Regression Model for Level-1/2 Maintenance Plan of China Railway High-speed Electric Motor Unit

The total operation mileage of China High-Speed railway network is more than 30 thousand kilometers, and over 3600 High-Speed Electric Motor Unit (EMU) carry out the operation, making China the largest High-Speed railway network all over the world. It benefits a lot to improve the operation efficiency and reduce the maintenance cost of High-Speed railway by scientifically planning the daily operation and repair process of EMUs. Taking the specific EMU depot as the research object, this paper proposed a regression model which includes two steps to estimate the Level-1/2 maintenance plan of EMUs. In the first step, the influence factors of average daily running mileage of EMUs are analyzed, and a multiple regression model with the EMU routing, allocated number of EMUs to the depot, on-line rate, and average fault rate as the independent variables was built to forecast the average daily running mileage. Based on the average daily running mileage of each EMUs as input, the second step constructed an optimized model with the minimum maintenance cycle waste mileage as the objective function to equalize the maintenance expiration time between multiple EMUs. To solve the optimization model, particle swarm optimization was applied to obtain the best solution. Finally, this paper predicted the repair plan of EMUs of the depot instance in the year 2021.