Design improvement for enhanced process safety in a biodiesel production unit using Fuzzy Bayesian network analysis

Abstract

The study presents a novel comprehensive approach for quantitative dynamic risk assessment in the reaction section of a biodiesel production plant. The methodology combines sequential established techniques (HAZOP analysis, Bow-Tie (BT) analysis) with a Fuzzy Bayesian Network (FBN) in process safety assessment of biodiesel production. Following a thorough process hazard identification, detailed design gaps were addressed and modified. Dynamic risk assessment was then performed by combining the BN with fuzzy approach to cover the limitations of conventional risk analysis and to deal with the uncertainty in the modeling due to lack of credible equipment failure rate data. Assessment on the effectiveness of design improvement revealed a substantial 90% risk reduction following the implementation of technical recommendations obtained from the HAZOP analysis. Sensitivity analysis using the FBN also revealed the reactor pressure transmitter and safety valves as the critical safety elements of the process. Implementing design modifications based on such analysis reduced the risk level by 95%. This study highlights the effective combination of the established methods with FBN for dynamic risk assessment and targeted design improvements, ultimately leading to enhanced safety in biodiesel production plants.