On-line fault diagnosis in a Petri Net framework

The paper addresses the fault detection problem for discrete event systems modeled by Petri nets (PN). Assuming that the PN structure and initial marking are known, faults are modeled by unobservable transitions. The paper recalls a previously proposed diagnoser that works online and employs an algorithm based on the definition and solution of some integer linear programming problems to decide whether the system behavior is normal or exhibits some possible faults. To reduce the on-line computational effort, we prove some results showing that if the unobservable subnet enjoys suitable properties, the algorithm solution may be obtained with low computational complexity. We characterize the properties that the PN modeling the system fault behavior has to fulfill and suitably modify the proposed diagnoser.