A fuzzy-based model proposal for risk assessment prioritization using failure mode and effect analysis and Z numbers: a real case study in an automotive factory.

Abstract

An integrated Failure Mode and Effect Analysis (FMEA) methodology combines the Delphi technique, Decision-Making Trial and Evaluation Laboratory (DEMATEL), and Vlsekrerijumska Optimizacija I Kompromisno Resenje (VIKOR) methods, enhanced with Z numbers. The approach applies to a real case in a bus production line in Turkey to improve risk assessment in human-robot collaboration. DEMATEL identifies interdependencies among failure modes, while VIKOR enables effective risk prioritization. Z numbers incorporate expert judgment under uncertainty in evaluating occurrence, severity, and detection. Severity and occurrence receive the highest weights (0.4062 and 0.2589). Hardware, process reliability, and robot technology level emerge as the top three failure modes. Fuzzy Cohen's κ ensures expert consensus and statistical reliability. A comparative analysis supports the precision, adaptability, and robustness of the methodology for decision-making in complex industrial environments.