pipelines can be expected to play a significant role in the built in Grade X70 or below. Pipelines supporting EOR
required transportation infrastructure.” This indicates it is tended to run through quite remote areas, with source sites
likely that many CO2 pipelines will be built for CCS-related selected, in general, as close as possible to the reservoir
service, which will develop in addition to the existing network being worked to minimize the length of the pipeline. For
of more than 30 pipelines supporting EOR6 [23]. this reason, well over half of these pipelines have a length
less than 100miles (160km), with almost all having lengths
Reference 23 designates three types of CO2 pipelines in less than twice that distance. In general, the diameter
EOR service circa 2007, with the discriminator being the of such lines is also small in comparison to natural-gas
transported product specification. What was termed Type I transmission pipelines, with well over half having a diameter
covered special, single-use pipelines with case-by-case (rather of 12in (305mm) or less, with almost all having diameters
open) specifications, with Type II being multiple source / less than twice that size.
user lines and ‘strict’ specifications (this type was noted as
typical of most of the North American network), and Type III In contrast, to the authors’ knowledge, the diameter of lines
represented hybrid lines, with relaxed but ‘controlled’ CO2 currently now built in the US for CCS service run up to
composition. These three scenarios differ somewhat from 24in7 (610mm), and are made in Grades up to X80 (551MPa).
the compositions moved by pipelines supporting early EOR Trunklines are under consideration for anthropogenic CO2
– say prior to 1985. Of the six or so CO2 pipelines operating that will run from the north to the south of the US, leading
circa 1985, about half moved nearly pure dome-sourced to distances the order of five to ten times that just cited.
CO2, while the others moved CO2 with small amounts of Currently constructed pipelines for CCS service, such as
hydrocarbons whose levels depended on the source field and the Green Pipeline, run through high-consequence areas
its variability. If history is to serve as a benchmark for current (HCAs) that pose a public concern given the asphyxiative
practices, the role of impurities must be considered. If that properties of gaseous CO2, and its terrain-tracking transport
same history is to be a benchmark for CCS applications – as a dense (heavier than air) vapour for onshore pipelines,
where CCS involves anthropogenic sources given its GHG but are much less an issue in an offshore context [25]. Its
roots – then differences in impurities and their levels must properties as a supercritical fluid lead to unique concerns,
which increase nonlinearly with diameter given the
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 7. As indicated above there are some quite large-diameter segments supporting EOR,
as a consequence of CCS, nor is it naturally occurring in the form it is transported. notably the Cortez pipeline [5].
functional dependence of running fracture on diameter [26]. illustrates this graphical scheme circa the 1970s. The curve
As such, the requirements for fracture control of pipelines labelled ‘fracture’ reflects the relationship between fracture
transporting nearly pure CO2 in the volumes associated with speed and pressure as a function of toughness, which was
EOR are much less demanding than for the typically larger- plotted by trial and error until the fracture speed for a given
diameter systems anticipated for service transporting CO2 toughness was tangent to the curve labelled ‘gas’ that reflects
containing impurities typically associated with CCS. Tables speed of decompression in the wake of the expansion waves
2 and 3 in Reference 14 identify possible trace impurities by propagating back into the pipeline. Arrest is ensured by use
source type and summarize their potential effects on capture, pipe steel that is specified with that or greater toughness.
compression, pipeline transmission, and injection, and so As becomes is evident later, arrest occurs rapidly once the
are useful assessing related safety implications. stress at the crack-tip drops below its critical value.
Each of the above three considerations points to the need By the late 1970s analytic approaches [27] and semi-analytical
to address running fracture in regard to CO2 pipelines in schemes [28, 29] appeared, as did empiricism in the form
CCS-related service. Likewise, it was evident that differences of curve-fits to full-scale test data [30]. The mid-1980s saw
between pipelines designed to support EOR versus transport continued attempts to analytically capture this phenomenon
mixtures of CO2 and the trace impurities common to CCS [31-33], including a comprehensive energy-balance
service that precludes use of existing pipelines as design formulation, and a large-scale numerical formulation that
benchmarks. On this basis, the next section considers made use of crack-tip opening angle (CTOA) as its measure of
approaches to quantify fracture propagation and arrest in cracking resistance. After significant work to refine the metric
CCS applications. As noted above, such considerations are for fracture resistance, the formulation solidified [34], but its
not new – they have been with us since the 1970s for natural blind application to predict arrest for the Alliance Pipeline
gas transmission [26], the 1980s in reference to dome-sourced full-scale tests led to values of CTOA the order of 25º for
near-pure CO2 pipelines [4-6], and more recently in regard arrest [35], well beyond the moderate toughness levels actually
to CO2 pipelines in CCS service [15, 18-20]. Consequently, required. This gave rise to redefinition of CTOA and some
what follows is a brief introduction and review, which suffices reformulation of the model, which resulted in a more correct
as background to assess the viability of the assumptions and outcome the order of 11-12º [36]. Subsequently, definition
the viability of the model and its predictions. of CTOA was further updated [37], and then revised again
in the late 1990s along with additional ‘tweaks’ [32], which
Modelling to quantify fracture contributed to its improved case-specific predictability. While
this work continues, primarily in Italy [34-38] and the UK
propagation and arrest [39], and holds much promise, day to day arrest-toughness
predictions still make use of the BTCM, except as modified
Technology to quantify running fracture spans from raw to address higher-toughness steels [40, 41]. Subsequent work
empiricism into numerical formulations which, while has addressed other issues, but relies on the same concepts
potentially elegant in concept, still embed empirical [42, 43]. In this context there are three approaches to make
calibration(s). While the rapid evolution of fracture- predictions for CO2 applications:
mechanics’ theory and technology development that began in
the 1960s and has continued since provides the foundation, • the BTCM as is [16, 44];
running ductile fracture is a complex phenomenon. Running • adaptations of GASDECOM with other elements
ductile fracture couples fluid and solid mechanics with of the BTCM largely intact [45, 46]; and
fracture mechanics, gas dynamics, and thermodynamics. • use of the concepts that underlie the BTCM with a
The coupled nonlinearities of these disciplines involve soil- return to first principles, as needed [22].
structure interaction for buried onshore pipelines, and its
parallel offshore, where the fact that pipelines can operate As for dense-phase (rich) natural gas, the expansions waves
at significant depths adds further complexity. leading to decompression for CO2 propagate through a one
or two/multi-phase medium. The gas dynamics’ aspects of the
Work through the mid 1970s led to what has been referred BTCM were formulated under the assumption that the flow is
to as the Battelle Two-Curve Model [25] (BTCM), which a homogeneous isentropic process. The model was formulated
reflects the significant efforts of Maxey [26]. This formulation using the Benedict-Webb-Rubin EoS as modified by Starling
capitalized on the basic fracture concepts, and coupled that (BWRS) [47], which was packaged as software that has become
with gas dynamics, and thermodynamics to characterize known as GASDECOM. Because this formulation was
driving force and the inherent fracture resistance. The BTCM developed for typical natural gas producing fields, its scope
is referred to as such because it quantifies the driving force included binary interactions for a range of NGLs, as well
reflecting the gas’ decompression speed versus pressure- as for CO2. Use of this technology in applications to quite
induced wall stress response and the fracture speed versus rich gases decades later showed it to be quite robust, well
pressure (stress) response through use of two curves, whose beyond expectations [41]. The basic GASDECOM model
iterative solution for computational reasons in the 1970s was has seen only modest changes since, which focused on the
done by plotting their speed versus pressure trends. Figure 3 solver and related algorithms early in 2000.
While adaptation of the scheme developed for methane- in a pipeline that suddenly suffers a guillotine break. At
dominated mixes was found necessary in Battelle’s 1980s the time of the break, the flow is zero at the exit plane
work involving nearly pure CO2 [10, 48], others have used and a sonic wave propagates upstream at the local speed of
GASDECOM for gases dominated by CO2 at least in their sound. The moving expansion wave converts fluid at rest
initial work [16, 44]. As Battelle continued to experience into fluid in motion. As the exit plane velocity increases
issues in adapting the source-code for GASDECOM to the from zero, another expansion wave propagates into the fluid,
CO2-impurity mixes typical of CCS, we have since adopted which is now moving at a speed, denoted u, at a slightly
alternative EoSs, depending on the application, and returned reduced pressure and temperature. In reality, this process
to a first-principles approach. The Span and Wagner EoS is continuous with expansion waves propagating upstream,
[49], which was developed and calibrated specifically for pure each moving a little slower than the preceding one, with
CO2 is used for this benchmark scenario. Both the P-R EoS the exit plane velocity continuously increasing until the
[21] and the GERG EoS [50] were initially considered when exit plane velocity equals the expansion wave speed and
impurities effects were involved, however, an adaptation the flow is considered choked. The exit plane speed can
of the GERG EoS is now preferred, with care still taken to be described by the following [51]:
assess the practical viability of the outcomes. The speed of
decompression is then inferred from gas dynamics referenced (1)
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, where ρ is the density, a is the local speed of sound, and t is
with the required arrest toughness being that needed to the time after the rupture. The speed of sound is a property
slow the speed of propagation to equal or less than that of the fluid and can be found in the literature for a wide
for decompression. variety of fluids. For pure CO2, the speed of sound is available
from a number of studies, with the results based upon the
Key assumptions and their Span and Wagner [49] EoS, which is widely considered as
highly accurate.
implications
In single-phase fluids, the process is considered isentropic
Fundamental to predicting fracture arrest is the assumption and the density path can be found by tracing an isentropic
that arrest occurs when the wall stress falls below some critical path on a thermodynamic state diagram from the initial
level due to decompression, because the rate of propagation pressure to a final pressure. Under this assumption, and using
has slowed such that the pressure-induced wall stress the necessary data, Equn 1 can be readily integrated and
decreases. The historic view with the BTCM is that arrest the exit velocity obtained. The focus here is the expansion
ensues within a diameter or so once toughness suffices to wave, which propagates at the local speed of sound in a
cause tangency between the gas and fracture curves in Fig.3. frame moving with the exit velocity. In a frame at rest, the
A second key assumption is that the decompression response propagation velocity is then a + u (where the direction is
can be calculated independent of the fracture response, positive in the upstream direction, so u < 0, a > 0). With this,
with these being brought together through an empirically the propagation velocity can be tracked and compared to
determined function that couples their response, that was the crack propagation velocity in the manner of the BTCM.
termed the backfill coefficient. Yet another key assumption
is that the flow during decompression is a one-dimensional, For a mixture of CO2 and impurities as dictated by the
homogeneous, isentropic process that develops in response product stream, the situation is more complex. First, the
to expansion waves that propagate back into the pipeline, properties must be determined by an appropriate EoS. As
the implication of which is that it is an isentropic, reversible noted above, these range from relatively simple expressions,
process, with relative motion of liquid and vapour neglected. such as the P-R EoS, to more-complex and realistic expressions
Finally, it is assumed that the speed of the expansion waves that are grounded in thermodynamics and fit with extensive
can be determined relative to the instantaneous local density data sets. One of the most accurate sets is based upon the
of the fluid. These assumptions are considered next, more GERG database, as modified by Lemmon [52] of the US
or less in the reverse order they have been cited. National Institute of Standards and Technology. These
expressions are widely used for hydrocarbon mixtures, but
Speed of expansion waves and flow response do not have extensive data for mixtures with high fractions
of CO2 and impurities, indicating that there is a need for
Expansion waves of single-component, single-phase fluids additional data to support CO2 mixture studies.
have a long history of study, are generally considered well
understood and thus provide a solid basis for understanding As the pressure drops in the pipeline, the fluid may undergo a
the more complex case of expansion-wave propagation in change in phase. Figure 4a illustrates an isentropic expansion
CO2 mixtures with impurities and phase change. Such that starts at 2160psia (149bar) and 90°F (32°C) for a CO2
analysis is formulated in regard to a fluid at high pressure mixture specific to one client. Such plots develop for pure
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 the dome being a mix of these two subcritical phases. For
differences being the locations of the phase boundaries and a supercritical CO2 pipeline, there is a spectrum of initial
associated properties. The y-axis in this figure is pressure temperatures and pressures ranging from the inlet to the
on a logarithmic scale, while the x-axis is density. The exit of the pipeline, so there can be a spectrum of responses
light layered trends are isotherms, while the heavy curved depending on position along the pipeline considered in
trend that runs from the upper right down to the lower regard to the guillotine rupture, and its local pressure and
left across the layered isotherms is the isentrope from the temperature. For the particular initial conditions considered
initial condition. The region above the critical point at the in Fig.4a, the CO2 starts as a supercritical fluid, passes into
top of the two-phase dome comprises supercritical fluid, the liquid region, and then encounters the liquid-vapour
while the liquid phase is to the right side of the figure, and saturation boundary, and remains therein through the
the vapour phase is to the left side, with the region within remainder of the process.
Fig.5. Speed of sound of CO2 and CO2

mixtures referenced to three EoSs.
Fig.6. Pressure versus time for a

location 12ins (~30cm) upstream of
the exit plane.
In reference to Fig.4b, the x-axis is density as it is in Fig.4a, Equn 1, starting from rest (highest density) and increasing
whereas the y-axis is speed in regard to each of the three to choked flow. The middle curve is the wave speed in a
curves shown in the figure. Each of these curves represent stationary reference frame and is simply the vector addition
a process that begins at 70°F (21°C) and 2200psia (152bar), of the upper and lower curves. The wave speed starts at the
which in this figure originates from the upper right corner sonic speed (highest density) and then decreases as the fluid
of the figure, as the local fluid density corresponds to that speed increases, finally reaching zero, which is the choked
prior to rupture. The upper curve is the speed of sound, flow condition.
relative to the moving liquid. As time passes, the local density
decreases as the fluid increases in velocity. As evident, the Each of Figs 4a and 4b are unique to the initial conditions, to
sonic velocity is strongly nonlinear with density initially, the fluid mix in terms of the fraction of CO2, and to the mix
with this dependence diminishing with time. The sonic of impurities and their relative levels. As becomes apparent
velocity is a property of the liquid and vapour determined next, they are also dependent on the EoS embedded in the
from the equation of state, with multi-phase sonic speeds process description, so a compendium of such behaviour
determined from the method outlined in Wallis [53]. The as a function of circumstances along a pipeline, much less
lower curve represents the outflow fluid velocity from for differing product streams is implausible.
Fig.7. Effect of impurities on

decompression wave velocity.
Figure 5 further illustrates trends in the speed of sound for flow is a strong but common assumption; among others it
CO2 as a function of density and fluid composition, to better assumes equality of liquid and vapour velocities, which at best
quantify the range of responses that can occur depending is true only for low bubble volume fractions. For the results in
on pressure and composition, and also the manner these Fig.5, the speed of sound of the liquid and vapour phases has
are related via the EoS that underlies the analysis. While been estimated in the context of REFPROP [52], while the
speed was the dependent parameter in regard to density in speed of sound in the multi-phase region has been estimated
Fig.4b, in Fig.5 the sonic speed is a function of pressure using an expression from Wallis [53].
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 The expansion process for the mixtures in Fig.5 has been
transported fluid exists as a liquid, and then drops sharply assumed isentropic – which is reasonable for the single-
as the isentrope crosses the liquid-vapour phase boundary, phase part of the expansion but is not generally valid for
reaching a much lower value as a gas-liquid mixture, and the multi-phase part. Essentially then it is the transition
even lower as a gas. This response is central to controlling region between the liquid and vapour speed of sound that
running fracture, because this behaviour underlies changes is open to question. However, it should be noted that the
that lead to decompression local to the rupture plane, such phase transition itself may not necessarily occur at the phase
that arrest can occur, with sufficient toughness supplied boundary, due to nucleation delays.
such that the crack propagation speed is equal or less than
the decompression speed. Consequently, understanding the Realizing that the underlying formulation can be used
speed of sound as a function of conditions is important to to quantify pressure response within the pipeline, it is
the designer, and especially understanding the boundary of instructive to consider that response and the rate of that
the phase change regions for CO2 mixtures with impurities. process in contrast to crack speeds approaching arrest.
Consider Fig.6 in this regard, which shows the pressure
Figure 5 provides insight into the role of the EoS that underlies response 12ins (30cm) upstream of the exit (rupture) plane
the locations of the phase boundaries and how these move as a function of time. This result is generated for pure CO2,
as a function of impurities, which taken together dictate the for isentropic expansion beginning at 2200psia (152bar)
breakpoints in the trends shown in this figure. The speed of and 70°F (21°C). It is apparent from this figure that there
sound is shown for three fluids: pure CO2, a 95% CO2 and is a strong drop in pressure until the liquid-vapour phase
impurities mixture, and a 98% CO2 and impurities mixture. boundary is reached, and that the response time is the
The mixes in both cases represent client estimates of what might order of milliseconds. While this outcome reflects the same
be transported in their pipeline, which might be considered assumptions as noted above, it is apparent that when such
proprietary and so are not elaborated. Results are presented events occur they do so at very high speeds. Because the
for these mixtures characterized by the GERG EoS and the timeframe is short the occurrence of subtleties arising due
P-R EoS, both of which were introduced previously along with to three-dimensional versus one-dimensional response are
the Span and Wagner EoS for pure CO2. These results reflect indicated to have only modest influence on where along
the assumption of homogenous one-dimensional flow in the the pipeline – relative to its diameter – the decompression
two-phase region: that is, the properties are a weighted average velocity is found to match the propagation speed, which in
based upon the liquid and vapour fractions. Homogeneous the BTCM framework defines arrest. As such, the minimum
Fig.8. Isentropic processes plausible in

