Journal of Pipeline Engineering

March, 2009

Vol.8, No.1

Journal of Pipeline Engineering

SA

M

PL

E

C O

PY

incorporating The Journal of Pipeline Integrity

Scientific Surveys Ltd, UK

Clarion Technical Publishers, USA

Journal of Pipeline Engineering Editorial Board - 2009

SA

M

PL

E

C O

PY

Obiechina Akpachiogu, Cost Engineering Coordinator, Addax Petroleum Development Nigeria, Lagos, Nigeria Mohd Nazmi Ali Napiah, Pipeline Engineer, Petronas Gas, Segamat, Malaysia Dr Michael Beller, NDT Systems & Services AG, Stutensee, Germany Jorge Bonnetto, Operations Vice President, TGS, Buenos Aires, Argentina Mauricio Chequer, Tuboscope Pipeline Services, Mexico City, Mexico Dr Andrew Cosham, Atkins Boreas, Newcastle upon Tyne, UK Prof. Rudi Denys, Universiteit Gent – Laboratory Soete, Gent, Belgium Leigh Fletcher, MIAB Technology Pty Ltd, Bright, Australia Roger Gomez Boland, Sub-Gerente Control, Transierra SA, Santa Cruz de la Sierra, Bolivia Daniel Hamburger, Pipeline Maintenance Manager, El Paso Eastern Pipelines, Birmingham, AL, USA Prof. Phil Hopkins, Executive Director, Penspen Ltd, Newcastle upon Tyne, UK Michael Istre, Engineering Supervisor, Project Consulting Services, Houston, TX, USA Dr Shawn Kenny, Memorial University of Newfoundland – Faculty of Engineering and Applied Science, St John’s, Canada Dr Gerhard Knauf, Mannesmann Forschungsinstitut GmbH, Duisburg, Germany Lino Moreira, General Manager – Development and Technology Innovation, Petrobras Transporte SA, Rio de Janeiro, Brazil Prof. Andrew Palmer, Dept of Civil Engineering – National University of Singapore, Singapore Prof. Dimitri Pavlou, Professor of Mechanical Engineering, Technological Institute of Halkida , Halkida, Greece Dr Julia Race, School of Marine Sciences – University of Newcastle, Newcastle upon Tyne, UK Dr John Smart, John Smart & Associates, Houston, TX, USA Jan Spiekhout, NV Nederlandse Gasunie, Groningen, Netherlands Dr Nobuhisa Suzuki, JFE R&D Corporation, Kawasaki, Japan Prof. Sviatoslav Timashev, Russian Academy of Sciences – Science & Engineering Centre, Ekaterinburg, Russia Patrick Vieth, Senior Vice President – Integrity & Materials, CC Technologies, Dublin, OH, USA Dr Joe Zhou, Technology Leader, TransCanada PipeLines Ltd, Calgary, Canada Dr Xian-Kui Zhu, Senior Research Scientist, Battelle Pipeline Technology Center, Columbus, OH, USA

❖❖❖

1st Quarter, 2009

1

The Journal of Pipeline Engineering incorporating

The Journal of Pipeline Integrity Volume 8, No 1 • First Quarter, 2009

PY

Contents

C O

Peter Tuft .................................................................................................................................................................... 5 The Australian approach to pipeline safety management

PL

E

Dr Érika S M Nicoletti and Ricardo Dias de Souza ............................................................................................... 19 A practical approach in pipeline corrosion modelling: Part 1 – Long-term integrity forecasting

M

Dr John Beavers, Patrick Vieth, and Dr Narasi Sridhar ....................................................................................... 29 Ethanol transportation: status of research, and integrity management

SA

Dr Chris Alexander .................................................................................................................................................. 35 Evaluating damage to on- and offshore pipelines using data acquired using ILI Professor Andrew Palmer and Dr Yue Qianjin ..................................................................................................... 49 Rethinking laybarge pipelaying H S Costa-Mattos, J M L Reis, R F Sampaio, and V A Perrut ............................................................................... 53 Rehabilitation of corroded steel pipelines with epoxy repair systems Assadollah Maleknejad ............................................................................................................................................. 63 Technical and commercial challenges in procurement and implementation of major international pipeline projects

❖❖❖

As part of an American Petroleum Institute study, experimental efforts were undertaken to assess the effects of wrinkle bends on the fatigue life of pipelines, and three 36-in x 0.281-in pipes were fitted with wrinkle bends having nominal depths of 2%, 4%, and 6% (wrinkle depth percentage calculated by dividing wrinkle depth by the nominal diameter of the pipe). OUR COVER PICTURE shows the pipe sample with 2% wrinkles, and details of this research are included in the paper by Dr Alexander on pages 35-47.

2

The Journal of Pipeline Engineering

T

HE Journal of Pipeline Engineering (incorporating the Journal of Pipeline Integrity) is an independent, international, quarterly journal, devoted to the subject of promoting the science of pipeline engineering – and maintaining and improving pipeline integrity – for oil, gas, and products pipelines. The editorial content is original papers on all aspects of the subject. Papers sent to the Journal should not be submitted elsewhere while under editorial consideration. Authors wishing to submit papers should send them to the Editor, The Journal of Pipeline Engineering, PO Box 21, Beaconsfield, HP9 1NS, UK or to Clarion Technical Publishers, 3401 Louisiana, Suite 255, Houston, TX 77002, USA. Instructions for authors are available on request: please contact the Editor at the address given below. All contributions will be reviewed for technical content and general presentation. The Journal of Pipeline Engineering aims to publish papers of quality within six months of manuscript acceptance.

Notes 4. Back issues: Single issues from current and past volumes (and recent issues of the Journal of Pipeline Integrity) are available for US$87.50 per copy.

SA

M

PL

3. Information for subscribers: The Journal of Pipeline Engineering (incorporating the Journal of Pipeline Integrity) is published four times each year. The subscription price for 2009 is US$350 per year (inc. airmail postage). Members of the Professional Institute of Pipeline Engineers can subscribe for the special rate of US$175/year (inc. airmail postage). Subscribers receive free on-line access to all issues of the Journal during the period of their subscription.

5. Publisher: The Journal of Pipeline Engineering is published by Scientific Surveys Ltd (UK) and Clarion Technical Publishers (USA):

C O

E

2. Copyright and photocopying: © 2009 Scientific Surveys Ltd and Clarion Technical Publishers. All rights reserved. No part of this publication may be reproduced, stored or transmitted in any form or by any means without the prior permission in writing from the copyright holder. Authorization to photocopy items for internal and personal use is granted by the copyright holder for libraries and other users registered with their local reproduction rights organization. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising and promotional purposes, for creating new collective works, or for resale. Special requests should be addressed to Scientific Surveys Ltd, PO Box 21, Beaconsfield HP9 1NS, UK, email: [email protected].

PY

1. Disclaimer: While every effort is made to check the accuracy of the contributions published in The Journal of Pipeline Engineering, Scientific Surveys Ltd and Clarion Technical Publishers do not accept responsibility for the views expressed which, although made in good faith, are those of the authors alone.

Scientific Surveys Ltd, PO Box 21, Beaconsfield HP9 1NS, UK tel: +44 (0)1494 675139 fax: +44 (0)1494 670155 email: [email protected] web: www.j-pipe-eng.com www.pipemag.com Editor and publisher: John Tiratsoo email: [email protected] Clarion Technical Publishers, 3401 Louisiana, Suite 255, Houston TX 77002, USA tel: +1 713 521 5929 fax: +1 713 521 9255 web: www.clarion.org Associate publisher: BJ Lowe email: [email protected]

6. ISSN 1753 2116

vvv

www.j-pipe-eng.com went live on 1 September 2008

1st Quarter, 2009

3

Editorial

Pipelines as a source of conflict

“R

Chapter 2 starts with a brief history of the many transit pipelines which have been associated with very negative experiences. In the past, they included those operating in the Middle East; more recently, attention has been focused on those in the former Soviet Union. The chapter then describes lines which can be viewed either as success stories or as having too recent a history for the outcome to be determined. This history helps in identifying which characteristics make for ‘good’ and ‘bad’ transit countries. These include:

SA

M

PL

E

ECENT EVENTS between Russia and Ukraine at the start of 2009, and Russia and Georgia in 2008, have brought transit pipelines back into the media spotlight. Any reading of the history of transit oil and gas pipelines suggests a tendency to produce conflict and disagreement, often resulting in the cessation of throughput, sometimes for a short period and sometimes for longer. It is tempting to attribute this to bad political relations between neighbours. This is certainly part of the story, but also important is the nature of the ‘transit terms’ – tariffs and offtake terms – whereby transit countries are rewarded for allowing transit. Put simply, the trouble with transit pipelines has a significant economic basis. The report addresses three questions:

a greater share. Even though this would apply to any commercial transaction, the key difference with transit pipelines is that there is no overarching jurisdiction. More transit pipelines will be needed in the future, since oil and gas reserves close to market are being depleted, and there is growing demand for natural gas in the world’s primary energy mix. In recent years, there has been a noticeable fragmentation of legal jurisdictions as the Soviet Union and former Yugoslavia both collapsed. Many of the new transit pipeline projects being discussed are essentially the result of gaming strategies between the various players and will fail to materialize.

PY

HE NEWLY-published report from Chatham House entitled Transit troubles – pipelines as a source of conflict* raises a number of interesting and important issues, and is worth studying in detail, and can be downloaded from the reference below. We are pleased to have the agreement of the report’s author, Professor Paul Stevens, to publish the report’s summary; a brief biography for him appears at the end of this article, from which it will be seen that he is eminently well-placed to be a commentator in this area of the pipeline industry, where engineering and politics either meet or clash, depending on the viewpoint.

C O

T

• Why will oil and gas transit pipelines become more important to global energy markets in the future? • Why has the history of such pipelines been littered with conflict between the various parties? • What might be done to improve this record in the future and make transit pipelines less troublesome? Chapter 1 defines transit pipelines as lines which cross another’s ‘sovereign’ territory to get the oil or gas to market. Such lines have a number of relevant, common characteristics which tend to generate conflict. Different parties are involved, each with different interests and motivations. This invites disagreement between the parties because of the benefits to be shared and the fact that mechanisms exist toï: encourage one or other party to seek *Transit troubles: pipelines as a source of conflict. Prof. Paul Stevens, 2009. A Chatham House Report – see www.chathamhouse.org.uk.

• the importance of foreign direct investment in the transit country’s development strategy; • the importance of the transit fee in the country’s macro economy; • the dependence upon offtake from the line; • the availability of alternative routes; • whether the transit country is also an oil or gas exporter in its own right. Chapter 3 seeks explanations for poor performance in terms of politics but with the main discussion focusing on the underlying economics which generate conflict. One obvious source of political disputes is a history of bad relations between neighbouring countries. As for the economics, the key explanation is that there is no reasonable, objective basis for determining ‘transit terms’. The only sensible reason for the existence of a transit fee is to allow the transit country to share in the benefits of the project. This share will reflect the relative bargaining power of the parties to the negotiations. Over time this changes and thus there are always pressures to change the transit terms. This

4

The Journal of Pipeline Engineering

lecturer in economics (1979–93); and as Professor of Petroleum Policy and Economics at the Centre for Energy, Petroleum and Mineral Law and Policy, University of Dundee (1993–2008) – a chair created by BP.

trend is greatly encouraged by the existence of the ‘obsolescing bargain’, the structure of pipeline costs, and the growing volatility of oil and gas prices. Chapter 4 considers possible solutions to help reduce conflict and supply disruptions. These include:

US companies explore ethanol pipeline through US Midwest

• a military solution; • encouraging the transit country into the global economy to make it dependent upon foreign direct investment; • making the transit country dependent upon its own gas and oil supplies from the pipeline, although this can be a double-edged sword; • considering alternatives to the transit country not only in terms of geographic routes but (for gas) the actual means of transport including, for example, the use of liquefied natural gas (LNG); • encouraging multilateral jurisdictional solutions such as the Energy Charter Treaty; • developing mutual dependence between the transit country and the producer/consumer country.

WO MAJOR US pipeline companies have announced their plans to assess the feasibility of constructing an ethanol pipeline through the Midwest. If built, the pipeline would the first one totally dedicated to transporting ethanol in the US. Oklahoma-based Magellan Midstream Partners and Pennsylvania-based Buckeye Partners have partnered to explore creating the 2720-km long pipeline to transport ethanol from plants in Illinois, Iowa, Minnesota, and South Dakota to major cities including Pittsburgh, Philadelphia, and New York. The project is estimated to cost more than $3bn.

SA

M

PL

• • • •

a clear definition and acceptance of the rules; projects driven by commercial considerations; credible threats to deter the ‘obsolescing bargain’; mechanisms to create a balance of interest.

However, it is difficult to turn this ‘wish list’ into a practical agenda. The only practical, realistic solution in the near term is to introduce ‘progressive’ transit terms to existing and new agreements. However, ultimately both consumers and producers must diversify as far as is economically practical.

”

Professor Paul Stevens is Senior Research Fellow for Energy at Chatham House, Emeritus Professor at Dundee University and Consulting Professor at Stanford University. He was educated as an economist and as a specialist on the Middle East at Cambridge University and the School of Oriental and African Studies, London. He taught at the American University of Beirut in Lebanon (1973–79), interspersed with two years as an oil consultant; at the University of Surrey as lecturer and senior

PY

The American Coalition for Ethanol’s 2007 report lists Illinois as the second largest producer of ethanol in the US, at 317m gall/yr, and corn grown in Illinois is used to produce 40% of the ethanol consumed in the US, according to the Illinois Corn Growers Association (ICGA). Nearly one-third of all gasoline in the US already contains low levels of ethanol – usually between 5.7% and 10%, and the ICGA reports that 95% of the gasoline sold in the Chicago area contains 10% ethanol. However, high levels of ethanol cannot be piped through existing gasoline lines without damaging them. Once ethanol has been transported through existing pipelines, they can’t be shared with other refined products. “In pipelines today, you can ship different materials through in batches, with plugs that separate the shipments. However, ethanol – because it absorbs water, and is a corrosive agent – is really difficult to use in a nondedicated pipeline,” John Urbanchuk, the director of expert-resources firm LECG, said.

C O

E

Finally, the report considers a new solution: basing the ‘transit terms’ on a progressive fiscal arrangement similar to the sort of systems which govern upstream oil agreements. The report concludes that there will be an increasing need for and dependence upon oil and gas transit pipelines but such pipelines are inherently unstable because of political disputes and also, of equal importance, as a result of commercial disputes over the transit terms. These commercial disputes arise because there is no objective, reasonable or fair way of setting the transit terms. Many of the apparent solutions to this problem are, on closer examination, at best ineffective, at least in current circumstances. More generally, history suggests that a good experience with transit pipelines requires certain bestpractice conditions to be met. These include:

T

Magellan and Buckeye may be years away from construction, simply because not much is known about transporting ethanol through pipelines. Studies on the technical issues and economic impact of creating an ethanol pipeline are continuing, as highlighted in Dr John Beaver et al.’s paper on pages 29-34; no ethanol pipelines exist in the US, though Brazil is in the process of constructing one and Houston-based Kinder Morgan is understood to have announced plans to test an ethanol pipeline in Florida this year. Changing the way ethanol is transported may have more of an effect on consumer costs than adopting alternate fuels or even falling oil prices. “If you looked at something in Illinois or maybe Iowa, sending it to the East Coast by freight is anywhere between 16 and 18 cents a gallon. If you concluded on p61

1st Quarter, 2009

5

The Australian approach to pipeline safety management by Peter Tuft Peter Tuft & Associates, West Pymble, NSW, Australia

T

HE AUSTRALIAN APPROACH to management of pipeline safety and risk differs from that used in most other parts of the world: there is a strong focus on identifying causes of failure and designing against them using a cause/control model of risk management, and little use of quantitative risk assessment.

PY

Oil and gas pipelines in Australia are designed, constructed, and operated in accordance with AS 2885. Since a major revision in 1997, this has been a risk-based standard. While it does contain numerous design rules, their application is flexible and to some extent dependent on the outcomes of a mandatory safety management study. Key elements of the standard include separation of wall thickness selection from pressure design factor, mandatory protection against external interference, special requirements for highconsequence areas, and a safety management study process including qualitative assessment of residual risks.

SA

M

PL

E

C O

The AS 2885 process has been shown to be workable and effective :. It results in a design which is optimized for safety at every point along the pipeline while not incurring costs for features that do not reduce risk. The process is oriented principally to design of new pipelines, but is equally applicable to management of older pipelines which are suffering degradation or subject to changed conditions such as urban encroachment.

The Australian pipeline industry Transmission pipelines in Australia are often long but of relatively-small diameter (maximum 34in, typically 12-18in). They traverse vast lengths of remote and sparsely-populated country, but there are also substantial lengths in semi-rural areas and urban outskirts, and some within urban areas. There is a growing problem of urban encroachment on pipelines originally constructed in rural locations. Most Australian pipelines are relatively young (80% built since 1975) and therefore in reasonably good condition as a result of being designed and built to modern practices and with modern coatings, as well as having had limited time to deteriorate.

Author’s contact details: tel: +61 2 9983 1511 email: [email protected] This paper was presented as part of the proceedings of the 7th International Pipeline Conference – IPC 2008 – held in Calgary on 29 September – 3 October, 2008, and organized by the ASME’s Pipeline Systems Division. It is published here by kind permission of ASME.

The Australian pipeline industry is relatively small by global standards. The total length of high pressure transmission pipelines is just under 30,000km, and there are only a handful of major pipeline operating companies. Nevertheless the industry is quite large enough to be vigorous and to support a healthy population of specialist pipeline engineers. Some of the larger Australian pipeline construction and engineering service companies have successful export businesses with projects in diverse locations around the world. The Australian Pipeline Industry Association (APIA) sponsors an active research programme and has a cooperative research agreement with PRCI in North America and EPRG in Europe; the most recent tripartite Joint Technical Meeting was held in Canberra in 2007. Also well supported by APIA is Standards Australia Committee ME38, responsible for AS 2885. This committee has been active in developing standards for pipeline design/ construction, welding, operation, and pressure testing. The committee and its working groups include representatives from all sectors of the industry as well the technical regulators from each state, and has been responsive

6

The Journal of Pipeline Engineering

C O

satisfactorily controlled by any means that are appropriate rather than by application of a narrow set of fixed rules.

PL

E

in meeting the needs of both groups. The industry and regulatory representatives have a very co-operative approach and opinions diverge only on peripheral issues. The APIA research programme includes a number of projects initiated in response to the needs of the ME38 committee and the research outcomes are incorporated in new revisions of standards.

PY

Fig.1. The Australian pipeline network.

SA

M

The considerable distances and small loads in Australia create economic pressure to minimize pipeline costs, which provides a driver for technical innovation. The unified approach of the APIA and the ME38 committee provides the means by which innovation can be relatively quickly incorporated into standards and applied to new and existing pipelines.

Basis for AS 2885 Prior to 1997, AS 2885 and its precedent standards had been developed from the ASME/ANSI B31.4 and B31.8 codes, although considerable differences from those codes had evolved over time. In preparation for the 1997 revision, the code committee recognized that, despite the best intentions, a rigid rule-based code would often produce designs that were less than optimum in terms of safety, economics, or both. There was particular concern about anomalies that arose from the rule-based approach at boundaries between different location classes (reflecting population density, often defined in a very arbitrary way), and also with the way that the rules handled changes in population density as a result of urban growth. AS 2885, of course, still includes many rules. However, they are more flexible than previously, and the overriding requirement is to assess risks and ensure that they are

A fundamental aspect of the standard is the safety management study (SMS), described in more detail later in this paper. Virtually all aspects of the design must be reviewed through the SMS. While this may appear onerous, it is the route to flexibility in the application of rules so that the design can be optimized for safety. The SMS includes a qualitative risk review process, with the objective of identifying threats which may cause failure and ensuring that they are managed so that the residual risk is tolerable. The intention is that safety and risk management should be done by the engineers responsible for the design and operation of the pipeline, rather than being outsourced to risk specialists. Pipeline engineering and risk management should be integrated, and a corollary of this is that engineering and risk management form an iterative process; the design and operating procedures affect the risk profile, and treatment of risk feeds back to the design and procedures. Since pipeline design and operation are generally not complex processes, it is eminently sensible that this loop be contained entirely within the small team responsible for pipeline engineering. Risk specialists have an occasional role in providing technical analysis of the consequences of a pipeline failure (for unusual cases where the standardised approach is not applicable), and also for those few cases where quantitative risk assessment may be required. There is little use of quantitative risk assessment (QRA) in the analysis of Australian pipelines. Attempts have been made to use statistically-based QRA, but for such methods to produce realistic results they must be based on defensible

1st Quarter, 2009

7

0.7

LOC = Loss of Containment

Incident rate, per 1000 km-yr

0.6

0.5

0.4

0.6

Australian data for 86- 03 Europe data for '97-'01 USA data for '98-'02 Europe & US data from W. Guijt, O&GJ, Jan 26 '04 0.3

0.3 0.22 0.2 0.13

0.12

0.1 0.015 0 Europe Gas LOC

Europe Oil LOC

USA Gas LOC

USA Oil LOC

(This is not because the external damage rate is unusually high, but because the relatively-young age of Australian pipelines means that to date they have experienced only a few corrosion-related failures.) Also, Australian pipelines tend to be thin-walled because of the relatively-small diameters and high-grade steels used, and this makes them more vulnerable to loss of containment should serious external damage occur. For these reasons AS 2885 places considerable emphasis on external interference protection (EIP). There are mandatory requirements for both physical and procedural protective measures, and these must be appropriate to the level of threat that is identified:

PL

E

failure rates. The Australian pipeline incident record is too sparse for meaningful data to be extracted. There are less than 200 recorded pipeline incidents in the APIA database, and only a small fraction of these involve loss of containment. This is far too little from which to develop average failure rates which account for the range of parameters that can affect pipeline failure (wall thickness, depth of cover, and location class, to nominate just three that are most critical). Some QRA studies have been done based on UK and/or European statistical failure-rate data, notwithstanding that that the pipeline may be in the remote outback, and the resulting predicted failure rates have been one or two orders of magnitude higher than the overall average Australian failure rate (which includes the higher rate of incidents from more populated areas).

