A new approach to chemicals warehouse risk analysis using computational fluid dynamics simulation and fuzzy Bayesian network.

This study aims to assess the risk of chemicals warehouse using a Bayesian networks (BNs) and computational fluid dynamics (CFD). A methodology combining Bow-Tie (BT), fuzzy set theory (FST), and Bayesian network was employed, in which the BT was drawn for chemical spill scenarios. FST was utilized for the estimation of the basic events (BEs) occurrence probability, and the probability of interaction among a set of variables was obtained using BNs. Pool fire scenario radiation heat flux was evaluated using CFD code, fire dynamic simulator (FDS), and the solid flame model (SFM). Fail in forklift brake system (BE1), was the most significant cause for a chemical spill. Based on the CFD model, the heat flux is 31 kW/m2 at a distance of 3.5 m from the fire, decreasing to 6.5 m gradually. The maximum safety distance of 4 m is predicted by the CFD for heat flux that exceeds 12.5 kW/m2; however, SFM predicts approximately 4.5 m. According to the results, the amount of posterior risk is higher than the prior value. The framework presented in the chemicals warehouse for consequence analysis and dynamic risk assessment (DRA) of pool fire could be used for preventing the accidents and domino effects in the chemicals warehouse.