design vary significantly.
arrest toughness that results from this analysis is a reasonable practical applications involving rich (dense-phase) gases
indicator of the toughness needed for fracture control. [41]. Furthermore, the lack of equality between liquid and
vapour velocities generally means that these flows are not
Figure 7 develops in the context of analyses for the conditions isentropic, since the relative phase motion results in inter-
that underlie Fig.5, presenting a plot of pressure shown phase drag, heat transfer, and mixing. In addition, viscous
on the y-axis and the decompression wave speed on the and inertial effects will become important as the flow speed
x-axis in regard to the pure CO2 and the 95% CO2 plus increases, which are not included here.
impurities cases. Comparing these trends shows a practically
quite significant shift in response due to the presence of Clearly a compendium of pressure-volume outcomes would
impurities. As can be seen, this shift for the impurities mix be useful from a design perspective, but is virtually impossible
is associated with a higher pressure – which occurs due to to generate because of the range of product compositions and
the increased saturation pressure, and the shifting phase initial conditions. That view follows from our experience in
boundary in the pressure-density space. At pressures above working with the product streams anticipated to be carried
the phase boundary, the decompression speed is slower for under conditions akin to common-carrier / commodity-type
CO2 with impurities compared to pure CO2 at the same service, which indicates the range of possible mixes could
pressure. At pressures below the phase boundary, these trends be very large. It also follows from the observation that the
slow as expansion occurs within the two-phase boundary results of analyses as those in Figs 5 and 7 are highly sensitive
and the temperature drops. Because expansions waves form to both the impurities present and to subtle differences in
continuously and there is feedback in the pressure-volume- this mix or the relative percentage of a given constituent.
time space consistent with the EoS, the decompression As such, the focus here is on cause-effect aspects, and the
velocity gradually decreases, with that decrease generally assumptions involved, and their implications, in lieu of
stronger for the mixtures than for pure CO2. attempts to trend such outcomes.
Some results indicate that a plateau develops once the The last point to make in this context involves multi-phase
phase boundary is crossed regardless of the EoS used, effects, which are as important to hydraulics and other design
which is evident (for example) in References 45 and 46. aspects (which are not considered herein) as they are to running
Clearly this is in conflict with the view shown in Fig.7, but fractures. Suffice it to note that multi-phase wave propagation
as yet the reason for such has not been identified. In this has been extensively studied for supercritical water-blowdown
regard the need for data to support the reliable prediction events related to a loss of coolant accident scenario in the
of phase boundaries, equations of state, and the transport nuclear research community. While CO2 mixture flows are
properties for these CO2 mixtures, is emphasized, as such more complex, the existing related knowledge could provide
data are generally lacking. In addition, the multi-phase flow a strong base for understanding such flow response.
regime is uncertain. The homogenous assumption has not
been verified for these fluids; essentially it assumes that Discussion and design implications
the liquid and vapour phases move at the same velocity
and these velocities will depart significantly as the bubble Prior sections have identified the dependence of the
volume increases. However, we note that the homogenous outcomes on the initial conditions used in the analysis
flow assumption has been successfully applied in many (pressure and temperature), the scope of the impurities, and
the assumptions made. The need for data to assess/establish diagram it is readily apparent that substantially different
the viability of the predictive schemes also exists, as illustrated, behaviour can develop over the length of a given pipeline.
for example, by the apparent disparity in a plateau forming This is evident, for example, in Fig.8, which illustrates the
beyond the phase boundary in other work [45, 46] versus thermodynamic state of a 95% CO2 plus impurities mixture
the steady decay in the pressure-velocity response evident in for isentropic processes that start at 40°F up to 160°F (4
Fig.7. These are considered next in regard to decompression to 71°C), all at 2160psia (104bar). Depending upon the
and arrest-toughness prediction, and then illustrated in initial temperature of the fluid, the expansion process can
regard to historic CO2 pipeline designs in contrast to the descend on either the liquid or vapour side of the critical
spectrum of potential EOR-CCS applications. point, or possibly through the critical point. Reality for a
pipeline is that the pressure changes along its length, which
Viability of decompression and arrest toughness also plays into the need for the designer to consider the full
predictions range of conditions that might develop across the full range
of possible compositions. On this basis there is need to
As identified above, a valid concern exists in regard to the identify worst-case scenarios for purposes of decompression
EoS used, which in turn impacts the saturation pressure assessment and arrest-toughness prediction.
and temperature, and the phase boundaries. It also impacts
aspects such as compression / pumping decisions, and
hydraulics issues that control the recompression distance The design space for historic EOR versus plausible
for longer CO2 pipeline systems. Finally, such concerns CCS applications
impact predictions of the decompression behaviour of CO2
and CO2 mixtures, and carry through to the viability of the Perhaps the most instructive contrast between the outcome
concepts that underlie the BTCM and its reliance on the of the design approach that underlies the early CO2 pipelines
Charpy V-notch (CVN) energy as the toughness metric. and has continued since – and is still considered by some to
These are considered next in regard to decompression and be mature technology [1, 2] – is to compare the required arrest
arrest toughness prediction, and then illustrated in regard toughness for a range of designs for pipelines transporting
to historic CO2 pipeline designs in contrast to the spectrum CO2 across a range of pipeline capacities and CO2 quality
of potential EOR-CCS applications. specifications. To simplify this comparison, the outcomes
are presented in a normalized format – which avoids making
Developing a viable database to address these concerns with this comparison specific to details such as inlet pressure
scope adequate to address the range of potential EOR-CCS and temperature. Suffice it to say that parameters typical
applications is a bit like a Christmas list where there is no of supercritical CO2 transport have been considered in
Santa Claus. While work is planned or continuing to address regard to inlet pressure and temperature. To manage the
these gaps, the scope is limited, so the outcomes tend to be scope of this comparison and focus the outcomes, a simple
marginal in contrast to what might be needed given the scope binary CO2 product stream is considered, with the second
of concerns involved. More critical is the observation that constituent being one common to both EOR and CCS
some EoS and/or predictive schemes are comparable under product streams whose effect on arrest toughness is neither
certain circumstances but quite different for others, which worst-case nor trivial. The CO2 content ranges from pure,
complicates defining the empirical basis to discriminate which is well above what is cited as the usual Permian Basin
between them. This inconsistent disparity is perhaps the EOR minimum CO2 content of 95% [54], down to 90%,
biggest concern, as empirical understanding in regard to: which is equally below that minimum. Finally, to avoid the
nonlinear influence of diameter on the outcome of this
• the EoS (saturation pressure and temperature, and analysis, a constant diameter pipeline is considered.
the phase boundaries);
• Multi-phase flow (compression/pumping decisions, The reference pipeline used to normalize the outcomes
hydraulics issues, recompression distance, in this comparison is an EOR-service pipeline that for
decompression behaviour); and the sake of simplicity transports pure CO2. To keep the
• the viability of the concepts that underlie the BTCM normalization benchmark representative, this benchmark
for CO2 applications could require a significant is sized slightly larger than ‘average’ relative to the outcomes
testing matrix. in earlier discussion in the section titled Transport volume,
distances, and routeing. On that basis, the benchmark has
Design implications for CO2 pipelines been chosen as a 20-in (508-mm) diameter pipeline made of
Grade X65 (448MPa), whose wall was sized using a design
An important consideration to the pipeline designer is factor of 0.72 for operation at a given inlet pressure and
the temperature, pressure, and composition of the CO2 temperature8. To address the apparent shift from the earlier
mixture across the range of potential rupture locations. EOR designs through what could emerge if CCS service
Given the random nature of threats such as third-party
8. While this benchmark design has been chosen to represent typical onshore CO2
damage, fracture initiation leading to possible running
pipelines, its dimensions and pipe grade are comparable to one of the early-design
fracture could occur most anywhere along the length of some EOR pipelines that moved dome-sourced CO2, parts of which were retrofitted with
pipelines. By following the isentropic processes on the state fracture arrestors in the 1980s.
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 scenarios, and much less demanding than the constituents
capacity (transported volume) is inferred by considering a in typical CCS-related streams.
range of linepipe grades (diameter is fixed as noted above)
that permit relatively higher stresses. The resulting wall Because the benchmark pipeline has been used to normalize
thickness for a given grade is determined by design factor, the outcomes, that result is evident in the lower left corner
pressure, and diameter, which are constant for all scenarios. of this figure at (1,1), which is emphasized by the solid
Based on earlier discussion, the range of grades in addition circular symbol. Pipelines that require lower arrest toughness
to the X65 benchmark includes X70 (482MPa) and X80 and are relatively smaller than the benchmark exist, as do
(551MPa), which in current use in CCS service. others that are somewhat larger and require somewhat
higher toughness, the scope of which is typical of many
Figure 9 presents the normalized results of this comparison pipelines supporting EOR. This group of designs tends to
in terms of required arrest toughness shown on the y-axis fall within the dashed box that is roughly centred at the
as a function of pipeline capacity shown on the x-axis. The benchmark pipeline. As can be seen from the x-axis labels,
normalized outcomes are trended as a function the impurity only a 50% relative increase in capacity has been inferred
level, which as noted above involves the variation of one for this analysis, which is less than some might anticipate.
constituent in a binary fluid, which has been selected from The scale along the y-axis runs up to a value of 10, with the
among the candidate impurities because it is less demanding worst-case considered in these analyses regarding capacity
than most from a fracture-control perspective for usual EOR and impurities leading to a required arrest toughness that
approaches ten-times the toughness required for arrest in Summary and conclusions
the benchmark pipeline. Archival records at Battelle located
for one of the EOR pipelines that had arrestors installed This paper has examined aspects determining arrest-
back in the 1980s indicates toughness levels (hereafter CVN toughness requirements to control running fracture in CO2
full-size equivalent energy at service temperature) in excess pipelines, and addressed the validity of the view that the
of 60ft-lb (81J) in the sections where arrestors were installed. design of these systems deploys mature technology, as has
While this historic design scenario is comparable to the been suggested in some widely distributed reports on CCS
benchmark pipeline used to normalize the results shown as the means to manage GHG.[1, 2] While this might be
in Fig.9, this actual toughness level is slightly larger than inferred from decades of experience in transporting CO2 to
the required arrest toughness for the benchmark, which was support EOR, this view is open to question when moving
the order of 40ft-lb (54J). large volumes of anthropogenic CO2 from four perspectives:
It can be seen from Fig.9 that the presence of impurities • some of the early CO2 pipelines were retrofitted to
drives the required arrest toughness much more so than provide for fracture control;
pipeline capacity. This is evident in regard to the benchmark • trace impurities in CCS and EOR product streams
capacity (x-axis value equal to one), where a ten-fold increase can significantly increase the arrest toughness
in toughness is needed to deal with impurity levels the order required for fracture control;
of 10%, as compared to the modest increase in toughness • the volume transported for CCS and its routeing
due to capacity. Realizing that the minimum purity is often implies risks that can be much greater than for many
set at 95%, the decision to establish this level as compared EOR applications; and
to a lower level helps to offset concern for running fracture. • the technology used to quantify fracture control
While there is a desire to control the minimum CO2 content requirements is neither mature nor is the empirical
(apparently due to miscibility for the benefit of EOR, not database supporting it complete relative to current
safety), there is a clear need to monitor and/or control the expectations.
product stream in regard to this parameter given its strong
impact on required arrest toughness. While not evaluated directly in regard to theses perspectives,
some have considered running fracture a greater risk for
Equally, there is a need to better understand sources of CO2 pipelines than for hydrocarbon pipelines [55]. This
variability in toughness for linepipe, as such variability can paper has considered the technological background to
diminish the all-heat average (AHA) toughness and opens the ensure pipeline fracture control, considering important
door to longer propagation, which undermines specification aspects like the EoS and the assumptions embedded in the
of toughness as the control for running fracture. The data analyses of fracture-arrest requirements to offset fracture
in Fig.10 help to illustrate this point, where the results show concerns. That technology was used to illustrate the role
the distribution of toughnesses for a recent large order of of trace impurities after which the select design scenarios
transmission linepipe. The toughness coordinates in this have been considered as the basis for evaluating practical
figure run from 50 to 400ft-lb, equally 68 to 542J. Whereas implications in regard to both fracture control and related
the AHA for this large pipe order exceeds 200ft-lb (271J), aspects. This discussion has considered a range of fluid
the lowest toughness measured was 56ft-lb (76J), with a properties for naturally occurring and anthropogenic CO2
significant fraction of the heats with toughness less than one sources, and the service conditions anticipated for a range
standard deviation below the mean. A tighter population, of pipeline designs. Finally, the historic design space for
or a population skewed to above the mean, would greatly many CO2 pipelines supporting EOR has been contrasted
improve fracture control, and could reduce the AHA – that to that for pipelines in CCS service.
might simplify making specifications where high AHA
levels are needed. Important conclusions that can be drawn from this work
follow:
Central to fracture control is a fracture control plan,
which should be developed during the FEED (front-end • the minimum CO2 level can cause significant
engineering and design), and thereafter serves as the basis swings in the arrest toughness, such that toughness
to specify steel for the pipeline. Such planning offsets the requirements for pipelines in use for dome-sourced
eventual need to retrofit arrestors – which opens to costs CO2 underestimate that for CCS service by a factor of
and maintenance concerns that otherwise can be avoided. two – possibly more in contrast to relatively pure CO2;
Without such planning, where the inherent toughness is
found to be inadequate after the line is commissioned and • minor out-of-specification swings below the usual
the operational parameters are fixed, requires the use of 95% minimum can further increase the required
retrofit arrestors. Their use and placement is motivated by arrest toughness – with a decrease in the minimum
risk assessment and other considerations, just as was done below that level of 2.5% increasing the required
in the 1980s, which is about where this paper began. 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
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 A comprehensive exhibition will be part of the
subjects will run over the two and a half day event event, allowing companies from around the world
(and presented by industry leaders). to showcase their products and services. Contact
• Planning, design, construction and materials us today to book your space.
• Operations and maintenance
• Asset integrity management NETWORKING
• Inspection and cathodic protection Throughout the event there will be ample
• Repair and rehabilitation opportunities to network with participants to
• Automation and control further your business relationships. Meet with
• Leak detection industry leaders from around the world.
Paper abstracts are now being accepted.
vent.
this landmark e
s u re yo u attend
e
0 11 – mak
pen in early 2
e g istra ti ons will o
R
www.pipelineconf.com
• minor differences in the relative proportions of a by regulating saturation arrest pressures. Oil and Gas Journal,
trace impurity can have a significant influence on the pp44-46.
required arrest toughness, with subtle differences in 11. A.B.Rothwell, 1988. Fracture control in natural gas and CO2
the constituents present also a major driver – highly pipelines. In: Microalloyed HSLA Steels. ASM International,
volatile constituents can complicate the analysis and pp95-108.
significantly drive fracture arrest requirements; 12. M.Mohitpour et al., 2003. Pipeline design and construction:
• on-line monitoring of the transported stream and/ a practical approach. 2nd edition, ASME Press, (Note: the 3rd
or the injected streams on a trunkline should be edition is now available but is not used here, to emphasize the
considered to help manage risk; timeline of this 2nd edition).
13. anon., 2008. Interim guidance on conveying CO2 in pipelines
• the BTCM adapted to CO2 applications has been in connection with carbon capture, storage and sequestration
validated in regard to near-pure CO2 applications projects. UK-HSE.
and cases involving rich (dense-phase) natural gas: 14. B.Sass, B.Monzyk, S.Ricci, A.Gupta, B.Hindin, and N.Gupta,
however, like most other aspects involved in fracture 2005. Impact of SOx and NOx in flue gas on CO2 separation,
control, like the EoS, it remains unvalidated in compression, and pipeline transmission. Carbon Dioxide
applications to typical CCS product streams; and Capture for Storage in Deep Geologic Formations, Vol.2,
D.C.Thomas and S.M.Benson (Eds.), Elsevier.
• an expanded empirical database is essential to ensure 15. J.Barrie, K.Brown, P.R.Hatcher, and H.U.Schellhase, 2005.
viable fracture arrest predictions. Carbon dioxide pipelines: a preliminary review of design and risks.
Proc. 7th Int. Conf. on Greenhouse Gas Control Technologies.
Acknowledgements 16. A.Cosham and R.Eiber, 2007. Fracture control in carbon
dioxide pipelines. J. Pipeline Engineering, 6, 3, pp150–158.
Useful discussions with Dr Bruce Sass of Battelle’s Energy 17. C.B.Farris, 183. Unusual design factors for supercritical CO2
Systems and Carbon Management product line are gratefully pipelines. Energy Prog. 3, 3.
acknowledged, as is support from Battelle’s Science and 18. P.N.Seevam, J.M.Race, M.J.Downie, and P.Hopkins, 2008.
Technology fund in modelling the EoS and the expansion- Transporting the next generation of CO2 for carbon, capture and
wave response. storage: the impact of impurities on supercritical CO2 pipelines.
Proc.7th Int. Pipeline Conf., IPC2008-64063, Calgary, October.
References 19. A.Cosham and R.J.Eiber, 2008. Fracture control in carbon
dioxide pipelines: the effect of impurities. Idem, IPC2008-64346.
1. Carbon dioxide capture and storage, IPCC Special Report, 20. H.Li and J.Yan, 2009. Impacts of equations of state (EOS) and
2005. B.Metz, O.Davidson, H.de Coninck, M.Loos, and impurities on the volume calculations of CO2 mixtures in the
L.Meyer, (Editors). applications of CO2 capture and storage (CCS) processes. J.
2. P.W.Parformak and P.Folger, 2007. Carbon dioxide (CO2) Applied Energy, Elsevier.
pipelines for carbon sequestration: emerging policy issues. CRS 21. D.-Y.Peng and D.B.Robinson, 1976. A new two-constant
Report for Congress, 2007: Updated in 2008 as A.Vann and equation of state. Ind. Eng. Chem., Fundam., 15, 59–64.
P.W.Parfomak, ‘CRS report for Congress: regulation of carbon 22. B.N.Leis, J.H.Saunders, and E.B.Clark, 2010. Issues in
dioxide (CO2) sequestration pipelines: jurisdictional issues,’ quantifying the expansion wave response in CO2 pipelines.
which notes pipeline safety in a US jurisdictional perspective, 9th Annual Conf. on Carbon Capture and Sequestration,
but does not otherwise consider the topic. DOE/NETL, Pittsburgh.
3. D.I.Marsili and G.R.Stevick, 1990. Reducing the risk of ductile 23. L.S.Meltzer, 2007. CO2 transport – building on the current
fracture on the Canyon Reef Carriers CO2 pipeline. SPE 65th framework to meet demands of widely deployed commercial
Annual Technical Conference, New Orleans, LA, Proceedings scale CCS systems. 6th Annual Conf. on Carbon Capture
311-20, September. and Sequestration, DOE/NETL, Pittsburgh.
4. J.Watts, 1983. Sheep Mountain CO2 pipeline to boost West 24. www.statoil.com/en/ouroperations/explorationprod/ncs/
Texas production. Pipeline & Gas J. Vol/Issue 210: 6, May 1. snoehvit/pages/default.aspx.
5. W.R.Quarles, 1983. Willbros nears completion on Cortez 25. B.N.Leis and R.J.Eiber, 2010. Fracture control technology for
CO2 trunkline. Pipe Line Industry, Aug. transmission pipelines. PRCI Catalog L51846, 2010: updates
6. C.Horner, 1985. Choctaw carbon dioxide line laid in Mississippi. Eiber, R. J., Bubenik, T. A. and Maxey, W. A., Fracture control
Pipeline & Underground Utilities Construction, 40, 4, pp4-6, April. technology for natural gas pipelines, NG-18 Report No. 208,
7. anon. §195.111 Transportation of hazardous liquids by pipeline. Pipeline Research Council International, Project PR-3-9113,
US CFR Part 195. Battelle, 1993.
8. anon., 2010. Design and operation of CO2 pipelines. 26. W.A.Maxey, 1974. Fracture, initiation, propagation, and arrest. 5th
Recommended Practice, DNV-RP-J202, April. Symposium on Line Pipe Research, American Gas Association.
9. G.G.King, 1981. Design of carbon dioxide pipelines. Energy- 27. H.C.van Elst, 1974. Criteria for steady state crack extension in
Sources Technology Conference and Exhibition, Houston, gas pipelines. Int. Conf. on Prospects of Fracture Mechanics,
January: see also G.G.King, Design considerations for carbon Netherlands, pp299–318, June 24-28.
dioxide pipelines. Pipe Line Industry, 1981, pp125–132. 28. P.A.McGuire, S.G.Sampath, C.Popelar, and M.F.Kanninen,
10. W.A.Maxey, 1986. Long shear fractures in CO2 lines controlled 1978. A theoretical model for crack propagation and arrest in
pressurized pipelines. American Gas Association, Catalogue Eng. Frac. Mech., 71, pp1997-2013, 2004.
No. L00033, November. 40. B.N.Leis, R.J.Eiber, L.E.Carlson, and A.Gilroy-Scott, 1998.
29. W.A.Poynton, 1974. A theoretical analysis of shear fracture Relationship between apparent Charpy V-Notch toughness and
propagation in backfilled gas pipelines. Crack Propagation the corresponding dynamic crack-propagation resistance. Int.
in Pipelines, Paper No. 14, published by the Institute of Gas Pipeline Conf., ASME, Calgary, pp723-732: see also Leis, B.
Engineers, Newcastle upon Tyne, UK, March. N., Relationship between apparent Charpy V-Notch toughness
30. G.D.Fearnehough and D.G.Jones, 1980. Toughness and the corresponding dynamic crack-propagation resistance,
specification for shear fracture arrest in pipelines. Int. Conf. 1997, Battelle Report to R J. Eiber, Consultant, Inc. Exhibit
On Analytical and Experimental Fracture Mechanics, Rome, B-82, Proceeding GH 3-97, National Energy Board of Canada,
June 23-27. See also Bonomo, F. et al., Survey and tentative 1997-1998.
revision of ductile fracture arrest criteria in pipelines for gas 41. R.J.Eiber, B.N.Leis, L.E.Carlson, N.Horner, and A.Gilroy-
transmission, Ibid. See also Vogt, G. H., et al., Toughness for Scott, 1999. Full-scale tests confirm pipe toughness. Oil &
crack arrest in gas pipelines. EPRG Report, 3R International, Gas Journal, Nov.8, pp48-54.
22, 1983, pp 98-105, and others. 42. D.Rudland, D.-J.Shim, H.Xu, D.Rider, P.Mincer, D.Shoemaker,
31. F.Abbassian, 1985. Long-running ductile fracture of high and G.Wilkowski, 2007. First major improvements to the two
pressure gas pipelines. Dissertation for PhD, University of curve ductile fracture model. DOT-PHMSA No. DTRS56-
Cambridge, November. 03-T-0007, May.
32. L.B.Fruend and D.M.Parks, 1980. Analytical interpretation 43. B.N.Leis, and T.P.Forte, 2007. New approach to assess running
of running ductile fracture experiments in gas-pressurized fracture in transmission pipelines. DOT/PHMSA DTRS56-
linepipe. Crack Arrest Methodology and Applications, ASTM 05-T-0003, February.
STP 711, G. T. Hahn and M. F. Kanninen, Eds., pp359-378. 44. anon., 2010. Material requirements for CO2 line pipe. JFE
33. G.Buzzichelli, F.Nicolazzo, G.Demofonti, G.Re, S.Venzi, Steel Corporation / Marubeni-Itochu Steel Inc Presentation,
M.F.Kanninen, J.W.Cardinal, E.Z.Polch, T.B.Morrow, Battelle, February.
S.T.Green, and C.H.Popelar, 1987. Second annual report on 45. A.Cosham, 2009. CO2: it's a gas, Jim, but not as we know
the development of a ductile pipe fracture model. Southwest it. 5th Pipeline Technology Conference, Ostend, Belgium,
Research Institute report to PRCI, AGA/PRC Contract Nos. October.
PR 182-527 and PR 15-527, May 1987. See also Kanninen, M. 46. A.Cosham, R.J.Eiber, and E.B.Clark, 2010. GASDECOM:
F., O’Donoghue, P. E., Cardinal, J. W., Leung, C. P., Morrow, carbon dioxide and other components. 8th Int. Pipeline Conf.,
T. B., Green, S. T., Popelar, C. F., Buzzichelli, G., Demofonti, IPC2010-31572, October.
G., Rizzi, L., and Venzi, S., Dynamic fracture mechanics 47. K.E.Starling, 1973. Fluid thermodynamic properties for
analysis and experimentation for the arrest of ductile fracture light petroleum systems. Gulf Publishing Co., Houston: see
propagation in gas transmission pipelines, Pipeline Technology also Hopke, S. W. and Lin, C. J., Applications of the BWRS
Conference, Belgium, October. equation to natural gas systems., paper presented at the 75th
34. G.Demofonti and I.Hadley, 1992. Review of fracture parameters for National AIChE Meeting, Denver, March 1974.
laboratory measurement of resistance to ductile crack propagation 48. W.A.Maxey, 1983. Gas expansion studies. AGA NG-18 Report
in line pipe steels. CANMET Pipeline Conference, Calgary. 133.
35. S WRI, 1997. Written private communication to Von 49. R.Span, and W.Wagner, 1996. A new equation of state for
Rosenburg, E. L., March. carbon dioxide covering the fluid region from the triple point
36. S WRI, 1997. Written private communication to Von temperature to 1100 K at pressures up to 800 MPa. J. Phys.
Rosenburg, E. L., May. Chem. Ref. Data, 25, 6.
37. G.Demofonti, S.Venzi, and M.Kanninen, 1995. Step by step 50. O.Kunz, R.Klimeck, W.Wagner, and M.Jaeschke, 2007. The
procedure for the two specimen CTOA Test. 9th EPRG/PRCI GERG-2004 wide-range equation of state for natural gases and
Symposium, pp18-1 through 18-10. other mixtures. GERG Technical Monograph 15. Fortschr.-Ber.
38. G.Berardo, P.Salvini, G.Mannucci, and G.Demofonti, 2000. VDI, VDI-Verlag, Düsseldorf.
On longitudinal propagation of a ductile fracture in a buried 51. H.W.Liepmann and A.Roshko, 1957. Elements of gas dynamics.
gas pipeline: numerical and experimental analysis. Proc. 2000 John Wiley and Sons.
Int. Pipeline Conf., 1, New York, ASME, pp287 –294. See 52. E.W.Lemmon, M.L.Huber, and M.O.McLinden, 2010.
also Salvini P., Fonzo A., and Mannucci G., Identification of NIST Standard Reference Database 23: Reference Fluid
CTOA and fracture process parameters by drop weight test Thermodynamic and Transport Properties-REFPROP,”
and finite element simulation, Engineering Fracture Mechanics, Version 9.0, National Institute of Standards and Technology,
70, 3-4, 553-566, 2003. See also Mannucci, G., Buzzichelli, G., Standard Reference Data Program, Gaithersburg.
Salvini, P., Eiber, R. and Carlson, L., Ductile fracture arrest 53. G.B.Wallis, 1969. One-dimensional two-phase flow. McGraw-
assessment in a gas transmission pipeline using CTOA. 3rd Hill.
Int. Pipeline Conf., Calgary, Alberta, Canada, 1, pp315-320. 54. anon., 2009. Table 2 in Guidelines and regulations for oxy-fuel
39. S.H.Hashimi, et al., 2004. A specimen for studying the resistance carbon dioxide capture, transport and storage. IEA Oxy-Fuel
to ductile crack propagation in pipes. 5th Int. Pipeline Conf., Working Group.
IPC04-0610, Calgary, 2004: see also Shterenlikht, A., Hashemi, S. 55. J.Barrie, K.Brown, P.R.Hatcher, and H.U.Schellhase, 2004.
H., Howard, I. C., Yates, J. R. and Andrews, R. M., A specimen Carbon dioxide pipelines: a preliminary review of design and
for studying the resistance to ductile crack propagation in pipes, risks. 7th Int. Conf. in GHG Control Tech., Vancouver.
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.
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.
I 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
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
started in November, 2008, and has now been completed: wall thickness of the pipe.
Fig.1 shows the proposed route, which will form part of the
basic infrastructure that will significantly reduce greenhouse The second set of requirements is based on the need to
gas emissions in the UAE from 2020 onwards. prevent longitudinal ductile fractures and has been developed
by the gas pipeline industry using empirical data obtained
Captured CO2 will be transported in a new national CO2 from full-scale pipe-burst tests with natural gas. This set of
pipeline network to onshore oil reservoirs throughout Abu requirements is used to establish pipe toughness and, if the
Dhabi where it will be injected for reservoir management required toughness is unachievable, it dictates the minimum
and sequestration. Masdar is working closely with the Abu required wall thickness or the use of crack arrestors.
Dhabi National Oil Co (ADNOC) and the Abu Dhabi Co
for Onshore Oil Operations (ADCO). The final set of requirements is related to the optimization of
total project owning and operating costs. Cost optimization
The project will have a threefold benefit: it will reduce provides a rational methodology for choosing between
greenhouse gas emissions in the UAE, make CO2 available the apparently conflicting results of the first two sets of
for enhanced oil recovery (EOR), and free-up natural gas requirements, and leads to the selection of a pipeline that
that is currently being injected to maintain pressure in is both safe and economic.
some of the fields.
Nomenclature
Three different and sometimes conflicting sets of
requirements are needed to select wall thickness and A = area beneath Charpy notch, m2
toughness of CO2 pipelines. The first set of requirements CS = velocity of sound in CO2, m/s
limits stresses in the pipe wall under steady and transient CV = Charpy notch toughness, J
D = pipe outside diameter, m
E = modulus of elasticity, Pa
*Author’s contact details EN = normalized toughness parameter, (-)
tel: +1 403 398 3858
email: graeme.king@tensorengineering.com F = hoop stress design factor, (-)
Fig.1. Route map for Abu Dhabi CO2 pipeline network.
FJ = weld joint factor (ASME B31.4 Table 402.4.3) operating pressure is influenced by the composition of the
Pd = decompressed pressure at phase boundary, Pa CO2 mixture it carries and specific details of cost and economic
SMYS = specified minimum yield strength models used for the project.
T1 = temperature at time of installation, °C
T2 = maximum operating temperature, °C The dense phase was originally defined as a single phase
t = pipe wall thickness, m separating the gas and liquid phases immediately above the
α = linear coefficient of thermal expansion, 1/°C two-phase region, where fluid properties transition between
∆v = change in velocity of the fluid, m/s those of a gas and a liquid without any change of phase, and
∆P = surge pressure, Pa where fluid properties can be distinctly different from those
ν = Poisson’s ratio for steel, (-) of either a gas or a liquid [1]. Figure 2 shows the dense phase
ρ = density of the fluid, kg/m3 region on a pressure-enthalpy chart developed using the BWRS
σd = hoop stress due to internal pressure Pd, Pa equation of state for a design mixture of 95% pure CO2.
σf = pipe steel flow stress, Pa
σh = hoop stress, Pa Specific heat is one of the properties of dense-phase fluids
σL = longitudinal compressive stress, Pa that is different from either a gas or a liquid. Figure 2 shows
specific heat for gas, dense, and liquid phases (single-phase
Process design 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
CO2 can be transported most efficiently over long distances phase the specific heat is higher than 2.5kJ/kg-K and can
in the dense phase. The Cortez pipeline, for example, reach values as high as 10kJ/kg-K.
transmits CO2 from Wyoming to Texas and operates at
pressures up to 18MPa (2600psi), and the Weyburn-Souris Another unusual property of dense-phase fluids is their
pipeline in Montana and Saskatchewan has a maximum volumetric sensitivity to changes in temperature. For
operating pressure of almost 21MPa (3000psi). 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
These high operating pressures are due in part to the need for at 15°C and 10MPa is 800kg/m3 (see Point 2 on Fig.2). If the
a high minimum pressure to maintain single-phase operation, fluid followed the gas law, this 25°C change in temperature
and in part to the optimum frictional pressure loss along CO2 from 40°C (313K) to 15°C (288K) would cause the density
pipelines. As a result the optimum operating pressure can be to change only 8% from 500 to 540kg/m3. But dense-phase
higher than 21 MPa. For any specific project, the optimum CO2 is an order of magnitude more sensitive to changes
Location Dwellings per Mile General Description Design Factor

