Risk Assessment of Petroleum Pipelines Using A Com-Bined Analytical Hierarchy Process - Fault Tree Analysis (AHP-FTA)
Risk Assessment of Petroleum Pipelines Using A Com-Bined Analytical Hierarchy Process - Fault Tree Analysis (AHP-FTA)
Risk Assessment of Petroleum Pipelines Using A Com-Bined Analytical Hierarchy Process - Fault Tree Analysis (AHP-FTA)
1 Introduction
491
Dawotola, van Gelder, Vrijling: Risk assessment of petroleum pipelines
Risk assessment of petroleum pipelines entails the study of failures and consequences of
pipelines in terms of possible damage to property, human hazards, and the environment.
Ideally, most pipeline operators ensure that during the design stage, safety provisions are
created to provide a theoretical minimum failure rate for the life of the pipeline. There are
approaches such as corrosion control and routine based maintenance to ensure reliability of
pipelines during service.
Transmission pipelines are complex in nature, and their risk analysis could be simplified
by using a hierarchical approach, Huipeng Li (2007). However, little has been achieved on
hierarchical risk analysis of petroleum pipelines, as an aid to decision analysis, which is
required in making inspection and maintenance decisions. AHP is a promising method for
this application. AHP, developed by Saaty fundamentally works by using opinions of ex-
perts in developing priorities for alternatives and the criteria used to judge the alternatives
in a system, Saaty (1999). The outcome is a relative scale which gives managers a rational
basis for decision making. It has found applications in diverse industries, such as agricul-
ture, Quresh and Harrison (2003), oil and gas, Dey et al (2004), Al-Khalill (2005), Nonis et
al (2007), Brito and Almeida (2009), and the public sector, Dey (2002). Majority of AHP
applications in pipelines sector, Dey et al (2004), Al-Khalill (2005), Brito et al (2009) have
concentrated mainly on qualitative risk analysis and there has not been adequate applica-
tion of AHP in achieving quantitative risk assessment.
The focus of this paper is the introduction of a framework that implements combined ana-
lytic hierarchy process and fault tree analysis for quantitative risk assessment of petroleum
pipelines. The combined AHP-FTA is particularly suitable where sufficient data is lacking
to carry out a full scale Fault Tree Analysis. The case study of cross-country oil and gas
pipelines is used to illustrate the proposed methodology.
And the set, R of the above three questions can be mathematically represented as:
492
Van Gelder, Proske & Vrijling: Proceedings of the 7th International Probabilistic Workshop, Delft 2009
• Problem formulation – the ultimate goal of the AHP is defined. In this paper, it is the
determination of risk due to oil spillage from petroleum pipeline. After the goal defini-
tion, contributing factors to the failure are then identified. If applicable, these factors
are further divided into 1 or 2 sub factors. The major factors and sub factors responsi-
ble for product loss from a pipeline are presented in Fig. 1.
• Facility segmentation – The pipeline is divided into different sections based on their
peculiar similarities. The different sections are the alternatives of the decision making
process.
• Collection of pipeline information - Required features for the pipelines is divided into
physical data, construction data, operational data, inspection data and Failure history.
This information is documented for the hierarchical analysis.
• The next step is the development of a hierarchy structure, which consists of the goal of
the risk assessment, the failure factors and subfactors, if applicable and the pipeline
stretches.
• In the last step of the analytical hierarchy process, data of the pipelines are made avail-
able to a number of experts who the carry out a pairwise comparison of the pipeline
segments with respect to each risk factor (failure criteria). The outcome of the compari-
son is a matrix that ranks the pipeline stretches in order of the likelihood of failure.
• Consistency check: AHP provides the possibility of checking the logical consistency
of the pairwise matrix by calculating the consistency ratio (CR). AHP judgement is ac-
ceptable if CR is less than 0.1
493
Dawotola, van Gelder, Vrijling: Risk assessment of petroleum pipelines
G 1 n
λmax is the average of the elements of vector c : λmax = ∑ ci
n i =1
(4)
λmax − n
Consistency Index is then calculated using, CI = (5)
n −1
Where n is order of the decision matrix and λmax is obtained from equation (4) above.
