Abstract: To solve the problem of strong subjectivity and unreasonable weight distribution in traditional failure mode effect analysis (FMEA), this paper proposes an FMEA method based on fuzzy theory and D-S evidence theory. First of all, this paper refers to the software FMEA method, introduces the fault detection into hardware FMEA and uses the fuzzy theory to establish corresponding fuzzy evaluation grade and fuzzy number. Then, based on D-S evidence theory, an expert weighting method is proposed in order to allocate weight rationally. Finally, risk priority number (RPN) calculation is carried out to get the risk ranking of various failure modes. The case indicates that the method can distribute weights reasonably and accurately according to the evaluation deviation, make the evaluation result more reasonable, the RPN value more accurate, and the failure mode risk ranking more objective. This method can meet the practical demands, which is worth popularizing and applying in engineering practice.

Key words: D-S evidence, fuzzy theory, failure mode effect analysis, weight allocation