Modeling a system using observations in discrete event system for failure diagnosis

Model based failure diagnosis in discrete event system is based on state based sensor outputs. With the advancements in sensor technology, sensor outputs related to each event is available for monitoring. In this paper event based sensor outputs are taken into account in component modeling in a formal way. A modeling framework using the event based observations is proposed and a diagnoser is developed using this model. The modeling and diagnoser are based on finite state machines (FSM). Hidden failures in power system relays and circuit breakers are considered for the study.