Sparsely populated wasteland, wilderness, grazing and
Class 1 <10 0.72
farmland
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
Cement factories, CCGT Power stations, fertilizer, petro-
Tier-2 0.5 million – 1 million
chemical complexes
Table 1. Basic design factor (F).
in temperature, and in this particular case a temperature parameters subject to all the relevant technical restraints
change of only 25°C causes the density to change more than imposed by the requirements of hydraulic performance,
100% from 400 to 800kg/m3. governing codes, and good engineering practice.
The UAE has long summers with air temperatures rising Allowable stress
to about 48°C between May and September, and short
moderate winters between December and March with air Requirements for CO2 pipelines are included in ASME B31.4
temperatures rarely dropping below 6°C. The maximum Pipeline transportation systems for liquid hydrocarbons and
design temperature of CO2 from aerial coolers after other liquids [2], in which paragraph 402.3. uses a design
compression has therefore been set at 55°C. The ground factor (F) of 0.72 to limit the allowable hoop stress (σh)
temperature at pipeline depth fluctuates between 13°C in under steady flow conditions:
late winter and 38°C in late summer.
σh ≤ 0.72 FJ SMYS (1)
The volumetric sensitivity of dense-phase CO2 to changes in
temperature means that during summer when the ground Metallurgical investigations conducted as part of the
temperature at pipeline depth is close to 40°C, the density of optimization work looked at a range of steel strengths and
carbon dioxide at the inlet to intermediate booster stations selected API 5L X65 (L450) linepipe as the most suitable
and down-hole injection facilities would be approximately for the project. Table 402.4.3 of ASME B31.4 sets the joint
400kg/m3, but at low flow during winter when the ground factor (FJ) in Equn 1 for pipe made from this steel equal
temperature at pipeline depth is around 15°C, the density to unity so that the allowable hoop stress for the pipeline
of CO2 at pump suction would be approximately 800kg/ network under steady flow conditions is 72% SMYS. In
m3. Wide swings in density between summer and winter, addition, paragraphs 402.3.2(c) and 419.6.4(b) of ASME
and between high and low flows, can create challenges to B31.4 limit the combined stress of buried pipelines:
the smooth operation of the pipeline unless the magnitude
of the swings is identified early in the design and solutions σh + σL ≤ 0.90 SMYS (2)
are found and implemented.
In Equn 2, σL is the net longitudinal compressive stress
If centrifugal equipment is used to pump the dense-phase due to the combined effects of temperature rise and fluid
fluid, wide swings in fluid density cause proportionally wide pressure computed from:
swings in differential pressure between suction and discharge.
Speed control can be used to manage the swings by allowing σL = Eα(T2 – T1) – νσh (3)
the equipment to be run faster during summer when density
is low and slower in winter when density is high. Variable- Paragraph 419.6.4(b) of ASME B31.4 makes it clear that
frequency and variable-speed drives (VFDs and VSDs) are bending stresses only need to be included in combined
now widely used throughout the pipeline industry, so that stress calculations when designing above-grade portions of
once variations in density have been properly evaluated it is restrained lines and do not need to be considered when
not difficult to size and configure variable-speed centrifugal designing buried portions providing, of course, that the
equipment to handle the full range of operating flows, pipeline is built using proven pipeline construction practices,
pressures, and temperatures throughout the year. good engineering, and other more-specific rules included
in Chapter V of ASME B31.4.
The overall goal of pipeline design work therefore is to
configure the proposed system to handle a nominal flow of If a design factor of 0.72 is used, Equns 2 and 3 allow
20 × 106t/a of CO2 with a maximum design temperature of a maximum temperature differential of 160°C between
55°C, ground temperatures between of 17°C and 38°C, and a construction and operating conditions for buried pipelines
maximum design pressure of around 21MPa, and to optimize before extra wall thickness is required to keep the combined
Fig.2. Pressure-enthalpy diagram for

95% pure CO2 showing specific heat in
liquid, dense, and gas phases.
stress below 0.90 SMYS. The maximum temperature density over the normal range of operating conditions is
differential between construction and operation for the approximately 800kg/m3.
proposed pipeline is less than 50°C, and the combined
stress limitations of ASME B31.4 therefore do not require Figure 4 shows the velocity of sound in CO2, the typical value
additional wall thickness. of which over the normal range of operating conditions is
approximately 500m/s. Both density and velocity of sound
In order to enhance safety, the design factor of 0.72 has been in CO2 are slightly lower than for oil, so that the magnitude
modified depending on its location by additional restrictions of surge pressure waves in CO2 pipelines can be expected
of ASME B31.8 Gas transmission and distribution piping to be less than in oil pipelines.
systems [3]. Table 1 summarizes the design factors for the
four location classes specified in Table 841.114A of ASME Figure 5 shows the unit surge pressure defined as the change
B31.8. These additional requirements over and above the in pressure for a change in flow velocity of 1m/s. The
requirements of ASME B31.4 increase wall thickness of the typical value for unit surge pressure over the normal range
CO2 pipeline network in populated areas. of operating conditions is approximately 0.4MPa/(m/s).
Surge pressures Surge pressure is proportional to change in flow velocity,

and maximum surge pressure therefore depends on the
Surge pressures are produced by changes in velocity of the maximum flow velocity. The optimum flow for CO2 systems
moving fluid that result from shutting-down pumps or is less than 4m/s and therefore, for cost efficiency, it is best
pump stations, closing valves, or otherwise blocking the to design and build CO2 systems so that they operate at
flow. Internal viscous effects and inelastic properties of the velocities less than 4m/s. The maximum surge pressure in
backfill and pipe wall attenuate surge pressure waves as they a properly designed CO2 pipeline, if a valve suddenly closes
move away from the point of origin. or a pump station suddenly stops working, is therefore less
than 1.6MPa (that is, 4m/s x 0.4MPa/(m/s)).
Paragraph 402.2.4 of ASME B31.4 requires pipeline designers
to make surge calculations and to provide adequate controls Since the design pressure of the proposed pipeline network is
and protective equipment to prevent pressure rise due to 24.5MPa, the maximum increase in pressure due to sudden
surges and other variations from normal operations from valve closure or station outage is no more than 6.5% above
exceeding the internal steady state design pressure anywhere the design pressure. Because the maximum surge pressure
in the pipeline system and equipment by more than 10%. is less than allowed by the code (that is, less than 10%
above the design pressure), no increase in wall thickness is
The maximum surge pressure (∆P) can be precisely calculated required. The next section will comment on the effect of
from the fundamental equation: surge pressure on ductile fracture arrest in CO2 pipelines.
∆P = ρ CS ∆v (4) Ductile fractures

Figure 3 shows the density of dense-phase CO2 over a broad Paragraph 402.5.1 of ASME B31.4 requires the designer of
range of operating pressures and temperatures. Typical fluid CO2 pipeline systems to consider the possibility of ductile
Fig.3. Density of CO2 with highlighting

showing the normal range of operating
conditions.
fractures and provide reasonable protection to limit their Decompression and crack velocity
occurrence and length throughout the pipeline, with special
consideration at crossings and other appropriate locations. In the unlikely event that a dense-phase CO2 pipeline
More specifically, Paragraph 402.5.3 requires the designer bursts and a ductile fracture starts to run, the sudden loss
to minimize ductile fracture propagation by the selection of gas causes the CO2 to decompress isentropically into the
of pipe steel with appropriate fracture toughness and/or two-phase region. The decompression is rapid and highly
by the installation of suitable fracture arrestors. Design turbulent, and the gas and liquid components do not have
considerations include pipe diameter, wall thickness, fracture time to separate. The two phases behave like an homogenous
toughness, yield strength, operating pressure, operating single phase with a very low sonic velocity (less than 100m/s).
temperature, and the decompression characteristics of CO2
with its associated impurities. The low sonic velocity causes a sustained pressure at the
moving tip of the fracture equal to the pressure at which
The possibility that very long ductile fractures can occur the decompression crosses the phase boundary into the
in CO2 pipelines was first identified in the late 1970s [4] two-phase region. This high sustained pressure acts on the
almost 10 years after the first CO2 transmission pipelines unrestrained flaps of the fractured pipe just behind the
had been built in North America. Crack arrestors were moving fracture tip, and concentrates stresses at the tip of
subsequently retrofitted to existing CO2 pipelines to help the fracture large enough to tear the steel wall and drive the
reduce risk. Since that time new CO2 pipelines have been fracture along the length of the pipe. Pipe steels need to be
built with crack arrestors. both strong and tough to resist the forces and prevent the
crack from propagating.
Thicker walls and tougher steels can be used instead of crack
arrestors effectively to control long ductile fractures. If a In order to calculate the pressure at the tip of the moving
longitudinal subcritical crack in the pipe wall grows during fracture it is necessary to understand how CO2 decompresses
operation, it can initiate a longitudinal tear in the pipe wall. and the effect of phase behaviour. The actual composition
If the pipe wall is not strong enough or tough enough to resist of CO2 in the proposed pipeline is important in this regard.
the force of the decompressed pressure pushing against the CO2 from identified sources is relatively pure but during
unrestrained walls of the pipe on either side of the crack, a design it is not possible to identify all the future sources of
ductile fracture will form and run along the pipe. CO2. Therefore a worst-case estimate of composition from
present and future sources needs to be used during design
The solution is to increase either wall thickness or fracture to protect the pipeline system against long ductile fractures
toughness of the pipe steel, or both. Increasing wall thickness over its full operating life.
helps by reducing the stress at the tip of the fracture, and
increasing toughness helps by enabling the steel to absorb Of the possible impurities, hydrogen has the largest effect
more energy when it tears. Alternatively, crack arrestors can on the phase boundary and decompression behaviour
be installed to limit the length of the fracture. of CO2. Nitrogen has a lesser but still large effect while
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. two-curve method [9]. Each figure shows two sets of curves:
The effect of hydrogen on the properties of CO2 has been the first is a set of decompression pressure-wave velocity curves
determined accurately by experimental and theoretical work for the design-case mixture for a range of temperatures from
of Prausnitz and Gunn [5] and others [6, 7, 8]. Comparison 15 to 55°C (59 to 131°F) when the pipeline decompresses
of experimental results with predictions of equations of following a rupture. The decompression pressure wave
state using parameters developed by Prausnitz and Gunn curves define the speed that each pressure level propagates
[5] have shown that the PR equation of state is sufficiently back along the pipe from the initial fracture site as the
accurate for design work over a wide range of pressures and pipeline decompresses. Figures 6, 7, and 8 are for sudden
temperatures. and that the BWRS equation of state was decompression from operating pressures of 17.5, 21, and
sufficiently accurate for temperatures greater than -20°C. 24.5MPa (2538, 3046, and 3553 psi) respectively, and were
calculated using the BWRS equation of state.
After studying a range of possible compositions from a
variety of sources that might need to be transported in the The second set of curves in Figs 6, 7, and 8 is a set of Battelle
proposed network, a design-case mixture of 95% CO2 with fracture-velocity J-curves for a range of toughness values. For
1% hydrogen and 4% nitrogen by volume was chosen as illustrative purposes the J-curves were calculated for NPS 18
the basis for design work to prevent long ductile fractures (18-in diameter) Grade L450 (X65) pipe with wall thickness
over the operating life of the project. for the three different operating pressures calculated using
a design factor of 72%. The fracture-velocity J-curves define
Figures 6, 7, and 8 illustrate how the pipe toughness that is the relationship between the pressure driving the fracture
needed to arrest ductile fractures is found using the Battelle along the pipe and the velocity of the fracture. The toughness
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 the dense phase intersect the bubble-point line, whereas
sets of curves are tangential (just touching). decompression paths from Points 3 and 4 in the gas phase
intersect the dew-point line. Because bubble-point and dew-
Figures 6, 7 and 8 show that higher operating pressures require point lines have different slopes, dense-phase decompression
lower-toughness pipe steel. Figure 6 for decompression from from high-pressure Point 1 intersects the phase boundary
17.5MPa shows a required toughness of 120J; Fig.7 for at a lower pressure than decompression from low-pressure
decompression from 21MPa shows a toughness of 60J; and Point 2. On the other hand, gas-phase decompression from
Fig.8, for decompression from 24.5MPa, shows a toughness high-pressure Point 3 intersects the phase boundary at a
requirement of 40J. higher pressure than decompression from low-pressure Point
4. This difference explains why it is easy to control ductile
An interesting observation from these figures is that, fractures in dense-phase pipelines by increasing the design
when the initial operating temperature is kept constant, pressure, but not so easy in gas-phase pipelines.
the sustained pressure plateau is slightly lower for higher
initial operating pressures. This means that to prevent long Toughness calculations
ductile fractures in CO2 pipelines, unlike gas pipelines,
they need to be designed to handle low operating pressures The decompression path from 24.5MPa and 40°C for
rather than the maximum allowable operating pressure. If design-case CO2 is shown in Fig.8. It shows a sustained
CO2 pipelines are designed to prevent ductile fractures at pressure plateau extending between 60m/s and 330m/s at
low operating pressures they are generally safer from the a pressure of 8.4MPa. The toughness requirement where
point of view of controlling ductile fractures when they the decompression path just touches the fracture velocity
are operated at higher pressures, up to the maximum hoop curve is determined from the low-velocity end of the pressure
stress allowed by the code. Another conclusion that can be plateau. at a velocity of 60m/s.
drawn from this counterintuitive result is that increases
in operating pressure due to surge waves do not need to At velocities as low as 60m/s, the fracture-velocity J-curve is
be considered when selecting pipe toughness to arrest horizontal; it can be seen from Fig.8 that if the decompressed
ductile fractures. pressure is kept constant at 8.4MPa there is no practical
difference between the toughness required for a fracture
This counterintuitive behaviour of dense-phase pipelines is speed of 60m/s and the toughness required for a fracture
distinctly different from gas pipelines where the sustained speed of 0m/s.
pressure plateau increases markedly as operating pressure
increases. Figure 9 illustrates the reason for the difference, For CO2 pipelines, this observation allows a simplification,
and shows decompression paths from four different with no loss of accuracy for steel toughness up to
operating points on a pressure-enthalpy diagram developed approximately 200J, by setting fracture velocity equal to zero
using the BWRS equation of state for a mixture of 95% and fracture arrest pressure equal to the highest pressure at
pure CO2. Decompression paths from Points 1 and 2 in which the decompression enters the two-phase region [4].
Steel Wall @ 175 bar & 72% Wall @ 210 bar & 72% Wall @ 245 bar & 72%
Diameter
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
For many CO2 pipelines the fracture arrest pressure will There are approximations in the development of the
be equal to the cricondenbar of the mixture selected for Battelle fracture-arrest equations and uncertainties in their
the design work. If the fracture velocity is set equal to zero application to CO2 pipelines that need to be considered.
and the arrest pressure is set equal to the highest pressure The underlying theory was originally calibrated using
at which the decompression path intersects the phase results from full-scale pipe-burst tests to advance the design
boundary (Pd), the Battelle fracture-arrest equations [9], of pipelines carrying natural gas. The theory has been
after rearranging terms, become: extended without additional experimental validation to
CO2 pipelines operating at much higher pressures than
natural gas pipelines. Full-scale burst tests with proposed
(5) pipe, proposed operating pressure, and proposed pipeline-
quality CO2, are recommended to validate the applicability
of the theory to new CO2 pipelines.
where
Future variations in product quality have been accounted
(6) 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
(7) order to account for uncertainties in the theory itself and
its extension to dense-phase CO2. This has been done by
As the ratio of hoop stress at the fracture tip to flow stress adding 0.4MPa to the arrest pressure (cricondenbar). The
(σd/σf) approaches 0.30, the normalized toughness (EN) in cricondenbar of the design-case mixture is 8.4MPa, so that
Equn 6 approaches infinity and becomes highly sensitive the arrest pressure selected for this project for use in Equn
to small errors in the estimation of decompressed pressure 7 was therefore 8.8MPa.
and flow stress. When the ratio exceeds 0.28, wall thickness
rather than toughness should be increased to control Table 2 shows the minimum wall thickness required to
ductile fractures, or crack arrestors should be used. As a satisfy the hoop-stress requirements of ASME B31.4 with a
result, the practical limit of applicability of Equns 5, 6, design factor of 0.72 as well as the Charpy V-notch toughness
and 7 is given by: required to prevent long ductile fractures using fracture
arrest Equns 5, 6, and 7 when the pipeline is transporting
(8) the design-case mixture. Crack arrestors are indicated in
Table 2 for cases where the decompressed stress ratio (σd/
σf) exceeds 0.28.
Equations 5 to 8 can be used to calculate the toughness
(CV) required for the arrest of ductile fractures using Optimization
wall thickness (t), pipe diameter (D), decompressed
plateau pressure (Pd), and flow stress (σf). Alternatively, Table 2 defines three long CO2 transmission pipelines with
the equations can be solved iteratively to find the wall design pressures of 17.5, 21.0, and 24.5MPa using pipe sizes
thickness (t) needed to arrest ductile fractures for any from NPS 8 (8.625ih) to NPS 24 (24in). The three different
given toughness (CV). systems can be characterized as follows:
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 different maximum design pressures, compressor facilities
thinnest wall of the three alternatives using crack for all the systems were designed to take CO2 from source at
arrestors instead of toughness to arrest ductile the minimum pipeline operating pressure and deliver it to
fractures; injection facilities at the same minimum operating pressure.
• pipelines with design pressure of 21.0MPa using Initial and intermediate compressor stations were located
relatively high toughness steel to arrest ductile as required to boost pressure from the minimum operating
fractures; pressure to the system-design pressure.
• pipelines with design pressure of 24.5MPa and the
thickest wall of the three alternatives using relatively The pipe costs were based on current prices for API 5L
low-toughness steel to arrest ductile fractures. L450 PSL2 pipe and do not include a premium for higher-
toughness steels. It was assumed that crack arrestors could
Optimization on a cost basis can be used to choose between be installed at no additional cost so that the lower-pressure
these alternatives. To make a fair comparison of systems with systems would not be unfairly penalized. Figure 10 shows
Fig.9. Decompression paths for 95%

