A Rapid Diagnosis Method For Multiple Compensation Capacitor's Fault Of Jointless Track Circuits

The compensation capacitor is mainly used to improve the signal transmission quality in jointless track circuit.Its' failure will directly affect the normal work of the train control system,thereby threatening the safety of the train operation.This leads to the lack of real-time and completeness of fault detection in the field of compensation capacitance.In this paper,based on the influence of the compensation capacitors' fault on the amplitude envelope of the signal induced voltage of the locomotive,the on-line fault diagnosis of the compensation capacitor is carried out.The main contents of this paper include the following aspects:(1)A fault diagnosis algorithm based on BP neural network was proposed,this algorithm using the locomotive signal induced voltage amplitude envelope curve segmented according to the position of the compensation capacitors as the input of the BP neural network to train the network.Finally,this method uses the validation sets to test the BP neural network after training.(2)A fault diagnosis algorithm based on amplitude characteristic was proposed,regarded the cab signal received signal amplitude trend as the features to diagnose the fault of multiple compensation capacitors in a track circuit.Extracting from the actual data subsection processed according to the position of compensation capacitor by using this algorithm,then seek the specific value in the feature sequence from the reader to the learning in order to realize the fault diagnosis and detection.For particularly failure mode,using the first derivative as the constraint condition to achieve the fault identification accurately.Finally,using the simulation data to verify the algorithm.(3)The remote condition monitoring software of compensation capacitor is achieved.So the corresponding software modules developed based on this algorithm has been integrated into the remote monitoring platform.Finally,compared with the existing fault diagnosis algorithms.In this paper,based on the remote monitoring platform of compensation capacitor,a real-time monitoring algorithm of multi compensation capacitors' fault is proposed.The experimental results are better than the existing methods,and the algorithm has been used in part of railway bureaus.