Quality function deployment based failure mode and effect analysis approach for risk evaluation

Abstract

The traditional failure mode and effect analysis (FMEA) aims to define the risk priority number of each failure mode by multiplying occurrence, severity and detection parameters of each failure mode. Many studies in the literature commented the traditional FMEA due to its limitations such as equal importance degrees, discrete ordinal scales and multiplication of O, S and D criteria. This study suggests coping with these drawbacks a new method, which combines the quality function deployment (QFD) method under interval type-2 fuzzy set, Bayesian network and VIKOR method. The previous studies generally handle to rank the failure modes as only goal. This paper investigates concurrently the failure modes and the reasons of failure modes. QFD method provides us this benefit by employing customer requirements (CRs) and design requirements (DRs). CRs are the reasons of failure modes while DRs are failure modes. This study presents a new perspective to examine the correlations of CRs and DRs as linguistic terms by using Bayesian network. The VIKOR method deals with ranking the failure modes by aggregating the weights of O, S and D criteria. The developed method is utilized for risk evaluation of tile and marble works in building construction. The result shows that the most significant design requirement or failure mode is DR14 (unsuitable plug-socket utilization).