Research On Fault Diagnosis Of Choke Adaptor Transformer Fuzzy Neural Network Based On Wolf Pack Algorithm

Railway transportation is one of the most important modes of transportation in the country.Its safety and reliability are of great significance to the development of the railway industry.The choke adaptor transformer is an important component of the track circuit in the station and has been widely used in electrified railways.Installed in the combination of strong and weak electricity,not only provides the return flow for traction current and signal current,but also the large unbalanced traction current in the traction power supply system,which has an anti-interference effect on the normal operation of the track circuit,ensuring stable,safe and reliable operation of the train driving.However,due to the complexity of the structure of the choke adaptor transformer,the variability of the working environment,and the concealment of the location of the fault,when the fault occurs,the location of the fault cannot be timely due to the experience of the railway field staff.Accurate positioning,and can not accurately determine the cause of the failure,has a serious impact on railway operating efficiency and driving safety.Therefore,this thesis proposes an IWPA-FNN(Improved Wolf Pack Algorithm-Fuzzy Neural Network)model for fault diagnosis of choke adaptor transformer.This thesis mainly completes the following contents:(1)The important role of choke adaptor transformer in electrified railways is expounded.Through the research and application of choke adaptor transformer and fault diagnosis met hods by domestic and foreign scholars,the intelligent fault diagnosis research on choke adaptor transformer is expounded importance.(2)Through the analysis of the structural characteristics and working principle of the choke adaptor transformer,combined with the different working states of the track circuit system,the common faults of the choke adaptor transformer are sorted out,the causes and failure mechanisms of the faults are analyzed.According to the actual working environment of the 25Hz phase-sensitive track circuit on the railway site,the model is built under the test conditions,then the corresponding faults of the choke adaptor transformer are simulated,and the fault data in different states are collected,combined with FNN(Fuzzy Neural Network)is to build a smart fault diagnosis model for choke adaptor transformer based on fault characteristics.(3)Introduced WPA(Wolf Pack Algorithm),the basic ideas and steps of the algorithm were introduced,and the optimization performance was verified by comparison with standard test functions.In order to improve the late search ability of the algorithm,an IWPA(Improved Wolf Pack Algorithm)is proposed,and the feasibility of the improved algorithm is analyzed.(4)Aiming at the shortcomings of FNN early maturity and slow convergence,the simulation of wolf hunting process,using IWPA optimal output results to optimize FNN parameters,and constructing IWPA-FNN choke adaptor transformer fault diagnosis model;from the perspective of algorithm,to improve the fault diagnosis accuracy of choke adaptor transformer,the different models are simulated and compared with the fault samples to verify that the IWPA-FNN model meets the accuracy requirements,which provides selectivity for further research on fault diagnosis of railway signal equipment.