A Novel Method for Bearing Fault Diagnosis Based on Novel Feature Sets With Machine Learning Technique

Abstract Bearings often experience small and medium raceway damage due to operating and loading conditions, which induces abnormal dynamic behavior. The rotor-bearing system is tested at various conditions, and the influence of each fault has been presented in this study. The fundamental bearing characteristics frequencies and statistical features withdrawn from a vibration response are utilized for fault identification using a machine learning algorithm. Extreme learning machine (ELM) and the supervised machine learning method K-nearest neighbor (KNN) network were utilized to classify vibration data collected experimentally under various operating conditions. Bearing characteristics frequencies and statistical features are applied to both proposed approaches and compared regarding their prediction quality. The result shows that the ELM has better performance over the KNN in precision of fault recognition up to 99% and thus feels promising for condition monitoring of industrial rotating machines. This work provides valuable insights for operation, maintenance, and early fault warning related to bearings.