Research On Fault Diagnosis Method Of 25Hz Track Circuit Based On Improved Evidence Theory

As one of the basic equipments of the railway signal system,the track circuit plays an important role in the normal operation of the entire railway signal system.The 25 Hz phase-sensitive track circuit is a type of station track circuit commonly used in railways in China.Because of the harsh environment of its equipment and long-term high-intensity continuous operation,track circuit failures often occur.Once the track circuit fails,it will adversely affect the entire signal system,and reduce driving efficiency,even lead to safety accidents.At present,the fault handling of the track circuit on site is the empirical judgment of the electrical maintenance personnel after measuring the fault characteristic parameters through the instrument and meter.However,the structure of the track circuit is complex,and there are many factors affecting the operation of the track circuit.There is no clear correspondence between the cause of the fault and the parameter of the fault characteristic.The manual diagnosis method has low precision and poor efficiency.It is necessary to use the artificial intelligence method to diagnose the fault of the track circuit according to the actual situation.However,most of the existing fault diagnosis methods adopt a single intelligent method.Due to the imperfection of the knowledge base and the one-sidedness of considering the angle,the diagnostic accuracy is low.And there are often limitations in the diagnosis process.Therefore,aiming at the 25 Hz phase-sensitive track circuit of this research object,a fault diagnosis method based on improved evidence theory is proposed to meet the actual needs of the track circuit.The research content of this paper includes the following parts:Firstly,the structure principle of 25 Hz phase-sensitive track circuit is introduced.According to its different working states,the common fault circuit types and the mechanism of fault occurrence are summarized.The four-terminal network model of track circuit is established,and the critical voltage of track circuit is obtained by calculating the analog value of track section.Secondly,a preliminary diagnostic model is established.Three kinds of intelligent methods,such as grey relational analysis,fuzzy comprehensive evaluation and BP neural network,are used to diagnose the fault diagnosis of the track circuit respectively,and the corresponding fault diagnosis model is established.The Basic Probability Assignment(BPA)of the evidence theory is constructed by using the output results of each preliminary diagnosis method.Finally,a fault diagnosis model based on improved evidence theory is established.The reliability of the evidence is determined by the misjudgment rate of the diagnostic method,and the trust degree of the evidence is determined by the evidence distance and the uncertainty.The weight is obtained by combining the reliability and trust of the evidence,and the BPA of the evidence is reconstructed to eliminate the evidence conflict between the BPA obtained by each method.The evidence theory synthesis rule is used to fuse the reconstructed BPA to determine the fault conclusion.The diagnosis results show that compared with the single intelligent diagnosis method,the fault diagnosis method based on improved evidence theory information fusion has higher diagnostic accuracy and the diagnosis conclusion is more credible.