Australia, LOC

PY

Australia, All Incidents

C O

Fig.2. Comparison of Australian and non-Australian data.

SA

M

Such misuse of QRA methods has done much to damage their credibility in the Australian pipeline industry, which is unfortunate because there are applications of quantitative methods that are valid and useful. In particular, modern reliability-based quantitative methods have considerable potential but have not yet been adopted. AS 2885 acknowledges that QRA potentially has a role in assisting the evaluation of risk-treatment alternatives, to permit comparison of the risk-reduction benefits of various options. Statistical QRA may also have a role in assessing the risks associated with pipeline facilities which comprise standard process plant components and can therefore call on the extensive process plant failure data. A tacit feature of the AS 2885 principles is that while pipeline safety is the overriding priority and cannot be compromised, there is also flexibility to avoid incurring costs that do not add any safety benefit. The optimization of both safety and cost is a recurring theme in this paper.

External interference protection Damage by external forces is a major contributor to pipeline incidents worldwide, but is particularly dominant in Australia where it accounts for at least 80% of all incidents.

“A pipeline shall be designed so that multiple independent physical controls and procedural controls are implemented to prevent failure from external interference by identified threats. “The purpose of physical controls is to prevent failure resulting from an identified external interference event by either physically preventing contact with the pipe, or by providing adequate resistance to penetration in the pipe itself. “The purpose of procedural controls is to minimise the likelihood of external interference activity, with potential to damage a pipeline, occurring without the knowledge of the pipeline operator, and to maximise the likelihood of people undertaking such activity being aware both of the presence of the pipeline and the possible consequences of damaging it.” (Clause 5.5.1) The standard requires that all practicable controls be applied, with a minimum of one physical measure in rural locations and two in urban locations, and always a minimum of two procedural measures (see Table 1). There is considerable detail in the standard on the minimum requirements for each type of control to be considered effective. The overall effectiveness of the external

8

The Journal of Pipeline Engineering

Physical controls

Separation

Burial Exclusion (e.g. fencing) Barrier (e.g. vehicle crash barrier for an above-ground pipe section)

Resistance to penetration

Wall thickness Penetration barrier (e.g. concrete slab or encasement)

Procedural controls

Pipeline awareness

Liaison with landowners and third parties (local government, utilities, etc) Community awareness program One-call service ("Dial Before You Dig") Pipeline marking (warning signs and buried marker tape) Agreements with other users of shared corridors

External interference detection

Patrolling Planning notification zones Remote intrusion detection

Table 1. Physical and procedural controls (from AS 2885).

interference protection design must be reviewed as part of the safety management study.

resulting relationships are quite adequate for estimating the maximum size of excavator capable of penetrating any given pipe, and these formulae have been incorporated in AS 2885.

C O

AS 2885 mandates penetration resistance as a physical protection measure in the higher location classes. It is optional in rural areas, but the standard expects that the penetration resistance calculations will always be done in order to provide reference data that can be used in the SMS for assessing failure mode and consequences.

Allowances (corrosion etc)

Manufacturing tolerance (seamless pipe only)

Constructability

A key feature of the design for penetration resistance is that it is divorced from the pressure design factor. Pipeline wall thickness has traditionally been based on a pressure design factor of 0.72 in remote areas, with progressively-lower design factors as population density increases. However for a pipeline of small diameter and low pressure rating even a

Others as per Clause 5.4.2

"No rupture"

Penetration resistance

Internal pressure

Required WT

Nominal WT

SA

M

APIA sponsored research to determine the relationships between penetration force, pipe properties (grade, wall thickness), and excavator parameters (tooth dimensions, bucket force, excavator mass). (Previous work had been done by others, particularly in Europe, but was not directly applicable to typical thin-walled Australian pipelines.) It was found that for a given pipe and tool dimensions there is excellent agreement between experimental and finiteelement results for the force required to penetrate, but of course some variability enters the relationship between machine size and bucket force capability. Nevertheless the

Because there is huge uncertainty about the actual impact conditions the equations include an empirical parameter based on limited full-scale field trials. Adjustment of the parameter permits calculation of an upper-bound value (penetration quite likely) and lower-bound value (penetration not credible) for the size of excavator that may result in puncture.

Hydrostatic testing

PL

E

The last line of defence against loss of containment caused by external damage is the resistance of the pipe itself to penetration, and for this reason AS 2885 gives considerable attention to penetration resistance as a physical protection measure. Emphasis to date has been on resistance to excavators, given both their ubiquity and the state of knowledge, but it is recognized that other equipment – such as boring rigs – can also pose a significant threat.

PY

Penetration resistance

Fig.3. An illustration of a situation where penetration resistance is the governing influence.

1st Quarter, 2009

9

R1

Broad rural - rural or undeveloped land with only isolated dwellings

R2

Rural residential - single-residence blocks of area typically 1 - 5 ha

T1

Residential - suburban and associated infrastructure such as small shopping centres

T2

High density - predominantly multi-storey or large numbers of people present (e.g. major retail centres)

Table 2 (above). Primary location classes (from AS 2885). Table 3 (below). Secondary location classes (from AS 2885). Sensitive - schools, hospitals, jails, etc

Equivalent to T2

I

I ndustrial - light industry, car sales yards, etc

Equivalent to T1

HI

Heavy I ndustrial - heavy or hazardous industries

Varies from R2 to T2, latter if pipeline incident may escalate

CIC

Common I nfrastructure Corridor shared with other services such as transport and buried or overhead utilities

Particular attention to liaison and agreements with other parties in corridor

S

Submerged - water crossings and flood plains

Special design

C O

overpressure are addressed explicitly through consideration of the other factors influencing wall thickness. Hence in principle it is acceptable to operate a pipeline at 72% or 80% SMYS in an urban area if the wall thickness can meet all the other requirements without any increase above that for pressure containment. The SMS provides a thorough review of these issues before a design is finalized

Wall thickness

SA

M

PL

E

design factor as low as 0.4 will yield a wall thickness that can be penetrated by a backhoe. Conversely, at the maximum design factor of 0.72 a pipeline of large diameter and high pressure rating will have a wall thickness that cannot be penetrated by a machine of any size, so imposing a low design factor adds very considerable cost without any significant improvement in the risk of failure due to external interference. By separating the design for internal pressure from the design for penetration resistance, the overall pipeline can be optimized for both safety and cost.

PY

S

AS 2885 explicitly de-couples wall thickness determination from the traditional location class/design factor formula. It specifies that the required in-service wall thickness at each location along the pipeline shall be the greatest of the thicknesses required by whichever of the following factors are applicable at that location: • • • • • • • • • •

pressure containment penetration resistance “no rupture” (discussed later) other stress and strain criteria control of fast-running fracture special construction (such as bridges) vehicle loads at road and rail crossings mitigation of stress-corrosion cracking fatigue life external pressure

The design factor for pressure containment is independent of the location classification. Failure modes other than

Figure 3 (simplified from the Standard) illustrates one example of how this approach is applied in a case where penetration resistance happens to be the governing influence. The intent of this approach is again the principle that the design can be optimized for both safety and cost at each point along the pipeline route.

Location classification Virtually all pipeline codes use some concept of location classification to identify areas where the risks both to and from a pipeline are increased by higher population density. AS 2885 is no different, but has refined the concept in two ways and also applies it quite differently. Firstly, location classification is based on the area that would be seriously affected by an ignited full-bore rupture. Location classes are determined from the land use (as a proxy for population density) within a radiation contour of 4.7kW/m2 (1500BTU/hr.ft2). This is the generally-accepted radiation level at which an unprotected person will suffer second degree burns after 30s exposure. It will vary with the diameter and MAOP of the pipeline and, in principle, could be calculated for each pipeline on the basis of release rate and flame radiation correlations. However for most gas

10

The Journal of Pipeline Engineering

pipelines the calculation can be standardized to the extent that it can be simply read from graphs provided in the standard. (In fact for gas pipelines with the common MAOP of 10.2MPa there is an even-simpler rule of thumb: the 4.7kW/m2 radiation contour in metres is equal to the pipe diameter in millimetres – 300m for a DN 300 pipeline.)

For existing pipelines there are other requirements to be applied when the location class changes as a result of urban development.

No rupture The no-rupture requirement can be achieved by either of two means. Firstly, the hoop stress may be limited to less than 30% SMYS (the approximate level at which there is insufficient elastic energy in the pipe for any defect to propagate); for some pipelines this may lead to uneconomically-large wall thickness. Alternatively, through the SMS, the largest credible threat to the pipeline must be identified, the resulting maximum defect length determined, and the linepipe selected so that the critical defect length (above which the pipe will rupture) is at least 150% of this maximum hole size. Detailed guidance is provided for the calculations.

Basing location classification on the actual worst-case radiation damage permits the pipeline design to be realistically optimized for both safety and cost. The use of the full-bore rupture radiation distance still applies even for a “no rupture” design (see below), because it is the possibility of serious impacts on surrounding people that gives rise to the “no rupture” requirement in the first place.

SA

M

PL

E

As has already been made clear, AS 2885 does not directly link wall thickness to location class. Location class is instead used to adjust certain requirements of the safetymanagement system. In particular there are higher demands for external interference protection in higher location classes, and special requirements for high-consequence areas.

C O

A second refinement of the location classification system is the addition of five secondary location classes to highlight special features that may not be adequately identified by the primary location classification, as shown in Table 3.

High-consequence areas

AS 2885 uses the concept of high-consequence areas, although the definition and approach differ from those in the USA. A high-consequence area is formally defined as “a location where pipeline failure can be expected to result in multiple fatalities or significant environmental damage”. In practical terms this includes (but is not necessarily limited to) residential, high-density, sensitive, and industrial location classes. For a new pipeline there are two requirements that must be met in order to limit the consequences of any failure (Clause 4.7): • No rupture – “The pipeline shall be designed such that rupture is not a credible failure mode.” • Maximum discharge rate – “... the maximum discharge rate shall not exceed 10GJ/s in residential, industrial and sensitive locations, or 1GJ/s in highdensity locations.” (This brief extract omits other qualifying requirements.)

For example, a pipeline of DN 450 (18in NB) and 6.8mm (0.268in) wall thickness in X70 steel operating at 10.2MPa (1480psi) has a critical defect length of 64mm (hoop stress 72% SMYS). In a suburban area it is plausible to expect that the largest excavation machinery would not exceed 30t, and such a machine fitted with sharply-pointed penetration teeth is capable of penetrating this pipe. The resulting hole from the penetration tooth would be around 70mm long, which exceeds the critical defect length, and thus rupture would be possible. Hence this pipe is not acceptable in a high-consequence area. If the wall thickness is increased to 9.8mm, the same machine can no longer penetrate at all, so “no rupture” is achieved although the hoop stress is still well above 30% SMYS. Clearly this design process depends on the threats that apply to the particular pipeline and the conclusion from this example is not generally applicable (for example, there has been no consideration here of the threats posed by boring machines).

PY

AS 2885 defines four primary location classes, summarized in Table 2 (using very abbreviated definitions). There is an inevitable element of subjectivity in the allocation of location class in borderline areas. However this matters little, given the nature of the SMS process outlined later and the flexible approach to achieving an adequate level of safety.

Discharge rate limit The limitations on discharge rate were derived from generic QRA studies for suburban and high-density areas. These studies determined the magnitude of the largest ignited gas release that would fall within tolerable criteria for societal risk. The rate of discharge from a punctured pipeline depends mainly on the size of hole and the operating pressure, so these are the only parameters that the design engineer can adjust in order to comply with the limits. Even the scope for adjustment of wall thickness is quite constrained because penetration by excavators is largely a binary outcome – a given machine will either penetrate or it won’t, and if it does penetrate then, to a first approximation, the hole size is unaffected by the wall thickness. So the options are to either: • increase wall thickness until penetration by the largest identified threat is not possible, or

1st Quarter, 2009

11

Table 4. Hole sizes corresponding to discharge rates for two operating pressures

10.2MPa (1450psi)

2.7MPa (400psi)

1GJ/s

45mm

90mm

10GJ/s

145mm

280mm*

*Effectively full-bore rupture for many pipelines

disproportionate to the benefit gained from the reduced risk that could result from implementing any of the alternatives.” (Clause 4.7.4, emphasis added.)

• reduce the maximum allowable operating pressure until the discharge rate from the largest credible hole is within the specified limit.

Safety management study process

PL

E

Clearly the “no-rupture” and discharge-rate limitations cannot easily be applied retrospectively to existing pipelines, particularly those which are affected by urban encroachment. However in order to maintain consistency and integrity in the approach to pipeline safety management, the committee revising AS 2885 felt it was necessary to introduce a requirement that goes as far as possible towards achieving equivalent results.

PY

Change of location class

The list of alternatives to be considered indicates that the assessment of risk level and the determination of ALARP is to be taken very seriously. This is one situation where QRA studies may be of some value in helping with comparison of the alternatives; even if the absolute values of the quantitative risk predictions are questionable, there may be much use in their comparative rankings, to be assessed alongside the cost of each alternative.

C O

Table 4 shows, as an indication, the hole sizes corresponding to the specified discharge rates for two illustrative operating pressures.

M

These requirements for changed location class are best summarized by quoting almost in full:

SA

“Where land use ... changes along the route of existing pipelines to permit ... [high consequence areas] in areas where these uses were previously prohibited, ... [it] shall be demonstrated that the risk from a loss of containment involving rupture is ALARP [As Low As Reasonably Practicable]. “This assessment shall include analysis of at least the alternatives of the following: (a) MAOP reduction (to a level where rupture is non-credible). (b) Pipe replacement (with no rupture pipe). (c) Pipeline relocation (to a location where the consequence is eliminated). (d) Modification of land use (to separate the people from the pipeline). (e) Implementing physical and procedural protection measures that are effective in controlling threats capable of causing rupture of the pipeline. “For the selected solution, the assessment shall demonstrate that the cost of the risk reduction measures provided by alternative solutions is grossly

As noted previously, a formal safety management study is a fundamental requirement for any pipeline designed to AS 2885. Overall pipeline safety review is essentially a two-step process: • Design review: identify every potential threat to the integrity of the pipeline, and if possible apply controls so that “failure as a result of that threat has been removed for all practical purposes”. • Risk assessment: rank any remaining threats that are not fully mitigated, and ensure that the residual risk is reduced to a tolerable level. The intention, and general experience, is that the vast majority of threats are eliminated by application of controls at the design review stage and only a small number progress to risk assessment. While the SMS process is defined in terms of pipeline design, it is equally applicable (and mandated) for regular review of existing pipelines.

Threat identification A threat is “any activity or condition that can adversely affect the pipeline if not adequately controlled”. Identifying threats is conceptually similar to a HAZOP, although not

12

The Journal of Pipeline Engineering

Identify Threats

Initial Design

DESIGN REVIEW

Review Controls Failure?

No

Yes

Consequences & Likelihood

RISK High Intermediate ASSESSMENT No ALARP?

Low Yes

Accepted Design

office-based engineers may have guessed (although just sometimes the opposite occurs, which is an equallycompelling reason for gathering the data).

as formally structured. It helps to have very experienced people involved, and it also helps to use checklists. Most importantly, it helps to think laterally about anything that might go wrong. Even threats that are believed to be already mitigated should be included in the documentation, partly because it may lead to identification of weaknesses in the existing mitigation, and partly because it forms a sound basis for the safety and operating plan (which is mandated by AS 2855).

PY

For a new pipeline it is also vital to consider the design for future land use, and hence liaison with the local government or other planning authority is necessary.

PL

SA

M

It is fundamental that a threat exists at a location; sometimes that “location” may be the entire pipeline and the threat is considered to be non-location-specific (such as corrosion, some design defects, etc). However, the great majority of threats are associated with activities or events that occur at a particular location along the pipeline route. This may be a single point, such as threats associated with road maintenance at a road crossing, or may be more extended, such as threats associated with logging activities in a forest. Identifying external interference threats requires particular attention, including real data from the field. Field personnel involved in landowner liaison and pipeline patrol are invaluable aids, to the extent that an SMS that does not include their input is seriously devalued. Such people can provide details of the type and size of excavation machinery likely to be used at every point along the route (for consideration in the context of penetration resistance), and can often provide background information on other threat types as well. Gathering this information may appear onerous, but with adequate planning and support (such as a brief “land user survey” form) good field personnel can acquire it in the course of their ordinary duties. An additional benefit of gathering the real data is that the maximum equipment used is not uncommonly found to be rather smaller than

Threat control

C O

E

A checklist of potential threat types may include over 100 items, ranging from all sorts of external interference events, through a wide range of defects in design, materials, and construction, through to diverse mishaps involving corrosion, natural events, and operations and maintenance.

Fig.4. The SMS process.

Control of external interference threats has already been discussed. For other threats, appropriate controls must be put in place, and these may range from standard corrosioncontrol measures to quality-assurance procedures for design, manufacturing, and construction. In all cases, the key question to be asked is “are the controls sufficient to prevent failure as a result of the identified threat?” This may appear to be subjective, but there is usually a definitive answer if there is a clear understanding of the identified threat and the controls. (Of course, another threat that may require consideration is failure of the controls, but that can be addressed as a separate threat in its own right.) Once sufficient controls are in place, the threat is accepted and requires no further consideration, other than ensuring that the controls are documented and implemented. Threats which cannot be controlled by the application of external interference protection and other design measures become hazardous events which required risk assessment.

Risk assessment The risk-assessment phase involves qualitative estimation of the likelihood and consequences of failure leading to a ranking of risk on a scale of extreme, high, intermediate, low, or negligible. Extreme and high risks are intolerable and must be reduced (but they are also very uncommon if the pipeline is well designed in the first place). Intermediate risks are acceptable only if formally shown to be ALARP (discussed below).

1st Quarter, 2009

13

Risks are ranked according to a standard frequency/severity risk matrix (see Appendix 1), which includes considerable guidance on the ratings for likelihood and severity.

parties (including major landowners, developers, also possibly part-time), owner’s representatives, and the technical regulator or other government representatives.

Severity of a failure is most commonly assessed in terms of thermal radiation effects on people, assuming that any loss of containment will ignite. Given the basis for the location classification described previously it is usually straightforward to make this type of qualitative assessment. Consideration is also required for the effects of a failure on the environment and continuity of supply.

The SMS process is defined in Part 1 of AS 2885 (Design and Construction) but Part 3 (Operation) mandates that it be reviewed every five years, or more frequently if there is a change in circumstances surrounding the pipeline (such as a proposed development nearby). This means that for practical purposes the SMS remains “live” for the life of the pipeline. It is of course mandatory that all SMS deliberations be recorded in full, and because it is “live” it is desirable for the documentation to be readily updated.

• what else can we do to reduce risk? (adjust the route, for instance) • why haven’t we done it?

The SMS documentation forms part of the Safety and Operating Plan that is mandated by AS 2885 Part 3, as is entirely appropriate since risk management tacitly or explicitly underlies almost all pipeline operations and procedures, other than those involved in scheduling and commercial metering of the pipeline contents.

E

ALARP is achieved when either the answer to the first question is “nothing” or the answer to the second is “because the cost is grossly disproportionate”.

