Research On Wheel Wear Analysis And Re-Profiling Strategy Optimization Of EMUs Based On Data-Driven

The wear degree and service health status of EMU wheels have an important impact on the running safety of trains.Although more advanced automatic detection methods are currently used for EMU wheels,a large number of daily inspection data and wheel wear patterns have not been fully utilized.At the same time,the wheel re-profiling operation mainly relies on a fixed time limit or empirical judgment,and there are problems of excessive re-profiling or unreasonable re-profiling,which affects the operating efficiency of EMU trains and brings potential safety hazards.Therefore,based on EMU wheel shape detection data,this paper conducts in-depth research on wheel wear analysis and optimization of reprofiling strategy.In order to analyze the distribution range and wear characteristics of wheel shape wear,a fourth-order polynomial model was established to describe the relationship between wheel tread wear and running mileage,and the causes of wheel wear were analyzed from three factors: running route,wheel type,and wheel wear parameters.Correlation analysis and Gaussian distribution test were used to obtain the influence law of wheel shape parameters on the wear rate,thus laying a foundation for considering the restoration of tread shape in subsequent wheel re-profiling operations.Based on the historical data of wheel shape detection,a data-driven wheel wear prediction model is proposed.The data samples of the prediction model are constructed from the wheel shape detection data by the C-C method and the principal component analysis method,the LM algorithm is used to quickly adjust the weights of the prediction model,and the hidden nodes are reduced by the OMP method,thereby constructing a simplified and accurate wheel.Wear prediction model.Compared with other classical forecasting methods,the advantages of the forecasting model are confirmed.Aiming at the problem of "when to re-profiling" the wheel of the EMU,a decision model of the optimal wheel re-profiling cycle with closed-loop alternation of wheel wear and re-profiling loss is constructed.On the basis of the data-driven wheel wear prediction model,combined with the approximate model of the amount of re-profiling and the thickness and height of the rim before and after the re-profiling,it is determined that the wheel is in a closed-loop alternate state of wear and re-profiling loss under different re-profiling cycles.Whether the diameter value is used to the limit,so as to obtain the expected service life of the wheel and the number of re-profiling.Aiming at the problem of "how to re-profiling" the wheels of EMUs,this paper proposes an optimization strategy for the economical reprofiling of the wheels of the whole vehicle.The re-weighted scale nearest iteration method is used to register all affine deformed vehicle profiles,and different thin rim tread profiles are introduced to establish a single-wheel overhaul optimization model.On this basis,the driving requirements and experience of wheel diameter difference are considered the influence of the restored wheel shape on the wear characteristics can be obtained from the economical wheel re-profiling results of the whole vehicle.Compared with the actual re-profiling workshop,the re-profiling amount can be reduced by 21.1%.On the basis of the above work of establishing the theoretical model,the optimization decision-making system for train wheel re-profiling was developed and completed based on C# and SQL server database.Through the construction of wheel detection data access layer,function realization business logic layer,and operation result visualization interface layer,functions such as wheel wear analysis,wheel wear prediction,and rapid generation of vehicle wheel re-profiling plans are integrated,so as to formulate wheel re-profiling strategies for the entire vehicle.Provide intelligent decision-making.