An Intelligent System for Bearing Fault Diagnosis of Induction Motor using Wavelet Transform Based Deep Learning Framework

Induction motor (IM) plays a significant role in the production sectors, owing to its high durability, rugged nature, high efficiency and low cost. Electrical motor condition monitoring is a revolutionary technique for detecting defects in induction motors. As a result, bearing degeneration has become the most common cause of induction motor rotor failures. The detection of these defects has gained momentum globally since it can lessen the number of unexpected breakdowns of a crucial system. Aiming to the objective, this paper proposes an intelligent system for classification of bearing faults in an IM using wavelet transform aided deep learning (DL) technology. Continuous wavelet transform (CWT) has been employed to generate the time frequency (TF) representations from the bearing fault signals. The TF images are fed to a cascaded dense convolutional network (CAD-CN) for necessary feature extraction followed by classification. The primary aim of this article is to analyze various TF patterns, obtained from four mother wavelets, and process those patterns for necessary diagnosis. The proposed module extracts feature maps in several stages and strengthen features by concatenation operation unlike classical DL frameworks. Experimental findings reveal the superiority of the proposed module in terms of performance metrics, and it is found that CWT using Morlet performs the best.