Bayesian Belief Network Used in the Chemical and Process Industry: A Review and Application

With the increasing growth of the chemical and process industries, it is necessary to ensure the safe operation of their complex and often hazardous installations, given their proximity to residential areas. Several techniques, such as fault tree analysis (FTA), bow-tie analysis (BTA), and Bayesian belief networks (BBNs), have been developed for adequate probabilistic risk assessment and management. The current work is aimed at performing a brief statistical review of the use of Bayesian networks in the chemical and process industry within the last decade. The review reveals that Bayesian networks have been used extensively in various forms of safety and risk assessment. This trend is attributable to the complexity of the installations found in this industry and the ability of BBN to intuitively represent these complexities, handle uncertainties, and update event probabilities. The paper is concluded with an illustrative example of the use of BBN to investigate the effectiveness of the safety barriers of a gas facility.