Integrating FOC and SVPWM for Enhanced Induction Motor Control with Machine Learning Insights

This paper presents the design and implementation of Field-Oriented Control (FOC) integrated with Space Vector Pulse Width Modulation (SVPWM) for managing a 4 kW, 50 Hz squirrel-cage induction motor. The system effectively regulates rotor speed despite load disturbances, with significant responses observed at 3 and 7 seconds, marked by oscillations and instability. The actual $i_{sq}$ current stabilizes over time, closely tracking the reference current, while rotor flux remains stable with spikes reflecting load changes. Torque responses demonstrate oscillations characteristic of an under-damped system but stabilize post-disturbance. Line-to-line voltages confirm efficient PWM signal generation, maximizing DC bus voltage utilization. Performance validation through machine learning models reveals varied accuracy scores: Linear Regression (72%), Polynomial Regression (85%), Decision Trees (83%), Random Forest (92%), Support Vector Regression (88%), and Neural Networks (96%). These findings highlight the effectiveness of FOC and SVPWM, enhanced by machine learning, in improving the operational efficiency of induction motors under dynamic conditions.