IEEE Access (Jan 2023)
Decentralized Marking Fault Diagnosis of Labeled Petri Nets
Abstract
A decentralized marking fault diagnosis method is proposed to solve the problem that increasing the number of unobservable transitions may result in the inability to diagnose faulty markings, in a class of decentralized systems modeled by labeled Petri nets. Assuming that each local site knows the structure of the labeled Petri net, and the subnet induced by unobservable transitions is acyclic. The decentralized architecture consists of a set of local sites communicating with a coordinator that determines whether the faults have occurred in the system, which are modeled by markings. First, each local site constructs the corresponding dual verifier and calculates the local marking fault diagnosis state according to the local observations. Then, it exchanges the corresponding information with the coordinator according to the two proposed diagnosis protocols. Finally, the coordinator calculates the global diagnosis state according to the received information. In addition, the marking diagnosability under both protocols is analyzed. A sufficient and necessary condition for marking fault diagnosis in the decentralized architecture under the second protocol is proved.
Keywords