滚动轴承, 小波包变换, Hilbert变换, 故障诊断

 ,

For the modulation and energy concentration feature of fault ball bearing vibration signals, a fault diagnosis study based on wavelet packet energy and Hilbert transform was put forward. The vibration signals of ball bearing were decomposed and reconstructed using wavelet packet transform. And energy of every frequency band was calculated, then the signals of the frequency band with maximal energy were analyzed by applying Hilbert transform. Finally, the characteristic frequency of fault signals was extracted. The computation of fault features was accomplished artificially. A new method which can select fault features automatically was presented herein. Through processing and analyzing the practical ball bearing experimental data, it is shown that the fault diagnosis study can diagnose different running states of ball bearings due to surface damage accurately and quickly.