pure CO2 showing why burst pressure
inside gas pipelines is higher when
initial pressure is higher, but in dense-
phase 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 Conclusions

lifetime owning and operating costs for the three systems: it
can be seen that the lowest unit transportation cost (lowest ASME B31.4 sets out the equations required for calculating
lifetime owning and operating cost) is obtained with a minimum wall thickness of CO2 pipelines for any given
design pressure of 24.5MPa for all pipe sizes. This means it operating pressure and the Battelle fracture arrest Equns 5
is more cost-effective for long CO2 transmission pipelines to 7 in combination with Equn 8 can be used to calculate
to use pipe with thick walls, high operating pressures, and toughness to arrest ductile fractures, as well as the need
low toughness steels, rather than thin walls, low operating for extra wall thickness or crack arrestors. Finally, cost
pressures, and high-toughness steels or crack arrestors. optimization provides a rational methodology for choosing
between the different alternatives produced by the first two
A further increase in operating pressure is not expected sets of requirements, leading to the selection of a pipeline
to further reduce cost of transportation because 24.5MPa system that is both safe and economic. For this project a
is close to the pressure rating of 1500 Class flanges at the design pressure of 24.5MPa was selected as the most cost-
maximum pipeline operating temperature, and the high effective alternative resulting in the adoption of thick walls
cost of heavier flanges and fittings (2500 Class) would put and low-toughness steels for the arrest of ductile fractures.
a step in the optimization curves. Although it is more cost effective to use the pipe wall rather
than crack arrestors to arrest ductile fractures, the main 4. J.M.Prausnitz and R.D.Gunn, 1958. Volumetric properties of
benefit is the increase in safety that comes from limiting the nonpolar gaseous mixtures. A.I.Ch.E Journal, 4, 4, December,
length of ductile fractures to much shorter lengths than is pp430-435.
practical with crack arrestors. A secondary benefit of using 5. C.Yokoyama, K.Arai, S.Saito, and H.Mori, 1988. Bubble-point
the pipe wall rather than crack arrestors to control ductile pressures of the H2-CO-CO2 system. Fluid Phase Equilibria, 39,
fractures is that the pipeline is built with a thicker wall 101-110.
making it more resistant to third-party damage. 6. J.O.Spano, C.K.Heck, and P.L.Brick, 1968. Liquid-vapor
equilibria of the hydrogen-CO2 system. J. Chem. Eng. Data,
References 13, 168-171.
7. C.Y.Tsang and W.B.Streett, 1981. Phase equilibria in the H2/
1. D.L.Katz and G.G.King, 1973. Dense phase transmission of CO2 system at temperatures from 220 to 290 K and pressures
natural gas. Energy Processing Canada, November and December. to 172 MPa. Chem. Eng. Sci., 36, 993-1000.
2. American Society of Mechanical Engineers, 2006. ASME B31.4- 8. R.J.Eiber, T.A.Bubenik, and W.A.Maxey, 1993. Final report
2006 Pipeline transportation systems for liquid hydrocarbons on fracture control technology for natural gas pipelines.
and other liquids. Project PR-3-9113, NG-18 Report No. 208, Pipeline Research
3. Idem, 2007. ASME B31.8-2007 Gas transmission and Committee, American Gas Association, December.
distribution piping systems.
G.G.King, 1981. Design considerations for CO2 pipe line.
Pipe Line Industry, November, pp125-132.
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.
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, John Tiratsoo,

Clarion Technical Conferences Tiratsoo Technical
Call +1 713 521 5929 or visit www.clarion.org

A dynamic boundary ductile-

fracture-propagation model for
CO2 pipelines
by Prof. Haroun Mahgrefteh*, Solomon Brown, and Peng Zhang
Department of Chemical Engineering, University College London, UK
T 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 crack-
propagation 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.
T HOUSANDS OF KILOMETRES of pressurized pipelines

are used to transport large amounts of hydrocarbons
across the world. Although this method of transportation
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.
is generally considered to be safe, pipeline failures do occur
with some leading to catastrophic consequences (see for Fractures can initiate from defects introduced into the
example Refs 1, 2). In the US alone, despite having one of pipe by outside forces such as mechanical damage, soil
the most stringent safety requirements across the globe, over movement, corrosion, material defects, or adverse operating
202 pipeline incidents were reported during 2005 – 2009 [3]. conditions. Fractures propagate when the stresses acting
These resulted in an estimated $2 billion of damage leading on the defect overcome the fracture initiation tolerance
to 69 deaths and 254 serious injuries. of the pipe, reaching a critical size based on the pipeline
material properties and operating condition. As such it is
Propagating factures are considered as by far the most highly desirable to design pipelines such that when a defect
catastrophic type of pipeline failure. Such failures reaches a critical size and fails, the result is a leak rather than
involve the rapid axial splitting or tearing of the pipeline, a long running facture.
sometimes running over distances of several hundred
meters resulting in massive loss of inventory in a very The above requires a two-tiered design approach involving:
short time. Deservedly, understanding and modelling of
the mechanisms responsible for such type of failure has • providing sufficient fracture initiation resistance,
led to a large number of studies (see for example Refs 4, mainly via specifying the required pipe toughness,
5). Such interest has intensified recently [6-8] given the wall thickness and operating conditions
prospect of using pressurised pipelines for transporting
captured CO2 from fossil plants for subsequent storage. • ensuring sufficient fracture propagation resistance
such that if a running fracture occurs its length is
This paper is based on one presented at the First International Forum on Transportation limited to a short distance
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.
Notwithstanding cost implications, fracture initiation can
be largely controlled a priori by specifying the required
*Author’s contact details fracture initiation toughness, minimum wall thickness and
tel: +44 (0)20 7679 3835
email: h.mahgerefteh@ucl.ac.uk the maximum stresses acting upon the defect.
HLP
Parameter ECSC Alliance
A1 B1 C2
Rich Gas (see Rich Gas (see
Inventory Air Air Air
table 2) table 2)
Internal diameter (m) 1.182 1.182 1.182 1.4223 0.8856
Pipe thickness (m) 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
Table 1. Pipeline characteristics and prevailing conditions used for the full-scale burst tests.
However, controlling fracture propagation once a leak has Additionally, given the significant drop in the speed of
formed is more complex, presenting a unique set of challenges. sound and hence the depressurization rate during the
As well as the fracture toughness of the steel and the backfill transition from the gaseous to the two-phase region [12],
conditions, the fracture-propagation velocity and arrest length such analysis must also account for real fluid behaviour
depend on the depressurization rate, the thermal stresses, and through the use of an appropriate equation of state. Also,
the minimum pipe wall temperature relative to its ductile- non-isentropic effects such as the fluid/pipe wall friction
to-brittle transition temperature. To model the above and and heat transfer must be accounted for as these also directly
hence develop methodologies for overcoming such a type of affect the depressurization rate. Finally the temperature
failure, we need to understand the nature of the processes drop as a result of the Joule-Thomson expansion cooling
taking place once a fracture has been initiated. [13] of the fluid within the pipeline during discharge can be
significant. In the case of CO2, depending on the starting
The onset of a leak in the pressurized pipeline results in a conditions, such temperatures can reach as low as -70°C
series of expansion waves that propagate from the rupture resulting in very significant localised cooling of the pipe
plane towards the intact end of the pipeline at the speed of wall in contact with the escaping fluid.
sound [10]. As the main driving force for crack propagation
is the crack tip pressure [11], the precise tracking of the The minimum pipe wall temperature reached relative to
expansion waves, and their effect on the pressure profile its ductile to brittle transition temperature will dictate
along the pipeline, is essential for the proper modelling of whether the pipeline will fail in the ductile or brittle
fracture propagation. fracture manner. The modelling of brittle fractures in
Fig.1. Schematic representation of the

experimental setup used in the HLP
full-scale pipe burst tests [20].
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’ decompression model must also incorporate a suitable
previous publication [13]. Ductile fractures are the focus equation of state. This is especially important in the
of our attention in this work. case of CO2 pipelines given the unique depressurization
thermodynamic trajectory of CO2 [8].
The so called Battelle Two-Curve (BTC) approach by
Maxey [5] was the first used to express the criterion for the Crucially none of the studies reviewed simulate the dynamic
propagation of a ductile fracture in terms of the relation interaction between the rapidly changing crack tip opening
between the fluid decompression-wave velocity and the area and the pressure loading as the crack propagates.
crack-propagation velocity. If the fluid decompression-
wave velocity is larger than the crack velocity, the crack tip In this paper, we report the development and validation of
stress will decrease, eventually dropping below the arrest a rigorous dynamic boundary ductile-fracture-propagation
stress and causing the crack to arrest. Conversely, if the model which takes into account all of the important transient
decompression-wave velocity remains smaller than the crack fluid/structure interactions governing the fracture process.
velocity, the crack tip pressure will remain constant resulting The performance of the model in terms of predicting the
in indefinite propagation. crack-propagation velocity and arrest length is tested by
comparison against real data. These full-scale burst tests
Several studies have since been conducted for modelling conducted by the High-Strength Line Pipe Committee[20],
ductile fractures based on the BTC approach (see for ECSC X100 [21] and Alliance[22] for pipes containing either
example Refs14, 15). Some employ sophisticated finite- air or rich gas mixtures.
element methods for simulating material deformation
but use over-simplistic transient fluid flow models for The validated model is used to test the propensity of a
predicting the rupture plane pressure and hence the crack hypothetical but realistic pressurized CO2 pipeline to ductile
driving force (see for example Refs 16, 17). Others, on fracture propagation failure, paying particulate attention
the other hand, although accounting for the transient to the impact of the starting line temperature. The latter
depressurization profile within the pipeline, do not investigation was prompted by the findings of Cosham
deal with the impact of pipe wall heat transfer and and Eiber [7] indicating the significance of the starting
friction effects on the fluid decompression behaviour temperature on the CO2 depressurization trajectory relative
(see for example Refs 18, 19). Additionally a reliable to its phase transition boundary.

crack length for test B1 south-running
crack.
Inventory:
temperature = 6°C.
Curve B: DBFM prediction.
Theory
where, fw is the Fanning friction factor and D the pipeline
The full background theory of the fluid flow model employed diameter.
in this study to predict the fluid decompression velocity
and the crack tip pressure for a given opening area is given Also,
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 where θ is the angle of inclination of the pipeline to the
are respectively are given by: 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
where ρ, u, P and h are the density, velocity, pressure and diffusion associated with a finite-difference approximation
specific enthalpy of the homogeneous fluid as function of of partial derivatives is reduced.
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: 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)

crack length for test C2 south-running
crack.
Inventory:
rich gas (Table 2), initial pressure =
104bara, initial temperature = -5°C.
energy [5]. However, in the full-scale pipe bust tests conducted
by the High-Strength Line Pipe Committee (HLP) [33], the Component HLP C2 Alliance Test 1
BTC theory is used in conjunction with the drop-weight CH4 89.57 80.665
tear test (DWTT) energy, as this is shown to provide a more
accurate indication of the pipeline resistance to fracture. C2H6 4.7 15.409
Consequently this is the model applied in this work. C3H8 3.47 3.090
iC4H10 0.24 0.232
The two-curve model for the crack propagation velocity, vc,
and crack arrest pressure, Pa, are respectively given by [33]: 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 tip pressure Pt is taken to be the choked pressure at the
mean value of the tensile and yield stresses), pre-cracked pipeline release plane.
DWTT energy and ligament area of a pre-cracked DWTT
specimen. On the other hand Pt and tw are the crack tip For brevity, full details of the coupling of the fracture and the
pressure and pipe wall thickness respectively. The crack fluid decompression models will be reported in a separate

crack length for test ECSC X100
south-running crack.
Inventory:
temperature = 20°C.
study. The calculation algorithm automatically corrects Table 1 shows the pertinent conditions relating to each
for the effective pipeline length and hence the crack tip test. Table 2 on the other hand shows the rich gas feed
pressure as crack opens. The required fluid decompression compositions for HLP C2 and Alliance tests.
velocity and the crack tip pressure are determined from
the numerical solution of the conservation equations (1-3) The full-burst-test pipelines used comprised several sections
using our CFD computational package, PipeTech [34]. The of differing toughness for which the corresponding DWTT
Peng-Robinson [35] equation of state (PR EoS) is used for energy may be calculated. In all simulations, the pipe wall
the prediction of the pertinent fluid phase equilibrium data roughness and heat-transfer coefficient are taken as 0.05 mm
for both air and rich gas mixtures. In the case of CO2, the and 5 W/(m2 K) respectively. The latter correspond to the
Modified Peng-Robinson [36] EoS is used. As compared uninsulated pipeline exposed to still air in all simulations.
to the PR EoS, this equation has been shown to produce
better predictions of the phase equilibrium data during the An equidistant grid system comprising 100 nodal points is
most part of the depressurization process [37]. employed for the fluid dynamic simulations using PipeTech.
The corresponding discretisation time element is determined
Results and discussion using 90% of the Courant, Friedrichs and Lewy value [38].
The HLP full-scale experiments involved three series of

Validation burst tests, referred to as test series A, B, and C using X70
API grade pipelines containing air and a rich gas mixture.
The following shows the results relating to the validation of the Pipeline fracture was initiated using an explosive charge.
dynamic boundary ductile-fracture model presented above, Figure 1 shows a schematic representation of the pipe setup.
hereby referred to as DBFM, by comparison of its predictions
against the following published experimental data: Figures 2 to 4 show the variation of the crack velocity with
crack length for the south-running A1, B1 and C2 tests
• HLP full-scale burst test [20] respectively for the HLP full-scale experiments. Curves A
• ECSC X100 pipe full-scale burst test [21] show the measured crack length; Curves B, on the other
• Alliance full-scale burst tests [22] hand are the simulation predictions. In all cases, the

crack length for test Alliance Test 1.
Inventory:
rich gas (Table 2), initial pressure
= 120.2bara, initial temperature =
23.9°C.
corresponding Charpy Energy, Cv, for each pipe section is Returning to the simulation data (curves A), given the
given in the figures. experimental uncertainties, it is clear that in all cases the
DBFM predictions produce reasonably good agreement
Figures 5 and 6 show the corresponding data for ECSC with the test data.
X100 [21] and Alliance full-scale burst tests [22], respectively.
Returning to Figs 2-6, as it may be observed, the crack CO2 pipeline ductile fracture investigation
velocity significantly decreases with increase in crack length.
This is due to the significant rapid decrease in the crack The following describes the results of the application of
tip pressure as the pipeline depressurizes. As an example, the validated DBFM to the rupture of a hypothetical CO2
such behaviour expressed in terms of the variation of the pipeline. To ensure practical relevance, the respective
cark tip pressure with time is shown in Fig.7 for the HLP pipeline internal diameter and wall thickness of 590.7mm
A1 south-running crack. and 9.45mm are employed in the simulations. Cosham and
Eiber [7] suggest that such dimensions are the most likely for
Also as expected, the crack velocity decreases as the crack CO2 pipelines to be employed in CCS. The same authors
propagates into the pipeline section with the higher also propose that a Cv of 50 J would be sufficient to arrest
toughness, eventually coming to rest in all cases. As expected, a fracture for typical operating conditions of 100barg and
the data in Fig.4 show the smallest crack length as compared 10oC. The same pipeline operating conditions are chosen
to the other tests due to the combination of the much higher in the proceeding simulations.
fracture toughness pipe material employed together with
the lowest initial pressure. Figure 8, curve A, shows the predicted variation of the crack
velocity versus crack length based on the above conditions
The initial rapid increase in the crack velocity observed in many for the CO2 pipeline. For the sake of comparison, the
of the test data is due to the finite time taken for the initial analogous data for methane (curve B) and natural gas
notch to fully develop into an open flap following detonation. (85% methane-15% ethane, curve C) inventories are also
This time domain is ignored in the present simulations. presented.
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 10°C rise in the line temperature results in a fast-running
pipeline (curve C) exhibits a relatively constant high-velocity propagating fracture which covers the entire length of the
crack which propagates through almost the entire length of the pipeline.
pipeline before coming to rest at approximately 89m. Similar
trends in behaviour but of a smaller magnitude is observed in To explain the above, Fig.10 shows the variation of the
the case of the methane pipeline (curve B) where the fracture crack tip pressure with temperature for the starting line
comes to rest at a distance of approximately 18m. Of the temperatures of 10, 20, and 30°C relative to the CO2
three cases examined, the CO2 pipeline (curve A) offers the saturation curve. The latter is generated using the Span
best resistance to ductile fracture. Here the fracture almost and Wagner [39] equation of state for CO2. The calculated
instantaneously comes to rest at a distance of only 6m. crack arrest pressure of 43bara is also indicated in the same
figure. In the case of the highest line temperature of 30°C
Impact of line temperature (curve A), soon after pipeline failure, the dense-phase CO2
inventory crosses the saturation curve at the maximum
Based on an analysis of the CO2 depressurization trajectory, pressure of ca. 60bara, some 17bara higher than the crack
Cosham and Eiber [7] postulate that the initial temperature arrest pressure of 43bara, thus resulting in a propagating
of the CO2 pipeline may have a significant impact on its fracture. The fracture comes to rest once the crack tip pressure
resistance to ductile fracture failure. 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
Figure 9 shows impact of the line temperature on the propagated through the entire length of the pipe.
variation of fracture velocity versus fracture length for the
CO2 pipeline at four selected temperatures of 30°C (curve Conclusion
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 The development and validation of a dynamic boundary
temperature results in a relatively modest increase in the ductile-fracture-propagation model for pressurized pipelines
fracture velocity and fracture arrest length. The data at 30°C was presented. The model, based on the coupling of a
(curve A) is an exception to this rule. Remarkably only a semi-empirical fracture model with the transient real fluid

