Reinforced Morlet wavelet transform for bearing fault diagnosis

Abstract

Morlet wavelet transform has been successfully used in signal de-noising for bearing fault diagnosis. However, the existing approach has shortcomings in optimal mother wavelet parameters selection criterion and envelope signal reconstruction. This paper presents two improvements to overcome the two shortcoming of the existing approach, and proposed the reinforced Morlet wavelet transform. Two case studies were conducted to compare the performance of reinforced Morlet wavelet transform with the existing approach. Preliminary results have shown the fault detection consistency improvement from existing 17.7% to 67.6%. Results also shown of the early-fault-detection ability of the reinforced Morlet wavelet transform.