Bearing condition evaluation based on the shock pulse method and principal resonance analysis

Abstract

Compared to vibration-based measurements, the acceleration sensor-based shock pulse method (SPM) has been proven to be more effective, direct and simple for evaluating the condition of rolling bearings. Following the principle of SPM, the sampling frequency should be high enough to cover the resonance response of the acceleration sensor. This will undoubtedly cause too much redundant information to be measured, which impedes the industrial application of the SPM. To solve this problem, this paper proposes a principal resonance analysis (PRA) method based on information entropy to identify the principal resonance component with a bearing fault vibration signal. A novel condition evaluation method based on SPM and PRA is developed by replacing the resonance response of the sensor with the principal resonance component. The PRA is utilized for identifying the frequency band of the principal resonance, based on which, a band-pass filter can be constructed to realize further extraction. The extracted component participates in the subsequent SPM analysis, and eventually the condition of rolling bearings can be quantified and evaluated using the SPM theory. Experimental data under different bearing health states were analyzed. The analysis results showed that the proposed method can effectively evaluate the condition of rolling bearings.