Bearing Damage Analysis with Artificial Intelligence Algorithms: A New Feature Extraction Approach for Current-Based Signals.

Three-phase induction motors are widely used in industrial facilities, where the maintenance of these machines is preponderant for industrial processes. Recent research reports that vibration-based data acquisition is the most common approach to perform bearing condition monitoring because it can extract more relevant information. However, the acquisition of vibration-based signals is expensive, requiring accelerometers and other external devices to transmit and process the signal information. Otherwise, current-based signals are directly measured by the supply system or inverters, enabling the current-based data acquisition in most industrial cases. In this context, this work introduces a new current-based method to identify bearing damages, applying artificial intelligence algorithms. Experimental and on-site tests present promising results, validating this approach for bearing damage diagnosis. [ABSTRACT FROM AUTHOR]