Research On Transmission Grid Fault Diagnosis Method Based On Time Series Abductive Neural Fuzzy Petri Net

With the rapid development of economy and society,electric energy has gradually become an indispensable part of national economy growth.As a bridge connecting power plants and users,the normal operation of transmission grid directly affects the reliability of power supply.In recent years,some experts and scholars have put forward many methods to diagnose transmission grid faults,among which the fault diagnosis method based on Petri net has become a research hotspot.On the basis of learning and summarizing the existing research results,this thesis further studies the transmission grid fault diagnosis method based on Petri net.(1)In order to reduce the complexity of the model and improve the efficiency of fault diagnosis,a fault diagnosis model based on Hierarchical Fuzzy Petri Net(HFPN)is established in this thesis.On the basis of the HFPN model,considering that the weight and action confidence of the existing fault diagnosis model of transmission grid based on Petri net are mostly obtained by manual experience,this thesis introduces the error back-propagation method in BP neural network into the HFPN model.The weight in the HFPN model is adjusted according to the correct action rate of protection and circuit breaker over the years.In the same time,the average value of the correct action rate of protection and circuit breaker over the years is taken as the action confidence of protection and circuit breaker.Case and model performance analysis show that this method can reduce the influence of subjective factors,improve the accuracy of fault diagnosis,reduce computational complexity.(2)Considering the temporal attributes between component fault and protection action,protection action and circuit breaker trip,this thesis introduces temporal information into the established HFPN model,and establishes a fault diagnosis method for transmission grid based on time series abductive Petri net.Firstly,the structural correlation characteristics and temporal correlation characteristics among components,protection and circuit breakers are analyzed according to the established time series abductive Petri net,in the same time,the protection and circuit breaker action confidence that do not have both structural correlation characteristics and temporal correlation characteristics are modified.Then,the component fault probability is calculated according to the established HFPN model.Finally,starting from the fault components,the thesis analyses whether the protection and circuit breaker fail to operate,malfunction,time mark error,and whether the data acquisition system fails to report information.(3)The HFPN model is combined with the error back-propagation method of BP neural network and the time series abductive Petri net.A visual fault diagnosis system is developed by using LabVIEW graphical programming language to assist maintainer in their work.