A database is the preferred means of recording the threats, controls, risk evaluation, and risk treatments. An appropriately structured database can also record and close-out various corrective actions that arise during the process. Attempts are sometimes made to use a basic spreadsheet but, except in the simplest cases, this rapidly becomes unwieldy because of the large quantity of information and explanatory comment that must be recorded.

PY

The concept of ALARP is the basis for determining whether a risk ranked intermediate can be tolerated. “ALARP means the cost of further risk reduction measures is grossly disproportionate to the benefit gained from the reduced risk that would result.” (definition, Clause 1.5.3). In practical terms ALARP can be assessed by asking:

C O

As low as reasonably practicable (ALARP)

PL

SMS implementation

M

The phases of an SMS are:

SA

• initial design (for new pipelines only) • data gathering, discussed under threat identification (above) • desktop design review (pre-analysis for workshop) • validation workshop(s) involving all stakeholders The workshop is mandated by AS 2885; for a major project, the workshop may take a week. A workshop on a single major encroachment problem for an existing pipeline may occupy a full day. The great value of a workshop is that it generates synergies from the interaction of a diverse group of stakeholders, identifying both threats and solutions that would not be apparent to an individual working alone. It also provides consensus and buy-in from all participants. The value of a workshop is demonstrated by the observation that no matter how well the engineering and risk team think they have prepared, the workshop will always produce new issues. There is a clear analogy with a HAZOP meeting. Stakeholders who should attend the workshop include design engineers, operations’ management, field personnel (land agents, patrol officers, etc.), construction management, relevant technical specialists (involved in corrosion, materials, etc., and possibly part-time), relevant outside

Past and future trends In Australia over the past 20 years there has been slow but deliberate movement away from “design by rules” towards risk-based “design by thinking”. The first edition of AS 2885 in 1987 recognized in principle a need to move away from rules based on location class and design factor (derived from US codes), but it did not achieve a practical change in the way pipelines were designed. The 1997 revision of AS 2885 was a substantial rewrite which introduced the concept of risk assessment as an underlying principle of pipeline design, and was a bold move that required significant change to the pipeline design process. While the principles were correct, the wording in the standard did not fully define all the requirements necessary to implement it effectively: conscientious players could do it well, but some just paid lip service to the new concepts. Nevertheless, the industry as a whole embraced the idea and gradually developed a broadly-agreed set of good practices. These have now been codified in the 2007 revision of AS 2885. In AS 2885.1-2007 the principles have not changed but the requirements are specified more explicitly with the aim of minimizing loopholes. The most substantial change is the addition of the high-consequence area requirements as described previously. This revision of the standard was published little more than a year ago so there has been only

14

The Journal of Pipeline Engineering

limited opportunity to assess how well the newer requirements are being accepted and implemented by the industry. However there was extensive public consultation and issue of drafts over a period of two years, so the industry generally has been improving risk-management practices and there is no evidence to date that the changes are causing major difficulties.

The cause/control model was adopted when the Australian pipeline industry became concerned that inappropriate statistical QRA methods may have been imposed on it by safety regulators who were comfortable with this approach in the management of hazardous industries, but who failed to appreciate its shortcomings when applied to pipelines in general and, particularly, pipelines in Australia. As a preemptive defence against any such moves, the industry sought to establish a safety-management strategy which it considered to be more appropriate, and which would provide genuine improvements in safety while not incurring costs that did not achieve practical reductions in risk. On the whole the AS 2885 approach has been well accepted by regulators, with only limited areas where QRA is imposed (and almost invariably done badly, using inappropriate data and methods, as noted previously).

Future trends in risk management of Australian pipelines are difficult to assess, especially since the latest version of the standard is still so new. The need for further changes may only become apparent after the new standard has been in use for a considerable time.

PY

In comparison with other pipeline codes, one feature of AS 2885 that may appear distinctive is the relatively-broad discretion permitted, bordering on the subjective. This is a deliberate strategy, and to date there are no indications that there has been any abuse of the flexibility that is permitted.

SA

M

PL

In the very long term, as the Australian pipeline network continues to age, quantitative methods may also find increasing application in prioritising deteriorating pipelines for repair.

Reliability-based quantitative risk methods have to date been barely recognized in Australia. There is clearly great potential in such methods, but it seems likely that they will be adopted only when driven by risk issues which are difficult to resolve in any other way.

Discussion Implicit in the AS 2885 approach to pipeline safety is a cause/control model of pipeline incidents: they have identifiable causes, and those causes can be controlled through design and operation so that the possibility of pipeline failure is either eliminated or reduced to a tolerable level. An alternative view is that incidents have random causes and can never be totally prevented; this underlies some QRA approaches which use statistical failure rates. In fact, all incidents do have causes, but there is uncertainty in the knowledge of those causes, so the cause/control and random views of risk management are really at opposite ends of a spectrum of knowledge. Nevertheless, AS 2885 is biased strongly towards the identification and management of specific factors that might lead to failure.

In fact, to the contrary, it appears that most pipeline engineers are quite conservative people who like to have rules and who are keen to be seen to be complying with the Standard. Hence the Standard is generally being applied conservatively. The review through the SMS workshop, the state technical regulators and (in some states) independent design validation go a long way to ensuring that the permitted discretion is properly applied.

C O

E

A key issue may be the extent to which quantitative methods are used. As discussed previously, the Australian pipeline industry in general does not believe that quantitative riskanalysis methods add value to routine design or risk assessment, and such methods are not currently used except where mandated by local regulations. However, as cities continue to expand around pipelines built to rural location class standards, it is perhaps increasingly likely that some intractable questions of risk versus cost may benefit from comparative numerical risk estimates. Any trend in this direction will be encouraged by the new AS 2885 requirement that the risks due to a pipeline subject to urban encroachment must be rigorously demonstrated to be ALARP, including comparison with alternatives such as reconstruction or relocation of the line.

There are some minor concerns that the requirements of the SMS process are not always well understood, but this is likely to fade as time passes and the industry becomes increasingly familiar with the new requirements. Having said all that, the new requirements for highconsequence areas have not yet been seriously tested in cases where there has been very extensive urban encroachment over pipelines built for rural conditions (minimum wall thickness, minimum cover, but now with houses within metres of the pipeline for many kilometres). At least one such SMS review is imminent at the time of writing, and the outcomes will be observed with interest. Overall, the Australian pipeline industry appears to be satisfied with the AS 2885 approach to pipeline safety and risk management. It works well for us in allowing both the safety and costs of pipelines to be optimized.

References 1. AS 2885.1-2007 Pipelines - Gas and liquid petroleum. Part 1: Design & Construction. Standards Australia, 2007 2. AS 2885.3-2001 Pipelines - Gas and liquid petroleum. Part 3: Operation and maintenance. Standards Australia, 2001

F R E Q U E N C Y

C O N S E Q U E N C E S

Hig h

Intermediate

Not anticipated for this pipeline at this location Theoretically possible but never on a similar pipeline

Remote

Hypothetical

Hig h

Unlikely to occur but possible

Unlikely

Extreme

Extreme

Expected to occur once per year or more May occur occasionally

Occasional

Frequent

Environment

M

NOTE: Significant environmental consequences may occur in locations which are relatively small and isolated.

C O

L ow

Intermediate

Hig h

Hig h

Extreme

E Hig h

Negligible

L ow

Intermediate

Intermediate

PY

Major off-site impact or longterm severe effects or rectification difficult

Long term interruption of supply

S up p l y

PL

Short term interruption or prolonged restriction of supply

Prolonged interruption or long-term restriction of supply

Multiple fatalities result.

People

Effects widespread, viability of ecosystems or species affected, permanent major changes

I njury or illness requiring hospital treatment

Few fatalities, or several people with lifethreatening injuries

SA

Severe

M aj or

Catastrophic

Typical severity classes

Negligible

Negligible

L ow

L ow

Intermediate

Negligible

Negligible

Negligible

L ow

L ow

Effect very localised (< 0.1 ha) and very short term (weeks), minimal rectification

No impact; no restriction of pipeline supply

Short term interruption or restriction of supply but shortfall met from other sources

Localised ( = 7

PY

To define the range of each defect’s environment, a vicinity parameter must be empirically determined, using the relationship expressed in Equn 1. Each defect will also have its associated characteristic length, as defined by Equn 21.

A number of additional simplistic assumptions have been made, and a general outline of them will be given in the following paragraphs. • Process characterization: irreversible, evolving at a constant rate, and at discrete time intervals. • Defect population: ILI reported metal-loss anomalies trimmed, based on the empirical criterion defined in Equn 3: Dj ≥

2Et 1.28

(3)

• Coating degradation lag: must be empirically defined based on coating data history and engineering best judgment. • Cathodic protection: is assumed to remain in a steady-state condition throughout the entire service life of the pipeline. • Probability density functions (PDFs): defect depth dimensions and corrosion rates are described by Gaussian PDFs. It is worth noting that, given that each defect’s corrosion rate is represented in terms of a local average, there is a normalizing effect on the overall depth corrosion rate data set2.

Mathematical framework Corrosion rates PDFs The probability density functions should be individually defined, taking account of the damage accumulated in each defect neighbourhood, according to the previously-outlined principle of local corrosion activity3. If a significant change in the system’s operating conditions takes place after any

where Et represents tool the measurement error and

2. Pipeline geometry, material features, and the axial and circumferential corrosion rates, have only been considered deterministically, as will be the allowable damage as a consequence.

1. The parameter n should be adjusted in order to obtain an average segment length not exceeding 1-2 km.

3. Clustering criteria should preferably be applied after a future morphology forecast, not before.

22

The Journal of Pipeline Engineering

particular event, then a factor Fh is defined accordingly, using Equn 4; otherwise, Fh is assumed as 14. Fh =

Δt1 + sΔt2 Δt s

while Equn 2 should be used to define its neighbourhood characteristic length.

Future defect morphology

(4)

The average dimensions of future defects can be calculated from Equn 6a, and Equn 6b is used to determine the associated dispersion.

Internal defects Individual defect growths (radial, axial, and circumferential) are determined by means of Equn 5a. The subsequent application of the local corrosion activity principle leads to the determination of the corrosion rate average for the defect population located in the adjoined region by using Equn 5b, while the dispersion is obtained from Equn 5c. Furthermore, the characteristic length associated with each defect neighbourhood (Lseg) can also be defined, as previously discussed.

D f = D i .RLi .Δt f

2 ⎛ E ⎞ σ Df = Δt f (σ Li ) + ⎜ t ⎟ ⎝ 1.28 ⎠

(5a)

j

∑ (R

j=i−n

PL − Rj )

2

Li

2n

M

(2n + 1) j=i+ n

σ Li =

(5b)

j=i−n

SA

RLi = Fh

∑R

(5c)

It is proposed that pipeline coating holidays are considered stationary. Thus, circumferential and axial growth rates are assumed as zero at all active sites located on the pipeline’s external surface. Equation 5d represents the depth growth rate, considering the lag in coating degradation. dj Δt s − Δt c

d a = f (l f , w f )

(7)

Probability of exceedance The future defect depth (df) shall not exceed its allowable depth (da) [22, 23], as represented by the limit-state function in Equn8:

External defects

Rdi =

There are a number of metal-loss assessment criteria that can be used to determine damage tolerance. The most simplistic and widely known is ASME B31.G, which only takes into account axially-oriented corrosion defects submitted to internal pressure loading. Depending on the particular system’s damage characteristics (which can include circumferential- or even helically-oriented defects), or the existence of axial loads (such as those geotechnically or thermally induced), an appropriate criterion should be chosen to deterministically find out the maximum allowable defect depth as a function of its forecast width and length, according to Equn 7.

PY

E

j=i+ n

(6b)

C O

Di Δt s

2

Damage tolerance

Hence, each flaw on a pipe’s inside surface will have one single PDF representing its depth corrosion growth rate, while axial and circumferential rates, as well as its neighbourhood characteristic length, are deterministically defined. Ri =

(6a)

(5d)

Equations 5b and 5c must therefore also be applied in order to characterize the defect’s depth corrosion rate PDF,

4. The scoring factor for changes in service conditions (s) should be determined based on historical data (coupons/probes, comparison of multiple ILI data or computation simulations) and engineering best judgment [20].

d f − da < 0

(8)

In the current approach, df is characterized by a normal distribution, while da is deterministic. This means that the probability of a pipeline exceeding the limit-state condition at each defect can be determined as the area on the righthand side of the allowable depth under the df PDF (see Fig.2)5.

Economic remediation rate The economic remediation rate which provides cost-effective operation must be ascertained by a pipeline’s own operator, considering each case individually. It is outside the scope of

5. Most commercial packages have standard functions to perform this.

1st Quarter, 2009

Table 1. Construction and operational data.

23

Pipeline 1

Pipeline 2

Pipeline 3

Pipeline 4

Diameter (in)

16

14

22

16

Minimum thickness (mm)

8.7

8.2

6.3

7.9

Pipe material

X60

X65

X40/46

X35

Length (km)

184

228

98

98

Service life (yr)

26

36

32

41

MAOP (kg/sqcm)

100

97

21-56*

31-41*

*worst case scenario hydraulic simulated range.

this work to accomplish a full perspective into problem, but some of the factors that must be taken into account in such an analysis include:

Case studies

SA

M

PL

As the whole model is based on averaging the behaviour of the local corrosion process, its application is not recommended to systems where hot-spot mechanisms (such as stray current, under-coat corrosion, etc.) are significant features.

In order to illustrate the model’s application, four case studies have been chosen, the input data for which is summarized in Table 1. A brief introduction is given for each, before the model results and overall performance are discussed.

PY

• Pipeline 2: an onshore gas pipeline that has been used to transport both wet and sour products. Accumulated internal corrosion is severe although, on the other hand, almost no external metal-loss indications have been reported as a result of the dryness of the of region crossed by this pipeline, in the NE of Brazil.

C O

Restriction on the model’s applicability

E

• technical and economic viability of alternative pipeline systems, or other modes of transportation • the ratio between the cost of a new pipeline and the estimated maintenance costs of the existing one • the impact of a possible delivery shortage on the local economy • current, and possible future, economic scenarios

• Pipeline 1: an onshore pipeline carrying dry gas since its operation began. Accumulated corrosion damage was slight on both the external and internal pipeline surfaces.

• Pipeline 3: a trunk line responsible for transporting all of one refinery’s crude oil supply. During its operational life, it endured production water pumped through recurrently, together with some high-BSW content product. Long shut-down periods were also a regular occurrence. Internal corrosion damage is quite severe and channelling damage is general. In order to meet an increase in demand, an increase in flow capacity was required. The resultant new service conditions were simulated by the worstcase hydraulic scenarios, and the maximum operational pressure profile was defined accordingly. • Pipeline 4: an onshore line which has been used to transport naphtha and crude oil, the latter usually

Fig.1. Local corrosion activity.

24

The Journal of Pipeline Engineering

Fig.2. The probabilistic limit-state function.

POE threshold range of 10-4-10-5, and the economic remediation rate specified for each case, it can be concluded that pipeline 3 could be safely operated for almost 30 years, while pipeline 4 would be cost-effectively operational for approximately 20 years at most.

SA

M

PL

E

The model’s output data from the four case studies are summarized in Table 2, while Fig.3 shows the overall normalized local corrosion activity. Figures 4 and 5 present the expected probability of exceedance for safe operations (at the required levels) without repair to the 200 most critical metal-loss areas in each case study, for the next 20 and 30 years, respectively. Pipeline 2 was not analysed for external corrosion, due the lack of significant indications on its external surface. In view of a desirable operational

C O

Results and discussion

PY

with a high BSW content. Again, production water transportation was a frequent occurrence together with extensive shutdown periods. The whole pipeline has bad channelling damage, as shown in Fig.1.

I N T E R N A L

As a result of applying these forecasts, the company’s board of directors has undertaken the following:

Pipeline 1

Pipeline 2

Pipeline 3

Pipeline 4

213

-

867

222

5

-

5

5

0.08-0.006

-

0.053-0.004

0.80-0.006

50

-

100

100

Remaining life Expected under historic conditions [years]

> 30

-

30

20

Population - filtered

337

10,370

50,325

23240

5

10

20

15

0.065-0.008

0.080-0.006

0.045-0.003

0.081-0.007

Pecuniary remediation rate Cost-effective coating repairs number

40

80

80

50

Remaining life Based under historic conditions [years]

30

15-20

10

5

Population - filtered

E X T E R N A L

Conversely, with the exception of pipeline 1, Figs 6 and 7 show that internal corrosion developing over 30 years would be a direct threat. Pipeline 4 is not expected to maintain its present use for long, while the operational reliability of pipelines 2 and 3 will not be cost-effective for more than 20 and 10 years, respectively.

Vicinity parameter (n) Local corrosion rates Gaussian distribution parameters [mm/year] Pecuniary remediation rate Cost-effective coating repairs number

Vicinity parameter (n) Local corrosion rates Gaussian distribution parameters

Table 2. Modelling parameters and output.

1st Quarter, 2009

25

25000

20000

15000

10000

5000

0 0,032

0,036

0,040

0,044

0,048

0,052

0,056

0,060

0,064

0,068

mm/year

Fig.3. Internal corrosion rate histogram for pipeline 3.

Individual

SA

M

PL

E

C O

PY

Local

Fig.4. 20-year POE forecast of the worst external metal-loss anomalies.

Fig.5. 30-year POE forecast of the worst external metal-loss anomalies.

0,072

0,076

26

The Journal of Pipeline Engineering

SA

M

PL

E

C O

PY

Fig.6. 10-year POE forecast of the worst internal metal-loss anomalies.

Fig.7. 20-year POE forecast of the worst internal metal-loss anomalies. • pipeline 4 was converted to specified diesel transportation; • a major rehabilitation project is being carried out on pipeline 3, together with several mitigating actions (including a new strategy regarding production water); • a brand new pipeline is under construction to replace pipeline 2 (mainly in order to supply the local market’s forecast rising demand) while an alternative use for pipeline 2 is being studied.

Conclusions Nowadays, new onshore pipeline systems must be planned well in advance. Sometimes almost a decade can pass

between the conceptual design and commissioning stages, mostly as consequence of the complexities concerning the legal agreements with landowners through whose land the pipeline will be routed, together with the tougher regulations regarding environmental and operational issues. Pipeline operators therefore need to forecast their systems’ remaining lives with reasonable long-term accuracy. Despite corrosion being the major time-dependent threat to ageing pipeline systems, there is little available guidance concerning corrosion modelling for real pipeline service conditions, and the subject remains controversial. The current work has been developed to support the operator’s long-term strategic planning, by providing a straightforward stochastic model to forecast the remaining

1st Quarter, 2009

27

The technique provides powerful information with no need for further expensive or laborious analyses. The use of the model has already proved to be particularly relevant to forecasting critical problems long before they present any real threat. The model has also been used to give rise to active mitigation planning, such as a review of inhibitor strategy and definition of the scope of coating rehabilitation projects.

PL

E

Additionally, if more-sophisticated mathematical packages are available, the model could be easily adapted to incorporate further refinements, incuding:

PY

A balance has been established between over- and underconservative assumptions, and the model had been considered suitable for forecasting periods of up to 30 years. Its algorithm is set out in detail and it can easily be implemented using standard commercial mathematical packages.

