Acta Polytechnica Hungarica Vol. 10, No. 3, 2013
– 17 –
On Finding Better Wavelet Basis for Bearing
Fault Detection
Lajos Tóth
Department of Electrical and Electronic Engineering
University of Miskolc
H-3515 Miskolc-Egyetemváros, Hungary
e-mail: elklll@uni-miskolc.hu
Tibor Tóth
Department of Information Engineering
University of Miskolc
H-3515 Miskolc-Egyetemváros, Hungary
e-mail: [email protected]-miskolc.hu
Abstract: This paper considers the comparision of the Meyer and Morlet wavelet for
bearing fault diagnosis. We created a wavelet based upon a transient vibration signal
model established for signals generated in deep-groove ball bearings with pitting (spalling)
formulation on their inner race. The wavelet creation used the sub-optimal algorithm
devised by Chapa and Rao that matches a Meyer wavelet to a band limited signal in two
steps. We tested the applicability of the matched wavelet for identifying this kind of bearing
failure. The Morlet wavelet was used as a benchmark for evaluating the performance of the
matched wavelet since many publications show its successful application. It was shown that
for analysing exponentially or near-exponentially damped vibration responses like the
vibration produced by spalling on the inner race of a deep-groove ball bearing, the Morlet
wavelet is a reasonable choice and gives better results than the Meyer wavelet.
Keywords: Wavelet analysis; bearing vibration analysis; wavelet matching; condition
monitoring
1 Introduction
It is known that the Discrete Fourier Transform (DFT) is most suitable for testing
finite-energy, periodic, time-discrete quantities. The reason why the DFT is still
used effectively for the vibration analysis of bearings is that most of the complex