CI
Using equation (4), Consistency Ratio is calculated as, CR = (6)
RI
Where RI is the random index, and its value is obtained from table 1 below.
n 3 4 5 6 7 8 9 >9
RI 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.49
Other measures of consistency have been defined. For example, J. Mustajoki and
R.P.Hämäläinen (2000) give a Consistency Measure (CM) of between 0 to 1 using the
Multi Attribute Value Theory inherent in their Web-HIPRE software. A CM of 0.2 is con-
sidered acceptable.
2 r ( i, j ) − r (i, j )
CM = ∑
n(n − 1) i > j (1 + r (i, j ))(1 + r (i, j ))
(7)
Where r ( i, j ) = max a(i, k )a(k , j ) , k ∈ {1,..., n} is the extended bound of the comparison
matrix element a(i, j ) , and r (i, j ) is the inverse of r (i, j ) . CM gives and indication of the
size of the extended region formed by the set of local preferences, when wi ≤ r ( i, j ) w j for
all i, j ∈ {1,..., n}
494
Van Gelder, Proske & Vrijling: Proceedings of the 7th International Probabilistic Workshop, Delft 2009
• Fault tree analysis of important failure factors: Fault Tree Analysis of the most impor-
tant failure factor is carried out to determine acceptability of risk. Using the FTA and
AHP results, the overall failure probability is then calculated.
Sabotage
Corrosion Error Defect
495
Dawotola, van Gelder, Vrijling: Risk assessment of petroleum pipelines
3.1 Introduction
Oil and gas pipelines located in 3 different parts of Nigeria: EL Pipeline (Escravos), AB
Pipeline (Warri), and AZ Pipeline (Benin) (Fig. 2) were considered as a case study. A
summary of the characteristics of the pipelines are shown in tab. 2 below. The goal of the
research is to conduct a risk assessment of given pipelines using Analytical Hierarchy
Process (AHP) method. This is achieved by determining the relative contribution of differ-
ent failure factors to the overall pipeline failure. Pipeline failure is defined as loss of struc-
tural integrity of pipeline which may lead to unintended products loss.
Based on literature research conducted, Adebayo and Dada (2008), L.P.E YO-Essien
(2008), NNPC (2008) five factors have been identified as being mostly responsible for
pipeline failures in Nigeria, namely external interference, corrosion, structural defects, op-
erational defects, and other minor failures.
A total of six pipeline experts participated in the expert judgement study on risk assess-
ment of the petroleum pipelines. The affiliations of the experts are in the following organi-
sations: Shell International, Chevron Exploration, BJ Services, Nigeria Petroleum
Development Company (NPDC), Nigeria National Petroleum Company (NNPC), and
SBM Offshore. Attributes of the pipelines and a pipeline historical failure records sheet
containing defining characteristics of the pipelines were made available to the experts with
an AHP questionnaire.
A hierarchy tree of the three pipelines is constructed using the Web-HIPRE software, J.
Mustajoki and R.P.Hämäläinen (2000). The tree (fig. 3) contains information on the goal
(Risk Based Pipeline selection), criteria (failure factors) and sub-criteria (sub division of
failure factors). The decision alternatives are the three pipelines under consideration.
496
Van Gelder, Proske & Vrijling: Proceedings of the 7th International Probabilistic Workshop, Delft 2009
Fig. 3: Hierarchy tree for the risk assessment of EL Pipeline, AB Pipeline and AZ Pipeline
Tab. 3: Pair wise ranking of failure criteria and likelihood of failure of the three pipelines
Factors Likeli- Sub-Factors Likeli- EL Pipeline AB Pipeline AZ Pipeline
hood hood
Others 0.044
497
Dawotola, van Gelder, Vrijling: Risk assessment of petroleum pipelines
The overall consistency measure (CM) of the group matrix is 0.195 which is considered
consistent with the 0.2 stated by R.P.Hämäläinen (2009).
Fig 4 shows that the most significant failure criterion for the pipelines is external interfer-
ence followed by corrosion, with relative likelihood of failure of 0.607 and 0.214 respec-
tively.
From Fig.5 it can be seen that sabotage remains the most significant failure criterion for the
three pipelines. EL Pipeline is the most vulnerable among the three pipelines. This is ex-
pected considering its proximity to the Niger-Delta region of Nigeria.
