Research On Fault Diagnosis Of Discrete Event Dynamic Systems

Fault diagnosis is very important for all kinds of large-scale systems in modern industry.Only the faults in the system can be diagnosed quickly and accurately can the system run safely and stably.Discrete event dynamic systems(DEDS)is a class of artificial dynamic systems,whose state migration is caused by discrete events according to certain operating rules.The fault diagnosis method based on automaton model has been widely used in discrete event dynamic systems.It uses the finite state automaton to model the system,and then construct the diagnoser by removing unobservable events from the model.The system is diagnosed according to the correspondence between automaton paths and observable events sequence in the diagnoser.The fault diagnosis process based on the automaton model includes three steps: modeling,diagnostic analysis and diagnosis.First of all,the system phenomena and characteristics will be described by the events,states and state transition relationships.And then the diagnosability of the system will be analyzed to determine the feasibility of fault diagnosis.Finally,the fault diagnosis method is proposed.In discrete event dynamic systems,there often exist timing sequence relationships among the occurrence of events.Adding the timing sequence to the model can reduce the difficulty of modeling and the scale of models and improve the diagnosis efficiency.Previous methods generally assume that faults can not be self-recovery.However,the actual system is not the case.The existence of this fault increases the complexity of the system.In this thesis,a discrete event dynamic system with timing sequence constraint and self-recoverable faults is analyzed,and an automaton model with timing sequence constraint is proposed.Based on the model,an online diagnosis algorithm is proposed.And an example is given to show the effectiveness of the algorithm.In this thesis,a fuzzy automata model is used to describe the uncertain discrete event dynamic system.And a diagnostic analysis method based on the observable events sequence is proposed.By reverse reasoning based on the received observable events sequence,all the paths which can generate the observation in the fuzzy automata model and the corresponding threshold events when the observation is obtained from these paths will be obtained.The diagnosability of these paths with respect to the threshold events is analyzed,and the definition of diagnosability based on observable events sequence is given.The diagnoser with respect to observable events sequence is constructed and the necessary and sufficient conditions for diagnosis are put forward.Finally,an example of animal treatment process is analyzed.Finally,a fault diagnosis software is designed for discrete event dynamic systems,which can build automaton model and diagnose systems to verify the methods proposed in this thesis.