Dynamic Performance of a Six-Phase Interior Permanent Magnet Generator under Faulty Conditions with Neural Network Control

This paper investigates the dynamic performance of a 16-pole, 35 kW six-phase Interior Permanent Magnet (IPM) generator designed for wind energy conversion system (WECS) applications. The study analyzes the generator’s behavior under both healthy and single-phase open-circuit fault conditions using the MATLAB/Simulink environment. Two converter control strategies are evaluated and compared: (i) a conventional state-space vector modulation (SVM) controller and (ii) an artificial neural network (ANN)-based control scheme. The IPM generator is modeled in the dq-xy-0102 reference frame, and its dynamic responses, including electromagnetic torque, rotor speed, phase current, and power, are thoroughly analyzed. Simulation results reveal that while the SVM method ensures stable operation under healthy conditions, the ANN controller demonstrates notable performance improvements under faulted scenarios: torque ripple is reduced by 9.8%, current unbalance by 7.6%, and speed overshot by 11.4% compared to the baseline SVM control. Furthermore, the ANN-based scheme enhances fault-tolerant capability and accelerates transient recovery, improving system reliability and operational continuity. These findings confirm that the proposed ANN control significantly enhances the robustness and efficiency of six-phase IPM generators integrated into next-generation renewable energy systems.