机械故障信号小波包分解的时域特征提取方法研究

RESEARCH ON THE EXTRACTION METHOD OF TIME-DOMAIN SYMPTOMS BASED ON WAVELET PACKET DECOMPOSITIONS OF MECHANICAL FAULT SIGNAL

本文利用小波包分解机械故障信号 ,对分解后的不同频带内的分解系数进行时域重构 ,分别对重构时间序列用时域分析的方法提取对故障敏感的特征参数 ,实现对往复机械等复杂机械的故障诊断 ,并以某油田往复泵为例。验证了该方法对这类机械进行故障诊断的有效性。

In the paper the mechanical fault signal is analyzed by wavelet packet decomposittion. On the basis, by use of the decomposed coefficients a time series is reconstructed, from which some valid fault-identified parameters are extracted by time-domain analysis. The effectiveness of the method which is used to diagnose reciprocating machine is verified by a real example.