Information fusion framework for feature classification in machine fault diagnosis

On the basis of information fusion theory and machine Maintenance Information Fusion System, the process of information fusion and experimental results of various classifiers for equipment fault diagnoses are studied in detail. A new and general information fusion framework for feature extraction is suggested for the diagnosis, using MLP, RBFNN and KNN classifiers and Dempster-Shafer theory. The fault feature extraction and the application of the information fusion framework in fault diagnosis are carried out, respectively. It has been shown that the performance of the proposed framework is efficient in the computational experiment.