A methodology for the quantitative evaluation of NPP fault diagnostic systemsʹ dynamic aspects

A fault diagnostic system (FDS) is an operator decision support system which is imple- mented both to increase NPP eciency as well as to reduce human error and cognitive workload that may cause nuclear power plant (NPP) accidents. Evaluation is an indispensable activity in constructing a reliable FDS. We ®rst de®ne the dynamic aspects of fault diagnostic systems (FDSs) for evaluation in this work. The dynamic aspect is concerned with the way a FDS responds to input. Next, we present a hierarchical structure in the evaluation for the dynamic aspects of FDSs. Dynamic aspects include both what a FDS provides and how a FDS operates. We de®ne the former as content and the latter as behavior. Content and behavior contain two elements and six elements in the lower hierarchies, respectively. Content is a criterion for evaluating the integrity of a FDS, the problem types which a FDS deals with, along with the level of information. Behavior contains robustness, understandability, time- liness, transparency, e€ectiveness, and communicativeness of FDSs. On the other hand, the static aspects are concerned with the hardware and the software of the system. For quantita- tive evaluation, the method used to gain and aggregate the priorities of the criteria in this work is the analytic hierarchy process (AHP). The criteria at the lowest level are quanti®ed through simple numerical expressions and questionnaires developed in this work. these well describe the characteristics of the criteria and appropriately use subjective, empirical, and technical methods. Finally, in order to demonstrate the feasibility of our evaluation method, we have performed one case study for the fault diagnosis module of OASYSTM (On-Line Operator Aid SYStem for Nuclear Power Plant), which is an operator support system devel- oped at the Korea Advanced Institute of Science and Technology (KAIST).