A Novel Adaptive Fault Diagnosis Method for Wind Power Gearbox

In the noisy environment, fault characteristics of the composite faults of the wind power gearbox are coupled with each other, which makes the extraction features more difficult. In order to extract the characteristics of composite faults, a new fault diagnosis method for wind power gearbox is proposed in this paper, namely the modified Savitzky Golay Laplacian of Gaussian filter (MSGloG). The method can not only solve the defects that the scale parameters of the Modified Laplacian of Gaussian filter (MloG) filter are not adaptive, but also overcome the problems that the smoothing effect is too much affected by noise. Firstly, determining the Laplace model of Gaussian filter, and using the least square convolution smoothing process to improve the signal-to-noise ratio of the vibration signal. Secondly, a new marginal envelope spectrum entropy index is proposed to measure the complex fault characteristics. Finally, a new chaotic grey wolf optimization algorithm is proposed, which uses the marginal envelope spectral entropy as the fitness function, and the purpose is to make the MSGloG noise reduction adaptive. The method extracted the faults of the bearing outer ring and rolling elements successfully.

Automated summary

A new fault diagnosis method for wind power gearbox, named MSGloG, is proposed in this paper to extract the characteristics of composite faults, successfully detecting faults of bearing outer ring and rolling elements. The method overcomes the difficulties caused by noise and effectively improves the signal-to-noise ratio of vibration signals.