A quantitative LNG risk assessment model based on integrated Bayesian-Catastrophe-EPE method

Abstract

Increasing quantities of natural gas are transported as liquefied natural gas (LNG) worldwide, and LNG is also proposed for use as a clean fuel for ships and trucks. This scenario raises concerns for the safety of LNG bunkering and storage at ports, due to the potentially severe accidents that may arise from LNG leakage. In this paper, an integrated quantitative risk assessment model for LNG bunkering and storage at ports based on Bayesian-Catastrophe-EPE (Energy transfer theory, Preliminary hazard analysis and Evolution tree) method was proposed. The energy-based EPE model was used to derive Bayesian network (BN) causal structure, and the catastrophe theory was employed to deal with experts’ judgment to determine the conditional probability tables of BN. The proposed BN-based risk assessment model can provide a novel perspective to identify hazards and risks, and to assess the evolution process of LNG accidents from causes to consequences. The results of scenario analysis of typical LNG accidents demonstrate the soundness and applicability of the proposed model. Moreover, sensitivity analysis was implemented to identify critical hazards and quantify the correlations between each element considered in LNG accidents. The proposed risk assessment framework is of great significance to widen the technical tools available to support safety assessment and loss prevention of LNG bunkering and storage at ports.