ARMA模型盲辨识仿真研究及其在机械故障诊断中的应用

介绍一种新的基于高阶累积量的ARMA模型的递推盲辨识算法，并对ARMA模型的盲辨识方法进行了仿真研究，同时与常用的方法（残余时间序列法、q切片法、Newton法）进行了比较和分析。仿真结果表明，该算法具有良好的收敛性和准确性，运行速度快。特别是随着系统的阶次的增加，速度提高越明显。由于在计算中引入高阶累积量，因而该算法还可抑制高斯噪声的于扰。在此基础上，将该方法应用到机械故障诊断中，并进行实验研究，实验结果表明，该方法是有效的。

SIMULATION OF BLIND IDENTIFICATION WITH ARMA MODEL AND ITS APPLICATION TO MACHINE FAULT DIAGNOSIS

A new order-recursive method for blind identification based on high-order cumulants is introduced and compared with the traditional time series identification method,i.e.residual time series,q-slice and Newton method.The simulation results show that with this method only less calculation work is needed and good convergence and accuracy can be achieved,especially at higher system order.Because high-order cumulants are introduced into the calculation,this method can suppress the disturbance from gaussian noise and can be successfully applied to machine fault diagnosis.The experiment result shows the effectiveness of the method presented.