Evidence based risk analysis of fire and explosion accident scenarios in FPSOs

Abstract

The paper proposes a methodology for risk assessment and probabilistic modelling of fire and explosion accidents in Floating production storage and offloading (FPSO) units. The overall objective of the paper is to propose a Bayesian Network probabilistic framework towards quantification of fire and explosion events in FPSOs derived from a systematic analysis of incidental and accidental scenarios specific to FPSOs. A data set of around 800 incidents in FPSOs from various open-source agencies is prepared and analysed. The key features of the incidents are discussed and the relationships amongst the significant variables influencing the types of incidents are assessed. The circumstances of potential incidents related to fire and explosions are discussed. A detailed risk analysis is conducted using the risk matrix approach to screen and rank the major accidents occurring in FPSOs. Then, a Bayesian Network model of high-risk fire and explosion scenarios is developed based on evidences obtained from accident reports and expert opinions. The model uses a framework based on immediate causes, basic causes and causal factors to demonstrate various accidental scenarios specific to FPSOs. A sensitivity analysis is conducted to identify the most important causal factors and the aspects that need more research work for decision-making. The contribution of the present study is threefold: a methodology for comprehensive risk assessment in FPSOs is proposed; FPSO specific incidents and accidents are characterized; a probabilistic model for fire and explosion scenarios is developed from a causal framework. The results of the paper provide FPSO developers and operators with information to prevent and mitigate fire and explosion accidents.