498
Van Gelder, Proske & Vrijling: Proceedings of the 7th International Probabilistic Workshop, Delft 2009
The failure probability of petroleum pipeline is modelled as sum of failure probability due
to external interference, corrosion, construction defect, operational errors and other minor
failures.
However, based on table 3, once the probability of failure of one of the factors is known,
the total probability, P (Total failure of petroleum pipeline) can be determined.
For example, using
P (failure of pipeline due to corrosion) = 0.214* P (Total failure of pipeline) (9)
P (failure of pipeline due to corrosion) can be determined from parameters in the fault tree
(Fig. 6). The outcome from Fault tree is a quantitative data which would be valuable in-
formation in determining individual or societal risk acceptance, Vrijling et al (2004).
External Operational
Construction Corrosion Others
interference Error
499
Dawotola, van Gelder, Vrijling: Risk assessment of petroleum pipelines
4 Conclusions
This paper proposes a combined AHP-FTA methodology for risk assessment of cross
country pipelines. We achieve a number of important conclusions.
A ranking methodology has been presented which uses available data and structured judg-
ment to select operating pipeline based on the risk of failure. The results can be further
used to achieve a quantitative risk analysis through the implementation of fault tree analy-
sis. The approach is capable of reducing the rigour of quantitative risk analysis by focusing
on the most important criteria.
The fault tree analysis of AHP Failure criterion would be further investigated in future
works and be compared to available failure data. A major setback of the approach is the
subjective nature of the AHP methodology and a structured expert judgement study will be
investigated in the future.
References
[1] J. Aczel, T.L. Saaty, “Procedures for synthesizing ratio judgements”, Journal of Ma-
thematical Psychology 27 (1983), pp 93 - 10
[2] A. Adebayo, A.S Dada 2008 “An Evaluation of the causes of oil pipeline incidents In
Oil and Gas Industries in Niger Delta Region of Nigeria”, J. Eng. Applied Sci., 3(3):
pp 279-281
[4] E.N.Allouche and A.L. Bowman 2006 “Holistic Approach for Assessing the Vulner-
ability of Buries pipelines to Earthquake Loads”, Natural Hazards Review.Vol 7 No
1 pp12-18
[5] A.J. Brito and A.T de Almeida 2009, “Multiattribute risk assessment for risk ranking
of natural gas pipelines”, Reliability Engrg and Sys Safety 94 pp187-198
[6] P.K Dey 2002 “Benchmarling project management practices of Caribbean organiza-
tions using analytic hierarchy process”, Benchmarking: An international journal, Vol
9 No 4.
[7] P K Dey et al 2004 “Risk -based maintenance model for offshore oil and gas pipelines:
a case study”, Journal of Quality in Maintenance Engineering, Vol. 10 Number 3 pp.
169-183.
500
Van Gelder, Proske & Vrijling: Proceedings of the 7th International Probabilistic Workshop, Delft 2009
[9] Huipeng Li 2007, “Hierrachial Risk Assessment of water supply systems”, PhD thesis,
Lougborough University, Leicestershire, UK
[11] NEITI (2006) “Report on the process and audit 1999-2004 refineries and product im-
portation”, Appendix A: Schematics, NEITI Nigeria.
[12] NNPC 2008, “Annual Statistical Bulletin”, Corporate Planning and Development
Division (CPDD), NNPC, 2008.
[13] C.N Nonis et al 2007 “Investigation of an AHP Based Multicriteria weighting scheme
for GIS routing of cross country pipeline projects”, 24th int. sym. on aut. and rob. in
constr.
[14] M.E. Quresh, S.R. Harrison 2003, “Application of the Analytical Hierarchy process to
Riparian Revegetation Policy options”, Small-scale Forest Economics, Mgt. and Pol-
icy 2(3).
[15] T.L. Saaty 1999, “The seven pillars of the Analytic Hierarchy Process”, in proc.
ISAHP, Kobe.
[16] L.P.E Yo-Essien 2008, “Oil Spill Management in Nigeria: Challenges of Pipeline
Vandalism in the Niger Delta Region of Nigeria”, in Proc. Int. Petr. Env. Conf., 2008.
[17] J.K. Vrijling et al 2004 “A framework for risk criteria for critical infrastructures: fun-
damentals and case studies in the Netherlands”, Journal of Risk Research 7 (6), pp
569–579.
501