Failure diagnosis and recovery based on DES framework

Abstract

As many industrial systems become more complex, it becomes extremely difficult to diagnose the cause of failures. This paper presents a new failure diagnosis approach based on discrete-event systems (DES) framework. In particular, the approach is a hybrid of event-based and state-based ones leading to a simpler failure diagnoser with supervisory control capability. In our approach, we include the failure recovery events for failures in the system model in order to derive a diagnoser we refer to as a recoverable diagnoser. Further, in order to reduce the state size of the recoverable diagnoser, a procedure to construct a high-level diagnoser is presented. The design procedure for diagnoser is presented along with a pump-valve system as a illustrative example.