论文部分内容阅读
对基于EMD-AR模型的齿轮箱故障诊断进行了研究。对齿轮箱故障振动信号采用EMD方法进行分解,得到有限个平稳的IMF(本征模式函数),对其建立AR模型,将建立的每个AR模型残差的方差和自回归参数建立Mahalanobis距离判别函数,最后进行模式特征综合,确定齿轮箱的工作状态以及故障类型。研究表明,采用EMD-AR模型进行齿轮箱故障诊断是可行有效的,提高了齿轮箱故障检测的准确性。
The gearbox fault diagnosis based on EMD-AR model was studied. The vibration signal of gear box fault is decomposed by EMD method to obtain a finite number of stationary IMFs (eigenmode functions). The AR model is established and the variance and autoregressive parameters of each AR model are established to establish Mahalanobis distance discriminant Function, the last pattern feature synthesis to determine the work of the gear box and the type of failure. The research shows that the EMD-AR model is feasible and effective for gearbox fault diagnosis, which improves the accuracy of gearbox fault detection.