Modeling the impact of natural and security hazards in an LNG processing facility

Abstract

Development of accident models based on cause and effect relationships facilitates the formulation of accident prevention and mitigation plans in the Chemical Process Industries (CPIs). In this paper, failures of accident prevention barriers triggered by manmade and natural hazards are causally modeled using Fault Trees (FTs) models. Additionally, updated technique of FTs basic and top events failure probabilities was applied using Hierarchy Bayesian Approach (HBA) based on basic events precursor data. This updated methodology overcomes the uncertainty limitation in the determination of FTs reliability data, as well as converge them into their accurate values. Moreover, it provides valuable information supporting risk based decision. The methodology was applied to LNG pipeline and liquefaction plant Dispersion Prevention Barrier (DPB). The result shows the capability of the methodology to model natural and security hazards (NE&ISHs) in both qualitative and quantitative manners, as well as, to update FT events failure probabilities through the use of the precursor data to the HBA. Outcomes demonstrate that the average posterior failure probability of DPB of that particular case study increased from 0.0613 to 0.204232 which represents a 3.33 times increment compared with the prior.