Research On Condition Evolution Model And Preventive Maintenance Decision Methods For Repairable Components Of Emus Based On Stochastic Differential Equation

As energy becomes increasingly scarce,high-speed Electric Multiple Units(EMUs),with their energy-saving and environmentally friendly advantages,have become a popular choice for transportation,and the deployment scale of EMUs has grown steadily.Currently,the main maintenance method of high-speed EMUs in China is Time-Based Maintenance(TBM).However,as the technology for condition detection and the theoretical understanding of component degradation advance,limitations and drawbacks have emerged,such as under-maintenance,over-maintenance,wasted maintenance costs and so on.Additionally,high-speed EMUs are complex devices with a variety of faults and fault mechanisms,and the economic and fault relationships between components are interrelated.Repairable components of EMUs are also affected by external stochastic factors during operation,leading to dynamic and uncertain condition evolutions.Therefore,researching and analyzing the condition evolution models and preventive maintenance strategies of high-speed EMUs is of great theoretical significance and engineering application value in improving the maintenance level of repairable components and promoting the transition of EMU maintenance from the "TBM" to "TBM and Condition-Based Maintenance(CBM)",ensuring the safe,reliable,and economical operation of EMUs.This dissertation focuses on the key repairable components of high-speed EMUs,such as traction motors,traction transformers,pantographs,and mechanical components,which can be monitored for condition.The dissertation focuses on the construction,analysis and solution of the component’s condition evolution model,the preventive maintenance decision-making method,and other related topics.The main research achievements are as follows:(1)Aiming at the lack of effective model description and theoretical support for EMU preventive maintenance components in complex environment,the dissertation proposes a method for constructing and analyzing condition evolution models based on Stochastic Differential Equation(SDE).The SDE model is constructed using random theory.The internal stochastic factors are defined as the drift coefficient,which is approximated by polynomial function.External stochastic factors are defined as the diffusion coefficient,which is simulated by Brownian motion.Through the parameter solving process,it is revealed that the drift coefficient and diffusion coefficient are mutually independent.Three models of the Ordinary Differential Equation(ODE)model,the SDE model and the Generalized Stochastic Differential Equation(GSDE)model are analyzed.It is concluded that the ODE model and the GSDE model are the expectation model and refinement model of the SDE model respectively.After analyzing the common rule and progressive relationship between the TBM model and the CBM model,it is concluded that: the ODE model corresponds to the TBM model and describes the global condition monitoring data information in the process of component maintenance.The SDE model corresponds to the CBM model,which is the description of the local condition monitoring data information.It solves the technical measure of "TBM and CBM" in the maintenance system of EMUs.(2)Aiming at the coordination of multiple units of repairable components and the economic interdependence between components during operation,the dissertation proposes a decision-making method for Condition-Based Opportunity Maintenance(CBOM)based on SDE.By determining the optimal opportunity threshold for each component,the goal of optimizing maintenance cost rate is achieved,which alleviates the problem of high maintenance costs,high downtime,under-maintenance,and over-maintenance.The decisionmaking method ultimately improves the availability of repairable components,and thus provides valuable reference for on-site repairs of high-speed EMUs.(3)Aiming at the multiple-phase degradation characteristics of repairable components during operation and the possibility of fatal shock failures,the dissertation proposes a decision-making method based on GSDE and considering shock failures.The essence of the method is to divide the whole space into several local spaces,and use different stochastic differential equations to describe each local space.The mathematical expressions for the expected times of different maintenance methods are derived by analyzing the characteristics of different degradation phases and combining shock failures and imperfect maintenance.The optimal maintenance strategy is selected to optimize the maintenance cost rate objective.This method alleviates the difficulties in the scheduling of on-site repair resources and the implementation of repair plans,as well as the poor adaptability of repair strategies.It is more accurate and closer to the real situation of repairable components during operation.(4)Aiming at the issues of the lack of integration between the life laws of components and their actual status of the current maintenance strategy,excessive inspection and high maintenance costs,the dissertation proposes a joint optimization decision-making method for both TBM and CBM based SDE.By considering the internal and external degradation rates of repairable components,a comprehensive degradation rate function is established.On this basis,taking into account state fluctuations caused by external disturbances,a health deterioration model is constructed.Markovian and Brownian features of the model are analyzed,and its numerical solution method is studied.Furthermore,CBM maintenance decision-making model and joint TBM and CBM maintenance decision-making model are established and validates through examples,demonstrating their effectiveness.The joint maintenance decision-making method consideres the health condition of components during operation,makes appropriate maintenance plans,and provided necessary resources,fully utilizing the advantages of TBM and CBM models.This method addresses the problem of ineffective integration between TBM and CBM during practical maintenance processes.The analysis and research in this dissertation focuses on the condition evolution model of the key repairable components of high-speed EMUs.The dissertation proposes several preventive maintenance decision-making methods to effectively reduce the occurrence of under-maintenance,over-maintenance,and high maintenance costs during the maintenance process of EMUs,providing theoretical guidance and reference value for the maintenance work of repairable components of high-speed EMUs.