朔黄铁路基于车辆动力响应的轨道状态评估方法研究

Research on Track State Evaluation Method Based on Vehicle Dynamic Response of Shuozhou-Huanghua Heavy Haul Railway

基于轨道几何检测数据和动力学仿真模型计算的轮轨力、车体加速度等动力响应数据,首先利用相干函数分析轨道几何与车辆响应的相关性,然后利用深度学习中的长短时记忆网络建立预测模型,自动学习轨道几何与车辆响应的复杂关系,最后利用模型预测的车辆响应来识别轨道病害和评估轨道状态。结果表明:车体加速度、轮重减载率、脱轨系数主要与高低、轨向、超高的相关性显著;预测模型能够有效预测货车动力响应,预测值与仿真值的相关系数在0.8以上,为强相关;预测模型能够有效识别一些轨道几何不超限、但车辆响应超限的轨道几何隐形病害,实现基于车辆响应评估轨道状态。

Based on the track geometry detection data and the dynamic response data such as wheel rail force and vehicle body acceleration calculated by the dynamic simulation model,the correlation between track geometry and vehicle response was analyzed by using the coherence function,and then a prediction model was established by using the long short-term memory network in deep learning to automatically learn the complex relationship between track geometry and vehicle response,finally,the vehicle response predicted by the model was used to identify track diseases and evaluate track conditions. The results show that the vehicle body acceleration,wheel load reduction rate and derailment coefficient are significantly related to longitudinal level,alignment and cant. The prediction model can effectively predict the dynamic response of freight cars,and the correlation coefficient between the predicted value and the simulated value is above 0.8,which is a strong correlation. The prediction model can effectively identify some hidden track geometric diseases in which the track geometry does not exceed the limit but the vehicle response exceeds the limit,and realize the track state evaluation based on the vehicle response.