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 propagation. A relatively modest increase in the line
important fluid/structure interactions governing the fracture temperature from 20 to 30°C resulted in a running fracture
propagation and arrest process. which propagated through the entire length of the 100m
pipe. Analysis of the data revealed that upon crack initiation,
A particularly important feature is accounting for the change the pipeline inventory rapidly transforms from the dense
in the effective pipeline length as the pipeline unzips and phase into the saturated state, thereafter following a relatively
its impact on the crack tip pressure. Following its successful prolonged depressurization trajectory along the saturation
validation against real pipe burst data reported for air and curve. The crack will propagate for as long as the crack tip
rich-gas mixtures, the model is used to test the propensity pressure remains higher than the crack arrest pressure. In
of a hypothetical but realistic pressurized CO2 pipeline to the case of the CO2 pipeline at 30°C this cross over will
ductile-fracture-propagation failure. Such investigations are not happen before the fracture has propagated through the
particularly timely given the real prospects of using CO2 entire length of the pipe. Obviously such phenomenon will
pipelines as part of the carbon capture and sequestration have significant practical implications when transporting
(CCS) chain. CO2 at different ambient temperatures as part of the CCS.
Simulations conducted using a 100-m long pipeline The study assumes that the adopted rather simplistic but
containing methane, natural gas or CO2 at 100barg and nevertheless effective drop-weight tear test (DWTT) energy
10°C revealed that whereas for the natural gas and methane approach validated for air and rich gas mixtures is also
inventories the fracture propagated through most of the pipe applicable to CO2. This assumption is justified given the fact
length, in the case CO2 pipe, the crack length was limited that the only fluid parameter introduced into the DWTT
to only a short distance. approach is the crack tip pressure.
The change in the temperature of CO2 however was found Furthermore the fluid-flow model employed is based on
to have a remarkable impact on the resistance to fracture the plausible homogenous flow assumption in which the

crack length for the 100-m long, 100-
barg CO2 pipe at different starting line
temperatures.
constituent fluid phases remain at thermal and mechanical Safety and Environmental Protection, 154, pp869 - 879.
equilibrium during the fracture-propagation process. Manchester: IChemE.
7. A.Cosham and R.J.Eiber, 2008. Fracture propagation in CO2
Addressing the non-equilibrium phase behaviour, together pipelines. J. Pipe. Eng., 7, pp115-124.
with the impact of the typical impurities present within the 8. M.Bilio, S.Brown, M.Fairwheather, and H.Mahgerefteh, 2009. CO2
CO2 stream for the various capture technologies on the pipelines material and safety considerations. IChemE Symposium
ductile fracture behaviour of CO2 pipelines, is currently Series: HAZARDS XXI Process Safety and Environmental
being investigated by the authors. Protection, 155, pp423–429. Manchester: IChemE.
9. H.Kruse and M.Tekiela, 1996. Calculating the consequences of
References a CO2-pipeline rupture. En. Conv. Man., 37, 95, pp1013-1018.
10. D.J.Picard and P.R.Bishnoi, 1988. The importance of real-
1. A.Bartenev, 1996. Statistical analysis of accidents on the Middle fluid behaviour and non-isentropic effects in modeling
Asia-Centre gas pipelines. J. Haz. Mat., 46, 1, pp57-69. decompression characteristics of pipeline fluids for application
2. G.Papadakis, 1999. Major hazard pipelines: a comparative in ductile fracture propagation analysis. Cana. J Chem. Eng.,
study of onshore transmission accidents. J. Los. Prev. Proc. Ind., 66, 1, pp3-12.
12, 1, pp91-107. 11. G.Fearnehough, 1974. Fracture propagation control in gas
3. PHMSA, 2010. Pipeline incidents and mileage reports. Retrieved pipelines: a survey of relevant studies. Int. J. Pres. Ves. Piping,
from primis.phmsa.dot.gov/comm/reports/safety/PSI.html. 2, 4, pp257-282.
4. B.N.Leis, X.Zhu, and T.Forte, 2005. Modelling running fracture 12. Idem., 1987. Calculation of the thermodynamic sound velocity
in pipelines: past, present and plausible future directions. in two-phase multicomponent fluids. Int. J. Multip. Flow, 13,
Proc. 11th Int. Conf. Frac., Italy. 3, pp295–308.
5. W.A.Maxey, 1974. Fracture initiation, propagation and arrest. 13. H.Mahgerefteh and O.Atti, 2006. Modelling low-temperature-
In: Proc. 5th Symposium in Line Pressure Research. Houston. induced failure of pressurized pipelines. AIChE J., 52, 3, pp1248-1256.
6. H.Mahgerefteh, G.Denton, and Y.Rykov, 2008. CO2 pipeline 14. B.N.Leis, 1997. Relationship between apparent (total Charpy
rupture. IChemE Symposium Series: HAZARDS XX Process V-Notch toughness) and the corresponding dynamic crack-
Fig.10.The variation of crack tip

pressure with temperature for the 100-
m long, 100-barg CO2 pipe at different
starting line temperatures.
propagation resistance. Proc. Int. Pipe. Conf., Calgary: ASME. 22. D.M.Johnson, N.Horner, L.Carlson, and R.J.Eiber, 2000. Full
15. H.Makino, I.Takeuchi, and R.Higuchi, 2008. Fracture scale validation of the fracture control of a pipeline designed
propagation and arrest in high-pressure gas transmission to transport rich natural gas. Pipeline Technology, 1, 331.
pipeline by ultra strength line pipes. In: 7th Int. Pipeline 23. H.Mahgerefteh, O.Atti, and G.Denton, 2007. An interpolation
Conf.. Calgary. technique for rapid CFD simulation of turbulent two-phase
16. P.E.O’Donoghue, M.F.Kanninen, C.P.Leung, G.Demofonti, flows. Pro. Safe. Env.
and S.Venzi, 1997. The development and validation of a 24. H.Mahgerefteh, A.Oke, and Y.Rykov, 2006. Efficient numerical
dynamic fracture propagation model for gas transmission solution for highly transient flows. Chem. Eng. Sci., 61, 15,
pipelines. Int. J. of Pres. Ves. Piping, 70, 1, pp11-25. pp5049-5056.
17. Z.Zhuang, 1999. Analysis of dynamic fracture mechanisms in 25. H.Mahgerefteh, P.Saha, and I.G.Economou, 1999. Fast
gas pipelines. Eng. Frac. Mech., 64, 3, pp271-289. numerical simulation for full bore rupture of pressurized
18. H.Makino, T.Sugie, H.Watanabe, T.Kubo, T.Shiwaku, S.Endo, pipelines. AIChE J., 45, 6, pp1191–1201.
et al., 2001. Natural gas decompression behaviour in high 26. A.Oke, H.Mahgerefteh, I.Economou, and Y.Rykov, 2003. A
pressure pipelines. ISIJ Int., 41, 4, pp389-395. transient outflow model for pipeline puncture. Chem. Eng.
19. A.Terenzi, 2005. Influence of real-fluid properties in modelling Sci., 58, pp4591-4694.
decompression wave interacting with ductile fracture 27. M.J.Zucrow and J.D.Hoffman, 1975. Gas dynamics. Wiley,
propagation. Oil Gas Sci. Tech., 60, 4, pp711–719. New York.
20. T.Inoue, H.Makino, S.Endo, T.Kubo, and T.Matsumoto, 2003. 28. C.Bisgaard and H.Sørensen, 1987. A finite element method
Simulation method for shear fracture propagation in natural for transient compressible flow in pipelines. Int. J. Numer.
gas transmission pipelines. In: International Offshore and Methods Fluids, 7, pp291-303.
Polar Engineering Conf., 5, pp121-128. Honolulu. 29. E.Lang, 1991. Gas flow in pipelines following a rupture
21. I.Takeuchi, H.Makino, S.Okaguchi, N.Takahashi, and computed by a spectral method. J. App. Math. Phy., 42, March.
A.Yamamoto, 2006. Crack arrestability of high-pressure gas 30. K.Bendiksen, D.Maines, and R.Moe, 1991. The dynamic two-
pipelines by X100 or X120. In: 23rd World Gas Conference. fluid model OLGA: theory and application. SPE Production,
Amsterdam. 6, 6, pp171–180. Society of Petroleum Engineers.
31. J.R.Chen, S.M.Richardson, and G.Saville, 1995. Modelling 36. D.Wu and S.Chen, 1997. A modified Peng-Robinson equation
of two-phase blowdown from pipelines – 1: a hyperbolic of state. Chem. Eng. Comm., 156, 1, pp215-225.
model based on variational principles. Chem. Eng. Sci., 50, 4, 37. H.Mahgerefteh, G.Denton, and Y.Rykov, 2008. A hybrid
pp695–713. multiphase flow model. AIChE J., 54, 9, pp2261–2268.
32. idem, 1995. Modelling of two-phase blowdown from pipelines 38. H.Mahgerefteh, Y.Rykov, and G.Denton, 2009. Courant,
– 2: a simplified numerical method for multi-component Friedrichs and Lewy (CFL) impact on numerical convergence
mixtures.Ibid., 50, 13, pp2173-2187. of highly transient flows. Chem. Eng. Sci., 64, 23, pp4969-4975.
33. H.Makino, I.Takeuchi, M.Tsukamoto, and Y.Kawaguchi, 2001. 39. R.Span and W.Wagner, 1996. A new equation of state for
Study on the propagating shear fracture in high strength line carbon dioxide covering the fluid region from the triple-point
pipes by partial-gas burst test. ISIJ Int., 41, 7, pp788-794. temperature to 1100 K at pressures up to 800 MPa. J. Phys.
34. PipeTech. Pipeline rupture simulation software: www. Chem. Ref. Data, 25, 6.
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.
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 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 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
Background because of the different thermodynamic behaviour of the
contents, particularly for dense-phase transport. There is
With the current drive towards reducing emissions of carbon no public domain evidence that such a failure mechanism
dioxide (CO2) gas to the atmosphere, pipeline operators has occurred either in natural gas pipelines in service or in
and designers are investigating options for the transport of laboratory experiments.
bulk quantities of CO2. This will involve moving CO2 from
sources such as power stations equipped with carbon capture The postulated failure mechanism involves a complex
to storage sites such as aquifers or depleted oil or gas fields. interaction of the thermo-fluid mechanics of a leaking
This will require either the construction of new pipelines, pipeline, heat transfer driven by cold fluid escaping through
or the re-use of existing pipelines originally constructed to either a crack or a hole, crack initiation and propagation,
carry natural gas or other fluids. and crack arrest. These issues are discussed separately, after
a summary of the postulated failure mechanism in the next
A concern has been raised that leaks in a CO2 pipeline section. The issues considered are grouped as fluid flow
could escalate to a propagating fracture. This is due to and heat transfer at a leak, fracture initiation, immediate
the potentially large temperature drop associated with the arrest after initiation, and full-bore fracture propagation,
expansion of either gaseous or dense-phase CO2 to ambient and these are considered successively in the subsequent
conditions. It is suggested this cooling would lower the sections. This arrangement of the material does involve
pipe wall temperature to an extent that a brittle fracture some repetition, but it was considered the best approach
to ensure that all issues are covered. A general discussion
This paper was presented at the First International Forum on Transportation of CO2 and conclusions then follow.
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. It should be noted that this paper has been prepared in the
context of UK pipeline design practices, although most of
*Author’s contact details
tel: +44 (0)1509 621814
the analysis and the conclusions should be appropriate for
email: bandrews@fleetech.com other locations.
Terminology • were Kc is described as “the critical fracture toughness

