Classification of rolling element defect by extraction of defect features using wavelet transform and ANN

摘要

This paper presents a scheme for classification of rolling elements defect in a cylindrical roller bearing. Defect features are extracted from raw signal, and TMI graph. The extracted features are utilized to train the feed forward ANN. After training, test features are applied to ANN for the classification of defect. Accuracy of the proposed method in the classification of defects is 94%.