INTELLIGENT FORECASTING OF AUTOMATIC TRAIN PROTECTION SYSTEM FAILURE RATE IN CHINA HIGH-SPEED RAILWAY

Intelligent and personalized dynamic maintenance and spare parts configuration of high-speed railway have been the main trend to guarantee the safety capability of trains. In this paper, a new Automatic Train Protection (ATP) system failure rate calculation method is proposed, and the delay time and embedded dimension are determined by C-C algorithm. Then the phase space is reconstructed from one-dimensional time series to high-dimensional space. Based on chaotic characteristics of failure rate, a short-term intelligent forecasting model of failure rate of ATP system is established. The actual failure statistics from 2010 to 2018 are used as samples to train and test the validity of the model. From prediction results, it shows that the proposed chaos prediction model has an accuracy of 99.71%, which is better than the support vector machine model. Through the intelligent prediction of failure rate, this paper solves the maintenance inflexibility and imbalance of supply and demand of spare parts configuration.