Research on Prediction of Time Between Failures for Onboard Subsystem of Train Control System

The prediction of Time Between Failures (TBF) is one of the key issues of the fault prediction. There is little effective method to predict the TBF of onboard subsystem for train control system. In this paper, a combined prediction model of TBF with high accuracy is proposed. Firstly, single prediction method including Echo State Network (ESN), Back Propagation(BP) neural network, Support Vector Machine (SVM) is used to predict the time between failures of onboard subsystem. In order to enhance the performance of the prediction, a combined prediction model is established on the basis of Seasonal-Trend decomposition procedure based on Loess (STL). Finally, simulation is employed with the accuracy of 96.49%. This study is helpful for the implementation of Preventive Maintenance (PM) tasks.