2. R.Bea et al., 2003. Reliability based fitness-for-service assessment of corrosion defects using different burst pressure predictors and different inspection techniques. 22nd International Conference on Onshore Mechanics and Arctic Engineering, June 8-13, Cancun. 3. NACE RP-0775. Preparation, installation, analysis and interpretation of corrosion coupons in oilfield operations. 4. NACE SP0502, 2008. Pipeline external corrosion direct assessment methodology. 5. J.M.Race, S.J.Dawson, L.Stanley, and S.Kariyawasam, 2006. Predicting corrosion rates for onshore oil and gas pipelines. International Pipeline Conference, Calgary. 6. S.B.Cunha, A.P.F.Souza, E.S.M.Nicoletti, and L.D.Aguiar, 2006. A risk-based inspection methodology to optimize inline inspection programs. The Journal of Pipeline Integrity, pp133-144. 7. M.Ahammed, 1998. Probabilistic estimation of remaining life of a pipeline in the presence of active corrosion defects. International Journal of Pressure Vessels and Piping, 75, pp321329 8. S.L.Fenyvesi, H.Lu, and T.R.Jack, 2004. Prediction of corrosion defect growth on operating pipeline. Proc. International Pipeline Conference, October 4 - 8, Calgary, Canada. 9. A.Valor, F.Caleyo, L.Alfonso, D.Rivas, and J.M.Hallen, 2007. Stochastic modeling of pitting corrosion: a new model for initiation and growth of multiple corrosion pits. Corrosion Science, 49, pp559–579. 10. A.Ainouche, 2006. Future integrity management strategy of a gas pipeline using Bayesian risk analysis. 23rd World Gas Conference, Amsterdam. 11. P.J.Laycock and P.A.Scarf. Exceedances, extremes, extrapolation and order statistics for pits, pitting and other localized corrosion phenomena. Corrosion Science, 35. no 1-4, pp135-145, 193. 12. J.L.Alamilla and E.Sosa, 2008. Stochastic modelling of corrosion damage propagation in active sites from field inspection data. Corrosion Science, 50, pp1811–1819. 13. J.L.Alamilla, D.De Leon, and O.Flores, 2005. Reliability based integrity assessment of steel pipelines under corrosion. Corrosion Engineering, Science and Technology, 40, 1. 14. S.A.Timashev, 2003. Updating pipeline remaining life through in-line inspection. International Pipeline Pigging Conference, Houston. 15. S.A.Timashev et al., 2008. Markov description of corrosion defect growth and its application to reliability based inspection and maintenance of pipelines. Proc. 7th International Pipeline Conference, Calgary. 16. G.Desjardins, 2002. Optimized pipeline repair and inspection planning using in-line inspection data. Pipeline Pigging, Integrity Assessment, and Repair Conference, Houston. 17. B.Gu, R.Kania, S.Sharma, and M.Gao, 2002. Approach to assessment of corrosion growth in pipelines. 4th International Pipeline Conference, Calgary. 18. G.Desjardins, 2001. Predicting corrosion rates and future corrosion severity from in-line inspection data. Materials Performance, August, 40,8. 19. J.M.Race et al., 2007. Development of a predictive model for pipeline external corrosion rates. Journal of Pipeline Engineering, 6, pp15-29. 20. R.B.Eckert and B.Cookingham, 2002. Advanced procedures for analysis of coupons used for evaluating and monitoring internal corrosion. CC Technolgies, Doublin, OH, USA. 21. ASME B 31G. Manual for determining the remaining strength of corroded pipelines.

C O

life of corroded pipelines. As input data, the newly-developed model requires pipeline geometry and material properties, the worst-case scenario for operational pressure, a goodquality set of metal-loss ILI data, and also the economic threshold for the system’s future remediation. The pioneering concept of local corrosion activity was introduced, and the underlying simplistic assumptions are detailed together with the entire mathematical framework. The definitions and roles of the empirical parameters have also been described.

M

• non-Gaussian behaviour (for which an automatic best-fitting-distribution tool is required)

SA

• full limit-state approach: pipeline geometry and material properties could also be considered probabilistically (if convolution integrals can be easily solved) [2, 10, 24] • any specifics of a system’s history could be taken into consideration by making the necessary adjustments to the model’s premises and assumptions.

Acknowledgments The authors thank Petrobras Transporte SA for permission to publish this paper, and their colleagues Dr Sérgio Cunha, Carlos Alexandre Martins, and João Hipólito de Lima Oliver for many enlightening discussions and contributions.

References 1.

B.Gu, R.Kania, and M.Gao, 2004. Probabilistic based corrosion assessment for pipeline integrity. Corrosion 2004, NACE International, New Orleans.

28

The Journal of Pipeline Engineering

24. G.Pognonec, 2008. Predictive assessment of external corrosion on transmission pipelines. 7th International Pipeline Conference, Calgary.

PL

E

C O

PY

22. H.Plummer and J.M.Race, 2003. Determining pipeline corrosion growth rates. Corrosion Management, April. 23. F.Caleyo et al., 2002. A study on the reliability assessment methodology for pipelines with active corrosion defects. International Journal of Pressure Vessels and Piping, 79, pp77-86.

Mister Mech Mentor, Volume I: Hydraulics, Pipe Flow, Industrial HVAC & Utility Systems

by Trevor M. Young

by James A. Wingate

SA

M

Technical Writing A–Z: A Commonsense Guide to Engineering Reports and Theses, British English Edition

TITLES OF INTEREST FROM ASME PRESS

Topics include: format and content of reports and theses; copyright and plagiarism; print and Internet reference citation; abbreviations; units and conversion factors; significant figures; mathematical notation and equations; writing styles and conventions; frequently confused words; grammatical errors and punctuation; commonsense advice on issues such as getting started and holding the reader’s attention. 2005 256 pp. Softcover ISBN: 0-7918-0237-X Order No. 80237X $29 (list)/$23 (ASME member) Order sets of 10 copies at a special price. Order No. 80236S $199

American Edition: 2005 256 pp. Softcover ISBN: 0-7918-0236-1 Order No. 802361 $29 (list)/$23 (ASME member) Order sets of 10 copies at a special price. Order No. 80236S $199

Pipeline Operation and Maintenance: A Practical Approach by M. Mohitpour, J. Szabo, and T. Van Hardeveld Covering pipeline metering, pumping, and compression, the book covers day-to-day concerns of the operators and maintainers of the vast network of pipelines and associated equipment and facilities that deliver hydrocarbons and other products. It is a useful reference for veterans and a training tool for novices. 2004 600 pp. Hardcover ISBN: 0-7918-0232-9 Order No. 802329 $125 (list)/$99 (ASME member)

Order Now!

Gain practical knowledge from frank, colorful cases and learn to solve mechanical problems related to hydraulics, pipe flow, and industrial HVAC and utility systems with these organized solutions to the problems involving: water and steam hammer phenomena; gravity flow of liquids in pipes; siphon seals and water legs; regulating steam pressure drop; industrial risk insurers’ fuel gas burner piping valve train; controlling differential air pressure of a room with respect to its surroundings; water chiller decoupled primary-secondary loops; pressure drop calculations of incompressible fluid flow in piping and ducts; water chillers in turndown; hydraulic loops; radiation heat transfer; and thermal insulation. 2005 160 pp. Softcover ISBN: 0-7918-0235-3 Order No. 802353 $45 (list)/$36 (ASME member)

Pipeline Design and Construction: A Practical Approach, Second Edition by M. Mohitpour, H. Golshan and A. Murray This second edition includes updated codes and standards information, solutions to technical problems, additional references, and clarifications to the text. It offers straightforward, practical techniques for pipeline design and construction, making it an ideal professional reference, training tool, or comprehensive text. 2003 700 pp. Hardcover ISBN: 0-7918-0202-7 Order No. 802027 $110 (list)/$88 (ASME member)

North America: www.asme.org • Europe: www.ihsatp.com

1st Quarter, 2009

29

Ethanol transportation: status of research, and integrity management by Dr John Beavers1, Patrick Vieth1, and Dr Narasi Sridhar2 1CC Technologies, Inc (a DNV Company), Dublin, OH, USA 2 DNV Research and Innovation, Dublin, OH, USA

T

T

PL

E

C O

PY

HE PIPELINE INDUSTRY is undertaking considerable research to determine the best approach to manage the potential for internal stress-corrosion cracking (SCC) to occur while transporting fuelgrade ethanol (FGE). Based on the results to date, it appears that FGE meeting the ASTM D 4806 specification can cause SCC of carbon steel. The parameters that affect the potential for SCC (oxygen, water, etc.) are understood, and the research is now focused on methods to reduce the likelihood of SCC. The current state of the research is discussed.

SA

M

HE US Energy Policy Act of 2005 (amended in 2007) established a nationwide renewable fuels standard starting from 15 billion litres (4 billion gallons) of all biofuels in 2006 to 136 billion litres (36 billion gallons) in 2022. Ethanol will constitute almost 90% of this renewable fuel. As the price of gasoline has oscillated, the economic viability of alternative fuels also is undergoing re-evaluation. However, the long-term need for biofuels, both for reducing gasoline dependence and carbon footprint, is undeniable. Biofuels can be broadly classified into several generations depending on their status of commercial production readiness. The ethanol produced from sugars (such as sugar cane, beets, and grapes) and starch (corn, wheat) is considered to be a first-generation biofuel. The production techniques for these feedstocks are well established, and world-wide commercial production of ethanol from these sources was approximately 50 billion litres (13.2 billion gallons) in 2008. The ethanol produced from grasses (such as switch grass), Author’s contact information: tel: +1 614 761 1214 email: [email protected]

This paper was presented at the 21st Pipeline Pigging and Integrity Management conference held in Houston on 10-12 February, 2009, organized by Clarion Technical Conferences, Houston, and Scientific Surveys Ltd, Beaconsfield, UK.

agricultural/food processing wastes, and other cellulosic materials requires enzyme treatments that will require significant additional process development. The ethanol (or alcohol) from these sources is referred to as secondgeneration ethanol (and also as cellulosic ethanol). A further development in alcohol production will be the use of transgenic materials (low-lignin trees), which will require advanced enzymatic treatments, and the resulting alcohols will be the third-generation cellulosic fuel. A somewhat similar categorization exists for biodiesel production. The ethanol supply chain is illustrated in Fig.1. Ethanol is produced in bio-refineries and must be transported to terminals, where it is blended with gasoline to produce the most commonly used blends E-10 (10% ethanol) and E-85 (85% ethanol). As shown in Fig.1, rail and truck are currently the predominant means of transporting ethanol in North America. Brazil has a history of transporting ethanol via pipelines and ships. In terms of the volumes transportable by the different modes shown in Fig.1, one barge load is roughly equivalent to 15 to 20 rail cars or 6080 truck loads. In comparison, a 16-in pipeline can transport an equivalent of 15 barges on a daily basis. The number of new rail cars constructed has to rise substantially, and new terminals that can accommodate unit train shipments have to be constructed, to allow rail shipment of the future anticipated fuel volumes. Barge transport can benefit substantially by a pipeline delivery system, while increasing

30

The Journal of Pipeline Engineering

Ethanol plant

Oil Refinery

Pipeline Barge (2%)

Rail (30%)

Truck (67%)

Blending Terminal

Blend

SA

M

PL

Figure 2 shows that most of the bio-refineries in the US are located close to the middle of the continent, whereas the population centres are along the coasts. Most of the current hydrocarbon pipelines move products from the Gulf of Mexico region to the east and west coast and the midwest. Thus, new pipelines will be required to transport the fuelgrade ethanol (FGE) from the bio-refineries to the population centres.

Background A survey of published literature and service experience with SCC in FGE was published by the American Petroleum Institute (API) in 2003 [1]. Documented SCC failures of equipment in users’ storage and transportation facilities have dated back to the early 1990s: the majority of the cracking has been found at locations near welds where the primary stresses leading to SCC have been residual welding stresses. No cases of SCC were reported in ethanol manufacturer facilities, tanker trucks, railroad tanker cars, or barges, or following blending the FGE with gasoline. All occurrences of SCC were at the first major hold point (the FGE distribution terminal) or in the subsequent end-user gasoline-blending and -distribution terminals. An example of SCC in terminal piping is shown in Fig.3: note that the leak is near a girth weld adjacent to a piping tee.

The API survey did not pinpoint what causes ethanol SCC, but the failure history suggests that the SCC may be related to changes in the FGE as it moves through the distribution chain over a period of days, weeks, or months. These observations led to an industry-sponsored research programme to identify the causative factors. A ‘Roadmapping Workshop’ held in October, 2007, identified a number of research gaps in safely transporting FGE via pipelines [2], which were divided into four areas:

C O

E

the truck transportation poses significant logistical problems, including training a much larger number of drivers than anticipated to be available in the future. Thus, pipeline transportation is the most cost-effective mode of transporting large volumes of ethanol.

PY

Fig.1. Ethanol supply chain for North America.

(i) ethanol sources and quality, (ii) pipeline operations, (iii)standards, guidelines, and training, and (iv) pipeline integrity. Ethanol SCC potentially impacts all four of these areas and research is continuing or planned to address these gaps. A number of factors were identified as contributors to SCC, and SCC-mitigation strategies are being developed. This paper summarizes the current state of the research.

Environmental factors affecting SCC in FGE The results of research on chemistry effects on SCC have demonstrated that FGE that meets applicable API standards (Table 1) is a potent cracking agent in the presence of oxygen [3, 4] . Several research programmes have examined the effects of the contaminants (such as chloride) or the denaturant in FGE on SCC behaviour in aerated ethanol solutions. Studies by Sridhar et al. [5] and Beavers et al. [6] showed that chloride significantly increased the susceptibility of carbon steels to SCC in ethanol.

1st Quarter, 2009

31

ASTM limits (ASTM D4806) Requirement Minimum

Maximum

92.1

-

Methanol, vol. %

-

0.5

Solvent-washed gum, mg/100 ml

-

5.0

Water content, vol. %

-

1.0

1.96

4.76

I norganic chloride, ppm (mg/L)

-

40 (32)

Copper, mg/kg

-

0.1

Acidity (as acetic acid CH3COOH), mass % (mg/L)

-

0.007 (56)

6.5

9.0

Ethanol, vol. %

Denaturant content, vol. %

pHe

PY

A statistically-designed study by Sridhar et al. [5] showed that dissolved oxygen is the most important factor affecting SCC in FGE: no SCC occurred under any circumstances without the presence of dissolved oxygen. Based on the oxygen concentration, a “critical” potential regime was identified for SCC [8], Fig.4, which depends on chloride concentration. In the presence of chlorides, SCC extends to lower corrosion potentials.

PL

E

Furthermore, the fracture mode changed from predominantly intergranular to predominantly transgranular as the chloride concentration increased from 0 to 40ppm. Methanol also appeared to increase SCC susceptibility [5]. The water content of the FGE has also been shown to affect the SCC behaviour: anhydrous ethanol will not promote SCC [4] and water contents above about 4.5% by volume completely inhibit SCC [7]. Between these limits, water does not appear to have a significant effect on SCC [5, 7].

C O

Table 1. Specification of fuel-grade ethanol.

SA

M

Factors that have not been shown to have a significant influence on SCC in FGE include acidity, one common general corrosion inhibitor, and the denaturant [4, 5]. SCC was observed in SSR tests over a wide range of pHE and acetic acid concentration. Sridhar et al. [5] showed that one common corrosion inhibitor added to FGE to protect against automotive corrosion (Octel DCI-11) did not have any effect on SCC of steel.

Fig.2. Ethanol production locations in the US.

At high corrosion potentials in some ethanols, SCC was not observed, and the reason for this behaviour is still unclear. Beavers et al. [6] showed that removing oxygen by chemical, mechanical, or electrochemical methods all resulted in suppression of SCC in slow-strain-rate (SSR) tests in a simulated FGE. Oxygen removal also caused a negative potential shift in the free corrosion potential, as shown in Fig.5. It is well known that dissolved oxygen increases SCC susceptibility of steel in other non-aqueous environments, such as ammonia and methanol. Therefore,

32

The Journal of Pipeline Engineering

Leak

Fig.3. SCC observed in terminal piping system containing FGE.

SA

M

PL

C O

E

The earlier research, funded by API, demonstrated that FGE as well as E-85 (85% ethanol – 15% gasoline) promoted SCC [4]. More recently, Beavers et al. [9] evaluated the effect of ethanol-gasoline blend ratio on SCC in a research project funded by PRCI. The study was performed with notched SSR specimens of an X-46 linepipe steel in a simulated FGE. No SCC was observed in gasoline or E-10 (10% ethanol blend), but SCC susceptibility increased rapidly with increasing ethanol concentration for E-20 and higher blends. Surprisingly, E-30 was nearly as susceptible to SCC as FGE, as shown in Fig.6.

Studies [8, 9] have shown that no two ethanols are created equally in terms of SCC tendency. Some ethanols do not cause SCC even at high dissolved oxygen levels; others cause significant SCC. Aging of ethanol samples appears to alter their SCC tendency significantly. The variations in the ethanol chemistry and their impact on SCC behaviour are the subject of ongoing studies.

Metallurgical factors affecting ethanol SCC Field experience and laboratory testing indicate that severe straining is required for ethanol SCC to occur. SCC of ethanol storage tanks has been observed only in severelystrained areas associated with non-post weld heat-treated welds and/or in tanks with design/installation issues [1]. For example, floor areas that were not adequately supported experienced SCC as a consequence of cyclic loading from filling and withdrawal of ethanol. Some of the earliest laboratory studies of SCC in ethanol were conducted using U-bend specimens. SCC was not observed in these tests

unless a “bad” welding bead perpendicular to the stressing direction and an extremely severe bending mode were included. In SSR tests with un-notched specimens, SCC was observed near the necked region of the specimen [6]; notched-SSR tests exhibited SCC at the notch root [9]. All these observations suggest that severe plastic deformation and the presence of dynamic plastic strain are necessary for SCC to occur. In more-recent crack-growth tests using compact tension specimens, the presence of a cyclic loading component has been shown to exacerbate SCC [7].

PY

it is not surprising that dissolved oxygen is a major contributor to SCC in FGE.

Recent studies reported by Beavers et al. [9] have shown that the extent of SCC was not dependent on steel grade ranging from X-42 high frequency electric-resistance welded pipe material to cast steel for pumps. For one grade, the weld area of a double submerged arc weld appeared to be slightly more resistant to SCC than the base metal.

Mitigation of ethanol SCC The field experience and research results specifically addressing ethanol SCC, as well as broader experience with other forms of SCC in the pipeline and other industries, point to potentially-effective methods for mitigation of ethanol SCC. The research on oxygen effects on ethanol SCC clearly demonstrates that, regardless of how the oxygen is removed, SCC can be mitigated. Both mechanical deaeration and one chemical oxygen scavenger were shown to be effective. It is probable that the oxygen is absorbed in the ethanol during the transportation process and it might be possible to minimize oxygen contamination, as opposed to removing it once it is already in the ethanol. True SCC inhibitors also potentially could be effective. Research by Beavers et al. [6] showed that some filmforming amines have an inhibiting effect on ethanol SCC. The identification of the best possible inhibitors or inhibitor

1st Quarter, 2009

33

With Cl

520

Without Cl Only Cl Only MeOH No Cl or MeOH Low H2O

Maximum Load, Kg

480

EtOH-10%Gasoline EtOH-15% Gasoline Wet Milling EtOH Dry Milling EtOH Reagent EtOH+air E85 Sample 1 Deaerated E85 Sample 1 Aerated High Potential EtOH Aerated Reagent EtOH Still Air E85 Sample 1 Still Air E85 Sample 2 Deaerated

No SCC SCC

440

400

360 -400

0

400

800

PY

Corrosion Potential, mV vs. Ag/AgCl EtOH

C O

Fig.4. SCC vs corrosion potential indicating a critical potential (dependent on chloride level) below which SCC was not observed [8]. pipelines. Post-weld heat treatment of all welds could minimize SCC, although the hoop stresses from the internal pressure in transmission pipeline might play a bigger role in the SCC process. Grit blasting prior to coating has been shown to play a role in the mitigation of external SCC of gas transmission pipelines [10]. The compressive residual

PL

E

packages, taking into consideration diverse issues such as toxicity, compatibility with combustion engines, cost etc., requires further research.

SA

M

The association of SCC in the terminals with residual stresses provides another mitigation avenue for new ethanol

Crack Growth Rate, mm/s

2.0E-06

No Deaeration

1.5E-06

1.0E-06

5.0E-07

Steel Wool Hydrazine

0.0E+00 -400

-300

-200

Nitrogen Deaeration

-100

0

Mechanical Deaeration

100

200

Average Corrosion Potential in Test, Ag/AgCl EtOH Fig.5. Crack growth rate as a function of average potential for SSR tests in simulated FGE with various deaeration methods [6].

34

The Journal of Pipeline Engineering

SCC Crack Growth Rate, mm/s

5.0E-06 4.0E-06 3.0E-06 2.0E-06 1.0E-06 0.0E+00 -1.0E-06 10

20

30

50

95

Ethanol Concentration, %

PL M

Summary

SA

Large increases in production and transportation volumes of FGE are expected as a result of new energy policies dictating substantially-higher usage of alternate fuels. Ethanol has been shown to cause SCC of steel in the presence of dissolved oxygen and chloride. Significant advances have been made in understanding the various parameters that can affect SCC and in identifying mitigation strategies. Going forward, this information will be needed to manage the integrity of ethanol pipelines.

References 1. R.D.Kane and J.G.Maldonado, 2003. Stress corrosion cracking of carbon steel in fuel grade ethanol: review and survey. API Technical Report 939-D, American Petroleum Institute, Washington, DC, September. 2. Energetics, Inc., 2007. Safe and reliable ethanol transportation and storage technology roadmapping workshop, October 2526, Dublin, Ohio.

