Fault Predictability of Stochastic Discrete Event Systems Based on Probabilistic Labeled Petri Nets

In this research, a fault prognosis method for stochastic discrete event systems (SDES) is presented using probabilistic labeled Petri nets (PLPN) . The goal is to enhance the accuracy of fault prognosis by using a Petri net model with probabilistic labels. Boundary and non-indicative markings are defined to develop this method. It can identify non-fault prefixes before faults occur, ensuring timely and accurate fault warnings within a limited number of steps. A verifier is constructed to transform the verification of fault predictability into a solvable model checking problem. This process yields a sufficient and necessary condition for the fault predictability of SDES. The proposed method offers a new perspective for fault prediction in SDES, enhancing system reliability and safety in practical applications.