below which a fracture propagates”, Y is a “shape
The following terminology has been used to distinguish factor depending on the crack length and geometry”,
between different possible types of defect. It is important a is the crack half length, and σ is “the sum of the
that these distinctions are maintained, as there can easily pressure and thermal stresses”.
be confusion between the different possible types of loss of
containment in the pipe wall. • Although not completely clear in the published
papers, it appears to be assumed in [2] that once
• Leak – a break in the pipe wall through which fluid the condition in Eqn (1) is satisfied, a propagating
can escape. This is used as a generic term. A leak fracture occurs immediately. In [2] this is described
is stable at the current length, applied stress and as “a secondary more catastrophic running brittle
metal temperature. fracture”; in [1] both ductile and brittle propagating
fractures are described.
• Crack – a sharp-ended feature which can be analysed
by fracture mechanics methods which characterize No experimental evidence is given to support this
the crack tip stresses as a singularity. mechanism. In [3] it is suggested that this mechanism was the
cause of the gas pipeline failure at Ghislenghien, Belgium,
• Hole – a rounded feature, typically (but not in July 2004. At the time of writing this incident is still the
necessarily) circular. Fracture mechanics’ methods subject of litigation and the official Belgian government
are not applicable to holes. enquiry has not issued a report, so it is not clear if this
speculation is correct.
• Rupture – a break in the pipeline which has an
opening equivalent to at least the full-bore area. Heat transfer and
A rupture may remain stable or may escalate to a
propagating fracture. fluid flow at a leak
• Propagating fracture – a fracture moving continuously As noted above, heat transfer between the leaking fluid and
along the pipeline at a velocity of around 100 m/s the pipe wall is essential if the material around the leak is to
or higher. be cooled and possibly reach the lower shelf of the ductile-
brittle transition curve. This section considers these issues.
Postulated failure due to local Heat transfer and flow at a leak in a pipeline
cooling at a leak
The model in [2] assumes that flow both within the pipeline
The failure mechanism postulated in [1, 2] can be summarized and through the leak in the wall is turbulent and that heat
as follows: transfer is by forced convection. The leak is considered to
be a 5-mm diameter hole, with a crack extending from it
• Fluid (usually a gas but possibly a dense-phase fluid) for 50mm. It is not clear if leakage is assumed through only
escapes through a leak in the pipeline wall, expanding the hole, or through both hole and crack. Heat transfer
and cooling. coefficients are estimated using coefficients from the
literature. Although not clearly stated, it appears that the
• The low temperature fluid cools the pipe wall at flow through the leak is treated in the same way as flow
the leak. within the pipe. The model also does not allow for heat
recovery from the surrounding soil as [2] refers to natural
• Cooling a carbon steel reduces its toughness as the and forced convective heat transfer “to ambient”. Whilst in
material becomes brittle at lower temperatures. the immediate vicinity of the leak the pipe may be exposed,
Additionally, the temperature differential between even here it will be surrounded by escaped gas rather than the
cold material at the leak and warmer areas remote ambient. Remote from the leak the pipeline will be buried
from the leak induces thermal stresses. and heat transfer would be expected to be by conduction
from the soil, not by a convective mechanism.
• The combined effect of reduced toughness and
thermal stress initiates a fracture. The condition for An analysis has been carried out for leaking ethylene pipelines
fracture is set in reference [2] by a pure linear elastic in [4]. This analysed flow and heat transfer at circular holes
fracture mechanics (LEFM) approach, with fracture in the wall of ethylene pipelines, with contents in both
assumed to occur when: the gas and dense phases. In most cases the flow through
the hole was choked and the bulk of the fluid expansion
(1) actually occurred in a shock wave outside the pipe wall.
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.
The fluid flowing through the hole was thus at a similar greatest cooling occurs when the pipeline is shut-in so that
temperature to the bulk fluid and so there was little local the bulk fluid pressure and temperature fall. For the more
cooling of the material around the hole. The exterior of likely case of a small undetected leak, where flow continues,
the pipe was surrounded by a cloud of cooled gas, but the the temperatures return to ambient over shorter distances as
heat transfer from this gas to the pipe wall was relatively the pipe wall is heated by the warm flowing bulk contents.
low. Other points from this study were:
Experimental studies on leakage from full-scale pipelines
• Steady-state pipe metal temperatures were reached are rare, as they are difficult and expensive to perform. One
within 1 minute, rather than taking up to 10 minutes relevant study was carried out by SZMF [5] on a 1067-mm
(the timescale predicted in [2]) to reach a steady state. diameter, 33.5-mm wall thickness grade L555 (X80) vessel
simulating a storage vessel for compressed natural gas (CNG)
• Figure 1 shows predicted wall temperatures from transport applications. Leakage of lean natural gas through
[4] as a function of pressure during blowdown a fatigue crack produced a lowest measured temperature (on
of a supercritical ethylene pipeline. The lowest the outer surface near the crack tip) of -70°C, in agreement
predicted temperature is -58°C but this occurs at with predictions made by an unknown method, and these
only approximately 25% of the initial pressure, predictions are reproduced in Fig.2. They show a significant
which would considerably reduce the driving force through-wall temperature gradient, with the crack at inner
for fracture initiation. surface of the vessel at around -10°C. The cooling is greater
than [2] predicts for natural gas, albeit starting from a higher
• For supercritical conditions, Fig.1 shows that the pressure of 180 bar. This result suggests that leakage through
location of the hole affects the temperatures and cracks may be more severe than leakage through holes.
behaviour once the pressure falls to the boiling point
and two-phase flow is established. This is because Flow through cracks
– with a hole at the bottom of the pipe – relatively
warm liquid is forced through the hole; with a hole Workers at Sheffield University have studied the flow of
at the top, colder gas escapes from the hole. fluids through narrow cracks, rather than through a hole
[6-8]. Whilst they did not study heat transfer effects, and
• The model in [2] appears to assume the pipe is above the pressure differentials were small, they did identify an
ground, as there are references to convective heat effect of surface roughness on the flow regime. At very small
transfer. In contrast, Saville [4] explicitly models a openings, comparable to the surface roughness, the flow
buried pipeline and includes heat recovery from the was essentially laminar and “followed” the local surface
soil. Crater formation at the leak due to displacement roughness. With increasing separation of the crack faces,
of the soil by the escaping fluid was also considered. the flow could move between the peaks of the surface and
transitioned to turbulent flow at higher velocities.
Although there are differences in the treatment of the local
fluid flow and heat transfer, both models show that the metal This work shows that there may be complex effects of the
temperature recovers to the bulk fluid temperature over relatively surface and crack opening; the heat transfer between the
short distances. Saville’s model predicts a more rapid recovery fluid and the crack would be affected by the flow regime
as there is less local cooling. Both models also predict that the and the fluid velocity.
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.
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
Fig.2. Predicted temperatures at a through-wall fatigue crack reduce the toughness.
due to leakage of natural gas. Figure 3 of [5], Reepmeyer,
Lothe,Valsgaard, Erdelen-Peppler and Knauf: Full-scale gas The model in [2] assumes the presence of a sharp crack
leak test at a large-diameter X-80 DSAW pipe. ASME in all cases. Work on the failure of volumetric corrosion
International Pipelines Conference 2006, IPC2006 10005. defects in low-toughness linepipe [12, 13] has shown a
Copyright ASME 2006, used with permission. substantial influence of defect geometry on the failure
Nuclear industry studies behaviour. The most significant factor appears to be
the local defect geometry, in particular the acuity of the
The nuclear industry has studied the flow of fluids through defect. The tests were carried out using material that
cracks as part of ‘leak before break’ (LBB) arguments, which was on the lower shelf of the Charpy transition curve at
aim to show that a through-wall crack will leak before room temperature, but the behaviour of blunt machined
breakage or rupture occurs. A part of these arguments is to defects simulating volumetric corrosion at quasi-static
show that a leak can be detected before further sub-critical strain rates was predicted by conventional ductile-failure
crack growth leads to rupture; this requires an estimate of models. The trend of the results is shown in Fig.3, which
leakage rates, and much of this work is summarized in [9]. shows the transition temperature as a function of the
However, the nuclear LBB studies have concentrated on stress-concentration factor associated with the defect. The
isothermal leakage, presumably because the containment stress-concentration factor increases as the defect becomes
is effectively at constant temperature. This is not relevant sharper, showing that the effective transition temperature of
to the current issue for CO2 pipelines where heat transfer a rounded defect can be well below the Charpy transition
to the containment (the pipe) from the fluid expanding in temperature. Thus, for leakage through holes rather than
a non-isothermal manner is the key issue. cracks, it is unlikely that fracture initiation would occur,
even in low-toughness material.
Fracture initiation Crack initiation due to local cooling
This Section considers issues associated with the analysis
and prediction of fracture initiation from a locally cooled Fearnehough [14] carried out experiments using liquid
area of a pipeline. nitrogen vapour to cool steel plates containing a centre crack
of length 600mm and subject to remote axial tension stresses
Fracture analysis methods in the range 15 to 124N/mm2. The aim was to simulate the
effects of spilling LNG onto the outer containment of a
The model in [2] assumes LEFM applies, so that the initiation storage tank. The material tested was a grade 43A structural
of fracture is controlled by the stress-intensity factor; the paper steel, which is approximately equivalent in strength to a
also assumes a step transition between ductile and brittle Grade L290 (X42) pipeline steel. The reported 30-J Charpy
behaviour, although this is an oversimplification of the true impact temperature was -15°C; as the material thickness
behaviour of a pipeline steel. Ignoring the beneficial effects was 6.3mm, this would have passed the typical UK gas
of warm pre-stressing, discussed below, the behaviour of the transmission pipeline requirement of 27J at 0°C in a 2/3
pipeline should reflect the low constraint of a thin-walled specimen.
structure loaded in tension. Thus, even on the lower shelf, the
behaviour is likely to be better than would be expected from Local cooling to around 80°C below the bulk metal
simple LEFM considerations. Fracture initiation should be temperature was required to initiate cracks, which then
predicted using an elastic-plastic measure of toughness such as arrested. It was argued that the fracture was initiated by the
the crack-tip-opening displacement (CTOD) or the J-integral. additional thermal stress generated by the local cooling and
the applied remote stress level was not a factor. It should
There are other concerns with the fracture-analysis method be noted, however, that the stresses used in this study were
in [2], as the authors appear to assume that the thermal relatively low by the standards of many pipelines.
and primary stresses can simply be added to give a higher
uniform hoop tensile stress when calculating the stress- The experimental study on a Grade L555 vessel [5]
intensity factor. The thermal stress will decay remote from discussed above shows that fracture initiation will not
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 An arrested crack with a larger opening would be closer
is subject to significant local cooling. The fatigue crack to a hole than a crack. If, as the work considered above
in the vessel did not propagate and remained stable suggests, there are differences in the fluid flow and heat
throughout the experiment. This was a modern high- transfer of holes and cracks, a larger defect would behave in
toughness material, with an upper-shelf Charpy energy a manner similar to a hole and there would be less cooling.
exceeding 400J at -10°C, so it is possible that it was not This would reduce the likelihood of subsequent re-initiation
on the lower shelf at –70°C. 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
This section considers crack arrest after initiation from a with a higher toughness. The amount of crack extension, or
locally cooled leak. This is important, as the postulated equivalently the position of the crack arrest, was considered
failure mechanism described above does not clearly to be a function of both stress level and temperature, as the
allow for this. Whilst the enlargement of a small leak propagation distance was greater in the specimens subjected
to a bigger leak, or a full-bore rupture, is not desirable, to higher remote tension. This observation that arrest was
such an escalation is preferable to the formation of a a function of stress and temperature is consistent with
propagating fracture. pipeline experience with propagating brittle fractures [15].
The West Jefferson test was developed for brittle crack Thus it is considered that, provided a pipeline is operated
arrest tests on pipeline steels and is an example of brittle above its DWTT transition temperature, a propagating brittle
crack arrest occurring under a constant applied stress. In fracture should not escalate from a leak, even when operating
this test a through-wall crack is initiated in a pressurized at stress levels up to 72% SMYS. If an old pipeline has a
vessel (so the stress ahead of the crack is constant at the high DWTT transition temperature, it could be operated
hoop stress) and propagates along the length of the pipe. at lower stress levels using the relationship between arrest
The fracture appearance and the temperature at the point temperature and hoop stress developed by Fearnehough
of crack arrest correlate well with full-scale test results and [15] and included in Edition 2 of IGE/TD/1 [16]. This
the results of the Battelle drop-weight tear test (DWTT). relation is shown in Fig.4.
The West Jefferson methodology has been used for many
years and is not dependent on the pipe contents – in fact It is possible that cooling of the remaining contents
typically the tests fill the pipe section with 95% of water during blowdown following a full-bore rupture could
and pressurize the remaining space with air or nitrogen to lower the temperature of the pipe wall below the DWTT
reduce the stored energy in the vessel. transition temperature. However, this cooling would also
be accompanied by a reduction in pressure which would
Fracture propagation and arrest reduce the hoop stress. Again, the relationship between
arrest temperature and hoop stress shown in Fig.4 could be
The phenomena of both brittle and ductile fracture used if the pressure–temperature trajectory is known from
propagation in pipelines are well understood and can be simulations of the depressurization.
analysed by accepted methods. The two types of propagating
fracture are considered separately below. It should also be An alternative approach to predicting brittle fracture arrest in
noted that propagating fractures can originate from pipeline a blowdown would be that proposed by Battelle [17], which
damage, often due to mechanical interference or corrosion, relates the arrest behaviour to the Charpy transition curve.
that does not involve the postulated cooling and fracture
mechanism. Pipeline designers already design against Propagating ductile fractures
propagating fractures for this reason. It may be necessary to
modify the design methods to take account of the specific The phenomenon of propagating ductile fractures in gas
decompression behaviour of CO2, but these changes will pipelines has been known for over 30 years, and has been
be required for a CO2 pipeline in any case. extensively studied. Rothwell gives a good overview of these
studies [18]; most recent work has concentrated on testing
Propagating brittle fractures very high strength linepipe with yield strength over 690N/
mm2 [19] and rich gas mixtures [20]. Propagating ductile
These fractures are characterized by little plastic deformation fractures typically run at axial velocities in the range 200-
and are driven by the elastic stress in the pipe wall. The 300m/s, although in some full-scale tests velocities down to
fractures travel at axial velocities above the acoustic velocity in 100m/s have been measured. These velocities are less than
the fluid. As a result, the crack tip is continually propagating the acoustic velocity in the fluid, and the crack is driven
into a region where the pipeline contents are undisturbed. by the pressure of the escaping fluid acting on the ‘flaps’
These cracks cannot be affected by rapid cooling of escaping developing behind the crack tip as the pipe cracks. At these
fluid as they are moving faster than the fluid can depressurize velocities there will be negligible heat transfer to or from
and cool. The steel ahead of the crack will remain essentially the crack tip area, and so the cooling effects described in
at the ambient temperature prior to a leak occurring. Thus a Section 2 will not occur.
propagating brittle crack in a CO2 pipeline will behave in the
same manner as a propagating brittle crack in a natural gas The steps required to control propagating ductile fractures
pipeline. The standard fracture-control approach of ensuring in a gas pipeline are also well understood, so that formal
that the pipe steel is operating above its DWTT transition fracture control plans are now explicitly required in some
temperature will prevent propagating brittle fractures. pipeline design codes such as the Australian code AS 2885
A DWTT requirement has been routinely specified for gas [21]. UK codes such as TD/1 [22] and PD 8010 [23] do
transmission pipelines in the UK since the introduction not have such an equivalent explicit requirement, but the
of Gas Council specifications for steel linepipe in the late material toughness requirements in the codes will achieve
1960s, and most existing UK gas transmission pipelines a high level of resistance to propagating fractures.
will satisfy this requirement. Pipelines pre-dating this with
a high DWTT transition temperature would require special The special requirements for CO2 pipelines discussed in [1]
consideration, but the approach would be the same as has are well known in the pipeline industry. Maxey identified
been used for natural gas pipelines since the publication of them in [24] and by 1990 crack arrestors had been installed
Edition 2 of IGE/TD/1 [16]. For new build, any competent on the Canyon Reef Carriers CO2 [25] pipeline in west Texas
pipe supplier should be able to produce pipe meeting the because of concerns over ductile crack propagation. Whilst
usual DWTT requirement of 85% shear area for design some of the effects of rapid decompression of dense-phase
temperatures around 0°C. CO2 are counter-intuitive, they can be predicted [26]. The
effects of impurities in the gas have also been considered It was noted above that there were more-realistic fracture-
[27], but again the fundamental approach to fracture control assessment methods than the simple LEFM approach used
remains unchanged. The greatest uncertainty appears to be in [2]. A further consideration is the effect of local constraint.
in predicting the decompression behaviour of the contents. It is now widely accepted that pipelines are a ‘low-constraint’
If necessary, experimental confirmation of predictions by structure, as they are thin and defects are loaded predominantly
shock tube testing may be required. in tension. This has the effect of increasing the effective
toughness above that measured in a standard fracture-toughness
It is accepted that the control of propagating ductile fractures test. Constraint-based arguments have been used to assess the
in a CO2 pipeline may be more difficult than in a typical UK behaviour of longitudinal seam-weld defects [29] and pipeline
onshore natural gas pipeline, particularly for a dense-phase girth welds [30]. These methods can be used to assess the
pipeline. However, the methodology for control is known. behaviour of through-wall cracks in a pipeline subject to local
For a new build, these problems should be manageable cooling. Methods of including constraint effects in fracture
with correct specification of the material properties. If an assessments are given in Section III.7 of R6 Rev 4 [10].
existing pipeline is being converted to transmit CO2, then
an assessment of the toughness will be required to show The available evidence shows that if a fracture should initiate
that propagating fractures can be controlled. from a locally cooled crack, it will arrest as it extends into
warmer material. Fearnehough’s tests [14] conclusively
General discussion 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
Heat transfer stress intensity factor or when the crack grows into an area
of increasing fracture toughness. In the present case, the
The available work suggests there may be differences fracture toughness will increase as the crack is growing into
between the cooling experienced at a hole and that at a warmer material. A longer crack would have an increased
‘tight’ fatigue crack. If [4] is correct and the flow through stress-intensity factor simply due to the greater crack length,
a hole is choked so that the bulk of the expansion and but would have grown away from the high stresses associated
cooling takes place outside the pipe wall, there will be little with the locally cooled area.
cooling at a leak through a hole. If this is the case, the
likelihood of the postulated fracture mechanism occurring It is possible that a crack may grow due to the postulated
in practice reduces. local embrittlement mechanism and reach a length exceeding
the stable through-wall crack length for the pipeline. This
In contrast, the test with a fatigue crack reported in [5] would cause the leak to escalate to a rupture. The critical
suggests that expansion through a tight crack may generate through-wall crack length can be calculated using the
a substantial temperature drop in the material. The work standard methods for an axial crack in a pipeline.
discussed in above also suggests that there are differences
in flow behaviour for tight cracks which would affect heat Fracture propagation
transfer. This difference requires further investigation.
As noted above, it is considered that accepted techniques can
Fracture initiation be used to assess fracture propagation in CO2 pipelines, and
indeed these methods will be required to control propagating
The available evidence for fracture initiation due to local fractures originating from other forms of damage such as
cooling is conflicting. Fearnehough’s tests [14] using liquid external interference. It is understood that work is being
nitrogen vapour showed that fracture can be initiated at a carried out elsewhere to address issues specific to the control
locally cooled crack, and there are other cases known of of propagating fractures in CO2 pipelines, in particular the
failures occurring when components are cooled under load: rapid decompression behaviour of CO2.
for example the failure of a heat exchanger in the Longford
incident [28] was attributed to a combination of thermal Warm pre-stressing
shock and local embrittlement of cold material. However,
the test in [5] showed that fracture will not necessarily occur Warm pre-stressing is a phenomenon where the low
under these conditions, and tests on simulated corrosion temperature fracture toughness of a material is improved by
defects discussed above have shown that the defect acuity a prior loading at a higher temperature. Various explanations
appears to be a major factor. exist for the effect, such as crack-tip blunting at the higher
A combination of a substantial temperature drop, a sharp temperature, the generation of beneficial residual stresses at
crack, and a material with a high transition temperature a crack tip by the prior loading, or metallurgical effects such
is required to initiate a fracture due to local cooling of as the de-cohesion of brittle second phase particles during
a leaking pipeline. Quantifying these effects will require the high temperature loading. Guidance on the application
understanding of the temperatures generated at a leak, of warm pre-stressing arguments is given in Annex O of BS
together with fracture analyses. 7910 [11] and Section III.10 of R6 Rev 4 [10].
Rio
Pipeline
FP1
2011
Conference & Exposition
September 20-22
Rio de Janeiro • Brazil
Call for Papers:

11/10
Submission Deadline
December 17, 2010
Participation Organization / Realization
Information:
Phone.: (+55 21) 2112-9000
Fax: (+55 21) 2220-1596
e-mail: riopipeline@ibp.org.br
www.riopipeline.com.br
For a defect created in an operating pipeline, without of 0.56°C per bar, the maximum cooling associated with a
unloading and re-loading, crack tip residual-stress effects leak would be 21°C. If the operating temperature is +4°C,
cannot be significant. Thus the operating mechanisms are the metal temperature could reach as low as -17°C at a leak,
likely to be crack-tip blunting and, possibly, metallurgical if perfect heat transfer occurred and there is no heat recovery
effects. There is ample evidence that brittle and transition from the bulk contents or surrounding soil.
fracture toughnesses are strongly influenced by notch acuity:
all the fracture-toughness testing standards require a fatigue- Crack initiation
sharpened crack, because this gives lower-bound results.
Limits are set on the maximum stress intensity factor Kmax To assess the possibility of fracture occurring at a sharp crack
when fatigue pre-cracking fracture-toughness specimens to under these conditions, the fracture resistance of the material
ensure the crack is not artificially blunted before the main at -17°C must be estimated. Unless test data exist, it is only
fracture test. The stress acting on a crack before cooling possible to estimate the effects of cooling from the linepipe
can be considered to be analogous to a fatigue pre-cracking specification acceptance criteria. Such an approach should
cycle with a high Kmax, leading to crack-tip blunting and be conservative, as the actual material properties should
an elevation of the fracture toughness. be better than the specification minima. LX/1 compliant
material would have achieved 27J Charpy energy at 0°C
In the authors’ opinion it is not appropriate to claim credit in a 2/3 size specimen, and would have 75% shear area in
for the commissioning hydrotest as a prior warm load if the DWTT at 0°C. Unfortunately this information cannot
warm pre-stressing arguments are used. This is because most be used directly to provide the toughness at a different
defects will have been introduced after the hydrotest; for temperature. The approach developed by Battelle for brittle
example mechanical damage will have produced a gouge fracture arrest predictions [17] can be used, as follows.
or dent-gouge defect at some time after commissioning.
Hence the hydrotest will have stressed defect-free material, Figure 6 of [17] can be used to estimate that the Charpy
and there can have been no effect of the hydrotest on a transition temperature in a 2/3 size specimen would be 18°F
crack-tip region which did not exist at the time of the test. (10°C) below the DWTT transition temperature, which is
known to be 0°C or lower for LX/1 material. Hence the
It can be argued that the initial load due to internal pressure estimated Charpy transition temperature is -10°C, and so
before the onset of any cooling at a leak is the warm pre-stress, when fully cooled, the material is operating at 7°C below
and the additional local thermal stress generated by cooling the transition temperature. Using Fig.5 of the Battelle
is an additional load, giving a ‘load – cool – fracture’ case. report, the shear area in a 2/3 size Charpy specimen tested
The applied stress-intensity factor will be increasing during at 7°C below the transition temperature is 72%. At this
the cooling due to thermal stresses adding to the driving shear area Fig.7 then shows that the impact energy is 75%
force due to pressure loading (which will remain constant if of the maximum upper-shelf energy. Thus the minimum
the leak is small and undetected). As a result the simplified Charpy toughness of the material at a fully cooled leak
warm pre-stressing argument, that failure is avoided if the would be predicted to be 0.75 x 27 = 20J. In practice, the
stress-intensity factor is monotonically falling during cooling, material upper-shelf energy would be above the specification
will not hold, and a more-detailed analysis is required. minimum, and it is likely that the transition temperature
would be lower than that estimated, so the 20-J estimate
Case study is conservative. The effect of reducing the impact energy
from 27J to 20J is to reduce the critical through-wall crack
This section applies the arguments developed in this paper length calculated using the NG-18 toughness-dependent
to a case where an existing pipeline is being re-used to model [31] by about 10%, from 350mm to 320mm.
transmit CO2. Such re-use is environmentally beneficial as Given the conservative assumptions in the analysis, and
it avoids the carbon costs associated with the construction the likely benefits of warm pre-stressing, it is judged that
of a new pipeline. this difference is not significant and that brittle fracture
initiation is unlikely to occur for this specific case.
Basic data
The change in toughness predicted by the analysis in the
The pipeline is an existing gas transmission pipeline being previous paragraph is a function of the temperature change
reused to transmit gas phase CO2 at 38bar. This pipeline and, to a limited extent, the wall thickness. The pipe wall
is a 36-in (914mm) nominal diameter line with a nominal thickness affects the shift between the Charpy transition
wall thickness of 12.7mm, and was constructed from pipe temperature and the assumed DWTT transition temperature
complying with the former British Gas LX/1 specification. of 0°C for LX/1 compliant material. Smaller shifts will occur
Thus the material would be expected to have a DWTT for thinner pipe; other cases will require specific evaluations.
transition temperature of 0°C. The hoop stress under the For modern linepipe, the degree of cooling estimated in
proposed operating conditions is 137N/mm2. For an X60 this case would not be expected to be a problem, as the
material this is equivalent to 33% SMYS based on the material would be expected to still be on the upper shelf at
nominal wall thickness. Assuming Joule-Thomson cooling temperatures around -20°C.
It should also be noted that in Fearnehough’s tests on steel Concluding remarks