3. R.D.Kane, N.Sridhar, M.Brongers, J.A.Beavers, A.K.Agrawal, and L.Klein, 2005. Materials Performance, 44, 12. 4. R.D.Kane, D.Eden, N.Sridhar, J.Maldonado, M.P.H.Brongers, A.K.Agrawal, and J.A.Beavers, 2007. Stress corrosion cracking of carbon steel in fuel grade ethanol: review, experience survey, field monitoring, and laboratory testing. API Technical Report 939-D, 2nd Edn, American Petroleum Institute, Washington, DC, May. 5. N.Sridhar, K.Price, J.Buckingham, and J.Dante, 2006. Corrosion, 62, 8, pp687-702. 6. J.A.Beavers, M.P.Brongers, A.K.Agrawal, and F.A.Tallarida, 2008. Prevention of internal SCC in ethanol pipelines. NACE, Corrosion 2008 Conference, New Orleans, LA, March, Paper 08153. 7. J.A.Beavers and N.Sridhar, 2008. Unpublished results, PRCI SCC Program. 8. J.G.Maldonado and N.Sridhar, 2007. SCC of carbon steel in fuel ethanol service: effect of corrosion potential and ethanol processing source. Corrosion, Paper 07574, Houston, TX, NACE International. 9. J.A.Beavers, N.Sridhar, and C.Zamarin, 2009. Effects of steel microstructure and ethanol-gasoline blend ratio on SCC of ethanol pipelines. NACE, Corrosion 2009 Conference, Atlanta GA, March, Paper 095465. 10. J.A.Beavers, N.G.Thompson, and K.E.W.Coulson, 1993. Effects of surface preparation and coatings on SCC susceptibility of line pipe: Phase 1 – laboratory studies. Corrosion, NACE Paper 597, New Orleans, LA, March.

C O

E

stress imparted by the grit blasting process has been shown to effectively overcome the effects of residual stresses and the tensile hoop stress from internal pressurization. A similar process might be effective for the mitigation of ethanol SCC.

PY

Fig.6. Crack growth rate as a function of ethanol concentration for X-46 linepipe steel specimens tested in simulated FGEgasoline blends [9].

1st Quarter, 2009

35

Evaluating damage to on- and offshore pipelines using data acquired using ILI by Dr Chris Alexander Stress Engineering Services, Inc, Houston, TX, USA

E

VALUATING THE INTEGRITY of pipelines often involves assessing data acquired from an in-line inspection (ILI) run. ILI generates a range of data types, one of which is geometric data from a caliper tool. Once the data are collected, engineers are required to evaluate the relative severity of any indications that might have been found. With recent advances in storage capacity and instrumentation, the resolution of the acquired data is of sufficient magnitude to make relatively accurate assessments of the potential damage that might exist within a given pipeline system.

D

M

PL

E

C O

PY

In this paper a case study is provided that used data collected during an in-line inspection run of a damaged pipeline. The assessments included the development of finite-element models constructed using the geometric ILI data. Integral to the assessments were integration of actual pressure history data that, when used in conjunction with a cumulative damage assessment model, determined the remaining life of the selected anomaly. Additionally, the assessment used prior full-scale experimental data to confirm the accuracy of the models. A systematic approach for evaluating damaged pipelines using ILI caliper tool data is described.

SA

ENTS GENERATED IN onshore pipeline are typically the result of third-party damage, although rock dents are certainly a contributor for bottom-side defects. Damage to subsea pipelines typically occurs as the result of impact with an anchor. After the subsea incident occurs, ROVs (remotely-operated vehicles) are then deployed to survey the damage, followed by survey efforts to determine if the pipeline has been moved or laterally displaced. If it is believed that localized damage has been inflicted, it is essential that the profile of the dented region be determined, and in-line inspection is ideally-suited for collecting this data. From a geometry standpoint, the data collected includes points measuring radius, circumferential orientation, and longitudinal position (i.e. R-q-Z coordinates). Presented in this paper is a background section that discusses Author’s contact information: tel: +1 281 955 2900 email: [email protected] This paper was presented at the 21st Pipeline Pigging and Integrity Management conference held in Houston on 10-12 February, 2009, organized by Clarion Technical Conferences, Houston, and Scientific Surveys Ltd, Beaconsfield, UK.

how to evaluate dents considering previous research efforts and experience, following which is a discussion on how raw ILI geometry data are converted into the mesh for evaluation using the finite-element (FEA) method. FEA is used to calculate the alternating stresses in the dented region; once the stresses due to cyclic pressure are calculated, a fatigue curve is used to estimate the remaining life for the given dents. Results are presented from previous research on fatigue testing of pipes having plain dents, and the final section of the paper provides recommendations for industry in using ILI data to estimate the remaining life of damaged pipelines, and integrating previous test data where appropriate for validation purposes.

Background In the 1990s, a significant body of work on evaluating dented pipelines was performed under the direction of the Pipeline Research Council International, while other work was also performed on plain dents and related defects for the American Petroleum Institute. For the most part, this work focused on damage to pipelines involving plain dents and dents with gouges. Full-scale testing involving pipelines subjected to static and cyclic pressures were used to evaluate

36

The Journal of Pipeline Engineering

Pipe geometry

Grade and yield strength

Charpy impact (ft-lbs)

F i na l dent depth (d/D, %)

Failure stress (ksi)

B PU 2

36-in x 0.54-in

X60 (67.5 ksi)

44.2

3.4

67.2

Failed at 112% SMYS

BNO 2

36-in x 0.50-in

X60 (60.8 ksi)

19.1

4.5

67.7

Failed at 113% SMYS

BIE 1

24-in x 0.38-in

X52 (53.1 ksi)

14.0

5.4

24.9

Failed at 48% SMYS

B LV 1 (1)

30-in x 0.31-in

X52 (52.7 ksi)

19.2

3.5

71.1

(see note 2)

EUY 1

36-in x 0.66-in

X65 (68.9 ksi)

31.7

4.8

26.5

Failed at 41% SMYS

FJB 1 (1)

30-in x 0.48-in

X52 (58.8 ksi)

22.8

3.2

18.9

Failed at 36% SMYS

FJB 1 (1)

30-in x 0.48-in

X52 (58.8 ksi)

22.8

4.9

12.6

Failed at 24% SMYS

Sample number

the effects of dents having varying degrees of severity on the integrity of pipelines, and interested readers are encouraged to consult the reference documents provided in this paper. The predominant conclusion from these research efforts is that to properly assess a defect’s severity, one must appropriately categorize the defect. The major defect classifications that typically arise when assessing pipeline damage are:

C O

PY

The following equation was developed Maxey [1] and correlates the relationship between initial dent depth and the residual dent depth as a function of applied pressure and yield strength.

E SA

M

The sections that follow discuss in detail experimental testing that has been conducted to address the several classes of dent listed previously by different research programmes around the world. Detailed in each section are the appropriate references, critical variables associated with the defect in question, and the effects of loading (static or cyclic) on failure behaviour.

Plain dents Plain dents are defined as dents having no injurious defects – such as a gouge – and possessing a smooth profile (they are often classified as smooth dents). The critical variables relating to plain dents are: • dent depth (depth after rerounding due to pressure) • pipe geometry (relationship between diameter and wall thickness) • profile curvature of the dent profile • pressure at installation • applied cyclic pressure range. While the effects of certain variables are not clearly understood, it is apparent that the denting process plays a critical role in determining the future behaviour of the dent. Early research recognized that dent depth was one of, if not the most important, variable of interest. The dent

Table 1. Burst pressures for plain dents. Note: (1) cracks detected on inside seam weld of the sample; (2) sample yielded but did not break.

created initially changes as a function of applied pressure (statically or cyclically).

Do =

plain dents constrained dents gouges mechanical damage wrinkles

PL

• • • • •

Notes

where:

DR

⎡ ⎛ ⎞⎤ σ ⎢ -0.5066 × log ⎜⎜ ⎟⎟ ⎥ ⎢⎣ ⎝ σ y + 10,000 ⎠ ⎥⎦

(1)

s = hoop stress at instant of damage (psi) sy = yield strength of pipe (psi) Do = dent depth at instant of damage (in) DR = residual dent depth after removal of damaging tool (in) A review of the preceding equation by Hopkins [2] revealed some levels of unconservatism because the above formulation is lower-bound and ignores the elastic springback of the dent at zero internal pressure. Later work by Rosenfeld [3] indicates that some degree of progressive rerounding occurs with pressure cycles. It is these changes in dent depth, and associated changes in dent profile, that determine the eventual long-term behaviour of the dent. When considering pipes with relatively-high diameter to wall thickness ratios, a significant level of rerounding occurs on pressurisation. Work conducted for the American Petroleum Institute (API) [4] showed that for 12.75-in x 0.188-in X52 pipes, it was not possible to achieve dent depths greater than 3% of the pipe diameter when the pipe was pressurised to the maximum allowable operating pressure, even though initial dent depths as great as 18% were initially established. As will be discussed later in this paper, this rerounding reduces the severity of the dent. The behaviour of plain dents in static and cyclic pressure environments differ. The sections that follow provide insights on these differences.

1st Quarter, 2009

37

Table 2. Cyclic pressure tests on plain dents.

Sample number

Pipe geometry and grade

Initial dent depth (d/D, %)

Note: (1) no pressure in pipe sample US6A-2 12.75-in x 0.188-in, Grade X52 6 during indentation. (2) residual dent measured with no UD12A-3 12.75-in x 0.188-in, Grade X52 12 pressure in pipe after sample was pressurised to a 65% SMYS stress UD18A'-28 12.75-in x 0.188-in, Grade X52 18 level. (3) cycles to failure listed based upon Miner’s Rule in combining results from two applied pressure ranges (36% and 72% SMYS). (4) sample did not fail. Testing terminated due to excessive number of applied pressure cycles.

1,307,223

2.5

684,903

0.7

101,056

curvature reduce fatigue lives of pipes more than dents that are shallow with relatively-smooth contours. Work conducted for the American Gas Association [6], American Petroleum Institute [2], and by EPRG [5] all validate this position. The second factor determining the severity of plain dents is the range of applied pressures. In general, a fourth-order relationship can be assumed between the applied stress range and fatigue life: in other words, a dented pipeline subjected to a pressure differential of 200psi will have a fatigue life that is 16 times greater than if a pressure differential of 400psi were applied. Barring the effects of rerounding (which change the local stress in the dent), the fatigue lives of plain dents are reduced to a greater degree when increased pressure differentials are assumed. Table 2 provides several data points extracted from the API research programme showing the effects of dent depth on fatigue life. As noted in the data, the 6% dent never failed and had a fatigue life that exceeded the fatigue life for the 18% dent by one order of magnitude.

SA

M

PL

E

The definitive conclusion based on all available research is that plain dents do not pose a threat to the structural integrity of a pipeline other than the potential for reduced collapse/buckling capacity associated with the induced ovality. A discussion on the subject matter will follow in a later section of this paper. However, the classification of a plain dent assumes that no cracks, gouges, or material imperfections are present in the vicinity of the dent. Interaction of plain dents with weld seams, especially girth welds and submerged arc welds (SAW), can significantly reduce the burst strength of the damaged pipeline [4]. The primary cause of the reduction is crack development at the toe of the welds during pressurizing the pipe and associated rerounding of the dent.

1.3

PY

The response of plain dents to static pressure loads deals primarily with the effects of the damage on the burst strength of the pipe. In addition to concerns relating to dent depth and profile, the mechanical properties of the damaged pipe material are also important. Work was reported in the 1980s that correlates burst pressure with dent depth and material properties for pipes with different geometries and grades [5]. The tests involved pipe ring samples that were dented prior to pressure testing; Table 1 provides a summary of the test results.

Cycles to failure

C O

Response of plain dents to static pressure loading

F i na l dent depth (d/D, %)

In assessing the overall impact that plain dents have on pipelines subjected to cyclic service, one must consider both the applied pressure range and geometry of the dent. A given dent may not be serious in gas service, but could pose a detriment to fatigue life when considering the service requirements of liquid transmission pipelines.

Response of plain dents to cyclic pressure loading While plain dents do not pose a threat to pipeline integrity in a static environment, cyclic pressure applications can reduce the life of a pipeline. A survey of several gas and liquid transmission companies revealed the number of applied pressure cycles that can be expected for the respective fuel types [6]. A gas transmission line can be expected to see 60 cycles per year with a pressure differential of 200psi; however, the same pressure differential can occur over 1,800 times on a liquid pipeline in the course of a year. For this reason, liquid pipeline operators are considerably more concerned with fatigue than gas pipeline operators. The impact that a plain dent has on the fatigue life of a pipeline is directly related to two factors, the first of which concerns the dent geometry in terms of shape and depth. Dents that are deeper and possess greater levels of local

Dents with gouges While plain dents may be regarded as rather benign in terms of their impact on structural integrity, dents with gouges are a major concern for pipeline companies. The leading cause of pipeline failures is mechanical damage, which often occurs during excavation of pipelines, and the United States Department of Transportation (US DOT) has specific criteria for reporting outside incidents. The rate of reportable incidents for gas pipelines from 1970 to June 1984 was 3.1 x 10-4/km-yr, while the rate was approximately 6.8 x 10-5/km-yr for the period from July 1984 to 1992. A more-conservative estimate assumes that the actual incident rate may be as high as 10-3/km-yr due to unreported incidences. Regardless of the assumed incident rate, world-wide efforts have focused on the need for mechanical damage research. In the United States, most of

38

The Journal of Pipeline Engineering

Gouge depth (a/t, %)

Initial dent depth (d/D, %)

Burst pressure (psi)

Percent SMYS (Pburst / SMYS)

B1-1N

5

5

2,165

141

B1-3N

10

5

1,985

120

B1-6N

10

10

1,479

96

B1-7N

15

15

820

53

B1-8N

10

12

1,517

99

B1-11N

5

15

775

51

Sample number

the pressures at which they failed. As noted in the table, dent and gouge combinations that exceed 10% of the pipe diameter and wall thicknesses (respectively) are likely to have burst pressures that are less than the pressure corresponding to SMYS. The pipes used in testing had relatively-good ductility and toughness (32% elongation and Charpy V-notch impact energy of 51ft-lbs at room temperature); however, pipes without such material qualifications will fail at lower pressures. Work conducted by the Snowy Mountains Engineering Corporation in Australia) validates the importance of having sufficient ductility and toughness in reducing the potential for low failure pressures.

C O

E

Response of dents with gouges to static pressure loading

PY

the experimental work has been conducted by Battelle Memorial Institute and Stress Engineering Services, Inc, and has been funded by the American Gas Association and the American Petroleum Institute. In Europe, testing has been conducted primarily by British Gas and Gaz de France with funding from the European Pipeline Research Group. The severity of mechanical damage is rooted in the presence of microcracks that develop at the base of the gouge during the process of dent rerounding due to pressure (and to some extent elastic rebound). As with plain dents, dents with gouges respond differently to static and cyclic pressure loading. The discussions that follow provide greater details regarding the associated responses.

SA

M

PL

Unlike plain dents that do not severely affect the pressurecarrying capacity of pipelines, the deleterious nature of dents with gouges requires careful investigation. The failure patterns of dents with gouges that are subjected to static pressure overload involve the outward movement of the dent region, while development and propagation of microcracks at the base of the gouge occur with increasing pressure levels. Hopkins et al. [5] conducted numerous ring tests to address the failure pattern of dents combined with gouges and concluded that the failure mechanism was ductile tearing within an unstable structure.

Testing was conducted by Kiefner & Associates, Inc/Stress Engineering Services, Inc [4] for determining the burst pressure of dents containing gouges. All testing was conducted using 12-in NPS X52 pipes. Machined V-notches were made at various depths in the pipe samples, which were pressurised to 920psi (60% SMYS) and then dented with a 1-in wide bar. Table 3 lists six of the test samples and

Residual dent depth (% pipe diameter)

Based upon a review of the data and the experience of the author in experimental testing, it is difficult to envision a closed-form solution for predicting the failure pressure due to static overload of dents containing gouges. Although attempts have been made to do so, a paper written by Eiber and Leis [7] shows that the current models (developed for the PRCI and EPRG) do not satisfactorily predict burst pressures. Several of the primary reasons for the complexities in predicting burst pressure of dents with gouges are: • material properties (especially ductility and toughness) • sharpness and depth of gouge • pressures at indentation and during rerounding • dent profile and depth as well as resulting plastic deformation of pipe • local work-hardening and variations in throughwall properties due to denting The key to future experimental testing is only to address one variable while holding all others constant; the above list represents a satisfactory starting point.

Gouge depth (% pipe wall thickness)

Fatigue life

20%

Greater than 145,500 cycles

4%

None

Less than 6,930 cycles

4% (in pipe weld)

None

Less than 789 cycles

20%

Less than 199 cycles

None

4%

Table 3. Burst tests for dents with gouges. Note: (1) dents installed with an internal pressure of 920 psi. Dents permitted to reround after pressurisation. (2) material properties: 53.6ksi yield strength; 72.1ksi UTS; 51 ft-lbs CVN.

Table 4. Fatigue life for gouges, plain dents, and dents with gouges

39

C O

PY

1st Quarter, 2009

Research efforts funded by AGA and API indicate that when dents are installed in ERW seams the fatigue resistance is on the same order as plain dents [6, 8]. This assumes that good-quality seam welds are present in the pipe material. The presence of girth welds was shown to reduce the fatigue life of dents to a level less than ERW seams, but more than SAW seams. As an example, consider that the research programme for API tested a dent in a SAW weld seam that failed after 21,603 cycles, while the same dent in a girth weld failed after 108,164 cycles [8].

E

PL

Response of dents with gouges to cyclic pressure loading

Fig.1. 36-in diameter pipe with 2% wrinkles.

SA

M

Initial efforts in the pipeline research community focused on static burst testing of mechanical damage, but once a basic level of understanding of the fracture mechanisms were developed, efforts focused on fatigue testing. Cyclic pressure tests have been conducted on pipe specimens with a variety of defect combinations [4, 5, 6]; the research efforts conducted for the EPRG, AGA, and PRCI indicate that if the fatigue life for plain dents is on the order of 105 cycles, then the presence of gouges (in dents) reduces this value to be of the order of 103. Table 4 summarises data from the research conducted for the EPRG on ring-test specimens for relating plain dents and dents with gouges subjected to cyclic pressure service [5]. As noted, the presence of a gouge significantly reduces the fatigue life of a plain dent, although a gouge by itself is non-threatening (an observation validated by Fowler et al., [6]).

Response of dents in welds to cyclic pressure loading In addition to considering interaction of dents with gouges, efforts to assess the interaction of welds with dents have been conducted. Testing on submerged and doublesubmerged arc welds indicated that the dents in seam welds could significantly reduce the burst pressures and fatigue lives of the affected pipelines. The recommendation by Hopkins et al. [5] is that these defects should be treated with extreme caution and immediate repair considered.

Experimental study of strains in dented pipes While numerous studies have addressed the failure patterns of plain dents and dents with gouges, less effort has been made to evaluate the strains in dented pipes. Obviously, the complex nature of dent mechanics is a contributing factor; also, the use of finite-element analysis (FEA) permits engineers to accurately understand the stress/strain distribution in dents as will be discussed later in this paper. Lancaster et al. has conducted numerous tests directed at developing an understanding of strains caused by pressurization of pipes with dents, employing the use of both strain gages and photoelastic coatings. His work provides several useful findings, • During the process of rerounding the dents with internal pressure, approximately 60% of the dent

The Journal of Pipeline Engineering

Depth after cycling (inches)

40

0.00 Sample Configuration 2 percent buckle 4 percent buckle

0.50

6 percent buckle

1.00 0

5

10

15

20

25

Longitudinal Position (inches) Fig.2. Wrinkle profile for the three test samples. nominal diameter of the pipe). Figure 1 shows the pipe sample with 2% wrinkles, while Fig.2 shows the corresponding profiles for the three wrinkles that were tested.

had been recovered at a pressure equal to 70% of the yield pressure. There was evidence of creep at pressures above yield.

SA

M

PL

PY

E

• The highest strain measured on the rim of the dent was 7000me, and the maximum hoop stress concentration (SCF) was calculated to be 10.0. In comparing this SCF with those generated by finiteelement methods (FEM) for the API research programme, the maximum FEM SCF was calculated to be 7.2 for an unconstrained dome dent having a residual dent depth of 10% [8].

Pressure-cycle testing was performed where the samples were pressure cycled to 100% of the operating pressure. The following fatigue results were obtained:

C O

• The locations having the highest strains are on the rim of the dent. Interestingly, this location was consistent with the failure location for unconstrained dome dents in the API research programme that resulted in longitudinally-oriented cracks that developed on the exterior of the pipe [8].

In addition to the work conducted by Lancaster, Rosenfeld [3] developed a theoretical model that describes the structural behaviour of plain dents under pressure. His efforts also involved dent rerounding tests for validation purposes.

