基于小波多分辨分析的往复机械故障特征提取与识别  

Extraction and Recognition of Fault Features of Reciprocating Machinery Based on Wavelet Analysis

作　　者：王江萍[1,2] 王鸿飞[1,2] 王素英 

机构地区：[1]西安石油学院机械工程系 [2]石家庄广播电视大学

出　　处：《西安石油学院学报》1998年第1期30-32,52,共4页Journal of Xi'an Petroleum Institute

摘　　要：本文从往复机械故障诊断领域中特征信号处理的应用角度，探讨了利用小波多分辨分析与信息熵相结合，对往复机械故障进行诊断识别的方法．首先应用小波分解，将监测信号映射到由一个小波伸缩而成的一组基函数上去，在通频范围内得到分布在不同频段内的分解序列；在此基础上，对各分解序列进行ＦＦＴ变换，建立信号的小波特征熵，以此作为故障识别的特征参数，对往复机械运行故障进行诊断识别，并以压缩机振动监测信号为例，实现了不同频段范围内特征信息的提取与故障识别。The authors approach the method of fault diagnosis for reciprocating machinery by combining multi resolution signal decomposition of wavelet with information entropy.First by using wavelet decomposition,the monitored signal is cast upon a set of basic functions get from a wavelet by extending,so a set of decomposition sequences distributing in different freguency intervals are get within frequency domain,then by making FFT for each sequence,the characteristic entropy of wavelet of the signal is established,and it is used as characteristic parameter for fault recognition.As an example,the monitored signal of the air valve in a compressor is processed with the method,and it is shown that the method is effective for extraction and recognition of fault information.

关 键 词：往复机械 故障诊断 小波分析 空气压缩机 

分 类 号：TH45[机械工程—机械制造及自动化]