Radial Ball Bearing Inner Race Defect Width Measurement using Analytical Wavelet Transform of Acoustic and Vibration Signal

Abstract

In the present work, an experiment is carried out with a customized test setup where the seeded defects are introduced in the form of an axial groove on the inner race of a radial ball bearing. The nature of the acoustic and vibration signal bursts, and their correlation with the inner race defects, are established and estimated. Experimental investigation reveals that the analytical wavelet transform (AWT) is an effective tool for analyzing the acoustic and vibration signals, transmitted from the bearing, in order to characterize and measure the defect size. In the recent work, AWT followed by the time marginal integration (TMI) have been implemented on acoustic and vibration signals of a defective radial bearing. Size of the defect in the inner race of bearing is corroborated well with AWT scalogram. The segregation of the defect is carried out on TMI graph across the highest amplitude spike, which is due to signal burst (due to a contact of ball with bearing inner race defect). This manual demarcation on TMI graph in time axis provides the time duration (contact between a ball and the inner race defect). Using this time duration of the ball passed over bearing inner race defect, RPM of shaft mounted across bearing, and the fundamental train frequency, the defect width is estimated. The deviation of the measured width from the actual, using the proposed method, is sought below 5%. Summarizing, the proposed method can be reckoned a suitable and reliable measurement of radial bearing inner race defect width from acoustic and vibration signal.

Automated vibration-based fault size estimation for ball bearings using Savitzky–Golay differentiators

“…Two different forms of entry/exit observations have been reported in the literature, namely, step-impulse observations and double-pulse observations. Step-impulse observations have been reported by Sawalhi and Randall (2011) and later by Jena et al. (2012), Kogan et al.…”

“…These frequencies indicate the existence of a spall and they depend on operating speed and bearing geometry, all of which are measurable. However, the spall size cannot be directly extracted through frequency analysis (Randall, 2011b; Jena et al., 2012). The ultimate benefit of estimating the spall size is to allow the service life of a ball bearing to be extended even after detection, as long as the spall size has not yet reached a critical threshold (Sawalhi, 2007; Ismail et al., 2016).…”

“…Newly emerging studies (Sawalhi and Randall, 2011; Jena et al., 2012; Kogan et al., 2015; Moustafa et al., 2016) have shown that the fault response of a bearing in fact contains features reflecting two events that are associated with the spall width and how a ball enters and exits the spall. These points are: the entry point into the spall, “A”, and the exit point from the spall, “B”, as shown in Figure 1(a).…”

An effective health indicator for bearing using corrected conditional entropy through diversity-driven multi-parent evolutionary algorithm

“…Various methods are proposed by researchers to address these issues using the concept of decomposition and filtering of vibration signal. 7 These methods are wavelet decomposition and its variants, [8][9][10][11][12] enhanced empirical wavelet transform, 13 Wigner-Ville distribution (WVD), 14 intrinsic time-scale decomposition (ITD), wavelet packet decomposition (WPD), 15,16 local mean decomposition (LMD), 17 empirical mode decomposition (EMD) 18,19 and ensemble empirical mode decomposition (EEMD). 20 Since these methods have limited ability to process the non-stationary signals.…”

Force Analysis Due to Local Defect in Rolling Bearings for Fault Diagnosis

“…Researchers in Refs. [10][11][12] proposed different signal processing techniques to achieve the same objective.…”