Bearing degradation assessment based on Weibull distribution and deep belief network

The most important task of Prognostic and Health Management (PHM) in industrial company is to prevent catastrophic failures and reduce maintenance costs, where degradation assessment plays a significant role in schedule required preventive actions. In this paper, to assess the bearing's performance degradation, a new method which is based on the Weibull distribution and deep belief network is proposed. A health state and five degradation states are determined according to the fitted vibration feature by Weibull distribution, which is used to avoid areas of fluctuation of the statistical parameter. Deep belief network has the ability to model nonlinear time series, so it is used to classify the different states of the bearing. Finally, the experimental cases show that the proposed method has a good performance in degradation assessment.