Broken Rotor Bars Fault Detection in Induction Machine Using Machine Learning Algorithms

This paper aims to diagnose the Broken Rotor Bars (BRBs) fault in a three-phase induction machine using seven Machine Learning Algorithms (MLAs), which are respectively Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Naive Bayes algorithm (NB), Decision Tree (DT), Random Forest (RF), Discriminant Analysis (DA) and Extreme Learning Machine (ELM). The extracted features by the application of the Fast Fourier Transform (FFT) on Hilbert Modules (HM) are used as inputs to train the used MLAs. To evaluate the performance of these algorithms, we use several predefined models for each algorithm. The obtained results show that three algorithms (SVM, KNN, and ELM) gave a high performance with an accuracy of 100%.