基于LSTM算法的ZPW-2000A型轨道电路故障预测模型研究

Research on Fault Prediction Model of ZPW-2000A Track Circuit Based on Improved ELM

轨道电路是保证列车安全可靠运营的重要基础设备。随着铁路信号设备“故障修”向“状态修”的转变,为了对轨道电路监测信息分析进而制定科学合理的维修策略,提出一种基于LSTM算法的故障预测模型方法。首先,对轨道电路结构原理进行了分析,在此基础上提出使用LSTM模型对其故障进行预测,为验证模型的预测效果,提出用平均绝对误差(MAE)、平均绝对百分比误差(MAPE)均方误差(NMSE)和相对误差百分比(REP)对预测结果进行度量;其次,结合预测实例将改进算法与传统的SVM算法、BP神经网絡,RNN算法的预测效果进行对比,结果显示LSTM各指标均最优;最后,结合一个实例采用LSTM进行预测,验证了模型的高效性。

Track circuit is an important basic equipment to ensure the safe and reliable operation of trains.With the transition from"fault repair"to"state repair"of railway signal equipment,in order to analyze track circuit monitoring information and formulate a scientific and reasonable maintenance strategy,a fault prediction model method based on LSTM algorithm is proposed.First,the principle of the track circuit structure is analyzed,and on this basis,it is proposed to use the LSTM model to predict its failure.In order to verify the prediction effect of the model,it is proposed to use the average absolute error(MAE),average absolute percentage error(MAPE);and average absolute error(MAPE).Square error(NMSE)and relative error percentage(REP)are used to measure the prediction results;secondly,combined with prediction examples,the improved algorithm is compared with the traditional SVM algorithm,BP neural network,and RNN algorithm.The results show that the indicators of LSTM are all Optimal;Finally combined with an example,LSTM is used for prediction^which verifies the efficiency of the model.