The Use of AHP in Risk Ranking and Maintenance Planning for Cross-country Pipelines

1. The Use of AHP in Risk Ranking and Maintenance Planning for Cross-country Pipelines ARE 520 TERM PAPER PRESENTATION (Term 021) PREPARED FOR DR. SADI A. ASSAF BY AHMED D. IBRAHIM (210185) 06 JANUARY 2003

2. AGENDA • INTRODUCTION • OBJECTIVE OF THE PAPER • PREVIOUS STUDIES • RESEARCH METHODOLOGY • STUDY DATA • DISCUSSIONS OF RESULTS • CONCLUSIONS AND RECOMMENDATIONS

3. Introduction • Cross-country pipelines are the most efficient, safest, environmentally friendly, and economical way to ship hydrocarbons over long distances. • Transports significant portion of many nations’ energy requirements • Economies depend on their smooth and uninterrupted operation • Failure could lead to substantial losses • to the overall national economy and • may even pose environmental hazards to humans For Example: • Saudi Aramco operates around 17000KM of cross-country pipelines • Most are over 25 years old and passes through residential areas of the Eastern province of Saudi Arabia.

4. Introduction • Failure can be very catastrophic. Hopkins, (1994) reports that: • 51 people were burnt to death in Venezuela in 1993 • A 914mm pipeline in New Jersey in 1994 caused a loss of one life and injuries to over 50 people • Similar failures have occurred in the UK, Canada, Russia, Pakistan, and India • Operators are continually developing in-house maintenance policies to replace the rules-of-thumb inspection and maintenance practices. • There are modern methodologies which ensure the structural integrity of an operating pipeline without taking it out of service [Jamieson, 1986].

5. Objective of the paper • Develop a model for ranking the risk level associated with operating cross-country pipelines in order to be able: • to predict the likelihood of failure and estimate their severity • For the purpose of annual maintenance planning

6. Previous Studies Causes of pipeline failure: Source: Caseley, 1994

7. Previous Studies Risk-based Decision-making Biagiotti and Goose (2000):

8. Previous Studies • The total failure cost of pipelines consists of several components: • Cost of lost product • Loss of revenue • Cost of the pipeline repair or replacement • Liability costs • Property damage costs • Socio-political costs • Benefit of replacing the pipe segment. • The first six items are added and the seventh is subtracted therefrom.

9. Previous Studies • Methods of risk-based planning of pipeline maintenance • The approaches can be classified into two major categories: • Subjective Index Methods – qualitative approaches • Quantitative Risk Assessment • Consequence estimation • Probability estimation

10. Research Methodology • A comprehensive listing of failure causes were compiled from • Literature • Interview with 5 engineers having over 30 years experience in pipeline engineering, operation and maintenance. • Evaluation data such as repair history, inspection records, design parameters and current operating conditions were collected. • AHP model of these failure causes was built • A top-down pairwise comparison was conducted by the engineers to determine the likelihood of pipeline failure. • Elements at the first criteria level were compared, followed by pairwise comparison at the sub-criteria level and then finally the alternative level

11. Research Methodology • The costs of failure of the pipelines under study were estimated for each failure cause: • Based on previous failure costs • Other industry resources • Monte Carlo simulation method, using the RandDiscrete function of a RiskSim spreadsheet, was utilized to determine the total possible cost of failure using • Estimated failure costs • Probabilities generated for each pipeline by the AHP model

12. Study Data • The pipelines under study, are operated and maintained by Saudi Aramco.

13. Study Data The major categories of pipeline failure are: • Corrosion – internal or external depending on the location of initiation on the pipeline. • Mid-wall defects - stress corrosion cracking and hydrogen induced cracking. • External interference – malicious (sabotage or pilferage), third party or natural calamities • Construction and Materials defect • Operational problems – human and operational error, equipment malfunction. • Loss of ground support • Others

14. AHP Model

15. Results and Discussions Pairwise comparison for the 1st level shows: • Corrosion contributes 48.3% to the likelihood of a pipeline failure. • This corroborates with the real situation as indicated by the company experts and with global trend. • External interference, sometimes called 3rd party damage, follows with 20.7% likelihood of a failure.

16. Results and Discussions Pairwise comparison for the 2nd level shows that: • External corrosion (80%) has more impact on pipeline failure than internal corrosion (20%). • Hydrogen-induced cracking contribute 83.3% of the pipeline failure due to mid-wall defects whereas stress corrosion cracking contributes 16.7%. • Failure of protection devices is more likely to cause a pipeline failure than operational or human error because of sophistication in the protection systems and strong operator training programs • Material defects showed a percentage of 80% over construction defects (20%). This is explained by the strong controls that are well established for pipeline construction industry.

17. Results and Discussions Synthesis of the AHP model was performed.

18. Results and Discussions • Summary of the final outcomes of each pipeline against the risk factors are shown in the next slide • Both the local and global probabilities of each of the nine pipelines were computed against each of the risk factors.

20. Results and Discussions

21. Results and Discussions

22. Results and Discussions Summary of the likelihood and severity (in terms of cost) of pipeline

23. Conclusions • This paper describes the use of AHP / Monte Carlo simulation to determine the risk level associated with operating the cross-country pipelines. • The study reveals the effect of certain risk factors on the failure of pipelines/pipeline sections. • The developed model will enable management formulate cost-effective, customized, flexible and systematic maintenance program

24. Recommendations • The risk level obtained in terms of monetary values should be a basis for maintenance planning and prioritization of pipelines • More attention should be given to pipelines that have higher failure consequences even though they have less likelihood of failure. • A semi-annual meeting should be held with representatives of operation, maintenance, and engineering departments to share ideas and in deciding the criteria and for pairwise comparison. • A further work on the calculation of the benefits versus cost ratio of the recommended maintenance strategy is recommended. • The sensitivity analysis feature available in the expert choice can be used to resolve the conflicts that may occur when a team is assigned to conduct the pairwise comparison.

25. THANK YOU COMMENTS