A New Strategy for Detecting Gear Faults Using Denoising With the Orthogonal Discrete Wavelet Transform (ODWT)

This paper investigates a new strategy for early detection of defects in a power transmission pair of spur gears. Sensitivity to local defects is enhanced by processing the signal as follows. The orthogonal discrete wavelet transform (ODWT) of the band-pass filtered averaged signal is first obtained. This is followed by thresholding in the wavelet domain, thereby removing the low amplitude noise contribution. The inverse wavelet transform then essentially reconstructs the component of the signal that is due to the defect. Experimental results demonstrate the efficiency of this procedure.