System Importance Measures Using MCS with Extended Saphire Applications

In order to address some technical challenges raised in the space shuttle PRA development, we introduce in this paper, a definition of group level importance measure (GIM) to resolve the issue of risk ranking for mission phase, system (or element) within various level of the integrated shuttle PRA model. The GIM measures defined here are extensions of Fussel-Vesely (FV) type importances, and can also be extended to calculate other conventional type of importance measures, particularly at the gate importance level within a coherent system fault tree model. Algorithms are also developed for extended SAPHIRE [1, C. Smith, T. Wood, 1996] implementations. Examples are given to show that the GIMs can be obtained for every level of contributors modeled in a PRA. Furthermore, these importances can be easily obtained and organized based on a top down hierarchy to show the safety significance of general contributors to specific category of mission risk profiles. Such measure can be easily used to produce PRA risk ranking at all level of details, such as by phase, element, sequence bins, failure type or any combinations of specific group of event type and failure mode. In the development of the underlying GIM importance measures, a complete set of extended mathematic functional relationships among the most commonly used importance measures are also derived and presented in the first section of this paper for the applications of PRA practitioners [2, F. Hsu et al., 2002].