轮轨垂向力识别的时域方法

Wheel-rail vertical contact force identification based on inverse time-domain model

基于虚拟样机技术建立CRH_2高速检测车的动力学仿真模型(正演模型),并利用试验数据进行了模型验证。利用动力学仿真模型计算得到轮轨垂向力及轴箱垂向加速度。结合车辆—轨道系统的特点,将车轮看作一个单自由度振动系统,建立单自由度轮轨垂向力识别模型。基于杜哈密积分,在时域内推导了在线轮轨力识别算法。将正演模型的轴箱加速度作为输入,利用单自由度轮轨垂向力识别模型对轮轨垂向力进行了识别。识别结果与正演模型的轮轨力进行比较,两者相关系数达到0.82。该识别模型基于轴箱加速度逐点运算,可以快速高效地对轮轨垂向作用力进行识别,并实时监测轮轨状态。基于车轮动力学理论,建立多自由度轮轨垂向力反演模型:首先建立了轮对的横向、沉浮、侧滚等轮对运动方程,然后对部分参数加以简化,最后基于车辆—轨道系统的运动特点,将轮对运动方程解耦成单自由度进行求解。利用该模型识别的轮轨垂向力与正演模型轮轨力的相关系数达到0.97。单自由度车轮模型较为简单,但识别精度稍低;多自由度轮对模型较为复杂,但识别精度较高。两种模型的识别结果均满足工程需求,可以根据不同的需求选用不同的模型。

Based on Dynamic Simulation,the CRH_2 inspection vehicle-track model was established,and was validated by inspection data. The results showed that the CRH_2 inspection vehicle-track model was precise. Wheel-rail vertical contact force and the acceleration of axle-box were simulated by using the dynaic simulation model. Based on the characteristics of the vehicle-track system,the wheel-rail vertical force identification model of single degree of freedom was established. Based on Duhamel integral,on-line Wheel-rail vertical force identification algorithm was deduced.Using the simulation axle acceleration data,the vertical force was identified based on the identification model above.The vertical identified force was compared with the simulation results. The correlation coefficient of identified force and simulation force was 0. 82. The identification model above identifies the force point by point,and it was efficiently. It can monitor the status of wheel and rail in time. Based on the vehicle system dynamic theory,the identification model of multiple degrees of freedom was established. Firstly,the lateral,vertical and roll motion equations were established,then some parameters were simplified. Finally,three motion equations were uncoupled.Inputting the bogie and axle acceleration to the model of multiple degrees,the vertical force were identified. The correlation coefficient of identified force and simulation force was 0. 97. The results showed that this model was precise. The identification model of single degree of freedom has lower accuracy than the multiple degrees model,but it is simple. These two models could be used according to engineering requirement.