非线性自回归模型在轨道车辆转向架运行状态辨识中的应用

State Identification of Rail Vehicle Bogie System Based on GNAR Model

提出一种基于非线性自回归模型(GNAR模型)的轨道车辆转向架运行状态辨识方法。对不同状态的转向架系统输出信号建立GNAR模型,通过模型特征量,辨识系统不同状态,实现转向架运行品质实时动态监测。在ADAMS/Rail平台下的虚拟样机仿真实验说明了该方法的有效性和可行性,GNAR模型的特征量能够体现车辆转向架系统运行状态信息变化,用于判别其是否偏离正常功能,通过对这些特征量的监测和辨识,可以保证车辆系统的运行安全。

A state identification method based on a general expression for linear and nonlinear auto-regressive model （GNAR model） was presented for rail vehicle bogie system. GNAR models were established on the signal data from bogie system of different states and the system different states were detected and classified by means of eigenvector analysis. The virtual prototype simulation results in the ADAMS/Rail platform illustrates that the pro- posed method is effective and feasibile and it can be applied to monitor the technical conditions of vehicle mechanical assemblies.