An application of BN to envisage potential accidents in FLNG facility

The majority of existing fire and explosion consequence models currently used in the offshore LNG process facilities focus on modelling individual accidental hazard. However, such undesired events can occur simultaneously leading to multiple consequences. This study focuses on developing a model-based Bayesian network (BN) approach to envisage the most probable accidental scenario in complex offshore process facilities by considering the evidence of primary causes. The proposed methodology comprises of evolving scenario identification, accident consequence framework development, accident scenario likelihood estimation, and ranking of the scenarios. The approach was tested and demonstrated with a case study for an offshore LNG process facility (Floating LNG). The developed methodology using the BN approach was found to be effective in determining the most credible accidental scenario.