基于隐半马尔可夫模型的补偿电容维修机制优化

通过分析轨道电路信号传输机制,结合其电气化特性提出了一种增设轨面信号采集点的方式实现轨道电路补偿电容的故障定位并实现剩余寿命的估计。首先结合历史数据和仿真试验数据构建各补偿电容采集点的隐半马尔可夫模型,采用训练后的隐半马尔可夫模型对补偿电容实时运行状态进行监测;其次对轨道电路补偿电容运行数据偏移状态进行分析,通过安全偏移距离对各监测点的归一化分路电流进行安全分析,对超出安全偏移状态的补偿电容进行泄漏定位,结合监测数据与训练后的隐半马尔可夫模型估计补偿电容的剩余寿命。最后结合仿真试验数据表明增设补偿电容轨面采集点可实现补偿电容泄露定位并对其进行剩余寿命的估计,且剩余寿命估计的准确性可达90%以上,达到补偿电容维修机制优化的目的。

Optimization of Compensation Capacitor Maintenance Mechanis Based on Hidden Semi-Markov Model

Based on the analysis of signal transmission mechanism of track circuit combined with its electrification characteristics, this paper proposes a new method of adding track signal acquisition points to realize fault location of track circuit compensation capacitor and estimate its residual life. Firstly, a hidden semi-Markov model of each compensation capacitor acquisition point is constructed by combining historical data with simulation test data, and the real time operation status of compensation capacitor is monitored by the trained hidden semi-Markov model. Secondly, through the analysis of track circuit compensating capacitor operation data offset state, the normalized shunt current of each monitoring point is analyzed by safety offset distance, the leakage location of compensation capacitor beyond the safe offset state is carried out, and the residual life of compensation capacitor is estimated by combining the monitoring data with the hidden semi-Markov model after training. Finally, combined with the simulation test data, the rail acquisition point of compensating capacitor can be used to locate the leakage of compensating capacitor and estimate its residual life, and the accuracy of residual life estimation is more than 90%, optimizing the maintenance mechanism of compensating capacitor.