Machine Learning and Artificial Intelligence-Based Fault Diagnosis of Bearings and Rotating Machinery: A Comprehensive Review

Fault diagnosis of bearings and rotating machinery is a critical aspect of condition monitoring and predictive maintenance in modern industries. Bearing failures alone account for a significant proportion of breakdowns in rotating equipment, leading to economic losses and safety risks. In recent years, machine learning (ML), deep learning (DL), and reinforcement learning (RL) techniques have emerged as powerful alternatives to traditional signal-processing- based diagnostic approaches. This review presents a comprehensive analysis of recent research on ML- and AI-based fault diagnosis methods for bearings and rotating machinery, based strictly on the selected literature. Conventional machine learning methods, deep learning architectures, and advanced reinforcement learning frameworks are critically examined with respect to feature extraction, classification performance, dataset characteristics, and practical challenges. Comparative analysis highlights the advantages and limitations of different approaches, while key research gaps and future directions are identified, particularly in relation to data imbalance, generalization to real industrial conditions, and model interpretability.