Risk assessment of a liquefied natural gas process facility using bow‐tie and Bayesian networks

Abstract

AbstractIn this paper, a risk assessment methodology is presented to identify, analyze, and evaluate the risk associated with different consequences of a liquefied natural gas (LNG) process facility from west Algeria. Bow‐tie analysis (BT) is used to identify the possible causes of hazardous events and the dangerous outcomes and consequences resulting from the failure of the safety barriers. Moreover, the bow‐tie is mapped into Bayesian networks (BNs) to overcome some limitation of dependencies and updating in the events. Finally, the risk‐reducing measures implemented in the process are examined using BNs. The performances of the safety barriers are investigated considering the occurrence of fire and explosion in the facility. BN uses diagnostic (posterior) and predictive (prior) analysis to calculate the probabilities of the components and to predict the occurrence frequency of the consequences and the effectiveness of safety barriers. The results from the BN model indicate that pool fire is the most likely scenario that potentially causes damage to the LNG process.