Application of Daubechies 44 in Machine Fault Diagnostics

This research focuses on the application of Daubechies 44 (db44) for gearbox vibration signals. Vibration signals of a sophisticated motorcycle gearbox system were collected in four conditions: Normal Gearbox, Slight-Worn gear, Medium-Worn gear and Broken-Tooth gear. To monitor the gearbox failures, new features were introduced based on four statistical features: standard deviation, variance, kurtosis, and fourth central moment of continuous wavelet coefficients of synchronized vibration signals (CWC-SVS). Variance of CWC-SVS was selected as the pattern for finding the most similar mother wavelet function across gear vibrations. Among 324 mother wavelet candidates, results show that Daubechies 44 (db44) has a distinctive pattern across gearbox signals. In sum, drawbacks of mother wavelet selection in gearbox diagnostics have been developed in this research.