高速铁路路基不均匀沉降的动力学识别

高速铁路路基不均匀沉降直接影响列车的动力特性.本文建立了车辆轨道路基空间耦合动力学模型,对沉降区车体振动、轮轨力、钢轨加速度和轨道板加速度等动力特性进行了分析.在车辆动力响应和轨道动力响应中,车体垂向振动加速度受路基不均匀沉降影响最为明显,且最有规律可循.将车体垂向振动加速度作为输入量,基于RBF神经网络对路基不均匀沉降的弦长和幅值进行识别,通过网络逼近性能和输出结果的训练不断优化神经网络模型,最后可得预测效果误差小于2％,可用于路基不均匀沉降的识别.

Dynamic recognition method of subgrade differential settlement on high-speed railway

The subgrade differential settlement directly impacts the train＇s dynamic characteristics when it travels on high-speed railway.In this case,this paper analyzes a series of dynamic characteristics by building the vehicle-track-subgrade space coupling dynamics model.These dynamic characteristics include the vehicle vibration of the settlement area,the wheel-rail power,rail acceleration,ballastless track bed acceleration and other dynamic characteristics etc.Vehicle vertical acceleration follows the law and is more obvious under the influence of subgrade differential settlement in the vehicle dynamic response and rail dynamic response.Regarding the vehicle vibration as an input,we can recognize chord length and amplitude of the subgrade based on the RBF neural network.What＇s more,we optimize the neural network model through training the result and approximation performance.Finally,we can predict result whose error will be less than 2 ％,which can be used to recognize the differential settlement of the subgrade.