Wrinkle bends Wrinkle bends are associated with the bending of pipe that results in creating local indentations that may be regularly or irregularly spaced, along the length of the affected area. Wrinkle bends are not considered favourably by the pipeline codes and most operators. As a point of reference, ASME B31.8 841.231(g) states that wrinkle bends are permitted only on systems that operating at hoop stress levels less than 30% of the specified minimum yield strength. As part of the American Petroleum Institute study [8], experimental efforts were undertaken to assess the effects of wrinkle bends on the fatigue life of pipelines. Three 36in x 0.281-in pipes were fitted with wrinkle bends having nominal depths of 2%, 4%, and 6% (wrinkle depth percentage calculated by dividing wrinkle depth by the

2% wrinkle – no failure after 44,541 cycles 4% wrinkle – failure after 2,791 cycles 6% wrinkle – failure after 1,086 cycles

The above results were a significant find for the API research programme. The critical observations is that although depth of damage is important (wrinkle or dent), the more important factor is the profile shape of the damage. The change in radius of curvature along the length of the line is directly related to bending strains. As noted in the fatigue data, a wrinkle having a depth of 6% poses a significant threat to the integrity of the pipeline. Although intentional wrinkle bends are unlikely to occur offshore, the authors observed several anchor impact zones that clearly resembled the damage profile associated with wrinkle bends. For this reason, any damage in an onshore or offshore pipeline that resembles a wrinkle bend (i.e. defect having a sharp curvature, as in a kink) should be removed as soon as is prudent.

Summary of experimental work The information presented in this paper indicates that a significant level of research has been conducted world-wide in an effort to characterize and assess the severity of plain dents and dents with gouges. It can be concluded that a certain hierarchy exists in terms of defect severity, although unquestionable scatter is present in both the static and fatigue data. Empirical models and semi-empirical models have been able to predict with some success the failure pressure for dents with gouges; however, the large number of variables has so far precluded the development of a general model that can accurately forecast the burst and

1st Quarter, 2009

41

Circumferential position (every 12 degrees) 199.0881 199.5950 201.6023 204.6078 207.6003 199.0792 199.5974 201.6105 204.6081 207.5990 199.0838 199.6020 201.6066 204.6069 207.6073 199.0884 199.6066 201.6026 204.6057 207.6155 199.0876 199.6119 201.6034 204.6089 207.6123 199.0868 199.6172 201.6042 204.6122 207.6090 199.0859 199.6225 201.6050 204.6154 207.6057 199.0920 199.6131 201.6026 204.6171 207.6025 199.0981 199.6037 201.6001 204.6187 207.5992 199.1042 199.5944 201.5977 204.6203 207.5960 Data shown201.6014 (other than first 207.5948 199.1183 199.5910 204.6298 column) are201.6050 radial coordinates. 199.1323 199.5877 204.6393 207.5935 199.1311 199.5983 201.6018 204.6338 207.5974 199.1299 199.6089 201.5985 204.6283 207.6013 199.1287 199.6194 201.5953 204.6228 207.6051 199.1271 199.6203 201.6001 204.6252 207.6039 199.1256 199.6213 201.6050 204.6277 207.6027 199.1266 199.6162 201.6050 204.6283 207.5994 199.1277 199.6111 201.6050 204.6289 207.5962 199.1287 199.6060 201.6050 204.6295 207.5929 199.1381 199.6142 201.6148 204.6307 207.5978 199.1476 199.6225 201.6246 204.6319 207.6027 199.1541 199.6225 201.6205 204.6328 207.5954

PY

196.1399 196.1445 196.1445 196.1445 196.1469 196.1494 196.1518 196.1494 196.1469 196.1445 196.1506 196.1567 196.1606 196.1644 196.1683 196.1753 196.1823 196.1772 196.1721 196.1671 196.1720 196.1768 196.1768

209.8907 209.8935 209.8920 209.8904 209.8949 209.8994 209.9039 209.9008 209.8978 209.8947 209.8981 209.9014 209.8984 209.8953 209.8923 209.8959 209.8996 209.8927 209.8858 209.8788 209.8779 209.8770 209.8823

211.0764 211.0715 211.0837 211.0959 211.0951 211.0943 211.0935 211.0968 211.1000 211.1033 211.1057 211.1082 211.1114 211.1147 211.1179 211.1130 211.1082 211.1171 211.1261 211.1350 211.1262 211.1173 211.1167

C O

Axial

391300.1426 391300.1459 391300.1491 391300.1524 391300.1557 391300.1590 391300.1623 391300.1655 391300.1688 391300.1721 391300.1754 position 391300.1787 391300.1819 391300.1852 391300.1885 391300.1918 391300.1951 391300.1984 391300.2016 391300.2049 391300.2082 391300.2115 391300.2148

Fig.3. Raw in-line inspection data in cylindrical coordinates.

Fourier transform (FFT) routine. As a point of reference, where the raw data had 30 points circumferentially resulting in nodal spacing of 1.75in, the FFT-modified procedure produces 177 points circumferentially spaced at approximately 0.50in. The number of data points in the axial direction is adjusted to match the circumferential spacing so that the two are approximately equal (an element aspect ratio of 1:1).

E

Dent analysis

SA

M

PL

fatigue behaviour of all possible types of mechanical damage. Any evaluation involving numerical modelling based on ILI geometry data should be validated by referencing previous experimental work.

The primary focus of this paper is to specifically address the use of ILI data in evaluating dent severity, and the approach presented can be used for both on- and offshore pipelines. The presentation includes a discussion on converting raw ILI data into a format useful for generating a finite-element mesh, actually performing the analysis using FEM, and interpreting the data in terms of estimating future performance.

Converting raw ILI data The ILI data that are typically measured by an in-line inspection tool is presented in cylindrical coordinates (i.e. R-q-Z). Figure 3 provides a portion of an example data set taken from an ILI tool run: as can be seen, radial coordinates are provided as functions of circumferential and axial positions. In this particular data set the circumferential positions are provided every 12o, or approximately every 1.75in for the given pipe diameter. To generate accurate analysis results, this spacing is too large, and therefore an algorithm was developed to increase the mesh density and generate a more-refined mesh for the FEM based on a fast

Finite-element analysis Once the required level of mesh refinement has been made, the finite element model is generated, for which the R-q-Z coordinates serve as the nodes. A Fortran code was developed to read the reduced data and generate an Abaqus input file. The coordinates for each node were developed using the relationships shown below.

X = ( r + t 2 ) * sin θ Y = ( r + t 2 ) * cos θ Z =Z In these relationships, r is the inside radius from the ILI data; q, is the circumferential position relative to the pipe axis measured clockwise from the top of the pipe. The thickness of the pipe, t, is taken based on the pipe’s nominal wall thickness. The axis of the pipe was taken as the global z-axis. Figure 4 shows an overall view of a dent model, while Fig.5 shows an enlarged view of the region

42

The Journal of Pipeline Engineering

Fig.4. Global view of dent in finite-element model.

SA

M

PL

PY

E

Once the model pre-processing was completed, stresses were calculated based on the internal pressure loading. Although plastic strains are induced in any dented pipeline, experience has shown that after several pressure cycles a shakedown to elastic action occurs and the alternating stresses are typically within the elastic regime. Therefore, it is appropriate to elastically model cyclic stresses in dents. From the finite-element model, the principal stresses in the dented region of the model are calculated. From this stress state a stress concentration factor (SCF) is calculated by dividing the maximum principal stress by the nominal hoop stress. Figure 6 provides a contour plot showing the maximum principal stresses in a dent that resulted in a

maximum SCF of 3.58. It is noted in this figure that the maximum stress occurred on the outside surface of the model: these results are consistent with previous findings from experimental studies where fractures in plain dents subjected to cyclic pressures initiated on the outside surface of the pipe.

Data interpretation

C O

where the mesh density can be seen. The “S4” type shell elements were specified in Abaqus, and symmetry boundary conditions were specified at each end of the pipe model. For each analysis, a linear elastic analysis was performed where the internal pressure was the yield pressure of the pipe using the specified minimum yield strength (SMYS) of the respective pipe grade (for example, X52 has an SMYS of 52,000psi). A typical finite-element model has of the order of 25,000 elements.

Once the FEA model results are calculated and a representative SCF has been determined, the next step involves estimating remaining life. It should be noted that, for this particular discussion, the focus is on plain dents where failure due to static pressure overload is unlikely. If plain dents do fail, they are most likely to do so in the presence of cyclic pressures. Even if a large number of cyclic pressures is not likely, the process of calculating SCFs provides operators with a means for evaluating the relative severity among competing dents. From the author’s experience, the API X’ fatigue curve from API RP2A, Planning, designing, and constructing fixed offshore platforms, reasonably predicts the fatigue behaviour

Fig.5. Close-up view of dent in finite-element model.

1st Quarter, 2009

43

Resulting SCF of 3.58 on outside surface of dented region.

Units in psi

N = 2.978 x 1021 Ds-3.74

E

(2)

While the above presentation is certainly useful, for most operators an important unanswered question remains: how many years of useful service remain? In the absence of actual historical operating data, the 2,657 cycle number is not entirely useful. Therefore, to complete the analysis one must consider actual operating history. Listed below are the steps involved in evaluating the remaining life of a dented pipeline considering the ILI-based stress concentration factor used in conjunction with actual operating pressure cycle data.

C O

of plain dents subjected to cyclic pressure conditions. Provided below is the equation for the API X’ curve where Ds represents the stress range in units of psi.

PY

Fig.6. Maximum principal stresses on outside surface of FEA model.

SA

M

PL

As an example, consider the previously-presented dent analysis with the SCF of 3.58 (cf. Fig.6). If one assumes a cyclic pressure range of 36% SMYS for an X52 pipe, the nominal hoop stress range is 18,720psi. Including the SCF, the corresponding stress range in the dented region is 67,000psi. Using the API X’ curve, the resulting fatigue life is 2,657 cycles.

Fig.7. Historical pressure cycle data from an operating pipeline (pressure in psi).

1. Obtain pressure history plot similar to one shown in Fig.7.

44

The Journal of Pipeline Engineering

Pressure Cycle Histogram for Yellow Creek Suction Location From 11/1/06 - 11/1/07 (283 Days) 350

300 286

Cycle Count

250

200 159 150 101 100 70

60

51 40

50

48

40 39 19 20

27 16 12

10 10 9 6 5 5 5 3 1 6 2 0 2 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

15 00

14 40

13 80

13 20

12 60

12 00

10 80 11 40

10 20

96 0

PY

90 0

84 0

78 0

72 0

66 0

60 0

54 0

48 0

42 0

36 0

30 0

24 0

18 0

60

12 0

0

C O

Peak-to-Peak Cycle Magnitude (psi)

Fig.8. Pressure cycle histogram showing stress range cycle count.

PL

E

2. Use rainflow counting to develop a pressure cycle histogram similar to one shown in Fig.8.

M

3. Use histogram to determine a single equivalent cycle count such as 100 cycles at DP = 36% SMYS.

SA

4. Divide the calculated fatigue life by the annual cycle count to determine the remaining life in years. Referring once again to the previous example, we determined that for a stress range of 36% SMYS the fatigue life was 2,657 cycles. If a given pipeline annually experienced 100 cycles at DP = 36% SMYS, the remaining life in years would be 26.5 years.

Discussion The integrity of dents is related to not only the severity of the dent itself, but also to the possibility that the dent can interact with other features such as seam and girth welds. The author’s company was the principal investigator in a study conducted for the American Petroleum Institute to evaluate the severity of plain and constrained rock dents. Included in this study were evaluating the effects of seam and girth welds that interacted with dents. Listed below are the major dent groupings extracted from the dataset from this API study, and related data for these test samples are included in Table 5. Within these samples are groups based on a number of common characteristics.

These groups are important as they serve as the basis for some of the assumptions regarding dent performance. As an example, the test results associated with girth welds in dents provides information regarding the expected performance of plain dents versus those dents containing girth welds. Unless noted, all dents are unconstrained. • • • •

plain dents – samples 1, 3, and 28 constrained dents – samples 15, 26, and 27 dents with welds – samples 16 and 20 dents with welds subjected to hydrotest – samples 30 and 31 • double dents – sample 32 As noted in Equn 2 for the API X’ RP2A S-N curve, there is a numerical relationship of 3.74 between design cycles and applied stress range. This exponent will be used in developing empirical stress concentration factors for specific pipeline imperfections. One of the objectives of this study was to evaluate how the fatigue life of plain dents is reduced when considering features such as girth welds, seam welds, and double dents. The data presented in Table 5 is used to provide numerical correlation among these dents, as presented below. Stress concentration factor for dents interacting with ERW seam welds • Sample 16 (unconstrained dent with ERW) – 22,375 cycles

1st Quarter, 2009

45

Initial Dent Rebound Dent Final Dent N (DP=50% MAOP) (% pipe OD) (% pipe OD) (% pipe OD)

Sample

Description

1 3 15 16 20 21 26 27 28 30 31

Plain dent, unconstrained Plain dent, unconstrained Constrained dent ERW, Plain dent, unconstrained GW, dent, unconstrained GW 2" offset from dent, unconstrained Constrained dent Constrained dent Plain dent, unconstrained ERW, Plain dent, unconstrained, hydrotest GW, dent, unconstrained, hydrotest

6 12 12 12 12 12 24 18 18 12 12

22

Double dent unconstrained (dents 3.5 inches apart)

12

69 70 71 72

Plain dent, unconstrained (4-inch dome indenter) Plain dent, unconstrained (4-inch dome indenter) Plain dent, unconstrained (4-inch dome indenter) Plain dent, unconstrained (4-inch dome indenter)

6 12 18 24

4.9 6.8 N/A 7.7 7.6 6.8 N/A N/A 11.3 5.9 6.0 5.2 5.6 3.3 7.1 15.8 15.9

2.7 2.5 N/A 1.4 1.4 1.5 N/A N/A 0.7 0.7 1.0 0.8 1.2 0.7 2.3 4.9 5.0

1,307,223 684,903 426,585 22,375 2,020 38,972 98,483 235,008 101,056 277,396 213,876 217,976 359,350 263,910 204,246 234,934

Table 5. Test results for dents subjected to cyclic pressure fatigue testing.

C O

PY

Notes: (1) sample 1 (unconstrained 6% plain dent) and sample 15 (constrained 12% plain dent) did not fail even after extensive pressure cycling. (2) the final dent depth was measured after all phases of testing were completed. (3) observed failure pattern for unconstrained dents was an OD-initiated longitudinal flaw. (4) observed failure pattern for constrained dents was an ID-initiated circumferential flaw. (5) the tested cycles to failure, N, presented above assumed an applied pressure range of 50% MAOP (36% SMYS). For the 12.75-in x 0.188-in X52 pipe used in the testing the 50% MAOP value corresponds to 550psi.

Stress concentration factor for double dents

E

• Sample 3 (unconstrained dent) – 684,903 cycles

−1

SA

⎛ 22,375 cycles ⎞ 3.74 SCF = ⎜ ⎟ = 2.49 ⎝ 684,903 cycles ⎠

M

PL

A stress concentration factor is calculated using the above cycles to failure using a 3.74 order relationship between stress and cycle life. (3)

Stress concentration factor for dents interacting with girth welds • Sample 16 (unconstrained dent with ERW) – 20,220 cycles • Sample 3 (unconstrained dent) – 684,903 cycles A stress concentration factor is calculated using the above cycles to failure using a 3.74 order relationship between stress and cycle life. −1

⎛ 20,220 cycles ⎞ 3.74 SCF = ⎜ ⎟ = 2.56 ⎝ 684,903 cycles ⎠

Table 6. Fatigue life reduction factors.

(4)

• Sample 16 (unconstrained dent with ERW) – 217,976 cycles • Sample 3 (unconstrained dent) – 684,903 cycles

A stress concentration factor is calculated using the above cycles to failure using a 3.74 order relationship between stress and cycle life. −1

⎛ 217,976 cycles ⎞ 3.74 SCF = ⎜ ⎟ = 1.36 ⎝ 684,903 cycles ⎠

(5)

Using the calculated stress concentration factors, it is possible to develop a fatigue reduction factor, FRF, for each respective imperfection type. This value can then be used to estimate the effect that a particular anomaly has on the fatigue life of a plain dent. Several example calculations are provided. The FRF is calculated using the following equation, with results for the three anomalies tabulated in Table 6. (6)

FRF = (SCF)-3.74

Damage type

SCF

FRF

Dent with ERW weld seam

2.49

0.033

Dent with girth weld

2.56

0.030

Double dent

1.36

0.318

46

The Journal of Pipeline Engineering

6. J.R.Fowler, C. R.Alexander, P.J.Kovach, and L.M.Connelly, 1994. Cyclic pressure fatigue life of pipelines with plain dents, dents with gouges, and dents with welds. Prepared by Stress Engineering Services for the Offshore and Onshore Applications Supervisory Committee of the Pipeline Research Committee, PR-201-9324, June. 7. R.J.Eiber and B. N.Leis, 1995. Line pipe resistance to outside force. Paper 14, EPRG/PRC 10th Biennial Joint Technical Meeting on Line Pipe Research, Cambridge, April 18-21. 8. C.R.Alexander and J.F.Kiefner, 1997, 1999. Effects of smooth and rock dents on liquid petroleum pipelines, Phases 1 and 2. API Publication 1156, May, and October, respectively.

Conclusions

Bibliography

This paper has discussed methods for using ILI data to evaluate the severity of dents in pipeline systems. The most powerful feature of this technique is the ability for an operator to compare the relative severity of multiple dentlike defects in an effort to make decisions regarding which ones require immediate attention. In a world of unlimited resources, operators could evaluate and repair all defects; however, in the real world such options do not exist, and operators must prioritize their responses based on the best available sources of information.

C.R. Alexander, 2006. Assessing the effects of external damage on subsea pipelines. Paper No. IOPF2006-014, Proceedings of the ASME International Offshore Pipeline Forum, October 24-25, 2006, Houston, Texas. C.R.Alexander, J.R.Fowler, and K. Leewis, 1997. Analysis of composite repair methods for pipeline mechanical damage subjected to cyclic pressure loads. 8th Annual International Energy Week Conference and Exhibition, Houston, Texas, January. C.R.Alexander and L. M.Connelly, 1998. Analytical recreation of a dent profile considering varied soil, operating and boundary conditions. Energy Sources Technology Conference & Exhibition, Sheraton Astrodome Hotel, Houston, Texas, February 2-4. C.R.Alexander, 1999. Analysis of dented pipeline considering constrained and unconstrained dent configurations. Energy Sources Technology Conference & Exhibition, Sheraton Astrodome Hotel, Houston, Texas, February 1-3. American Society of Mechanical Engineers, 1991. Manual for determining the remaining strength of corroded pipelines. ASME B31G-1991, New York. American Society of Mechanical Engineers, 1992. Liquid transportation system for hydrocarbons, liquid petroleum gas, anhydrous ammonia and alcohols. ASME B31.4, New York. American Society of Mechanical Engineers, 1995. Gas transmission and distribution piping systems. ASME B31.8, New York. I.Corder and P. Corbin, 1991. The resistance of buried pressurised pipelines to outside force damage. Paper24, EPRG/PRC 8th Biennial Joint Technical Meeting on Line Pipe Research, Paris, France, May 14-17. D.G.Jones and P. Hopkins, 1983. Influence of mechanical damage on transmission pipeline integrity. Proc. Int. Gas Research Conf., London, June 13-16. P.B.Keating and R. L.Hoffman, 1997. Fatigue behavior of dented petroleum pipelines (Task 4), Office to the Office of Pipeline Safety, US Department of Transportation, Texas A&M University, May. J.F.Kiefner, W.A.Bruce, and D.R.Stephens, 1994. Pipeline repair manual. Prepared for the Line Pipe Research Supervisory Committee of the Pipeline Research Committee. J.F.Kiefner, C.R.Alexander, and J.R.Fowler, 1996. Repair of dents containing minor scratches. Proc. 9th Symposium on Pipeline Research, Houston, Texas, October.

SA

M

PL

C O

E

From the author’s perspective there is no standardized method for evaluating the severity of dents. It is hoped that the methods presented here can serve as a means for opening lines of communication between ILI companies, pipeline operators, and industry experts in formalizing a more systematic approach for evaluating dents. There is certainly ample evidence to suggest that a reasonable understanding of dent behaviour exists among subject matter experts. When this knowledge is coupled with a standardized analysis approach, the pipeline community at large will be well-served.

