Risk evaluation of submarine pipelines using improved FMEA model based on social network analysis and extended GLDS method under a linguistic Z-number preference relation environment

Abstract

As the lifeline of offshore hydrocarbon resources advancement and transmission, conducting risk evaluations is crucial to mitigate potential endangers related to submarine pipelines. The improved failure mode and effects analysis (FMEA) method has been widely applied in the risk analysis of submarine pipelines. But most studies evaluated the failure items directly, ignoring the risk attitudes of groups and individuals. Moreover, the impact of trust relationships in social network, consistency and credibility of risk references on final results are not considered. Thus, a novel FMEA method based on social network analysis and extended gained and lost dominance score (GLDS) is presented for risk assessment of submarine pipelines with enhanced reliability. Firstly, the linguistic Z-number preference relations (LZNPRs) are adopted to describe the risk preferences of experts through pairwise comparisons of failure items. And consistency checking and repairing algorithms are proceeded on contradictory evaluations. Secondly, a synthetic weighting model is raised to measure the expert weights considering the trust relationships in social network, the consistency and credibility of expert comments. Finally, the key failure modes considering the differentiated risk attitude are derived by extending GLDS with LZNPRs. The results of case analysis indicate that device failure is the most critical failure item, standing out with its maximum group utility value (1.0463), minimum individual regret value (0.1435), and corresponding maximum overall risk performance (0.2565). On this foundation, more targeted measures can be supervised to further improve safety level. The reliability and validity of the developed model are demonstrated through comparative and sensitivity analysis.