Abstract
Monitoring the health condition of industrial drives, especially, induction motors is crucial to maintaining the reliability and efficiency of industrial processes. Fault detection techniques tend to be invasive and expensive despite their effectiveness. This paper introduces a non-invasive methodology that leverages acoustic signals processed through sophisticated deep learning models, such as Convolutional Neural Networks (CNNs) and Long Short-Term Memory Networks (LSTMs). We aim to accurately detect and classify different fault conditions in induction motors by extracting frequency features from acoustic data. Our results show marked improvements over conventional methods, offering a valuable predictive non-invasive maintenance tool.