ABSTRACT
Traditionally in a system reliability analysis, one starts by a failure mode and effect analysis (FMEA). The purpose a FMEA is to find out the potential failure modes of a system and prioritize them to establish guide lines to a posterior probabilistic analysis. The establishment of these guidelines, depending on the complexity of the system under analysis, is based on the priority associated to each failure mode. Many studies have been developed to establish an adequate, or coherent, prioritization of failure modes. In this paper we consider an approach based on probabilistic composition of preferences for failure modes, considering the traditional criteria, i.e., occurrence, severity and detectability. This approach seeks to consider the uncertainties from experts’ knowledge through a probabilistic reasoning, differently of the traditional fuzzy approach. To verify the applicability of this approach for nuclear systems, we present an analysis of an AFWS from a typical two loop PWR.
