MTS-HMM for Rolling Bearing Health State Assessment

Health state assessment is a key technology for system Prognostic and Health Management (PHM) and an important basis for remaining useful life prediction and maintenance decision-making. Rolling bearings are key components of rotating machinery equipment and also one of the most vulnerable components. It has important theoretical and practical significance to evaluate the health state of rolling bearings. In this paper, the empirical mode decomposition (EMD) method is used to extract the vibration signal characteristics of rolling bearings, the dimension of the features is reduced by Mahalanobis-Taguchi system (MTS), and a Hidden Markov Model (HMM) is combined with Mahalanobis distance (MD) to complete health state assessment of rolling bearings. The experimental bearing life cycle data set provided by the Intelligent Maintenance Center of the University of Cincinnati is selected to verify the effectiveness of the proposed method. The experimental results show that the method can detect an early failure and has good sensitivity.