PY

A final comment concerns two factors that were not considered in the analysis efforts discussed here. The first concerns the presence of corrosion: if corrosion is expected, one can assume that the remaining life of the dent will be reduced relative to the non-corroded case. Secondly, no consideration of tool tolerance was included in the geometry of the finite-element models. On this second issue, readers are encouraged to interface with tool vendors regarding tolerances and what, if any, effect they would have on the resulting dent geometry.

References and bibliography 1. W.A.Maxey, 1986. Outside force defect behaviour. NG-18 Report 162, AGA Catalog no. L51518. 2. P.Hopkins, 1991. The significance of mechanical damage in gas transmission pipelines. Paper 25, EPRG/PRC 8th Biennial Joint Technical Meeting on Line Pipe Research, Paris, May 14-17. 3. M.J.Rosenfeld, 1998. Investigations of dent rerounding behaviour. Proc. Int. Pipeline Conference, 1, pp299-307, Calgary, Canada. 4. C.R.Alexander, J.F.Kiefner, and J. R. Fowler, 1997. Repair of dents combined with gouges considering cyclic pressure loading. 8th Annual International Energy Week Conference and Exhibition, Houston, Texas, January. 5. P.Hopkins, D.G.Jones, and A.J.Clyne, 1989. Significance of dents and defects in transmission pipelines. Proc. Int. Conf. on Pipework Engineering and Operations, London, February 21-22.

47

SA

M

PL

E

C O

PY

1st Quarter, 2009

48

The Journal of Pipeline Engineering

The Pigging Products and Services Association An international trade association serving the pipeline industry Our aims are to promote the knowledge of pigging and its related products and services by providing a channel of communication between the members themselves, and with users and other interested parties.

Services include:

PY

Free technical information service available to all Sourcing of pigs and pigging services

E

Pigging seminars – next one 19th November 2008 Aberdeen

C O

Complimentary Buyers Guide and Directory of Members

PL

PPSA newsletter, “Pigging Industry News”

SA

Training courses

M

PPSA’s book “An Introduction to Pipeline Pigging” PPSA web site – www.ppsa-online.com

Want to join? Full members - pigging manufacturers and service providers Associate members - pipeline operators, suppliers and allied industries Individual members - anyone with an interest in pigging To find out more visit our web site www.ppsa-online.com or contact the Secretary at [email protected]

Pigging Products and Services Association P O Box 2, Stroud, Glos., GL6 8YB, UK Telephone: +44 (0) 1285 760597 Facsimile: +44 (0) 1285 760470 Email: [email protected] www.ppsa-online.com

1st Quarter, 2009

49

Rethinking laybarge pipelaying by Professor Andrew Palmer*1 and Dr Yue Qianjin2 1 Centre for Offshore Research and Engineering, Department of Civil Engineering, National University of Singapore, Singapore 2 State Key Laboratory for Structural Analysis of Industrial Equipment, Department of Engineering Mechanics, Dalian University of Technology, Dalian, China

T

An improvement from 0.5km/day to 5km/day over the 28 years from 1976 to 2003 corresponds to less than 9% a year. Without for a moment wishing to downplay the courageous financial and engineering investments that made that improvement possible, it is clearly modest by comparison with other industries such as computing and electronics, where the progress over the same period has been several orders of magnitude.

Many aspects of the project are instructive. Everything possible was done to minimize work at sea. The pipeline was welded together onshore, tested, and wound onto reels in great lengths. No connections were made at sea. The only marine operation was to lay the pipeline, and that could be done in one night. There were of course very good reasons in the special context of the project: it was possible that at sea the operation could still be attacked.

Creative dissatisfaction ought to encourage us to look for further progress, perhaps through radical change. One inspiration to every pipeline engineer is the PLUTO project, more than 60 years ago. The military recognized that the

Alternative construction methods

Author’s contact details: email: [email protected]

The reeling method is of course still widely applied, principally for small-diameter lines, though existing

SA

M

PL

E

C O

PY

Productivity has much improved. The second Forties pipeline was laid 15 years later by one barge in one season, and had a much higher average layrate of 1.9km/day, and rates near the average were achieved after only a few days [1]. The average layrate had become a much higher fraction of the peak layrate. A BP press release for a 22-in, 105-km, line to the Shetlands from the North Atlantic, laid by the Solitaire in 2003, reported an average layrate of 6.9km/day and a peak of 7.8km/day. Layrates of the order of 5km/day are nowadays almost routine.

armies that would invade the European mainland from England would consume enormous quantities of gasoline (petrol), and sought the advice of Anglo-Iranian (the forerunner of BP) on how to transport the fuel across the English Channel. Two ideas were put forward: the first was a hollow submarine cable, laid from a cable ship; the other was a 3-in steel ERW pipe, which would be wound onto a floating reel and unwound as tugs towed the reel across the Channel, some 120km from Shanklin on the Isle of Wight to Cherbourg on the Cotentin Peninsula in France. The pipeline had no cathodic protection anodes and no coating, there was no tensioner, and the positioning of the pipe must have been very imprecise. The girth welds were made by flash-butt welding. The project had many impressive aspects that might make us wonder how much progress has been made in the intervening years. The first trial is said to have been carried out exactly one week after the first meeting, something no oil company could accomplish today: the papers would still be in the in-tray of the contracts’ lawyers. In 1944 a pipeline could be laid over 120km in 10hrs, which is not possible in 2008, presumably because we know so much more about it.

HIRTY YEARS AGO, pipelaying productivity was extremely low by present-day standards. The Forties 1 pipeline (32-in, 170km, maximum depth 125m) took two barges two seasons. The Frigg pipeline system (32-in, 730km, maximum depth 140m) had an average productivity somewhere around 0.5km/day: precise figures are not available. The first writer unwisely opined that the laybarge system was so slow and inefficient that it would have to be replaced by a different system. He was obviously wrong, which once more confirms how foolish it is to try to forecast the future.

50

The Journal of Pipeline Engineering

barcode

concrete

anode

buckle arrester

10 joints or more; 125 m+ ? Fig.1. Factory-fabricated length of pipeline. equipment can lay up to 18-in pipe. A limiting factor is that a large-diameter pipeline on a small-diameter reel has large bending strains, and may buckle unless the wall thickness is large. Another limiting factor used to be that conventional concrete weight coating cannot be reeled, but to resolve this problem a rubber-like flexible concrete has been developed. It is more expensive than conventional concrete but much cheaper than adding steel to gain additional submerged weight. Tests have demonstrated that flexible concrete can be applied to pipelines, and that it protects the anti-corrosion coating.

PL

SA

M

Another option is tow, but it has turned out primarily to be attractive for relatively-short lengths, up to about 8km, generally as bundles. It is of course possible to construct greater lengths by towing a number of long sections and then joining them together, but that has only occasionally been done and is generally thought not to be economically attractive.

Alternative laybarge schemes The conventional sequence includes a number of operations, carried out by different organizations and often at widely-separated locations. The sequential operations are fragmented and not vertically integrated. One company manufactures the pipes; other companies apply the anticorrosion coating and the concrete weight coating, and install anodes and buckle arresters. Lengths of pipe are transported to a laybarge, and there welded together in pairs in double joints, or sometimes into longer multiple joints. The girth welds are all made offshore. The laybarge needs to be large and expensive, and a great number of people are engaged. Each of them needs support, so that for each person working on the pipeline there are several others, controlling and positioning the barge, cooking, cleaning, managing, inspecting, transporting people to and

PY

An alternative – potentially better and cheaper – concept is to integrate the process, and to adopt the PLUTO principle of doing as much as possible onshore and as little as necessary at sea. Imagine an integrated onshore process that manufactures long lengths of pipeline, say 125m (10 joints) but perhaps longer still. The conceptual model is the high-technology factory, not the construction site or the offshore barge. The factory workers travel to the factory for each shift, and go home afterwards: they themselves look after and pay for all the accommodation, food, transportation, and recreation issues that would have to be provided by an offshore constructor. They work full time and are paid accordingly: it does not occur to them to expect to work 15 days on and 15 days off.

C O

E

Reeling is attractive because the bulk of the work required is carried out at the shore make-up site. This is particularly attractive if the application of internal lining is expected to slow down welding. Once the pipeline has been loaded it can be laid very rapidly, typically at 1m/s. However, nobody has yet been prepared to make the serious investment needed for a reelship or reelbarge that could lay 36-in pipelines, for example.

fro, doing laundry, organizing exercise rooms and putting on videos, and so on.

The lengths that the factory produces are complete, in the sense that when they leave the factory the only work that still has to be done on them is to connect them to the lengths on either side. Any necessary inspection, such as girth-weld gamma ray or UT, has already been carried out. The lengths include girth-weld anti-corrosion protection and infill, anodes, barcode internal and external identification, pre-installed connections for CP monitoring, and buckle arresters and internal coating (if required). Figure 1 illustrates a length: the lengths can be factoryhydrotested, and quality control in a factory environment is under much less time pressure than it is offshore. The lengths now have to the transported to the barge that will connect them into the pipeline and lower them to the seabed. The factory is at the shore, and the lengths are rolled from the factory onto specialized transportation vessels. The vessels sail to the laybarge, and unload them into a storage area. Figure 2 illustrates an S-lay version of the laybarge. It makes one weld at one station, gives the weld anti-corrosion coating and if necessary girth-weld infill, and lays the pipe over a stinger in the conventional way. How to make the girth connections is a separate question. Conventional ‘automatic’ welding is only one of many possible techniques [2]. There are many welding alternatives, and some of them are known to be much faster, cheaper,

1st Quarter, 2009

51

n connectio ramp

× st

in

g

er

Fig.2. Pipelaying schematic less dependent on skilled personnel, and more able to weld difficult materials such as some kinds of corrosion-resistant alloy. Among the alternatives are:

PY

friction-stir welding flash-butt welding homopolar welding electron-beam welding laser welding explosive welding

SA

M

PL

E

It is sometimes argued that these are newfangled innovations that are still in the early development stage. That is not correct. Flash-butt welding, for instance, is widely applied to rails, and in the former Soviet Union was used to construct some 30,000km of large-diameter pipeline; a recent Russian book says that it is not in use a present, though no reasons are given. McDermott bought the rights to the process from the Paton Welding Institute in Kiev, and reportedly spend $10 million on development for laybarge pipelaying in the West. McDermott was at one time enthusiastic [3], but was unable to find an operator willing to be the first to apply flash-butt, and appears now to deploy its enthusiasm elsewhere. Much effort has gone into the development of friction welding and electronbeam welding for laybarges. Homopolar welding was developed by the Center for Electromechanics at the University of Texas Austin: it makes a girth weld in a 12-in pipeline in 3sec by passing through the end butt a massive pulse of electricity, briefly reaching 15MW. The energy is kinetic energy stored in a flywheel brought back up to speed between one weld and the next, and so the power source does not need a continuous high-power service. At one time, commercialization of the homopolar system was in prospect, but it has gone quiet.

C O

• • • • • •

There is of course no reason why a pipeline should necessary be welded. High-quality premium threaded connections have been demonstrated to be reliable and able to withstand the moments and torques applied by pipelaying. Under the scheme proposed here, there would only be one threaded connection every 125m. The cost of a premium connection would become less significant, and would be paid for by the acceleration of pipelaying that it would make possible. A threaded connection can be made up in a matter of seconds.

Regrettably, the adherents of conventional welding are resistant to the welding systems that are successful in other engineering fields, and devote their energy to finding reasons not to examine alternatives. However, the perceived challenge of alternative welding have at least spurred the improvement of conventional techniques.

Conclusion Marine pipelaying can be made cheaper and faster by abandoning the traditional process of making up the pipeline from short sections on a laybarge, and instead making up much longer lengths onshore and joining them on a barge dedicated to that purpose, preferably applying a rapid process for making the girth connections between the lengths.

Acknowledgement An earlier version of this paper was presented at the Deepwater Operations Symposium in Singapore in November 2008.

References 1. C.J.London, 1991. Forties export pipeline project. Proc. Offshore Pipeline Technology Seminar, Copenhagen. 2. A.C.Palmer, J.Hammond, and R.A.King, 2008. Reducing the cost of offshore pipelines. Proc. Marine Operations Specialty Symposium, Singapore, paper MOSS-11, 275-284. 3. Flash-butt welding, video, McDermott, 1990..

The Journal of Pipeline Engineering

SA

M

PL

E

C O

PY

52

1st Quarter, 2009

53

Rehabilitation of corroded steel pipelines with epoxy repair systems by H S Costa-Mattos1, J M L Reis*1, R F Sampaio1, and V A Perrut2 1Programa de Pós-Graduação em Engenharia Mecânica, Laboratório de Mecânica Teórica e Aplicada, Universidade Federal Fluminense, Niterói, Brazil 2 Centro de Pesquisas e Desenvolvimento da Petrobrás – CENPES, Ilha do Fundão, Rio de Janeiro, Brazil

T

T

SA

M

PL

E

C O

PY

HE REHABILITATION OF corroded pipelines using epoxy repair systems is becoming a well-accepted engineering practice and an interesting alternative to the classic repair methods for metallic pipes in the oil industry, both saving time and allowing safer operation. In these repair systems, a pipe segment is reinforced by wrapping it with concentric coils of composite material after the application of epoxy filler at the corrosion defect. The technical specification ISO 24817 [1] gives requirements and recommendations for the qualification and design, installation, testing, and inspection for the external application of composite repairs to corroded or damaged pipework. Nevertheless, so far, composite repair systems are not totally effective for through-thickness corrosion defects because generally they cannot avoid leaking. The present paper presents a simple and systematic methodology for repairing leaking corrosion defects in metallic pipelines with epoxy resins. The focus is to ensure an adequate application of the epoxy filler such that the pipe will not leak after the repair. Such a procedure can be associated with a composite sleeve that will ensure a satisfactory level of structural integrity. Examples of repair systems in different damage situations are presented and analysed, showing the practical use of the proposed methodology.

HE REHABILITATION OF corroded pipelines with epoxy repair systems is becoming a well-accepted engineering practice and an interesting alternative to the classic repair methods for metallic pipes, mainly in the oil industry, saving time and allowing safer operation [2]. Since offshore platforms are hydrocarbon atmospheres, any repair method that uses equipment that produces heat and sparkling is forbidden: type B sleeves, leak clamps, and hot tapping are therefore excluded from the list of allowable repair methods. According to Ref.2, only Bolt-On Clamps with seals are allowed for leak repairs on offshore platforms. Corroded pipelines can be repaired or reinforced with a composite sleeve system, in which a pipe segment is reinforced by wrapping it with concentric coils of composite material after the application of epoxy filler at the corrosion defect. Generally, the composite sleeve is not only used as repair system itself (mainly to avoid or to restrain the

Author’s contact information: tel: +55-21-2629-5565 email: [email protected]

propagation of internal flaws), but also as a complementary procedure to enhance the reliability of weldments, eliminating the necessity of heat treatment (in the welding operation there is always a possibility of metallurgical changes in the parent metal in the vicinity of the weld). Technical specification ISO 24817 [1] gives requirements and recommendations for the qualification and design, installation, testing, and inspection of the external application of composite repairs to corroded or damaged pipework. Nevertheless, so far, composite repair systems are not effective for through-thickness corrosion defects because generally they cannot avoid leaking. The present paper presents a very simple and systematic methodology for repairing leaking corrosion defects in metallic pipelines with epoxy resins. The focus is to ensure that the pipe will not leak after a repair, and such a procedure can be associated with a composite sleeve that will further ensure a satisfactory level of structural integrity. The study is focused on what ISO 24817/TS defines as a defect type B – where the substrate requires structural reinforcement and sealing of through-wall defects (leaks) – and all three classes of repair, although mainly Class 3

The Journal of Pipeline Engineering

SA

M

PL

E

The main motivation for the study presented on this paper is leaking defects found in the produced water pipelines used in offshore oil platforms. The damages derived from corrosion process in produced water pipelines in platforms cause very important economic losses because the operation must be stopped while the repair is being performed (Fig.1). Although the operation pressure of these pipelines is not very high, the water temperature is between 60oC and 90oC, which can be a major shortcoming if polymeric materials are used as repair systems.

The objective is to ensure the pipe will not leak under the operational pressure and temperature after the repair. The maximum time allowed between the beginning of the repair and the return to operation is 75mins. Hydrostatic tests were carried out with water at room temperature and at 80oC to validate the epoxy repair systems that are applied in offshore produced water pipelines, and the experimental tests were aimed at analysing the performance of different epoxy resins in real offshore platform repair situations. Examples of repair systems in different damage situations are presented and analysed, illustrating the possibilities of practical use of the proposed methodology.

Epoxy resins Two different commercial fast-curing epoxy resins were analysed: both are two-component systems consisting of a base and solidifier. The first one (System A) is designed for leak repairs on tanks and pipes, as well as for other emergency applications, and is based on a silicon steel alloy

Fig.1. Corrosion damage in produced-water pipelines.

blended within high molecular weight polymers and oligomers. It is partly cured (machining and/or light loading) after 35mins at 25oC and is fully cured after 1hr at this temperature. Further technical data for System A includes:

C O

which is appropriate for systems transporting produced fluids.

PY

54

• • • •

flexural strength: tensile shear on steel: compressive strength: heat-distortion temperature:

59.3MPa 17.2MPa 55.8MPa 51oC

The second system (System B) is also a polymer-based system specially developed for repairs, and consisting of a mixture of epoxy resin and aluminium powder. It is partly cured (machining and/or light loading) after 18mins at 25oC and is fully cured after 40mins at this temperature. Further technical data for System B includes: • • • •

flexural strength: 67MPa tensile shear on steel: 19MPa compressive strength: 104MPa heat-distortion temperature: 120oC

Since the heat-distortion temperature for System A is very low (51oC) it was only tested at room temperature. The hydrostatic tests with pipes repaired with System 2 were performed at two different temperatures: room temperature and 80oC.

Methodology for the epoxy repair system Since epoxy repair systems do not necessarily avoid leakage, even if a composite sleeve is used, the following methodology

1st Quarter, 2009

55

Fig.2. Types of failure.

C O

In a repair of a pipeline with through-thickness defects with epoxy resins, two mechanisms of brutal failure can occur when pressure is applied, see Fig.2. The experimental procedure was designed to minimize the possibility of such failure modes.

PY

was created to improve the effectiveness of such repair systems in the produced-water pipelines used on offshore oil platforms. The experimental set-up in the laboratory was designed to approximate a real repair operation, where the resin has to be applied in field conditions (which affect the quality of the resulting epoxy repair). To optimize the process, avoiding stopping production for a long period, a maximum repair time of 75mins is suggested from the beginning of the repair procedure to the return to operation.

Through-thickness defect

E

Defect sizing

SA

M

PL

Defect sizing is important in order to define the limits for an effective use of the repair procedure. The dimensions of the defect should be determined by the smallest ellipse, with one axis parallel to the axis of the pipe, which fully contains the area of the flaw (see Fig.3). The maximum allowable defect size for the proposed repair procedure is

Fig.4. Surface preparation.

Fig.3. Defect sizing.

56

The Journal of Pipeline Engineering

⎧R ⎫ amax ≤ max ⎨ , t⎬ ⎩10 ⎭

PL

E

(1)

SA

M

where R is the inner radius of the pipe and t is the wall thickness. This means that the maximum allowable dimension for the semi-major axis a is the greatest value of either the wall thickness t or 1/10 of the inner radius R.

Proposed repair procedure

The repair methodology can be described as follows: Surface preparation Surface treatment often involves chemical reactions which produce surface modifications on adherends, or mechanical procedures, which improve adhesion by increasing mechanical interlocking of the adhesive to the adherend. In this way, the primary objective of a surface treatment is to increase the surface energy of the adherend as much as possible, and/or improve the contact between the adhesive and the adherend by increasing the contact area. Increasing roughness, or an increase in surface area, has been shown to give good results in improving adhesion. Subsequently, a relationship exists between good adhesion and bond durability.

Fig.6. Metallic wedge for smaller defects.

substrate. A final rinse with solvent was made to provide a surface free of oil, grease, and dirt surface. After this, the adhesive was mixed according to the manufacture’s procedure, and applied to the pipe. It is important to point out that, in a real situation, the pipe may be so corroded that sandpaper should be used with extreme care (see Fig.4). Also, since offshore platforms are hydrocarbon atmospheres, any method of mechanically roughening the surface that may produce heat or sparking (such as sandblasting, cutting, grinding), is unacceptable.

C O

defined by the semi-major axis of the ellipse, a, which is given by:

PY

Fig.5. Rubber cap to avoid adhesive spillage.

