Intelligent fault diagnosis and classification of induction motors based on Relax-1DCNN and Relax-RF

Abstract Induction motors (IMs) are prone to diverse faults during operation, and untimely detection can lead to catastrophic failures. To address this issue, this study develops a dual-module intelligent fault diagnosis system integrating the RELAX algorithm with 1D convolutional neural network (1DCNN) and random forest (RF). The RELAX algorithm first suppresses fundamental frequency interference to highlight intrinsic fault features. The RELAX-1DCNN framework realizes efficient fault detection with 99.75% average accuracy, while the RELAX-RF framework achieves precise fault classification with 99.19% accuracy. This integrated system provides a reliable solution for IM fault diagnosis, balancing high efficiency and classification precision.