plates [14] the highest stress was 124N/mm2, comparable
to the nominal stress of 137N/mm2 in this case. A cooling The issues regarding a postulated failure mechanism for
of over 100°C was required to initiate fracture in this test. leaking CO2 pipelines have been reviewed. The main
The cracks in these tests were 600mm long, greater than conclusions are:
the tolerable through-wall crack length for this pipeline.
• It is considered possible that under a restricted range
Crack propagation of circumstances a leak in a CO2 pipeline may enlarge
due to local cooling causing a brittle fracture. Such
Using the British Gas relation between hoop stress enlargement is only likely to occur for a sharp crack
and temperature shift below the DWTT transition in a material with a high ductile–brittle transition
temperature, shown in Fig.4, it can be seen that the temperature in a pipeline operating at a high pressure.
operating point at a hoop stress of 33% SMYS and a
temperature shift of 17°C below the DWTT transition • The main uncertainty in assessing if such enlargement
temperature is just above the 95% probability of arrest could occur is in the understanding of fluid flow,
line. Thus it is considered that even if a long length heat transfer, and cooling associated with the leak.
of pipeline were cooled to a low metal temperature, a This will be strongly influenced by the leak geometry.
propagating brittle fracture is not a credible event under
the proposed operating conditions. • Industry-standard fracture-assessment methods
can be used to determine if fracture will occur at
Further work a leak. These methods can take account of elastic–
plastic fracture and constraint effects to eliminate
The greatest unknown in assessing whether it is possible for unnecessary conservatism.
a leak to escalate to a rupture due to the postulated local
cooling mechanism is the amount of cooling generated at the • The beneficial effects of warm pre-stressing should
leak. Once the cooling is quantified, given knowledge of the be quantified and included in future assessments.
material’s toughness transition curve, fracture-mechanics’
methods can be used to assess if brittle fracture will occur. • For the specific case study of an existing pipeline
The published work suggests that the amount of cooling being re-used to carry gas phase CO2 at 38bar, it is
is strongly influenced by the geometry of the leak. Work is considered that crack extension by the postulated
required to address this issue. cooling and embrittlement mechanism will not occur.
It is suggested that the heat transfer and local cooling issue Acknowledgements
will require both numerical and experimental work. Whilst
numerical studies can give an indication of trends and the This study was funded by National Grid Gas Transmission,
important parameters, it is considered that experimental which is gratefully acknowledged.
verification is required.
References
An initial study using computational-fluid-dynamics’
(CFD) methods could be used to estimate leakage rates 1. M.Bilio, S.Brown, M.Fairweather, and H.Mahgerefteh,
and heat-transfer coefficients; this would be combined with 20009. CO2 pipelines material and safety considerations. In:
finite-element heat-transfer calculations for the steel pipe Hazards XXI - Process Safety and Environmental Protection
wall and stress analysis to determine the thermal stresses. in a Changing World. IChemE, Manchester, pp423-429.
The analyses should consider a range of hole sizes and 2. H.Mahgerefteh and O.Atti, 2006. Modelling low-temperature-
crack geometries, up to a limiting length of the critical induced failure of pressurized pipelines. AIChE Journal, 52, 3,
through-wall crack length for the geometry and material. 1248-1256.
Fracture analyses would be carried out using the results 3. Ibid., 2006. An analysis of the gas pipeline explosion at
from this study. Ghislenghien, Belgium. American Institute of Chemical
Engineers, Orlando, FL, USA.
The results of the numerical work would be used to design 4. G.Saville, S.M.Richardson, and P.Barker, 2004. Leakage in
an experimental programme. This is likely to require tests ethylene pipelines. Process Safety and Environmental Protection,
on a tough material, where crack growth is not expected, 82, 1, 61-68.
and a low-toughness material with a high ductile–brittle 5. O.Reepmeyer, P.Lothe, S.Valsgard, M.Erdelen-Peppler, and
transition temperature where brittle crack extension G.Knauf, 2006. Full scale gas leak test at a large diameter X-80
is predicted. The tests should also include a range of DSAW pipe. Paper IPC06-10005, vol. 3, Part A, American
defect acuity, from a round hole to a tight fatigue crack. Society of Mechanical Engineers, Calgary, AB, Canada, pp1-8.
Instrumentation of the tests should include temperature, 6. N.M.Bagshaw, S.B.M.Beck, and J.R.Yates, 2000. Identification
strain, and crack growth. of fluid flow regimes in narrow cracks. Proc. Institution of
Mechanical Engineers, Part C. Journal of Mechanical Engineering 20. D.M.Johnson, N.Horner, L.Carlson, and R.J.Eiber, 2000. Full
Science, 214, 8, 1099-1106. scale validation of the fracture control of a pipeline designed to
7. S.B.M.Beck, N.M.Bagshaw, and J.R.Yates, 2005. Explicit transport rich natural gas. In: Proc 3rd Int. Pipeline Technology
equations for leak rates through narrow cracks. Int.J. Pressure Conference, I, Bruges, Belgium, Elsevier Scientific, Amsterdam,
Vessels and Piping, 82, 7, 565-570. pp191-209.
8. L.V.Clarke, H.Bainbridge, S.B.M.Beck, and J.R.Yates, 1997. 21. Standards Australia, 2007. Pipelines - gas and liquid petroleum,
Measurement of fluid flow rates through cracks. Idem, 71, 1, Part 1: design and construction. AS 2885.1: Standards
71-75. Australia, Homebush, New South Wales.
9. J.P.Taggart and P.J.Budden, 2008. Leak before break: studies in 22. IGE, 2008. Steel pipelines and associated installations for high
support of new R6 guidance on leak rate evaluation. J.Pressure pressure gas transmission. IGE/TD/1 Edition 5. Institution
Vessel Technology, Transactions of ASME. 130, 1, 0114021- of Gas Engineers and Managers, Loughborough.
0114026. 23. BSI, 2004. Code of practice for pipelines - Part 1: Steel pipelines
10. Anon., 2001. Assessment of the integrity of structures on land PD 8010-1. British Standards Institution, London.
containing defects. R6 Revision 4. Barnwood. British Energy 24. W.A.Maxey 1986. Long shear fractures in CO2 lines controlled
Generation. by regulating saturation, arrest pressures. Oil and Gas Journal,
11. BSI, 2005. Guide to methods for assessing the acceptability 84, 31, 44-46.
of flaws in metallic structures BS 7910:2005 incorporating 25. D.L.Marsili and G.R.Stevick, 1990. Reducing the risk of
Amendment 1, September 2007. British Standards Institution, ductile fracture on the Canyon Reef Carriers CO2 pipeline.
London. In: Society of Petroleum Engineers Annual Conference, New
12. R.M.Andrews, M. Martin, and V.Chauhan, 2006. Assessment Orleans, Vol. 3. Society of Petroleum Engineers, Richardson,
of corrosion defects in old low toughness pipelines. IPC06- Texas, pp311-319. 26. A.Cosham, 2009. “It’s a gas, Jim, but not
10140. Proc.International Pipeline Conference, ASME, New as we know it”. Pipeline Technology 2009. R.Denys, editor,
York, pp1-14. Ostend, Belgium, Scientific Surveys Ltd, Beaconsfield, UK.
13. G.Wilkowski, D.Rudland, D.Rider, P.Mincer, and 27. P.N.Seevam, J.M.Race, M.J.Downie, and P.Hopkins, 2008.
W.Sloterdijk, 2006. When old line pipes initiate fracture in Transporting the next generation of CO2 for carbon capture
a ductile manner. IPC06-10326. Proc. International Pipeline and storage: the impact of impurities on supercritical CO2
Conference. Calgary, Canada September, ASME, New York, pipelines. IPC2008-64063. Int. Pipeline Conference Calgary.
pp1-12. ASME, New York, pp1-13.
14. G.D.Fearnehough, 1985. Progressive crack extension due to 28. D.M.Dawson and B.J.Brooks, 1999. The Esso Longford gas
local cooling of a crack in LNG storage tank material. Int.J. plant accident report of the Longford Royal Commission.
Pressure Vessels and Piping, 19, 283-292. Melbourne. Government Printer for the State of Victoria.
15. G.D.Fearnehough, D.W.Jude, and R.T.Weiner, 1971. The 29. R.M.Andrews, G.C.Morgan, and W.J.Beattie, 2004. The
arrest of brittle fracture in pipelines. In: Practical application significance of low toughness areas in the seam weld of linepipe.
of fracture mechanics to pressure vessel technology. Institution IPC04-0422. Proc. International Pipeline Conference. ASME,
of Mechanical Engineers, London, pp156-162. New York, pp1-10.
16. IGE, 1984. Recommendations on transmission and 30. Y.Y.Wang and D.J.Horsley, 2003. Tensile strain limits of
distribution practice: steel pipelines for high pressure gat pipelines. Paper 35. 14th Joint Technical Meeting on Linepipe
Transmission. IGE/TD/1 Complete Edition 2. Institution Research, Berlin, May. European Pipeline Research Group,
of Gas Engineers, London. Duisburg, Germany, pp1-14.
17. W.A.Maxey, J.F.Kiefner, and R.J.Eiber, 1983. Brittle fracture 31. J.F.Kiefner, W.A.Maxey, R.J.Eiber, and A.R.Duffy, 1973.
arrest in gas pipelines. NG-18 Report 135 PRCI Catalogue Failure stress levels of flaws in pressurized cylinders. Progress
L51436. In: PRCI Report, Washington. Pipeline Research in flaw growth and fracture toughness testing ASTM STP 536.
Committee. American Society for Testing and Materials, Philadelphia,
18. A.B.Rothwell, 2007. Evolution and current status of approaches pp461-481.
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.
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
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 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
analysis substantially reported in two publications by two
of the authors [1], [2].
basis the GHG savings offered are likely to be more modest The performance of the fossil-fuel-based CCS systems is
thanks to significant resource consumption in upstream and compared to that of several electricity generating renewable
downstream processes. This paper summarizes results from a energy technologies. Subsequently the potential offered
series of lifecycle analysis investigations of hypothetical fossil- by combining biomass and CCS technologies, with the
fuel-based electricity-generating CCS plant, emphasizing ultimate objective of producing net carbon dioxide capture
the role, albeit relatively small, that transport systems play from the atmosphere, is examined. The results from this
in contributing to the overall emissions. The discussion section of the paper are necessarily indicative as there are
encompasses both downstream transport systems (i.e. for many uncertainties about the GHG impacts of biomass
carbon dioxide) and upstream systems (i.e. for fuel and production and combustion, let alone the complexities
consumable materials). Much of this discussion draws on introduced by combining biomass combustion with carbon-
capture technologies. In consequence, considerable care
by Pipeline, organized in Newcastle upon Tyne in July, 2010, by Tiratsoo Technical
must be exercised in quantitatively comparing the various
and Clarion Technical Conferences, and with the support of the University of results presented throughout the paper. Nevertheless, the
Newcastle and the Carbon Capture and Storage Association. qualitative trends are clear enough to allow some useful
*Author’s contact details conclusions to be drawn.
email: t.cockerill@imperial.ac.uk
FP2
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 month-
ly 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
Fig.1. Summary of system boundaries for lifecycle analysis.
Overview of the lifecycle analysis System specification

approach
System boundaries
Lifecycle analysis (LCA) is a standardized method for
evaluating the environmental impacts of a given process or Figure 1 summarises the extent of the system considered
different competing processes. Greenhouse gas emissions, by the lifecycle analysis results presented here, with further
other air and water emissions, resource consumption, detail of individual elements in Table 1. The key element
and energy use are evaluated using energy and material of course is the power plant itself, and our calculations
balances. The evaluation procedure covers all sub-processes include direct emissions from combustion, plant internal
within the lifecycle of the system, starting from raw material energy consumption, energy used in operating an
production and ending with product and waste disposal. By monoethanolamine-based capture system, and energy for
evaluating the environmental impacts of different systems, plant maintenance activities. Note that the impacts of
recommendations can be made to reduce possible effects. electricity transmission beyond the power station are not
The work reported here makes use of LCA in studying the included, and thus the results quoted here are for electricity
impacts (with emphasis on GHG emissions) of fossil-fuel produced rather than delivered.
power generation with and without CCS. The main objective
is to evaluate the actual reduction in GHG emissions that Upstream process direct emissions include those arising from
can be realized by CCS in various configurations. fuel production activities, encompassing the energy used to
operate the machinery and transport systems required. Also
Broadly speaking, each of the systems studied consists of included within upstream processes is the production of
fuel production, its transportation to the power plant, other consumable materials such as limestone, ammonia, and
power plant construction, power plant operation, and monoethanolamine (MEA). For gas-cycle systems, leakage
any processes related to power/capture plant operation. from pipelines is accounted for. Indirect upstream emissions
For all CCS technologies, the analysis also includes the take account of equipment manufacturing, recognizing that
capture plant construction and operation in addition to production facilities are not dedicated to servicing the power
CO2 transport and storage. plant and the associated emissions should be shared across
Fig.2. Breakdown of lifecycle GHG

atmospheric emissions from power
plant technologies.
Category Coal-based systems Natural gas-based systems
• Power plant construction • Power plant

Construction • Capture plant construction • Gas pipeline
• CO2 transport pipeline • Capture plant
• CO2 transport pipeline
Fuel combustion Direct CO2 emissions Direct CO2 emissions
Mining (Equipment manufacture, Mining Gas extraction (platform construction,

Fuel production activities, Methane leakage, Coal cleaning, gas sweetening and flaring, methane
Land recovery for surface mining) leakage)
• Limestone/ammonia production • Ammonia production

• SCR catalyst production • SCR catalyst production
Other material production • Water treatment and distribution • Water treatment and distribution
• MEA production • MEA production
• NaOH/activated C production • NaOH/activated C production
• Boiler/ESP/Gasifier ash • SCR catalyst waste

Waste disposal • FGD waste • MEA re-claimer waste
• SCR catalyst waste
• MEA re-claimer waste
• Coal transport • Gas transport

• Local by rail • gas compression
• International by ship • onshore processing
Transport • Limestone transport by truck • Methane leakage
• Chemical transport by rail • Chemical transport by rail
• Waste transport by truck • Waste transport by truck
• CO2 compression and injection • CO2 compression and injection
Table 1. Details of the sub-processes included within the lifecycle analysis described in the paper.
all uses. Downstream processes consider waste transport and of only 37.5% are achieved due to commissioning and
disposal in relatively nearby locations. As with the upstream decommissioning activities respectively. Further outline
analysis where non-dedicated facilities are used, emissions details of the power plant are contained in Tables 3-6,
are attributed appropriately across all uses. but for full information reference should be made to the
authors’ previous work [1, 2].
As the figure shows, emissions arising from the construction
phases across all the supply chain elements are included in Overall results
the calculations, taking account of :
A breakdown of the results is shown in Fig.2 and Table
• materials 7. Unsurprisingly, the supercritical plant without CCS
• material production processes produces by far the largest atmospheric GHG emissions
• material transportation by truck over an average per unit of electricity output, dominated by the direct
distance of 50km emissions from combustion. All the CCS-fitted plant
• on-site energy consumption, comprising 80% diesel produce considerably less atmospheric emissions, with the
and 20% electricity taken from the UK grid. coal-fired IGGC plant generating the smallest quantity of
emissions per unit (kWh) of electricity produced.
Power plant decommissioning is accounted for, but
decommissioning of upstream and downstream equipment Emissions associated with transport processes
is not included.
Downstream
External factors
In all the CCS cases, emissions associated with CO2 transport
As well as the system boundary, another important influence are very small, representing between 1% and 1.8% of the
on the assessment is the location of the plant and the total per unit of electricity produced. This of course only
resources consumed in construction and operation. The applies for the system specification set out above. The
analysis below has been developed on the basis that the component of these emissions arising from powering any
plant is located in North-East England, and that fuel and recompression stations could change if energy sources with
other materials are sourced relatively locally so far as possible; differing carbon intensities were employed: the present study
Table 2 provides more comprehensive details. has assumed they are powered by electricity generated with
the UK grid average carbon intensity.
Life cycle emissions for fossil-
With longer or less-secure pipelines, the impact of CO2
fuelled CCS plant transport could increase considerably, as investigated below.
Other downstream processes include ash and other solid
Power plant types 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 Upstream emissions are included within the ‘operation’
precipitation (ESP) and flue-gas desulphurization components of Fig.2, which also includes the very small
(FGD) pipe-end clean-up technologies; downstream operational emissions associates with ash and
waste disposal. The upstream calculations assume coal is
• a similar supercritical coal-fired plant, but fitted produced from a nearby UK deep mine and subsequently
additionally with an MEA-based CO2 capture unit transported to the power plant by rail. This represents
having 90% CO2 capture efficiency; something of an idealized best case, as the limited number
of UK mines means transport will in general be over longer
• a natural gas fired combined-cycle (NGCC) plant distances. Limestone is also UK sourced and transported by
fitted with similar MEA-based capture unit; truck, with other consumables such as solvents transported
by rail. Natural gas is assumed to be sourced from the UK
• a coal-fired integrated gasification combined-cycle North Sea, carried via pipelines with the specification set
(IGCC) plant fitted with Selexol-based carbon out in Table 10.
dioxide capture, again having 90% capture efficiency.
For the base coal-CCS configurations considered here,
The calculations assume that the plant has a rated capacity in general just less than 50% of upstream emissions arise
of 500MWe, an operating lifetime of 30 years, and take from mining, with a similar contribution coming from the
three years to construct. Load factors are taken as 75% production and transport of all other consumables. Coal
except in the first and last year of operation, where factors transport accounts for approximately 1.5% of upstream
Category Coal-based power plants Gas-based power plants

Power Plant Location Teesside
Surface mine: Maiden’s Hall Exten-
sion, Northumberland
Mine location –
Deep mine: Killingley Colliery,
North Yorkshire
North Yorkshire: 50 km from power
Limestone Quarry –
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

CO2 on-shore collection point 0
plant
CO2 pipeline 50 km on-shore, 150 km offshore 0.039
Bunter Sandstone-Southern North
CO2 storage 0
Sea, Closure
Table 2. Location of power plant and key material inputs.
emissions, though as will be seen later, this low value is Upstream transport processes are much more important.
largely a reflection of very optimistic assumptions. Importing coal from Russia, rather than relying on local
production, has a severe influence on the emissions for
For the natural gas CCS system considered, the majority of the PC and IGCC plant. Much of this is due to emissions
the calculated upstream GHG emissions arise because of from the transportation processes, though it has also been
escapes in the natural gas supply system, though it should assumed here that a poorer quality coal is delivered. Similarly,
be noted that we have assumed comparatively high leakage supplying the NGCC+CCS plant from a gas network with
rate of 1%. Smaller contributions are distributed over the two percentage points greater leakage increases lifecycle
production and transport of other consumables. GHG emissions by about one-third.
Across all the results for CCS plant, it is clear that the Also of great importance is the effectiveness of CO2
operational GHG emissions are much large than those from capture. A 5% reduction in the overall proportion of CO2
CO2 transport. Indeed, under the assumptions set out here, captured unsurprisingly gives a substantial increase in GHG
emissions from pipeline CO2 transport are almost negligible emissions in all cases. This result remains qualitatively true
compared to direct emissions, operational emissions, and irrespective of whether the reduction is due to less-effective
emissions associated with capture. In the carbon capture capture equipment, or increased leakage from a pipeline
transport and storage (CCTS) system, the transport element transport system.
appears to have a tiny impact on GHG emissions.
The key conclusion is that, in terms of GHG emissions of a
Sensitivity study CCS system, most of the details of the downstream processes
are relatively unimportant. This presents a stark contrast
Figure 3 illustrates how sensitive GHG emissions from each to upstream transport processes, which our results suggest
of the CCS plant are to system changes, with an emphasis on have a much larger impact on emissions. If a design objective
transport processes. It is immediately clear that the details of the is to minimize lifecycle emissions, CCS systems should in
CO2 transport system have relatively little effect, as increasing general be situated to promote ease and effectiveness of fuel
the pipeline network length by 100 km raises lifecycle emissions supply, and with little regard to the implications for the CO2
by between 0.05% (PC+CCS) and 0.08% (IGCC+CCS). The pipeline transport network. However, one important factor
details of the other downstream processes, and in particular impacting the lifecycle performance is the total CO2 captured,
plant ash waste disposal, also seem unimportant. and the downstream transport system has the potential to
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 potential of this technology is considered in the remainder
the overall configuration of the CO2 transport system has of this paper.
little impact on GHG emissions, it is vitally important that
transport is as secure as possible. Lifecycle analysis of BioCCS
Comparison with other low- Objectives
carbon energy sources
CCS with biofuel firing (BioCCS) offers the attractive
For comparison purposes, Table 8 shows ranges of values potential of producing a net removal of carbon dioxide
for GHG emissions from other low-carbon electricity- from the atmosphere, since the carbon dioxide released
production systems taken from the literature. In general it is by biomass combustion was originally absorbed from the
clear that, despite producing much lower carbon emissions atmosphere in photosynthesis. A further high-level study has
than conventional fossil-fuelled plant, CCS cannot produce examined the lifecycle implications of BioCCS drawing on
electricity that is as low carbon as most renewables. the outputs of the UKCCSC supported study, supplemented
with data from the literature as described in detail by Laczay
It should be noted, though, that the dividing line is rather [3]. Both pure-biofuel and coal co-firing cases have been
fuzzy and dependent on the location and system boundary examined, with the pure-biofuel cases considering both
of the renewable-energy technology. Solar PV in locations miscanthus1 and RC willow as fuels. For the co-firing case,
with poor resources can result in emissions per kWh higher only miscanthus was analysed.
than those calculated here for IGCC systems. Equally, most
of the renewable-energy systems’ assessments do not take Approach and assumptions
account of the impact of intermittency on the lifecycle
emissions. Including electricity storage facilities within the The study considers a circulating fluidized bed (CFB) power
system boundary, for example, can dramatically worsen plant comparable to the 550MWth / 240MWe facility
the environmental performance. This raises a number of operated by Alholmens Kraft in Pietarssaari, Finland [4].
complex issues that are beyond the scope of this paper, This plant operates at a typical thermal efficiency of 38%,
but it should be kept in mind that fossil-fuel-based CCS but it was assumed that a 90% effective CO2 capture system
systems have the potential to offer supply controllability would reduce the power output by 25%, giving an overall
and a geographical independence that certain renewables conversion efficiency for the BioCCS system of 28.5%.
find difficult to match without additional facilities.
1 Miscanthus is a tall perennial grass that has been evaluated in Europe in recent years
One possible way of further reducing the emissions of CCS as a new bioenergy crop. It is sometimes confused with elephant grass (Pennisetum
systems is by combining them with biomass fuels. The purpureum) and has been called both ‘elephant grass’ and ‘E-grass’.
Parameter Value Parameter Value

