FMEA에서 믿음 엔트로피와 증거거리를이용한 리스크 평가 방안

Abstract

Purpose This study is to demonstrate a novel risk evaluation approach using a combined weight of team experts in the Failure Mode and Effects Analysis (FMEA) model under uncertain circumstances that are caused by highly conflicting opinions between team experts and unknown information about Risk Priority Number (RPM) evidence. Methods To recognize any uncertainty due to conflicting evidence and unknown information between the team experts, a novel combined weighting factors of the RPN evidence is generated by the framework of Dempster Shafer Theory (DST). The total combined weighting schemes is composed two parts: One part consists of belief entropy and belief conflicting coefficient to consider team experts’ conflicting opinions. Whereas, another part is based on the evidence distance such as belief Euclidean distance and belief cross entropy to consider team experts’ uncertainty about RPN evaluation. Results A novel risk evaluation model using a combined weighting factors to address uncertainty in FMEA is proposed. The case study, Advanced Driver Assistance System (ADAS) consisting 7 components and 7 parts, is illustrated to show the effectiveness and usefulness of the proposed method. Conclusions The suggested FMEA model is convenient for the risk evaluation in the following uncertain situations: i) The team experts’ opinions are highly conflicting. ii) The information of RPN evidence are imprecise and unknown. The proposed FMEA method is advantageous for effectively making a group decision of team experts in the presence of uncertain RPN evidences.