城轨列车轮对磨耗预测计算方法研究

Research on Wear Law and Prediction of Urban Rail Train Wheelset

为研究城轨列车轮对磨耗预测的计算方法,实现轮对尺寸预测,以国内某地铁线路轮对尺寸实测数据为研究对象,分别提取轮径、轮缘厚度、轮径差及车轮位置四项特征参数,并分析特征参数对轮对磨耗率的影响规律,利用XGBoost算法建立轮径、轮缘厚度磨耗率的仿真模型,模型的测试结果显示平均计算误差在允许范围以内,表明本模型具备良好的磨耗率仿真计算效果。在此基础上,进一步建立轮径、轮缘厚度的预测模型,利用某节车半年实测数据对预测模型进行实用性验证,对比结果表明,基于XGBoost的轮对尺寸磨耗预测模型对走行里程8万公里内的轮径值、轮缘厚度值预测的平均误差分别不超过0.175mm、0.3125mm,具有较好的预测效果。

In order to actualizing the prediction for wheel diameter and flange thickness,this article focuses on the numerical method to compute wheelset wear prediction urban rail train.Taking the data of wheel set size from a metro line,this study collects four characteristic parameters:wheel diameter,wheel flange thickness,wheel diameter difference and wheel position,and analyzes the influence regular pattern of characteristic parameters on the wear rate of wheelset.With XGBoost method,this study establishes a simulation model of wear rate of wheel diameter and flange thickness based on these four characteristic parameters.The error of the result calculated with the model is within the allowable range,which shows that this model has a good effect when calculating wear rate.Further more,we establish a prediction model of wheel diameter and flange thickness,and carry out practical verification on prediction model with measured data which comes from eitht wheels of a carriage in half a year.The comparison result shows that the predicting results are agreeable with the measured data.The average prediction errors of the wheel diameter and flange thickness are less than 0.175mm and 0.3125mm respectively at the running mileage of 80000km.