Prediction Of Wheelset Tread Wear Based On Grey Theory

The wheel is one of the important components of train, which state is directly related to the quality of train operation safety. The faults of the train wheel tread and unpredicted failures not only affect repair strategy formulation, but also lead to disastrous accidents. Therefore, effectively predicting the wear state of the wheel and assess the remaining life of in-service wheel are helpful to understand the wheel health state and developing the optimization strategy of wheel remain, reduce security risks and economic losses of inadequate and excessive maintenance.Aiming at the actual problem of wheel state real time monitoring and repairing, the method of grey system theory is proposed in this thesis. The grey combination forecasting model is constructed on the basis of the improved grey model, accurate prediction of the wheel wear is realized and remaining life is analyzed. The main contents are as follows:(1) The relevant knowledge of the grey system theory and the most important grey GM(1,1) forecasting model of the grey system theory is introduced, the premise condition of application of grey model is analyzed, and the grey characteristics of wheel tread wear is discussed.(2) The fault of the grey GM(1,1) model forecasting precision is low in the long-term is proposed. In view of the model initial conditions can not make full use of new information, the method of the latest value of the accumulated sequence accumulative sequence instead of the initial value is proposed. At the same time, the metabolic model of continuous rolling sequence and updating the initial value, then, the improved model is accepted. According to the prediction results of the model evaluate wheel tread remaining life, and can be used to make maintenance strategy. The results of example show that the method is feasible and effective in predicting wheel tread wear.(3) Considering that the characteristics of grey uncertainty of the wheel tread wear forecasting and the influence of prediction accuracy for limitation of single models. The grey combination forecasting model is proposed, the improved grey GM(1,1) model and discrete GM(1,1) model and quadratic index model are introduced. Three models are selected as single model of the combination model, then, grey fixed weight combination forecasting model which uses entropy method to determine the fixed weight coefficients and grey variable combination forecasting model which uses error of each model to determine the variable weight coefficients are constructed to improve the accuracy of forecasting model.(4) This thesis regard wheel tread wear as the research object, constructe the fixed weight combination forecasting model and the variable weight combination forecasting model, and simulate actual data of wheel tread wear by using grey combination forecasting model. The results show that, the combination forecasting methods this thesis proposed can better reflect wheel tread wear trend, and has a higher prediction accuracy than any other single model, the results are more reasonable and scientific. At the same time, the results of the fixed weight combination forecasting model are compared with the variable weight combination forecasting model to verified feasibility and effectiveness of the grey variable weight combination forecasting model.This study is not noly helpful to improve the safety and reliability of train operation, but also to formula vehicle repair strategy scientifically. Reducing repair quantity and increasing the service life of wheels to improve using efficiency greatly of vehicles.Therefore, this study provides reference and a new idea for the wheel state monitoring and the realization of state maintenance.