Dynamic Risk Assessment of Furnace System Using Dynamic Bayesian Networks

ABSTRACT Preventing major accidents remains a critical priority in the process industry, given their potentially catastrophic consequences. However, the rare, stochastic, and dynamic nature of such accidents presents significant challenges for conventional risk assessment approaches. This study develops a novel dynamic risk assessment framework that addresses these challenges. The proposed methodology combines Bow‐Tie analysis (BT) with Dynamic Bayesian Networks (DBNs) to create a robust decision‐support tool for managing major industrial risks. Initially, the BT approach identifies basic events, safety barriers, and potential consequences across multiple equipment operating states. These elements are then systematically integrated into a DBN model, which provides enhanced flexibility in modeling the dynamic failure mechanisms of both basic events and safety barriers through continuous precursor integration. The framework's effectiveness is demonstrated through a practical application examining a furnace system in a gas treatment unit at Hassi R'mel, Algeria. This case study validates the model's capability to provide real‐time risk assessment updates, offering significant improvements over static risk evaluation methods for industrial safety management.