Enhanced Model Predictive Current Control in Variable-Speed Drives Through Online Compensation of Converter Nonlinearities

This paper presents an enhanced model predictive current control scheme for induction machine drives that compensates converter nonlinearities such as dead-time and semiconductor voltage drops. The proposed approach extends the machine-converter model to incorporate fundamental and harmonic distortion states, while an Iterative Kalman Filter provides online estimates of stator currents, rotor flux, and distortion components. These estimates are integrated into both the predictive model and the steady-state reference, allowing realtime compensation and improved tracking accuracy. The control framework, implemented in MATLAB/Simulink and PLECS, was evaluated under various operating conditions, including low mechanical speed scenarios. Dynamic tests involving speed steps and load torque perturbations demonstrate that the proposed scheme achieves faster transient response and enhanced steadystate performance compared to a conventional field-oriented control system with an internal model predictive current control loop. The results confirm the effectiveness of the method in improving current estimation, reducing harmonic distortion, and enhancing dynamic robustness in low-speed operation.