基于灰色关联度样本优化的高速列车轴箱轴承温度预测方法

Temperature Prediction Method of Axle Box Bearing of High-speed Train Based on Grey Correlation Degree Sample Optimization

建立了一种基于多基函数组合的最小二乘回归的轴温预测模型,并用灰色关联的方法对样本进行优化选择。采用灰色关联分析对轴温相关因素进行选择,并根据轴温与环温的温差分为高温差与低温差2个阶段,建立了一种多基函数组合的最小二乘预测模型,通过不断地对上一阶段模型的预测偏差加权校正,减小了模型的输出偏差。基于某型高速列车的履历轴箱轴承温度数据,运用文中所采用的模型对某一区间车辆从温升到开始制动这一阶段的温度进行预测,通过预测评价值对模型进行验证,证明了模型的有效性,基于多基函数组合的预测模型的最大绝对误差为0.65,最大相对误差为1.64%,绝对平均误差为0.35,相对平均误差为0.99%,其预测精度较单一基函数的预测精度改善。

In this paper,an axle temperature prediction model based on the combination of multi-basis functions and the least square regression is established,and the samples are optimized by grey correlation method.The related factors of shaft temperature are selected by grey correlation analysis,and according to the temperature difference between shaft temperature and ring temperature,it is divided into two stages:high temperature difference and low temperature difference,and a least square prediction model based on the combination of multi-basis functions is established.through the continuous weighted correction of the prediction deviation of the previous stage model,the output deviation of the model is reduced.Based on the temperature data of the axle box bearing of a certain type of high-speed train,the temperature of the vehicle in a certain section from the temperature rise to the beginning of braking is predicted by the model used in this paper,and the model is verified by the prediction and evaluation value.The validity of the model is proved.The maximum absolute error,the maximum relative error and the absolute average error of the prediction model based on the combination of multi-basis functions are 0.65,1.64%and 0.35,respectively.The relative average error is 0.99%,and its prediction accuracy is better than that of a single basis function.