基于小波分析的齿轮箱故障诊断

Fault diagnosis in gearbox based on wavelet transform

通过对齿轮箱噪声机理进行研究,建立了相应的数学模型,提出了一种基于小波变换的故障诊断方法。其基本思想是:首先利用小波变换的多分辨率特性,滤除信号的随机噪声,并将提纯后的信号分解到不同的频区;通过对各频区信号进行比较与分析,选取能量集中、包含了调制成分的部分进行复小波变换,利用多尺度小波包络谱获得信号的包络信息;最后,结合傅里叶频谱,确定齿轮箱的故障原因。研究结果表明:该方法能有效识别出齿轮箱系统中轴和齿轮的缺陷,并能锁定故障发生位置、衡量故障程度。

Through the study of noise mechanism of gearbox, the corresponding mathematical model was estab-lished, and a fault diagnosis method based on the wavelet analysis was processed. The principle is to take use of the multi - resolution of wavelet transform to remove noise and decompose purified signal into different scales. Then, through the analysis and comparison of the signal of every scale, the frequency area which concentrates energy and contains the modulation components can be selected. After that, the envelope of fault feature could be extracted from multi-scale wavelet envelope spectrum. Finally, combined with Fourier spectrum, the failure causes of gearbox is determined. The results show that this method can identify the defect of shafts and gears, and judge the position and degree of gearbox faults effectively.