一种适用于机械非平稳故障信号的分析方法

该文提出了一种适用于机械非平稳故障信号的特征提取方法,鉴于该类信号时域特征明显,首先利用小波包进行分解,对分解后的不同频带内的分解系数进行时域重构,分别对重构的时间序列用时域分析的方法提取对故障敏感的特征参数,通过简化特征参数,确定判别阈值,使该方法更便于现场工作人员对机器故障的识别.同时也为具有同类特征的非平稳信号提供了一种行之有效的分析方法.文中最后用某油田往复泵实测数据验证了该方法的有效性.

A New Analytical Method for Non-stationary Fault Signal of Mechanical System

In this paper a new analytical method for non-stationary fault signal is presented. Due to the distinct feature in time domain, the signal is first decomposed by wavelet packet. On the basis, the decomposed coefficients are reconstructed to form a new time series, form which some valid fault-identified parameters can be exlracted by time-domain analysis. With less symptom parameters and appropriate selected thresolds, the method is more convenient for workers to inspect faults in field. The effectiveness of the method is verified by a real example.