服役工况下的高速列车轴承温度预测模型构建

Construction and Optimization of Bearing Temperature Prediction Model for High-speed Train Based on Service Condition

轴承温度预测对高速列车服役状态的评估以及运维策略的制定具有重要作用。针对复杂服役环境下的高速列车轴温预测问题,首先构建了轴承热力学近似计算模型,并分析了引起轴承温升的服役工况敏感参数,再利用支持向量机回归的方法建立了基于服役工况参数的轴承温度预测模型。对高速列车轴承履历服役数据进行统计分析,构建轴承温升相对速度变化的延迟量,确定轴承与环境的温差对轴承温升的影响,并据此对预测模型进行优化。以某高速列车大齿轮箱滚动轴承为例进行预测模型构建及优化方法验证,测试结果显示:预测模型优化前的综合平均相对预测误差为1.64%,优化后为1.35%,降幅为17.7%,预测模型优化前的综合最大相对误差为35.6%,优化后为17.7%,降幅50.3%。

Bearing temperature prediction plays an important role in the evaluation of the service status of high-speed trains and the formulation of operation and maintenance strategies.Aiming at the prediction of high-speed train axle temperature in complex service environment,the bearing thermodynamics approximate calculation model was constructed firstly,and the service parameters sensitive parameters of bearing temperature rise were analyzed.Then the support vector machine regression method was used to establish the service-based working condition and the bearing temperature prediction model of the parameters.On this basis,the statistical analysis of the bearing history data of high-speed trains was carried out to construct the delay of the relative temperature change of the bearing temperature rise,determine the influence of the temperature difference between the bearing and the environment on the temperature rise of the bearing,and optimize the prediction model accordingly.Taking a large gearbox rolling bearing of a high-speed train as an example,the prediction model construction and optimization method verification were carried out.The test results show that the comprehensive average relative prediction error is 1.64% before the prediction model optimization,and 1.35% after optimization,the decrease is 17.5%.The comprehensive maximum relative error is 35.6% before optimization,and 17.7% after optimization,the decrease is 50.3%.