Research On Fault Detection Of Tracking Circuit Compensation Capacitor Based On Bayesian Network

Rail transportation has gradually become the main force of China's transportation and an important driving force for the development of the national economy.The safety of railways is the top priority of railway transportation,and it is the most important factor to be considered.If the safety of railway transportation is not considered in time,or the safety issues are not resolved in time,it will lead to threats to people's lives and property losses.Accident happened.Track circuits,signal machines and turnout switches are the three most important outdoor equipment to ensure the safe operation of trains.Among them,track circuits are important safety devices that can detect train occupation,the current running position of the train,and its integrity.Once the track circuit fails,the safety of train operation cannot be guaranteed,and the consequences are unimaginable.Therefore,the fault detection of the track circuit is a guarantee for the safety of railway transportation.The timely detection and processing of the faults occurring in the track circuit can ensure the maximum safety of the train operation.The compensation capacitor is an important part of the track circuit,which can reduce the sensitivity of the rail,slow down the signal attenuation,and greatly improve the quality of the transmitted signal.Therefore,if it fails,there will be problems in the signal transmission process,resulting in poor signal quality or even code drop problems during train operation,which brings great hidden dangers to the safe operation of the train and the train control system.At present,the main method of on-site detection and compensation of capacitor faults is to use time-consuming and labor-intensive inspection vehicles to regularly perform periodic inspections along the track.This method is very inefficient,and the timeliness is very low,and failure conditions cannot be detected in time.Scholars at home and abroad have conducted general researches on compensation capacitor failures,but there will be problems such as long processing time and the need for offline processing.This paper uses Bayesian networks to analyze and study the existing problems.The main research contents are as follows:Bayesian network has been a popular method for dealing with fault detection in recent years.Unlike other intelligent methods,it is based on the method of probability theory and statistics and the idea of graph theory.It can be inferred that under large data,the complex probability of various situations In the case of fault diagnosis,better results can be achieved,but there are also some problems with Bayesian networks.For example,the diagnosis effect of Bayesian networks depends to a large extent on the construction of the network structure,and Bayesian networks Structural learning is a complex problem.The network built based on the expert system or the original algorithm has redundancy,which affects the diagnosis results to a certain extent.On this basis,the improvedbacterial algorithm is used to optimize the network learning of the Bayesian network.Bacterial algorithm is an emerging artificial intelligence algorithm in recent years,which is used to deal with the optimal solution of complex problems in multiple dimensions.However,the original bacterial algorithm has many limitations.For example,the distribution of the initial colonies may lead to a large difference in the number of iterations of the algorithm.The driving part of the algorithm is blind,and the global situation is not considered.To improve the algorithm based on such situations,the core innovation is to solve the limitations in the algorithm reasonably.Adaptive parameters are added to adjust the algorithm's drive process,and combined with the latter two parts of the algorithm,it can effectively save excellent bacteria and solve the problem of overall community optimization.This paper optimizes the network structure learning of Bayesian network based on bacterial algorithm,so as to obtain the optimized Bayesian network structure.Based on the obtained structural model,by using MATLAB on the Windows platform to diagnose the failure of the track circuit,the fault condition of the compensation capacitor can be clearly diagnosed,and the failure of each compensation capacitor can be analyzed to obtain the corresponding fault diagnosis results.Provide a new idea and method for railway intelligent operation and maintenance.