Study And Develop On The Real Time Prediction Model Of Axle Temperature Of High-speed Trains Based On Grey Theory

Bearings are the key components of the high-speed train bogie,and abnormal working conditions will seriously affect the safe operation of trains.At present,high-speed trains use on-board axle temperature monitoring systems to monitor the real-time changes in axle temperature.If the axle temperature exceeds a setting threshold,the status of the axle is diagnosed abnormally,and an early warning or alarm is issued,thereby effectively avoiding major safety such as the burning-axle and cutting-axle.However,because this system cannot predict the development trend of the axle temperature,it can only wait passively.Once an abnormal result occurs,the processing time left for the relevant personnel is relatively small and the processing method is relatively simple.Therefore,research on the real-time prediction method of high-speed train axle temperature has important research and engineering application value for safeguarding traffic safety and improving the economical operation.The prediction model based on grey theory has the advantages of small sample size,high computational efficiency,and it is suitable for real-time prediction of high-speed train axle temperature.However,the prediction results of the grey prediction model are monotonous.When the axis temperature changes non-monotonously concave or convex,the prediction error for the data after extreme point is large,and there is still a problem for long-term prediction accuracy.In order to solve these problems,the paper carried out the following work:(1)A classical grey theory time series prediction model is applied to real-time prediction of high-speed train axle temperature and its modeling parameters are optimized.By analyzing the on-board temperature monitoring data of high-speed trains and a variety of classical grey theory time series prediction models,the model of axle temperature time series prediction based on grey theory and its evaluation indicators are constructed.By evaluating the applicability and prediction accuracy of different grey theory time series prediction models for typical axle temperature variation characteristics,the GM(1,1)model is selected and its modeling parameters are optimized.(2)A time series prediction model for axle temperature based on deviation degree and grey theory is proposed.Reconstructing the GM(1,1)model and the quadratic polynomial,constructing a grey quadratic regression model using the least square method.Improved GM(1,1)model's prediction accuracy is not high when the axis temperature is non-monotonous.Based on the GM(1,1)model and the grey quadratic regression model,a time series prediction model based on the degree of deviation and grey theory was established.Improved the problem of insufficient stability in predicting the monotonically changing segment of temperature in the grey quadratic regression model.Five-minute prediction of the axle temperature shows that the axle temperature time-series prediction model based on the degree of deviation and grey theory is superior to the grey quadratic regression model and the GM(1,1)model.(3)A high-speed train axle temperature prediction model based on multi-factors and grey theory is proposed.Through the analysis of the influence of working factors on the axle temperature and the grey correlation analysis,the working factors that have a great influence on the axle temperature are selected.And then by comparing classical grey multivariate regression prediction models,the GM(1,N)model is selected.The division of acceleration,constant speed and deceleration operation phases was carried out.The model of axle temperature prediction based on the multi-condition factors and grey theory was constructed and optimized.The model improves the accuracy of the long-term prediction of the axle temperature time series prediction model.Ten-minute prediction of the axle temperature shows that the prediction model based on multi-condition factors and grey theory is superior to the axle temperature time series prediction model based on deviation degree and grey theory in terms of prediction accuracy and stability.(4)Formulation a real-time prediction and warning strategy for high-speed train axle temperature,and developed its system.Aiming at the deficiencies of the current on-board axle temperature alarm system,The three-level real-time warning strategy of high-speed train axle temperaturethe is designed based on Class 1 warning for abnormal temperature rise diagnosis,Class 2 warning for abnormal temperature rise 10 minutes predict overrun and Class 3 warning for abnormal temperature rise 5 minutes predict overrun.Through the continuous tracking and prediction of abnormal temperature rise,improve the axle temperature early warning capability,prolong the emergency treatment time.Developed high-speed train axle temperature real-time prediction and warning system based on the researched prediction model and early warning strategy.System application example shows that the system has achieved the expected design function.Through the above work,the paper provides a solution to the problem of using grey theory for high-speed train axle temperature prediction and warning,and it can also provide reference for the prediction and early warning of bearing temperature in related fields such as wind power.