Research On Condition Monitoring And Prediction Of Compensation Capacitor In Jointless Track Circuit

Track circuit as the key equipment of ground-train continuous information transmission in the train operation control system, cooperating with on board track circuit reader (TCR), plays an extremely important role in such aspects as the train occupying checking, integrity checking and ground-train continuous information transmission. Among them, the compensation capacitor is an important equipment to improve the signal transmission quality of jointless track circuit, whose failure will directly affect the normal operation of track circuit and train operation control equipments, which brings great hidden trouble to high efficiency and safety of train operation. At present, the existing methods and research achievements at home and abroad for compensation capacitor state monitoring, in aspects of timeliness, efficiency, cost and practicality in fault detection, have been unable to meet the requirements of the high-speed development of Chinese railway. Therefore, the paper analyses the influence law of how the amplitude envelope of the TCR induced voltage is affected by the compensation capacitor under different fault modes and different positions. And on the basis, the paper proposes the method of compensation capacitor state monitoring and fault prediction for jointless track circuit by making use of the train running data of the cab-signal remote monitoring. The main work of this paper is as follows:1. The method for state monitoring and prediction of compensation capacitor based on train running data is proposed. To prove the feasibility of the method, firstly, data source and configure of the time interval are fully analyzed; secondly, the model of amplitude envelope of induced voltage based on the transmission line theory is proposed and verified; finally, the influence law of how the amplitude envelope of the TCR induced voltage is affected by the compensation capacitor under different fault modes and different positions is systematically analyzed. The piecewise trend of TCR induced voltage amplitude envelope curve, the continuous determination coefficient and the fault angle of compensation capacitor are extracted as the key characteristic indexes of compensation capacitor state monitoring and fault prediction. And the influence law of how the fault features are affected by different fault modes is analyzed.2. A state monitoring method is proposed for the compensation capacitor on basis of symbolic directed graph (SDG). According to the extracted state features of the compensation capacitor, symbol directed graph model and dependency matrix of compensation capacitor fault are established using fault diagnosis theory of symbolic directed graph. The paper puts forward the online diagnosis algorithm of compensation capacitor under broken-wire fault and capacitor-decrease fault, which divides the states of compensation capacitor into five states and set up four corresponding risk warning levels. Moreover, the state monitoring algorithm of compensation capacitor is achieved.3. The state prediction algorithm of compensation capacitor based on phase space reconstruction and Kalman filtering is proposed for capacitor-decrease fault. The historical state feature sequences of compensation capacitor extracted from TCR induced voltage amplitude envelope are restructured by using the phase space reconstruction algorithm. At the same time, the residual life of the compensation capacitor is predicted by using trend extrapolation algorithm. And also the residual life curve distribution law of compensation capacitor whose capacity value reduces from the start to broken-wire is summarized.