High-speed train axle box bearing temperature prediction method based on optimization of deviation degree

The prediction of axle temperature has practical application value to ensure the safe operation of high-speed train. In this paper, according to the deviation between the predicted value and the actual value of grey quadratic regression and GM (1, 1), the weight coefficients of each model are determined for combined prediction. Firstly, the model is constructed with reasonable modulus and overlapping algebra, and the original axial temperature data is preprocessed. Then the features of grey quadratic regression and GM (1, 1) model are analyzed respectively. Finally, in order to avoid the mean weight can not adapt to the fluctuation of axial temperature, the two models are combined by the weight coefficient of deviation degree. Through weight allocation, the prediction deviation correction of the first mock exam model is continuously reduced, and the prediction bias of the combined model is reduced. Based on the axle temperature data of a certain type of high-speed train, the model used in this paper is used to predict the axle temperature of a certain line vehicle. The model is verified by the model evaluation index, and the results show that the optimization effect is effective.