Adaptive morphological filter to fault diagnosis of gearbox

Focusing on fault diagnosis extraction of gearbox, a novel approach is proposed according to the signal characteristics based on the adaptive morphological filter. Traditional linear filters have some limitations when extracting nonlinear features. As a nonlinear analysis method, the morphological filter has better performance on detail reservation and noise reduction, and can describe nonlinear morphological characteristics more clearly than linear filters. The structuring element (SE) of the morphological filter is similar to the window function of the linear filter. In order to avoid the drawbacks of the ambiguity of the selecting of SEs and the dependence on empirical rules, an adaptive morphological filter is proposed based on kurtosis maximization principle in this paper. The experimental results show that the adaptive algorithm can be more efficient to extract fault features of gearbox.