Abstract: In order to identify the harmonic source of grinding surface of angular contact ball bearing raceway, a spectrum analysis method based on wavelet decomposition and reconstruction was proposed, which was combined with the power spectral density (PSD) and correlation analysis. Firstly, the actual profile information of 7006C inner raceway in grinding test was collected by using surface profilometer based on wavelet decomposition and decomposed into 5 layers using “db5”wavelet basis, according to the roundness error, waviness and surface roughness demarcation index, the 5 layers of decomposition signals were reconstructed into low, medium and high frequency signals. Then, the reconstructed frequency bands were analyzed by the Power Spectral Density method (PSD), and the frequency characteristics of the main harmonic sources were identified. Finally, the frequency of error source was extracted, and the correlation between error source and machined surface harmonic was analyzed. The results show that,comparing with the unbalance of the motor rotor, the half-speed vortex of the oil film and the improper adjustment of the electromagnetic centerless fixture, the vibration of the grinding wheel spindle account for a larger proportion of the surface harmonics of the 7006C inner raceway processed by the grinding test. This method can be used as a supplement to the method of finding harmonic sources and comparing the correlation degree of harmonic sources in raceway grinding of angular contact ball bearings.

Key words: angular contact ball bearings;bearing raceway; raceway grinding;surface harmonic; wavelet analysis;power spectral density (PSD)