Analysis on accident-causing factors of urban buried gas pipeline network by combining DEMATEL, ISM and BN methods. (September 2019)

• Record Type: Journal Article
• Authors: Li, Feng; Wang, Wenhe; Dubljevic, Stevan; Khan, Faisal; Xu, Jiang; Yi, Jun
• Is Part Of: Journal of loss prevention in the process industries. Volume 61(2019)
• Page Start: 49
• Page End: 57
• Publication Date: 2019-09
• Subjects: Accident-causing analysis -- DEMATEL -- ISM -- Bayesian network (BN) -- Safety engineering -- Gas pipeline safety; Chemical industries -- Safety measures -- Periodicals; 660.2804
• DOI: 10.1016/j.jlp.2019.06.001
• Languages: English
• ISSNs: 0950-4230

Abstract

Abstract: The development of natural gas industry relies on safe and dependable pipeline networks. Gas pipe leakages can easily escalate to catastrophic events and result in tremendous losses of life and property in the urban context. It is therefore imperative to reduce the risk associated with urban buried gas pipeline network using reliable theoretical accident analysis. This study addresses this issue by systematic combination of three different approaches which include decision making trial and evaluation laboratory (DEMATEL), interpretive structure modelling (ISM) and Bayesian network (BN). The analysis methodology follows a two-stage procedure. First, a hierarchical network model represented by a cause-effect diagram is obtained using the combined DEMATEL-ISM method, which clearly confirms the coupling relationships among various accident-causing factors and the BN structure. It also identifies the most critical factors, which enables the owner/operators of the pipeline system to make decisions regarding the allocation of security management resources to reduce risks. Next, the hierarchical network model is mapped to a BN and expert judgments are further transformed into the conditional probability distribution, in order to quantify the strength of the coupling relationships among the accident-causing system, and determine main paths resulting in system failure. Moreover, it facilitates the analysis process by updating the developed BN model with given new information. The effectiveness and applicability of the proposed model has been validated in a case study, which indicates that the model is plausible in providing explicit risk information to support better safety management by prioritizing actions to prevent interrelated accidents. Highlights: The study originally visualizes the dependencies among various accident-causing factors of urban buried gas pipeline network. A hierarchical network model of the accident-causing system represented by a cause-effect diagram is given. The study points to critical accident-causing factors and main paths resulting in pipeline accidents. A Bayesian network model is established to quantify the strength of the interactions among the accident-causing factors. The case study validates the model to be plausible in providing explicit risk information to support safety management.