Research On Fault Diagnosis Of Tuning Area And Compensation Capacitance Of Jointless Track Circuit Based On Dynamic Inspection Data

Track circuit,as a vital component of the Chinese Train Control System,performs multiple functions such as transmitting train control information,detecting train occupancy,inspecting rail integrity,and so on.Its operational status is crucial for ensuring the safety of train operation.The compensation capacitor is arranged in the track circuit at equal intervals to reduce the loss of the inductance of the rail in the process of signal transmission and ensure the effective transmission of the signal.Tuning area,as a segment of jointless track circuit that substitutes the conventional mechanical insulation joint,functions to block signals in adjacent sections and realize the polarity crossing between sections.At present,fault diagnosis for compensation capacitance and tuning area issues predominantly depends on the signal detection system on the dynamic detection train.The current process from problem detection to problem resolution,especially on-site investigation,entails a considerable amount of human resources.To address the aforementioned issues,this thesis conducts fault diagnosis research on compensation capacitors and tuning area based on dynamic detection data.The specific research content is as follows:(1)The structure and working principle of ZPW-2000 A jointless track circuit are investigated.Based on the transmission line theory,the two-port network models of interference signals between the main track section and the adjacent section of the track circuit are established,and verifies the models through empirical detection data.(2)Based on the impact of parameter variations in the tuning area on adjacent interference signals,a fault diagnosis method for tuning area equipment is proposed by integrating Particle Swarm Optimization(PSO)algorithm and Genetic Algorithm(GA).By utilizing the maximum information coefficient between the established adjacent section interference signal model and the actual detection data,a multi-objective optimization model is constructed for the parameters of the tuning area equipment.Combining the advantages of the two optimization algorithms,the PSO-GA method is applied to solve the optimization model.The theoretical impedance variation of the tuning area equipment is computed using the model results,and the presence of equipment faults is determined based on this.The experimental results demonstrate that the PSO-GA method has faster computational speed and more optimal solution results in solving multi-objective optimization models for the tuning area.Simultaneously,this method can effectively simulate the performance variation process of the tuning area equipment and identify the location of abnormal parameter variations in the tuning area equipment.The method has been verified to be effective and reliable through analysis of measured data.(3)To further explore the compensating capacitor features contained in dynamic detection data and improve the efficiency of compensating capacitor fault diagnosis,this paper proposes a compensation capacitor fault identification method based on WPD-CNN,combining wavelet packet decomposition and convolutional neural networks.Determine the frequency band range of compensation capacitor characteristics through power spectrum analysis,then use wavelet packet decomposition method to decompose the original signal.Based on this,construct a state feature matrix that can simultaneously reflect the changing trend and the characteristics of compensation capacitors.Construct training and testing sets using dynamic detection data,input feature matrices from different fault types into convolutional neural networks for training,and validate them on the testing set.The experimental results show that the WPD-CNN method requires 5.9ms for feature extraction of a single signal,and the overall fault identification accuracy is 98.4%.It can effectively identify compensation capacitor faults at different positions,providing a basis for compensation capacitor fault diagnosis.(4)Combined with the method proposed in this paper,a fault diagnosis system for track circuit is proposed and developed.The system is embedded with track circuit modeling,tuning area fault diagnosis and compensation capacitance diagnosis models,as well as the functions of reading,analyzing and storing the original data.Through the user-friendly UI interface,it facilitates users to conduct various operations,and can directly display the signal waveform and basic characteristics for users.In the modeling and analysis module,users can flexibly adjust various data parameters of track circuit to achieve modeling of track circuit,while in the fault diagnosis module,users can customize various training parameters to complete automatic fault identification.