Application of KPCA and coupled hidden Markov model in bearing fault diagnosis
Liu Tao1,2, CHEN Jin2, Dong Guang-ming2
1.School of Mechanical and Electrical Engineering, Kunming University of Science and Technology, Kunming 650500;2.State Key Laboratory of Mechanical System and Vibration, Shanghai Jiao Tong University, Shanghai 200240, China
Abstract:The fusion of multi-channels bearing monitoring information can obtain more accurate result in bearing fault diagnosis. Therefore, a rolling element bearing fault diagnosis scheme based on KPCA and coupled hidden Markov model (CHMM) is presented. At first, the features are extracted from bearing vibration signal from multi-channels respectively. Then, the KPCA is utilized to reduce the feature dimension. At last, the new KPCA feature is input into CHMM to train and diagnose the bearing fault. The data acquired from bearing at normal condition, inner race fault, outer race fault and rolling body fault are analyzed. And the results demonstrate the effectiveness and validity of the proposed method.