Research On Fault Propagation Modeling Of Metro Signal System Based On Petri Net

With the rapid development of urban rail transit,the interval of subway operation is constantly shortened,and the operation safety requirements are increasingly stringent.The network of the structure and function of the subway signal system makes the multiple crosscoupling relationship between the subsystems and the internal components of the subsystems.When a local fault occurs in one or a few components or the connecting parts among them,which located in the subsystem will spread to the same level or other subsystems through the coupling relationship of the connecting parts,causing more faults.As the fault continues to accumulate and expand,that will lead to the loss of system function,affect the normal operation of the subway,and cause accidents in serious cases.At present,most of the fault disposal are based on manual experience,while a large amount of accumulated fault data were neglected or rarely used,resulting in inaccurate positioning,incomplete positioning,long positioning time and other problems in fault disposal.Therefore,combining with the historical fault data,this thesis studies a method of fault propagation modeling,which to simulate the fault propagation path,evaluate the risk of fault propagation path,and deduce the fault diagnosis process,that is of certain significance to the intelligent analysis of fault propagation.The main work of the thesis includes:(1)builded the hierarchical function-failure modelIn view of the signal system failure of structure layer properties and the spread of the gradation characteristics,based on a large number of fault information,carding the failure mode,failure logic relationship,which to summarize the signal system fault propagation characteristics and breeding rules,structures.Hierarchical function-failure model established the structure foundation of fault propagation Petri net model.(2)mined fault association relationship by using association rule algorithmBy selecting fault feature words,the fault data recorded in natural language are converted into a formal description of fault vector text.The parameters of support degree and confidence degree correspond to the correlation between fault type(mode)and fault characteristics,fault characteristics and association rules,which lays the parameter foundation of fault propagation Petri net model.(3)analied failure propagation Petri net modelGuided by the function-failure model framework,combined with the association rules mining parameters of support and confidence,borrowed from place,transition and arc ignition,auxiliary place,transition fire rules to describe the function-failure logic relations between signal system components.Introducing token,place,transition dyeing rules,in combination with fuzzy production rules to express the uncertainty information,which to build failure propagation Petri net model.Developed the forward search fault propagation algorithm of the fault propagation Petri net model,so that to obtain the fault propagation path.Meanwhile,the average influence value was introduced to evaluate the fault path risk.The reverse inference fault diagnosis algorithm of fault propagation Petri net model was developed to deduce the fault source and guide the fault disposal work of the system.Taking vehicle-mounted ATP as the research object,the correctness of the model and algorithm was verified.Thesis contains 44 figures,21 tables and 52 references.