Investigation of an induction motor control system based on machine learning methods for predicting operational modes

This study presents an investigation of an induction motor control system based on machine learning methods for predicting its operational modes. The research is motivated by the need to improve energy efficiency, reliability, and safety of electric drives in metallurgic, mining, and chemical industries' plants operating under harsh conditions such as dustiness and explosion hazards. A closed-loop scalar speed control system for an induction motor has been modeled in MATLAB Simulink, incorporating IR compensation and an intensity setter. Data collected from the simulation model have been used to train machine learning algorithms: k-Nearest Neighbors (kNN) and Random Forest. The models have achieved high classification accuracy for operational modes – 94.56 and 96.44 %, respectively. The results confirm the effectiveness of integrating machine learning methods into monitoring and diagnostic systems for electric drives, supporting the implementation of Industry 4.0 principles and digital transformation in industrial facilities.