As a key component of a mechanical system, the extraction and accurate identification of rolling bearing fault feature information are of great importance to guarantee the normal operation of the mechanical system. Aiming at that the extraction of rolling bearing fault features and traditional support vector machine parameters affects the overall accuracy of pattern classification, the improved CEEMDAN (complete ensemble empirical mode decomposition with adaptive noise) time-domain energy entropy-based model for fault pattern recognition is proposed. The ICEEMDAN method is developed to decompose the signal to obtain the IMF component series. Then, the particular IMF components are selected according to the trend of correlation coefficient and variance contribution rate; meanwhile, the information entropy (power spectral entropy, singular spectral entropy, and time-domain energy entropy) of the screened IMF components is calculated to construct the feature vector sets, respectively. Finally, the feature vector sets are input into the PSO-SVM (particle swarm optimization-support vector machine) based model for training and pattern recognition. The effectiveness of the proposed method of the improved CEEMDAN time-domain energy entropy and PSO-SVM model is testified through numerical simulation and experiments on rolling bearing datasets. The comparison proceeded with the other mainstream intelligent recognition techniques indicates the superiority of the method with the diagnostic accuracy of 100% as for the final validation.

As a key component of mechanical system, the extraction and accurate identification of fault characteristic information of rolling bearing is very important to ensure its normal operation. The diagnosis accuracy is occasionally low due to the limitation of information collected by a single type of data source. In this paper, the bearing vibration signal and acoustic emission signal are employed as analysis sources, a novel method based on ICCEMDAN (improved complete ensemble empirical mode decomposition with adaptive noise) with optimized SVM (support vector machine) is presented for the fault information fusion, feature extraction, and fault pattern recognition of rolling bearing. Firstly, ICEEMDAN algorithm is developed to decompose the rolling bearing vibration signal and acoustic emission signal for a series of IMF (intrinsic mode function) components. Secondly, the valuable components that can characterize the original signal status are selected based on the correlation coefficient-variance contribution criterion. Thirdly, the singular spectral entropy of the reconstructed component is calculated as the eigenvalue and the two signal eigenvectors are fused as a new eigenvector set. Finally, the feature vector set is input into the optimized SVM classifier model based on PSO optimization for training and pattern recognition, in which the accuracy and efficiency of the classifier model and SVM classifier model are compared. Study of model simulation and fault simulation experiments show that the presented model based on the singular value entropy fusion of ICEEMDAN and PSO-SVM can effectively extract the fault characteristics of rolling bearing signals and has a desired performance in the accurate pattern recognition.