Berthil Cepstrum: a Novel Vibration Analysis Method based on Marginal Hilbert Spectrum Applied to Artificial Motor Aging

Abstract

Motor age determination as a part of condition monitoring heavily employs vibration analysis. This study introduces a new method for such analysis, based on concepts of cepstrum and marginal Hilbert spectrum. This new method, named Berthil cepstrum may be applied in general signal processing, not only when vibration signals are concerned. Classical marginal Hilbert spectrum has also been applied to the artificial motor aging data with excellent results. Furthermore, a ranking of known spectrum-based methods for determination of motor age together with the new methods introduced in this study has been made based on SVM and RELIEF attribute ranking, showing quality of the new methods.