小波包—自回归谱分析及在振动诊断中的应用

Wavelet Packets-AR Spectrum Analysis and Its Application in Vibration Diagnosis

概叙了小波、小波包理论。由此提出一种称之为小波包—自回归谱分析的新方法，它将小波包和自回归模型结合起来。该方法在机械诊断中，实现不同频道范围内，不同零部件故障信息分离和提取。并以挖掘机提升系统齿轮箱的振动信号为例，说明该方法是提取弱故障信息并进行早期诊断的有效方法。

This paper briefly describes the basic principle of wavelets and wavelet packets theory. A new methodcalled wavelet packets-AR spectrum analysis is established via combining wavelet packets with AR model. The method can carry out multiresolution signal decomposition and extract feature information of different machine parts in condition monitoring and fault diagnosis of machinery. As an example,vibration singalof excavator elevating gear box is analyzed. It is demonstrated that this new method is very effective for extracting small fault feature information and early diagnosing latent mechanical faults.