An Improved FMEA Method Based on ANP with Probabilistic Linguistic Term Sets

Failure modes and effects analysis (FMEA), as a practical and easy-to-use reliability assessment tool, has been widely applied across various fields of researches. At the same time, it also receives criticisms for its limited selection of risk factors as well as its discrete numerical ordinal scales. In an attempt to fill in this gap, this paper proposes an improved FMEA method based on analytic network process (ANP) with probabilistic linguistic term sets (PLTSs). Firstly, the three risk factors of FMEA, namely occurrence, severity and detection, are broken down to more elaborate and specific sub-factors, and a network representing the influential relationship between these sub-factors is constructed. ANP is then utilized to derive the relative weights of factors, sub-factors and failure modes by making pairwise comparisons with the help of PLTS. To verify the rationality and applicability of our proposed method, a case study of hospital information system reliability assessment is carried out. Comparative analyses with other existing FMEA methods are also undertaken to highlight the differences and advantages of our proposed method.