Stochastic Activity Networks for the Modeling of Repairable Systems including Diagnosis Performance

The productivity and quality requirements have conducted the manufacturing systems to be more and more complex. Indeed, many sub-systems are in interaction such that the process system, the control system and the supervision or diagnosis system. The dependability and availability analysis in those systems is of a major importance, since they impact directly on system’s productivity and safety. The aim of this paper is to present a modelling framework based on an extension of Petri nets called Stochastic Activity Networks (SANs) that systematically includes the diagnosis performance for the dependability evaluation. The major advantage of such formalism is that it allows the modelling of dynamic systems by modelling all their possible states. And unlike tools such as automata and Markov processes, the modelling can be done simply and in a compact manner. Fault-tolerant systems are considered in here. That is systems including a diagnosis system to allow fault detection, and backup system(s) to allow fault-recovery. The systems under study are considered repairable. Monte-Carlo simulation study is conducted to show the impact of the diagnosis performance and corrective maintenance actions on the system’s availability.