Introduction of an internal rubber cap to avoid spillage of epoxy resin An elliptically-shaped rubber cap must be used to avoid resin spillage inside the pipe. Since the rubber is very deformable, it is easy to introduce the cap into the pipe, and it is maintained in position using a simple system of nylon strings. The cap should allow formation of and internal layer of adhesive with approximately the same thickness as the pipe wall, and with average dimension twice the size of the defect (see Fig.5). For through-thickness defects with the semimajor axis less than or equal to 5mm, it may be difficult to introduce the rubber cap, and a metallic wedge should be used instead (Fig.6). The following steps in the repair procedure are exactly the same if either the wedge or the cap is used. Application of the first external layer of epoxy adhesive

In order to obtain these properties, sanding with 120 or 150 sandpaper was used to achieve a white metal appearance and to remove some of the existing oxide layer in the

The epoxy adhesive layer applied externally should cover an area approximately five times that of the ellipse (Fig.7), and

1st Quarter, 2009

57

Smooth Finishing

Non-Smooth Finishing

C O

E

the thickness of this first layer must be at least equal to the thickness of the pipe; the layer should also have a smooth boundary for improved performance. After application, an initial epoxy polymerization time is allowed according to the manufacturer’s instructions (the maximum desirable being 20mins).

PY

Fig.7. External epoxy adhesive layer.

PL

Application of the second layer of epoxy adhesive

A second layer of adhesive must be applied without sanding. The repair procedure is considered adequate when:

SA

M

Fig.8. Equivalent system..

Epoxy Adhesive

Pipeline with throughthickness defect

Rubber Cap to avoid adhesive spilling

Metallic Clamp

Rubber Band Fig.9. Complete repair system.

58

The Journal of Pipeline Engineering

Fig.10. Test apparatus. a ⎞ ⎛ PR ⎞ ⎛ ⎜1 + 2 ⎟ ⎜ ⎟ ≥ σ y b ⎠⎝ t ⎠ ⎝

structural integrity of the pipe, but to prevent the two possible major failure mechanisms of the adhesive repair shown in Fig.2, mainly at the beginning of operation when the resin may not be fully cured.

(2)

E

)

SA

M

PL

(

C O

The stress distribution in a general through-thickness corrosion defect is very complex, but if the size of the defect is limited, a rough estimate of the magnitude of the permanent deformation close to the defect can be performed. The term on the left-hand side of Eqn 2 is the maximum stress in a thin-walled infinite plate with an elliptical defect with semi axes a and b subjected to traction of a uniform force per unit area S = PR/t (see Fig.8). The stress concentration factor in this case is Kt = 1 + 2 ab . The criterion in Ref.2 states that a permanent deformation close to the defect in a pipe can be neglected when KtS is less than the yield stress sy. For closed-ended pipes, the yield stress should be adjusted by a factor of 1.115 [3].

If this condition is verified, immediately after the application of the second epoxy layer a rubber sheet should be applied over the repair around the perimeter and a simple metallic clamp, similar to those used for garden hoses, can be attached (Fig.9). The clamp is not used to improve the

Under these circumstances, the proposed procedure is effective as a repair system by itself. Nevertheless, this procedure is intended to be used together with a composite sleeve (which is normalized, for instance, by Ref.1). The main objective is to ensure that composite repairs of leaking defects when qualified, designed, installed, and inspected using ISO/TS 24817 and the proposed procedure, will meet the specified performance requirements. The suggestion is to apply the epoxy resin as described in this paper and then apply a composite material sleeve, of a normalized thickness, to restrain the plastic strain and to assure a satisfactory level of structural integrity.

PY

where a and b are, respectively, the semi-major and the semiminor axis of the ellipse, R is the inner radius of the pipe, t the wall thickness and sy the yield stress of the pipe material.

An alternative method for defining the necessary thickness of composite material to ensure both the safety of repairs under operational conditions and the lifetime extension under operational conditions, can be found in Ref.4. This method, although simple, is acceptable for different failure mechanisms, including plasticity, fatigue, and fracture. The method meets the most widely-used criteria for the assessment of corrosion defects under internal pressure loading – a family of criteria described in [5] as the effectivearea methods. These include the ASME B31G criterion and the RSTRENG 0.85 criterion (also known as the

Fig.11 – Detailed temperature control system: 1 – the pressured water machine connection; 2 – the temperature control thermostat; 3 – the electrical resistance.

1st Quarter, 2009

59

Fig.12. 12-in SCH-0 steel pipe with a 10-mm repaired hole, before and after testing.

.

As a second step, the specimens were repaired with system B (no sleeves were used, only the clamp) and submitted to five pressure cycles (60mins at 30 kg/cm2) with the water temperature inside the specimen at 80ºC, increased while the water was at atmospheric pressure. The internal pressure was not increased until after the temperature had stabilized. After each pressure cycle, the specimen was cooled to room temperature, and each specimen was therefore also submitted to five temperature cycles during testing.

PL

E

Results and discussion

to check any eventual small leaks or reinforcement disbonding.

C O

modified B31G criterion). Nevertheless, this study is mainly focused on metal loss due to corrosion treated as a partthrough-wall defect in the pipe, and not on throughthickness defects.

PY

Fig.13. Deformed end cap after testing at 60kg/cm2 and 80ºC.

SA

M

An experimental set-up was designed to examine the effectiveness of the methodology, approximating to a real repair operation as far as possible. Five different specimens of API 5L grade B steel pipes, normally used in offshore platform for produced water, were used as the specimens for hydrostatic tests:

• specimen 1: 2-in diameter Schedule 80 pipe with a 3-mm diameter circular hole • specimen 2: 2-in diameter Schedule 80 pipe with a 10-mm diameter circular hole • specimen 3: 12-in diameter Schedule 20 pipe, 1300mm long, with a 10-mm diameter circular hole • specimen 4: 12-in diameter Schedule 20 pipe, 1300mm long, with a 30-mm diameter circular hole • specimen 5: 3.5-in diameter Schedule 20 pipe, 1000mm long, taken from the field with real corrosion defects (see Fig.4). Initially, all the repaired specimens (no composites sleeves were used, only the clamp) with the two systems were submitted to a classical hydrostatic test at room temperature to evaluate its strength and effectiveness. The maximum allowable time for each repair was 60mins, and all tests began exactly 75mins after the start of the repair process. In the tests, the pipe pressure was raised to 30kg/cm2 and maintained at this level for 60mins. After five cycles, if the repair did not fail, the specimen was unloaded and inspected

Once again, the maximum allowable time for each repair was 60mins, and all tests began exactly 75mins after the beginning of the repair. The temperature level of 80ºC was chosen in order to simulate average offshore fluid conditions. The system to control water temperature inside the specimens was designed specially for this procedure, and the whole system (including the electrical resistance) was installed at one end of the specimen, as can be seen in Figs 10 and 11. All the repairs performed with Systems A and B using the above methodology withstood the five pressure cycles with water at room temperature. The repairs also resisted the high-pressure tests; it was not possible to obtain a failure pressure since the pipe end caps were not designed for burst testing and they deformed plastically and failed before the repair failed, as can be seen in Fig.13. If the proposed procedure is not adopted, however, the repair may not be able to resist the loading. Table 1 shows the failure pressure obtained for specimen 2 – the 2-in

60

The Journal of Pipeline Engineering

F a ilu re p re ssu re (k g /c m 2 )

Te st 1

8 .9 2

2

17.64

3

16.17

4

18.35

5

14.27

Av e r ag e

1 5 .0 7

Table 1. Failure pressure for specimen 2 if the repair procedure is not adopted.

F a ilu re p re ssu re (k g /c m 2 )

Te st

20.18 (f ir st cycle )

2

4.92 (se con d cycle )

3

30.00 (f ir st cycle - af te r 10 m in )

4

13.92 (f ir st cycle )

5

9.84 (f ir st cycle )

E

diameter Schedule 80 pipe with a 10-mm diameter circular hole – repaired using system A (no cap and no clamp).

SA

M

PL

All the pipes repaired with System B at 80ºC resisted for the five cycles. In order to decide whether a given epoxy system can be used at higher temperatures, it is suggested that the same conditions are used as presented in Ref.1 for composite sleeves: “For a design temperature greater than 40oC the repair system shall not be used at a temperature higher than the glass transition temperature (Tg) less 30oC. For repair systems where Tg cannot be measured, the repair system shall not be used above the heat-distortion temperature less 20oC. For repair systems which do not exhibit a clear transition point, i.e. a significant reduction in mechanical properties at elevated temperatures, then an upper temperature limit, Tm, shall be defined (or quoted) by the repair supplier.” As an example, the failure pressures observed in hydrostatic tests performed with specimen 4 (which has heat-distortion temperature of 51oC) repaired using system A at 80oC are presented in Table 2.

It is interesting to note that the adhesive System A behaved surprisingly well when the proposed repair procedure was adopted, even at temperatures above the heat-distortion temperature. All the repairs resisted to five cycles at 80oC in tests performed on specimens 1, 2, and 3.

Conclusions The present work is a first step towards the definition of

Table 2. Failure pressure for specimen 4 at 80oC.

C O

PY

1

safer and more-reliable procedure for applying epoxy repair systems to through-thickness flaws caused by corrosion in metallic pipelines. This procedure is designed to be used together with a composite-sleeve repair system (which is normalized, for instance, by the ISO technical specification 24817). The proposal is to apply the epoxy resin as described in this paper and then to apply a composite material sleeve, with a normalized thickness, to restrain the plastic strain and to ensure a satisfactory level of structural integrity. The main objective is to ensure that composite repairs to leaking defects when qualified, designed, installed, and inspected using ISO/TS 24817, and also the proposed complementary procedure, will meet the specified performance requirements. The main requirements for epoxy resins to be used as repair systems are: fast curing, high heat-distortion temperature, and a thermal expansion coefficient similar to that of the material of the pipe. The full validation of this simplified repair methodology still requires an extensive programme of experimental investigation, mainly concerning fatigue, creep, ageing, and resistance to UV degradation and weathering.

References 1. ISO Technical Specification 24817, 2006. Petroleum, petrochemical and natural gas industries - composite repairs for pipework - qualification and design, installation, testing and inspection. 2. C.A.Jaske, B.O.Hart, and W.A.Bruce, 2006. Pipeline repair manual. Pipeline Research Council International, Inc. Virginia.

1st Quarter, 2009

61

Mechanical Sciences and Engineering, ISBN 978-85-8576930-7, pp485 – 496. 5. D.R.Stephens and R.B.Francini, 2000. A review and evaluation of remaining strength criteria for corrosion defects in transmission pipelines. ETCE2000/OGPT-10255, Proceedings of ETCE/OMAE2000 Joint Conference, Energy for the New Millenium, New Orleans, USA, 2000.

PY

3. A.T.de Mello dos Santos, 2006. Simplified analysis of the caps influence in elasto-plastic pipe burst tests. MSc Thesis, Universidade Federal Fluminense, January. 4. H.Costa Mattos, R.F.Sampaio, J.M.L.Reis, and V.A.Perrut, 2007. Rehabilitation of corroded steel pipelines with epoxy repair systems. In: Solid mechanics in Brazil 2007, Eds M.Alves and H.S.da Costa Mattos, Brazilian Society of

motivator is the Energy Independence and Security Act of 2007, which President Bush signed in December. The law requires that American fuel producers use 36bn gallons of renewable fuels by 2022, with is more than five times what is currently used.

SA

M

PL

E

take into consideration a refined product travelling the same distance, it would probably be under a nickel. So if you could get ethanol from Illinois to the East Coast for 12 cents a gallon cheaper than you can today, obviously a lot would change in the world, and the interest in ethanol would increase,” Robert White, director of operations for the Omaha, Nebraska-based Ethanol Promotion and Information Council, said. John Urbanchuk agreed, “The cost of shipping ethanol would be about the same as it is to ship gasoline through a pipeline.”

C O

Editorial (continued form page 4)

In the US, ethanol is primarily moved by rail and road tankers, which are costly and time-consuming methods of transportation. While a pipeline should help in the cost of distributing ethanol (and presumably in the cost of consuming it), it could also take away jobs from the rail and trucking industry, particularly in the Midwest. Most experts admit there’s something of a chicken-and-the-egg effect as companies consider shipping ethanol. The production of ethanol isn’t high enough today to create a desperate need for pipelines, but without a pipeline infrastructure in place, companies are hesitant to produce more ethanol. One

But the biggest challenge to Magellan and Buckeye now may not be moving products through an ethanol pipeline, but moving funding through the federal government pipeline. “They have reached out to Congress and said, ‘Some of this has merit, but we need some support and we need to know how much that support’s going to be, because we need to make business decisions on our end,’” Mr White said. “[This $3-bn project is] a substantial investment that you have to recapitalize somehow.” The companies’ press release repeatedly stresses the importance of government support: “Congressional support and assistance is necessary for a project of this nature given the changing federal policies associated with renewable fuels.” The Energy Act included a provision requiring that the US government undertakes its own feasibility study on ethanol-dedicated pipelines, and this study is due to be released in 2010.

The Journal of Pipeline Engineering

SA

M

PL

E

C O

PY

62

1st Quarter, 2009

63

Technical and commercial challenges in procurement and implementation of major international pipeline projects by Assadollah Maleknejad Vice President – Finance and Economic Affairs, Pars Oil Co, Tehran, Iran

O

PY

IL AND GAS pipelines have significant potential for enhancing stability and improving living standards in the host countries. Pipeline projects require significant investment to diversify upstream energy supply, downstream security of demand, and enhanced midstream transport infrastructure to increase market access and interconnectivity for both. Nevertheless, international pipeline projects face considerable technical and commercial challenges in their procurement and implementation. There are considerable risks attached to pipeline projects, pertaining to the techno-economic viability of the projects. However, technical risks can be seen as challenges that have to be faced to make the projects successful.

PL

E

the possibility of construction cost overruns the possibility of construction delays the possibility of operating cost overruns can the project be completed within acceptable performance levels? the possibility that the project will not operate continuously the possibility that the project will not operate according to environmental requirements

M

• • • • • •

C O

The technical challenges include:

SA

There are also other risks in respect of economic and security aspects as well as the investment climate in the region. However, these risks can also be seen as challenges that have to be faced to make the projects successful. These include: • • • • • • •

will the project fail due to economic pressuresor contract failure? is there a possibility of interest increases? is there a possibility of revenue reductions? is there a possibility of interest increases? will the project be affected by political events, such as legal changes, exchange rates, etc.? will the project operate under inconvertibility and limited transferability of currency? will the project operate under the risk that the government may unilaterally change the economic or fiscal conditions on which it is based?

This paper presents a risk analysis and risk mitigation methodology, and addresses the concerns of the stakeholders in relation to cross-border pipeline projects.

I

N STRUCTURING a finance package for an oil or gas pipeline project, it is essential that the project risks be identified, and how they will be allocated to the parties involved who include shareholders, banks, contractors, and governments. The ultimate allocation of these risks is a matter of negotiation, which can only be successfully concluded when all parties have agreed to what extent they are willing to commit funds and bear certain risks.

The key to successfully financing an oil or gas pipeline project is the equable allocation of risks among the project participants. Risks must be allocated to the party that is in the best position to manage them, and a rigorous methodology must be followed including: • risk identification • risk evaluation (sensitivity analysis)

64

The Journal of Pipeline Engineering

• risk mitigation • risk allocation (commercial negotiation), and • contractual documentation

operational performance. The construction risk is considered by the providers of finance to be relatively high, and potentially involves significant losses. It is therefore a most important financial risk; for example, should the project fail during the construction phase, the security over the assets of the project would be of little value. Thus, providers of finance do not want to take the construction risk, and normally ask for recourse to the sponsors’ other resources until the project is completed and tested.

The risks discussed below are described from a financial point of view and have been separated into the three phases of the project: pre-development, construction, and operation.

Pre-development

Operation phase

This is the phase in which relatively small amounts of money have been, or are being, spent. The oil or gas pipeline project may fail for a number of reasons, but the financial risks for the shareholders are limited.

SA

PY

M

PL

E

Of particular importance to the project during the predevelopment stage is how to deal with the perceived country risk exposure, and the ability of providers of finance to bear this risk. Providers of finance will generally not wish to commit money, time, or effort if the political risks of the project are not fully mitigated. Due to importance of this parameter, the political risks and required mitigations will be discussed further later in this paper.

• guarantees from suppliers of equipment for technical performance • supply guarantee by means of “supply or pay contract” from the supplier of the oil or gas • off-take guarantee by means of “take or pay contract” from the off-taker of the oil or gas • guarantees against political events (see below) • recruiting qualified operator • establishment of regional support and cooperation between countries in case of cross-border pipelines

C O

During this phase, providers of finance will form preliminary conclusions about whether the project is financeable. Due to the various uncertainties in this stage, the providers of finance will ascertain whether the project is based on solid technical, economic, and legal foundations. Once the primary parameters have been agreed, it will be necessary to establish the chances of survival of the project under various pessimistic conditions, including higher capital costs, delay of start-up, higher operating costs, higher financing costs, or combinations of any of these.

Once the pipeline is in operation, the main concern is that it may not operate on a continuing basis within acceptable economic, technical, and environmental parameters. Such operational risks are numerous, and are borne by the project and its limited-recourse providers of finance. However, the project can hedge against the risks through contractual and guarantee arrangements that in effect transfer some of these to other parties. The following are examples:

Construction phase During the construction phase, the main risk is failure to complete the project within acceptable performance levels and an acceptable timeframe and budget. In the case of oil and gas projects, one of the key risks is the ability of the contractor to construct and lay the pipeline according to acceptable standards: API certification and OA during construction, as well as independent technical due diligence, are normally sufficient mitigation for financiers. Furthermore, recognized experienced project management will be a requirement. Risks with respect to earthquakes and ground movement, as well as accessibility of the terrain, will be addressed prior to construction, and should be mitigated by technical solutions in line with the independent technical consultant’s recommendations. Construction risk falls on the project and its sponsors who, in turn, may be able to hedge their risk by purchasing various forms of insurance and obtaining guarantees from contractors regarding costs, completion schedules, and

Environmental issues It is important that the project should operate according to the latest environmental requirements, and providers of finance require an independently-performed environmental impact audit. This serves two purposes: firstly, it establishes a baseline environmental status, and secondly, it provides reassurance that (it is assumed) state-of-the-art standards are being implemented, meeting – as a minimum – both World Bank and local standards. With respect to a baseline report, this would mitigate exposure to damages resulting from existing environmental disturbances.

Political risks Throughout the project phases, political risk is of prime concern. Examples of political risk include (but are not limited to): • Inconvertibility and limited transferability of currency: the risk that sufficient hard-currency funds are not available to meet foreign currency obligations towards the project.

• Breach of contract by the host or transit government to the material rights of the project, such as through taxation terms, approvals, export rights, and production rights; restrictions on import or export of equipment; prohibition of the entry of personnel. Providers of finance will therefore, prior to committing to finance, evaluate the political risks compared to the exposure they are willing to take in combination with politicallyinsured funding from multilateral agencies (MLA) such as the World Bank and/or export credit guarantee agencies (ECGA).

An analysis of the risks from the perspective of the sponsors will consist of understanding the risks as identified by a number of independent experts and determining the ultimate effect on the viability of the project to service debt. The services provided by independent experts will include: • technical advice with respect to supply • technical advice with respect to procurement, construction, costs, timing, and environment • insurance advice • market advice with respect to offtake • legal opinions with respect to local jurisdiction, security aspects, taxation, recovery of foreign currency • legal opinions of the sponsors

Assuming a satisfactory outcome of this advice, a financial analysis will be completed, and a base case will be generated and the effect of possible construction cost overruns, construction delays, operating cost overruns, revenue reductions, and interest increases will be modelled for the purpose of selecting ratios in which the project may proceed for construction.

PL

E

As already discussed in this paper, the sponsors of the project will generally not be able to commit money, time, and effort on any international pipeline if the financial institutions (MAL and/or ECGA) are unwilling to establish exposure to political risks. Even though the sponsors may believe there will be opportunities for MAL and/or ECGA to participate in funding arrangements in the future, they may be unable at this moment to give a positive indication to proceed with the project.

Risk-analysis procedures

PY

• War, revolution, and civil war.

MLA and ECGA financing is available. At the same time, the MLA will require the host countries to provide guarantees against political events, legal changes, taxation changes, exchange rate fluctuations, and investment recovery by means of political support. If legislation for this is seen as unsatisfactory in the host countries, the guarantees will be based on offshore receivables, reserves, and disbursement accounts.

C O

• Expropriation or nationalization by the government: the risk that the government unilaterally takes control of a project or changes the economic or fiscal conditions (creeping expropriation) under which a project operates.

SA

M

Once MLA and ECGA funding is available in principle, more interest can be expected from commercial banks and capital markets to participate into the project. They would find reassurance in the political circumstances whenever

M

SA E

PL C O PY