Fault diagnosis for rolling bearing based on lifting wavelet and morphological fractal dimension

Abstract

The vibration signal of rolling bearing is complex and nonstationary. In this paper, in order to overcome the difficulty of rolling bearing fault diagnosis, lifting wavelet and morphological fractal dimension are combined, puts forward a method based on lifting wavelet and morphological fractal dimension for rolling bearing fault diagnosis. The step of the diagnosis goes as follows: firstly, decompose the vibration signal of rolling bearing into three layers by lifting wavelet transform and restructure it, then analyze energy spectrum of the reconstructed signal to get the energy distribution of signal in time-frequency domain. Secondly, calculate the morphological fractal dimension of energy in time-frequency domain to judge the status of bearing. Finally, the morphological fractal dimension of bearing vibration signal in time domain and time-frequency domain would be compared. The result shows that the status of bearing can be distinguished more accuracy through the propoesd method.