FRAM integrated fuzzy FMEA risk analysis for natural gas systems

Abstract

Natural gas installation systems involve complex technical, organizational, and human interactions where failures can lead to catastrophic safety consequences. Conventional Failure Mode and Effects Analysis (FMEA) is constrained by subjective scoring and the limited capture of systemic variability. This paper proposes an integrated framework combining the Functional Resonance Analysis Method (FRAM), fuzzy logic-based FMEA, and Pareto analysis to address these limitations. FRAM was applied to model system functions and identify variability sources across installation stages, which were translated into candidate failure modes. A multidisciplinary expert panel evaluated the occurrence, severity, and detection scores for each failure mode. A Mamdani-type fuzzy inference system was then used to calculate the fuzzy Risk Priority Numbers (RPNs), and Pareto analysis was applied to prioritize critical failure modes. The framework was validated in a real building natural gas installation project, identifying 81 failure modes. Fuzzy-FMEA produced substantially different risk rankings than classical FMEA. Specifically, improper tightening of valve connection points emerged as the highest-risk failure mode. Pareto analysis showed that 52 failure modes (64.2%) account for 80% of the cumulative system risk. The proposed framework enhances the completeness of failure identification and the reliability of risk prioritization, providing a systematic decision-support tool for natural gas installation safety management.