Maximum average impulse energy ratio deconvolution and its application for periodic fault impulses enhancement of rolling bearing

Periodic fault impulses, inevitably occurring along with a localized defect, are crucial for incipient fault diagnosis of rotating machinery. Whereas, they are awfully weak and masked by other interferences in industrial appli-cations. Blind deconvolution methods (BDMs) are diffusely used in enhancing periodic fault impulses submerged in vibration signals. Due to some drawbacks, the applications of traditional BDMs are restricted. Therefore, the maximum average impulse energy ratio deconvolution (MAIERD) method is proposed, where the maximization of a new index called average impulse energy ratio (AIER) is tailored as the objective function. AIER takes the sampling point with the largest amplitude in every fault period as the location of fault impulse. Also, the fault period is detected by the autocorrelation function of the envelope signal. Furthermore, the Morlet wavelet is appointed as the initial filter. The synthesized signals and experimental data collected from two different rolling bearing test rigs are processed for verification. The results show that the proposed MAIERD method is superior in enhancing periodic fault impulses compared with five popular deconvolution methods.

Automated summary

This paper proposes a maximum average impulse energy ratio deconvolution (MAIERD) method for enhancing periodic fault impulses in vibration signals. The method utilizes an objective function and autocorrelation function to locate and detect the fault period, with the initial filter being a Morlet wavelet. Experimental results demonstrate that the proposed MAIERD method is superior to other deconvolution methods.