Ambient temperature, ºC 15 Type of gasifier GE oxygen-
Ambient pressure, kPa 101 blown
Steam cycle heating rate, MJ/kWh 7.4 Gasifier temperature, ºC 1250
Excess air, % 20 Gasifier pressure, MPa 6
Temperature of flue gas exiting boiler, ºC 370 Steam input to gasifier, mol H2O / mol 0.45
C
Load factor, % 75
Carbon loss, % 1
Life time, years 30
Oxidant pressure (at outlet of ASU), MPa 4
ID fan efficiency, % 85
Oxidant composition, %O2 : % Ar : % 95 : 4: 1
Table 3. Key parameters for the supercritical-PC plant. N2
Particulate removal efficiency from 50
Parameter Value
syngas, %
Number of gas turbines 2
COS to H2S conversion efficiency, % 98
Excess air, % 180
H2S removal efficiency, % 98
NOx emissions rate, ppm 10
COS removal efficiency, % 40
Air compressor ratio 15.7
CO to CO2 conversion efficiency, % 95
Compressor efficiency, % 70
Sulphur recovery efficiency, % 95
Pressure loss across combustor, kPa 28
Steam added to shift reactor, mol H2O/ 1
Temperature into turbine, ºC 1330 mol CO converted
Turbine isentropic efficiency, % 85 Table 5. Key parameters for the IGCC plant with Selexol
Mechanical and generator efficiencies, % 98 capture.Table 5. Key parameters for the IGCC plant with
Selexol capture.
Table 4. Key parameters for the NGCC plant.
Due to the complexities associated with biomass lifecycle • All power plant operate with an annual load factor
analysis, this work used a simplified approach. In particular, of 90% for miscanthus yields are 18 wet tonnes per
emissions associated with power plant and CO2 pipeline hectare per year, having 30% moisture content. Once
construction processes have been neglected. As the latter harvested the feedstock is naturally dried in storage for
are relatively small this assumption is unlikely to have a 40 days reducing the moisture content to 10%. 60kg
significant influence on the results. The former are more of nitrogen fertilizer is used when establishing each
likely to have an impact on the detail of the calculations, hectare, and a production cycle lasts 15 years after which
but not the qualitative conclusions. Emissions associated the plantation must be cleared and re-established.
with the up-keep of the carbon capture system, for example
solvent replacement, have also been neglected. • For willow yields are 14 wet tones with 50% moisture
content, with the feedstock dried to 10% moisture
It should also be kept in mind that the study takes no content.
account of the whole system indirect impacts of wider
biomass use, such as induced land use change (ILUC) • Energy crops are transported by truck 100km from
which, it is argued, could have a devastating effect on the plantation to the storage/processing site, with a
the lifecycle sustainability of certain biofuels. There has further 100-km journey to the power plant.
been much recent debate about the GHG emissions that
should be associated with biofuel production, typified by • Losses of 7% occur, 11% during storage and 3% during
the Searching-Wang debate (see for example [5]) with the transport of energy crops.
Gallagher Review providing an excellent reference [6]. The
calculations reported here account for only the emissions • Coal is imported from South America, USA, Australia
that arise directly from biomass cultivation, processing and and South Africa
harvesting operations.
Results
The Biomass Environmental Assessment Tool (BEAT2) [7,
8] was used to calculate the energy yield and combustion Table 9 shows the calculated lifecycle GHG emissions for the
products in all cases, with the following parameters: pure-biomass based CFB power plant with CCS. Both cases
Fig.4. Lifecycle comparison of fossil

CCS plant with other low carbon
energy systems.
show strongly negative net GHG emissions, with the 90% Parameter Value
capture rate more than compensating for the emissions that CO2 removal efficiency, % 90
do reach the atmosphere. The net emissions are sufficiently
negative that the simplifications outlined earlier are very SO2 removal efficiency in capture plant, 99
unlikely to change the qualitative conclusion %
SO2 removal efficiency in FGD, % 98
More detail is shown in Fig.6, which compares the origins
of the emissions reaching the atmosphere. The upstream SO3 removal efficiency in capture plant, 99
processes for miscanthus and SRC willow show slight %
differences in emissions. Miscanthus has more emissions HCl removal efficiency in FGD, % 95
associated with cultivation/harvest compared to willow. This
NO2 removal efficiency in capture plant, 25
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 Ash removal efficiency in FGD, % 50
rather than chipped. Chips are more densely transported, MEA concentration, %w/w 30
and as a result, miscanthus transport emissions are nearly
twice that of chipped SRC willow. Lean CO2 loading, mol CO2/mol MEA 0.2
Blower efficiency, % 75
Results for co-firing with coal are shown in Fig.7, where the
Pressure across blower, kPa 15
vertical axis represents net atmospheric GHG emissions per
kWh of electricity produced relative to a supercritical coal Sorbent pump efficiency, % 75
power plant without a CO2 capture unit. Unsurprisingly, Pressure across pump, kPa 200
net emissions reduce almost in direct proportion to the
proportion by energy value of biomass in the fuel mix. A Compressor efficiency, % 80
useful observation is that miscanthus-based BioCCS appears CO2 outlet pressure, MPa 13.5
to become GHG neutral for a co-firing level of approximately
20%. Higher proportions of biomass produce net capture Table 6. Key parameters for MEA-based capture process.
from the atmosphere, though some care is necessary in
interpreting the values in the light of the simplifications the results in this paper suggest that there is some value in
outlined earlier. adopting an integrated approach to the design of the whole
system, as decisions made in one part of the CCTS chain
Discussion can have implications for the GHG emission of another.
Minimizing overall emissions requires that such interactions
are fully accounted for.
Emission minimization strategies
In the cases considered here, downstream pipeline-based
Presumably a key objective in the design of any electricity- CO2 transport does not have a substantial influence on
producing CCS system is to generate power with the lowest overall GHG emissions, and thus can largely be designed
achievable GHG emissions per unit. To reach this objective, independently from the rest of the system in this regard.
Source of GHG Emissions (gCO2e/kWh) Total

Plant Type
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.
Technology Range of GHG emissions References Comments

(gCO2/kWh)
Hydro 3-33 [11], [12], [13],
[14]
Geothermal 15-23 [13], [14]
Solar PV 39-217 [15], [14], [16], Highly location depend-
[17], [18] ent. Some higher values
included battery storage
Solar thermal (to electricity) 30-120 [19] Parabolic trough, central-
ized receiver & parabolic
dish
Wind 9.7-29.5 [20],[14] Higher values generally for
offshore
Wind with pumped hydro storage 20 [21]
Wind with compressed air storage 109 [21]
Nuclear fission 6-24.2 [12], [22], [14]
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

Upstream Process GHG Emissions (A) 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
NET GHG EMISSIONS (A+C+D) -1070 -1232
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).
Fig.5. Biomass with CCS system

summary.
Pipeline CO2 transport can influence system GHG emissions Diameter, cm Thickness, mm
via leakage, and hence minimizing escapes should be a
On-shore Off-shore On-shore Off-shore
primary design objective.
75 100 7.8 9
For countries of scales similar to the UK, pipeline length
Table 10. Diameter and wall thickness for natural gas
does not have a significant effect on overall GHG emissions. pipeline.
Hence the CO2 transport distance should not play major
role in CCS plant site selection. Upstream transport likely to exhibit much poorer efficiency than steady-state
processes, notably fuel transport, have a much stronger full-load operation. In consequence they have the potential
impact on emissions and should have an influence on site to substantially increase the overall GHG emissions of a
selection. Clearly any whole-system GHG-minimization CCS system, particularly if they are frequent events.
strategy should focus on simplifying fuel rather than
CO2 processing and transport, so long as any risk of CO2 New CCS plant will most likely be initially conceived for
leakage is avoided. This is particularly true with biomass- base-load operation, and thus it could be argued that their
based CCS systems. lifecycle performance will not be impacted by non-steady-
state operation. Over their lifetime, though, there will be
Areas of uncertainty substantial changes in energy and electricity markets, and as
they age, CCS plant are likely be moved towards a peaking
While undertaking this work we have identified a number role as is common with existing old fossil-fuel plant [9].
of areas where limited technical understanding constrains Moreover, expected increasing penetration of intermittent
the usefulness of LCA approaches for the analysis and renewables will push even CCS fossil plant towards operating
optimization of future CCS systems. The two most regimes that are more variable than those experienced by
important, in the opinion of the authors, are discussed in existing fossil plant.
this section.
Evaluating the impact of transient operation on GHG
Operational effects performance is hindered by poor understanding of both
Most LCA analyses assume that the systems they study CCS plant and carbon dioxide transport systems under
operate under steady-state, full-load conditions, and this such conditions. Further work is required in these areas
is true of almost all CCS studies. Where some account is in order to fully evaluate the ‘real world’ performance of
taken of variable loading, typically, analysts use only a load future CCS systems.
factor approach to account for periods of non-generation.
Where plant will be used predominantly to supply base load, Impact of biomass combustion products on CCS efficiency
this is a reasonable assumption, especially as construction
makes a relatively small contribution to overall emissions While the results in this paper suggest that biomass-to-power
even for CCS power plant. combined with CCS has the potential to produce negative
lifecycle GHG emissions, it is important to keep in mind
The vagaries of the electricity market mean that base-load that the underlying calculations assumed there was no
operation is unlikely for all CCS systems, in practice. As a detrimental interaction between the biomass combustion
result, some plant might be subject to substantial numbers products and both the capture system together with the CO2
of cold-start and shut-down procedures, as operators try to transport system. In general this would seem a reasonable
optimize their financial return. Operators may also wish assumption, it being widely accepted that co-firing reduces the
to run plant at part load. From a lifecycle GHG-emissions emission of pollutant elements (including sulphur, nitrogen,
perspective, non-steady-state and part-load operations are and mercury) in comparison to pure coal. However, biomass
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 Operational emissions depend on the source of the energy
acid in flue gases [10]. used to power recompressions stations. However any CO2
leakage from the pipeline system to the atmosphere has the
Again there is scant data available in the literature regarding potential to dramatically increase the impact of downstream
the effect of biomass combustion products on CO2 capture transport processes.
and transport processes. Further technical data are required
before the LCA can be taken forward. Upstream transport emissions, predominantly for fuel,
are more important, typically representing at least 2% of
Conclusions all GHG emissions. For biofuels, upstream emissions are
considerably more important. This is true even without
Transport emissions are a relatively small component of the taking account of the current uncertainty surrounding the
GHG emissions from CCS systems, though the quantities whole-system sustainability of biofuels, as typified by the
vary considerably with the assumptions underlying the Searchinger-Wang debate.
lifecycle analysis. As a general rule, downstream emissions
associated with pipeline CO2 transport are almost negligible, The results have implications for reducing lifecycle GHG
certainly with respect to the construction of short pipelines. emissions from CCS plant by optimizing plant location. In
particular, the lifecycle GHG impacts are much more 7. AEA. Biomass Environmental Assessment Tool Version 2 User
sensitive to fuel transport processes than downstream Guide., 2008. Available at http://www.biomassenergycentre.
Journal of Pipeline Engineering carbon dioxide transport. From the GHG perspective, it is org.uk/pls/portal/url/ITEM/5A8E649760A5B873E04014A
suggested that optimal plant location strategies for the UK C08044B4D.
8. AEA, 2007. The biomass environmental assessment tool
should focus on minimizing fuel processing and transport,
Editorial Board - 2010 and not be overly concerned about CO2 transportation
(BEAT2). www.biomassenergycentre.org.uk/portal/page?_
pageid=74,153193&_dad=portal&_schema=PORTAL.
distance. This is especially true for biomass-derived fuels, 9. DECC, 2009. Towards carbon capture and storage: Govern-
even if they are sourced from the UK. ment response to consultation. Technical report, UK Depart-
Obiechina Akpachiogu, Cost Engineering Coordinator, Addax Petroleum ment of Energy and Climate Change.
There are two key areas of uncertainty that have had rela- 10. M.F.G.Cremers, 2009. Technical status of biomass co-firing.
Development Nigeria, Lagos, Nigeria Technical Report IEA Bioenergy Task 32 Deliverable 4, In-
tively scant coverage in the literature and require further
Mohd Nazmi Ali Napiah, Pipeline Engineer, Petronas Gas, Segamat, Malaysia work. Firstly, most lifecycle studies of CCS systems make ternational Energy Agency, August.
Dr Michael Beller, NDT Systems & Services AG, Stutensee, Germany substantial simplifications with respect to the operational 11. L.Gagnon and J.Van de Vate, 1997. Greenhouse gas emissions
from hydropower: The state of research in 1996. Energy Policy,
Jorge Bonnetto, Operations Vice President, TGS, Buenos Aires, Argentina regime of the plants under consideration. Secondly, the
25, 1, 7–13.
Mauricio Chequer, Tuboscope Pipeline Services, Mexico City, Mexico impact that combustion of biomass fuels might have on the 12. L.Gagnon, C.Belanger, and Y.Ychiyama, 2001. Life-cycle assess-
Dr Andrew Cosham, Atkins Boreas, Newcastle upon Tyne, UK performance of carbon dioxide capture systems does not ment of electricity generation options: The status of research
appear to have been extensively considered in LCA studies in year 2001. Idem, 30.
Prof. Rudi Denys, Universiteit Gent – Laboratory Soete, Gent, Belgium of CCS systems. Such an extension can be readily included 13. Y.Uchiyama, 1996. Life cycle analysis of electricity generation
Leigh Fletcher, MIAB Technology Pty Ltd, Bright, Australia within the lifecycle methodology in principle, but there is and supply systems. In Proceedings of a symposium on elec-
Roger Gomez Boland, Sub-Gerente Control, Transierra SA, a scarcity of data to support such work. tricity, health and the environment: Comparative assessment
in support of decision making, pp279–291, Vienna, Austria,
Santa Cruz de la Sierra, Bolivia October. International Atomic Energy Agency.
Daniel Hamburger, Pipeline Maintenance Manager, El Paso Eastern Pipelines, Acknowledgements 14. H.Hondo, 2005. Life cycle GHG emission analysis of power
Birmingham, AL, USA generation systems: Japanese case. Energy, 30, 2042–2056.
This paper draws on CCS systems’ analysis projects un- 15. P.Denholm and G.Kulcinski, 2003. Net energy balance and
Prof. Phil Hopkins, Executive Director, Penspen Ltd, Newcastle upon Tyne, UK dertaken by the authors over the last eight years. Major greenhouse gas emissions from renewable energy storage sys-
Michael Istre, Engineering Supervisor, Project Consulting Services, support has been provided by the Tyndall Centre, and tem. Technical Report 223-1, Energy Centre of Wisconsin.
Houston, TX, USA the UK Natural Environment Research Council (NERC) Available at http://fti.neep.wisc.edu/pdf/fdm1261.pdf.
within the scope of the UK Carbon Capture and Storage 16. J.L.R.Proops, P.W.Gay, S.Speck, and T.Schroder, 1996. The
Dr Shawn Kenny, Memorial University of Newfoundland – Faculty of Engineering lifetime pollution implications of various types of electricity
and Applied Science, St John’s, Canada Consortium.
generation. Energy Policy, 24, 3, 229–237.
Dr Gerhard Knauf, Salzgitter Mannesmann Forschung GmbH, Duisburg, Germany 17. R.Kannan, K.C.Leong, R.Osman, and H.K.Ho, 2007. Life
Lino Moreira, General Manager – Development and Technology Innovation, References cycle energy, emissions and cost inventory of power generation
technologies in Singapore. Renewable and Sustainable Energy
Petrobras Transporte SA, Rio de Janeiro, Brazil 1. N.A.Odeh and T.T.Cockerill, 2008. Life cycle analysis of UK Reviews, 11, 702–715.
Prof. Andrew Palmer, Dept of Civil Engineering – National University of Singapore, coal fired power plants. Energy Conversion and Management, 49, 18. H.Schaefer and G.Hagedorn, 1992. Hidden energy and cor-
212–220. doi:10.1016/j.enconman.2007.06.014. related environmental characteristics of PV power generation.
Singapore 2. Ibid., 2008. Lifecycle GHG assessment of fossil fuel power Renewable Energy, 2, 2, 159–166.
Prof. Dimitri Pavlou, Professor of Mechanical Engineering, plants with carbon capture and storage. Energy Policy, 36, 19. C.Dey and M.Lenzen, 2000. Greenhouse gas analysis of
Technological Institute of Halkida , Halkida, Greece 367–380. electricity generation systems. In Proceedings of the ANZSES
3. S.A.Laczay, 2009. A comparative analysis of the economics and Solar 2000 Conference, pages 658–668, Griffith University,
Dr Julia Race, School of Marine Sciences – University of Newcastle, GHG emissions of fossil fuel, co-fired and dedicated biomass Queensland, Australia, November.
Newcastle upon Tyne, UK electricity generation systems with carbon capture and storage 20 L.Schleisner, 2000. Life cycle assessment of a wind farm and
Dr John Smart, John Smart & Associates, Houston, TX, USA in the UK. Master’s thesis, Centre for Environmental Policy, related externalities. Renewable Energy, 20, 279–288.
Imperial College London, September. 21. P.Denholm and G.L.Kulcinski, 2004. Life cycle energy re-
Jan Spiekhout, Kema Gas Consulting & Services, Groningen, Netherlands 4. Alholmens Kraft. The history. Website [accessed 15 July 2009] quirements and greenhouse gas emissions from large scale
Dr Nobuhisa Suzuki, JFE R&D Corporation, Kawasaki, Japan www.alholmenskraft.com/en/history/index.htm. energy storage systems. Energy Conversion and Management, 45,
Prof. Sviatoslav Timashev, Russian Academy of Sciences – Science 5. T.Searchinger, R.Heimlich, R.A.Houghton, F.Dong, 2153–2172.
A.Elobeid, J.Fabiosa, S.Tokgoz, D.Hayes, and T.-H.Yu, 2008. 22. R.Dones, T.Heck, M.F.Emmenegger, and N.Jungbluth, 2005.
& Engineering Centre, Ekaterinburg, Russia Use of US croplands for biofuels increases greenhouse Life cycle inventories for the nuclear and natural gas energy
Patrick Vieth, Senior Vice President – Integrity & Materials, gases through emissions from land use change. Science, 319, systems, and examples of uncertainty analysis: EcoInvent:
CC Technologies, Dublin, OH, USA 1238–1240. Energy Supply. Int.J. of Life Cycle Assessment, 10, 1, 10–23.
6. EGallagher, ed., 2008. The Gallagher Review of the indirect
Dr Joe Zhou, Technology Leader, TransCanada PipeLines Ltd, Calgary, Canada effects of biofuels production. Renewable Fuels Agency. Avail-
Dr Xian-Kui Zhu, Senior Research Scientist, Battelle Pipeline Technology Center, able at www.renewablefuelsagency.gov.uk/reportsandpublica-
Columbus, OH, USA tions/reviewoftheindirecteffectsofbiofuels.
❖❖❖

JPE CO2 Transportation by Pipeline Special Issue

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

JPE CO2 Transportation by Pipeline Special Issue

Hochgeladen von

Copyright:

Verfügbare Formate

December, 2010 Vol.9, No.

Great Southern Press Clarion Technical Publishers

Volume 9, No 4 • Fourth Quarter, 2011

James Watt ................................................................................................................................................................213

Graeme G King and Satish Kumar ....................................................................................................................... 253

Dr Robert Andrews, Dr Jane Haswell, and Russell Cooper ................................................................................ 275

Dr Tim Cockerill, Dr Naser Odeh, and Scott Laczay . ......................................................................................... 285

Guest editorial: CO2 transportation by pipeline – a special issue

T HIS SPECIAL ISSUE of the Journal of Pipeline

Carbon dioxide transport

AMEC Power and Process, Europe, Darlington, UK

Fig.1. North American CO2 pipelines

Capacity Length Pressure Year

Fig.3. Prescribed standards and codes

Fig.4. ISO 13623 petroleum and

Fig.5. BS PD 8010 steel pipelines

Fig.6. CO2 pipeline incidents

Fig.7.Typical orifice meter.

CO2 >95% >95%

<4% (for non-

H2O 0 <100ppm <500ppm

H2S 0 0.01-0.6% 0 <1450ppmv <200ppmv

Table 2. Examples of carbon dioxide stream composition.

Fig.8.Typical vortex meter and

1 Sui Northern Gas Pipelines Ltd, Pakistan

T HERE IS CURRENTLY interest in the techno-

Marriott Westchase Hotel

The international gathering of the global pigging industry!

Brief UK literature overview

Nearest UK offshore Oil & Gas

Bacton gas terminal

Table 2. UK existing onshore oil and gas terminals.

CO2 Emissions Range

Table 3. Classification of emitters according to emissions.

Table 4. Definition of pipeline design scenarios.

Selection of CO2 sources

Fig.2. Analytic approach used in

Fig.3. Approach to developing CO2

optimized to keep the calculated pressures of each source

For compression-based scenarios, pipeline diameters of the Main network characteristics

Emission Sources: Tier-0 Tier- 0 + 1 Tier- 0 + 1 + 2

1- Bacton Gas Terminal Mtpa 34 34 71 58 75 75

2-Easington Gas Terminal Mtpa 46 46 55 55 56 56

Table 6. Summary of main technical results for pipeline networks.

pipeline diameters must be used to meet the pressure

Transmission network capital costs

A summary of total costs required to establish the CO2

Considering the overall CO2 network costs (Table 7

Fig.6. Marginal tier costs, defined as

Fig.7. Range of transmission costs

Analysis Sensitivity analysis

For this study, sensitivities have been calculated to determine

Pipeline only based scenarios

Captured Capital Cost CO2 Transportation Cost

Pipeline + compression based scenarios

Captured Capital Cost CO2 Transportation Cost

T HE LITERATURE ON schemes to capture and store

1. With CCS defined in reference to managing GHG implies that re-injecting

Training courses – 2011

FEB 2011 February 1–

MAR 2011 March 17–18 Microbiological Corrosion in Pipelines (Houston)

APR 2011 April 25–29 Deepwater Riser Engineering Course (Houston)

Examination of some major governmental reports tends to

This question is simply answered in terms of three important