Probabilistic analysis of cascading events triggered by fire

Fires involving storage or process units were the most frequent initiating cause of severe cascading events in the chemical and petrochemical industry. Storage units were often the targets of escalation. Most industrial facilities adopt protection systems and procedural emergency measures to prevent escalation. Hence, the analysis of fire protection systems must be included in the assessment of escalation probability as well as in Quantitative Risk Assessment (QRA) procedures: the methodology illustrated in the present contribution was developed to the purpose. The methodology was based on a Layer Of Protection Analysis (LOPA) approach to account for implemented protection layers, whose performances were characterized in terms of both Probability of Failure on Demand (PFD) and effectiveness. Equipment vulnerability models were integrated with the LOPA results. Modified escalation probabilities, including the influence of safety barriers, were thus obtained. A case study was analyzed to exemplify the methodology implementation.