Analysis of bearing vibration signatures using empirical mode decomposition and Hilbert-Huang transform

Vibration-based damage detection has grown over the past decade with considerable attention paid towards monitoring of civil structures and machines. Much of the focus has been based on comparison of system properties 'before' and 'after' damage, with the premise that the system can be treated as linear in both states. This work uses a novel method for analyzing vibration signatures, aimed at monitoring structures and machines for incipient damage. This non-destructive method is based on a new technique, Empirical Mode Decomposition (EMD) and Hilbert-Huang Transform (HHT) for non-stationary and non-linear time series analysis. The technique essentially allows the decomposition of the time-domain signals into intrinsic oscillatory modes, providing a time-frequency distribution. Results from analysis of vibration signatures from anti-friction bearings will be presented. The data was obtained from experiments conducted on a lab test set-up specifically designed for this study. Analysis based on the time-frequency plots and Hilbert-Huang spectrum illustrate that this new approach may allow for the development of a reliable damage detection methodology for antifriction bearings.

Duke Scholars

Author Liming Salvino Civil and Environmental Engineering