用连续小波灰度图诊断齿轮故障

为了识别齿轮振动信号中的冲击性故障，利用连续小波变换对正常和故障齿轮的振动信号进行分析，将不同尺度下连续小波分解系数的绝对值用灰度图的方式表示出来，并利用特征矢量法估计信号功率谱进行验证，准确识别出了齿轮轴的不对中故障。分析结果表明，小波分析对信号具有多尺度分析能力，对振动信号中的冲击成分有很强的识别能力；连续小波变换灰度图不仅能直观反应齿轮的正常与故障状况，而且不会丢失冲击信号的时间信息，有利于寻找故障源，实现对齿轮故障的准确诊断。

Fault diagnosis to gears based on continuous wavelet coefficients plot

Continuous wavelet coefficients plot was used to analyze the vibrating signals of gears to distinguish the impacting fault.And the nicety of the result was validated by the power spectrum estimate using the eigenvectors method.The analyses show that the wavelet's capability of multi-resolution could distinctly recognize the impact components in signals.From this,whether the gears were in normal or fault states could be judged.Besides this,the time when the impact happened was embodied in the plots,which is very helpful to find the source of fault and so perform the accurate diagnosis.