A LOPA-based Approach To Risk Assessment and Management

Sonia Masmoudi

This article proposes an improved approach to risk criticality assessment by integrating fuzzy logic into Conventional Criticality Matrices (CCM) through a fuzzy inference system. Although conventional criticality matrices are widely used across various industrial sectors, they exhibit inherent limitations due to the subjectivity of expert judgments and the uncertainties associated with risk assessment data. The results highlight the potential of the rule-based fuzzy inference system to enhance conventional criticality matrices by flexibly representing and processing expert knowledge through the Fuzzy Criticality Matrix (FCM), thereby providing a better representation of uncertainties and improving the management of inherent risks. Furthermore, in order to reduce an unacceptable risk to an acceptable, or at least tolerable, level, the application of the conventional Layer of Protection Analysis (LOPA) method made it possible to assess risk reduction by analyzing the contribution of the different protection layers, verifying the adequacy of the existing independent protection layers, and identifying situations requiring the implementation of additional safety barriers. To conduct an inductive analysis, the use of the HAZOP method is essential, as it enables the accurate identification of possible causes and all potential observable consequences. The proposed approach, combining HAZOP, conventional LOPA, and fuzzy logic, was applied to a real industrial case study at SAROST S.A., demonstrating its effectiveness in supporting a more robust risk assessment and management process in industrial environments. Finally, as a perspective, a comparison between the two proposed fuzzy models namely the Fuzzy Criticality Matrix (FCM) and Fuzzy LOPA would be valuable to further improve the achieved level of risk criticality assessment.