Wheel Re-Profiling Policy For High-Speed EMU Based On Semi-Markov Decision Process

In recent years,with the rapid development of China’s economy and the improvement of social needs,high-speed railway is going into a comprehensive development era.The number of high-speed rail operations mileage and EMUs increases.It greatly improves the railway passenger transport capacity of China.It also has made tremendous economic and social benefits.Meanwhile,as more and more EMUs put into use,it puts forward higher requirements and challenges for the maintenance and management of the EMUs in service and effective maintenance.As one of the most important components of the EMU,the wheel is the most important part in maintenance research.The wear of wheel’s tread and flange is unavoidable due to the touching with track when moving,which threatens the safety and stability of EMU in service.To avoid performance degradation because of wear and ensure the safety of EMU,the wheel must be reprofiled or replaced.However,these processes generate significant material waste and high maintenance cost.With more recovery of the flange,the wheel must be replaced sooner.With less recovery,more maintenance cost increases due to re-profiling activities.Thus,the threshold of re-profiling activity and the recovery of the flange should be optimized,which is the key to make the re-profiling policy.To balance the threshold and recovery of the re-profiling activities,a dynamic re-profiling policy based condition monitoring at equidistant discrete time epochs is proposed.The policy can not only reduce the maintenance cost,but also extend its service life.Firstly,the wear rate of the tread diameter and flange thickness is determined based on the statistics and analysis of the wheel measurement data of a certain EMU,which reveals the wheel tread and flange wear law.Meanwhile,the distribution of re-profiling gain is obtained based on the result data of re-profiling activities.These research findings lay foundations to optimize the re-profiling policy.Secondly,to build the state space,the degradation process of the tread diameter and flange thickness is divided into finite degradation states.With the object of minimizing long-run expected maintenance cost per unit time,the re-profiling optimization model is established which is based on the semi-Markov decision process.We apply a policy-iteration algorithm to determine the optimal re-profiling policy.Thirdly,considering the influence of external shock and wear on the degradation process of tread diameter,the re-profiling optimization model is built,which is based on degradation threshold shock model and semi-Markov decision process.Finally,we combine the proposed policy with the fixed re-profiling policy based on the wheel measurement data of a certain EMU to demonstrate the method’s effectiveness.