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

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 verifi es the effi ciency of the model.