Remaining Useful Life Prediction And Health Management Of On-board System Equipment In High-speed Railways

In recent years,high-speed railways have made rapid development with their advantages of "large-capacity","high efficiency","low energy consumption",and "sustainability",meanwhile,it plays an important role on improving accessibility to broader geographic areas,increasing the mobility of residents and driving regional economic development.However,with the continuous increase of global high-speed railway lines and the total mileage,excessive maintenance costs have gradually become a key factor restricting the sustainable development of high-speed railways.The lack of mastery of the performance degradation laws of high-speed railways is the fundamental reason for lagging equipment maintenance timing,passive maintenance mode,and high maintenance costs.On the premise of ensuring the safe and efficient operation of trains,reducing maintenance costs as much as possible is one of the hot research issues on the sustainable development of high-speed railways.Based on the current fault disposal and maintenance plans for the on-board system in high-speed railway train control systems,this dissertation first analyzed the failure modes and effects of the on-board system,and calculated the risk priority number of the different functional units,and then selected the typical functional unit based on the risk priority number to establish performance degradation models for different failure mechanisms of the equipment,meanwhile realized the prediction of remaining useful life on equipment-level,functional unit-level,and system-level,finally designed the condition-based maintenance decision-making method under multi-equipment and multi-failure modes.According to the research content of this dissertation,the following innovates have been achieved:(1)Aiming at the problem that the root cause of failure in the mixed failure modes of the on-board system is difficult to find,a functional failure mode and effects analysis method based on the multilayer flow model was proposed.The sophisticated and multi-dimensional failure modes and effects clustering inside the on-board system were assorted,and the internal hidden association rules were drawn to describe the chain propagation of failure modes and effects.The propagation of failure modes from equipment-level,functional unit-level to system-level was demonstrated using two case studies,finally,the risk priority number of each functional unit in the on-board system was calculated,which solved the problem that traditional methods cannot reflect the on-board system functional structure and track faults.(2)Aiming at the problem that the degradation model for electronic equipment in the on-board system is difficult to establish,On the premise of multiple failure mechanisms and insufficient full life cycle data,a board-level physics of failure model under multiple components and failure mechanisms was proposed for the remaining useful life prediction,the equivalent system was established based on failure mode analysis,and then a board-level physics of failure model was established under multiple components and multiple failure mechanisms,and Balise transmission module in the typical functional unit was used as an example to verify the model,thereby the remaining useful life prediction was achieved under the multiple components,multiple failure mechanisms,and insufficient full life cycle data.(3)Aiming at the problem that life prediction of transmission equipment in the on-board system without life cycle data is hard,a fusion remaining useful life prediction algorithm combined particle filtering and neural network was proposed,several accelerated life tests under the load condition were designed to obtain the whole life cycle data,the optimal weights were allocated to balance the advantages between the physics of failure model and the data-driven algorithm,and D-cable in the typical functional unit was used as an example to verify the method,thereby the remaining useful life prediction was achieved under incomplete condition monitoring system and physics of failure model.(4)Aiming at the problem that unknown degradation process of equipment in the on-board system caused passive maintenance mode,the condition-based maintenance decision-making method under multi-equipment constraints and multiple failure modes was proposed,based on the remaining useful life from different levels of equipment,functional unit and system,and consideration about the economic and resource dependency constraints between multiple equipment,the cost-time function was constructed,and then the condition-based maintenance decision-making method under multi-equipment constraints and multi-failure modes was determined,the typical functional unit was taken as an example to verify this method,finally the problem of high maintenance cost caused by lagging repair time and passive maintenance mode was solved.This dissertation used the on-board system actual fault data and combined simulation tools to verify the effectiveness of the model and algorithm proposed.The results can provide a higher theoretical reference for the prognostics and health management system design of high-speed railways in China.