A Propane Tank Case Study: Dynamic Risk Assessment by Bayesian Approach

This study develops a dynamic risk assessment framework for fire and explosion hazards in chemical plants, focusing on a propane tank case study. Fault Tree Analysis (FTA) and Event Tree Analysis (ETA) were employed to assess root causes and safety barriers. The study also utilized the Exponential Smoothing Model to forecast precursor data over a five-year period, combined with Bayesian Networks to revise the safety barriers’ failure probabilities. The results revealed a gradual degradation of the Release Prevention Barrier (RPB) and Ignition Prevention Barrier (IPB) over the five-year span, with failure probabilities increasing by 79.15% and 118.52%, respectively. In contrast, the safety system’s ability to prevent accidents diminished significantly, with the probability of incidents rising by 624.89%. The final risk assessment demonstrated that dynamic approaches provide more accurate predictions of accident likelihoods compared to conventional methods. Dynamic assessment can identify the safeness of a plant over time by using historical or real-time data before occurrence of failures.