Robust Predictive Current Control Using Discrete-Time Integral Action for Induction Motors

Model predictive control (MPC) has gained attention in electric motor drives due to its dynamic performance, flexibility, and ease of implementation. This technique utilizes a mathematical model of the plant to predict its behavior and select the appropriate control signal. However, parameter mismatches and unmodeled dynamics can negatively affect the accuracy of the control performance. Researchers have proposed variations for the classic MPC structure in order to increase its robustness. This article proposes a new robust predictive current control (PCC) for induction motors using a discrete-time integral action. From the dynamic model of the machine, we develop a discrete-time predictive model for the stator voltage, including an integral term that improves the controller performance in the presence of parameter mismatches. The cost function in this study differs from typical MPC approaches by utilizing stator voltage vectors instead of current, torque, and magnetic flux. Experimental tests evaluate the performance of the proposed control method for stationary and dynamic behavior. The results demonstrate that the robustness of the proposed control is superior to classic PCC, presenting lower oscillations in situations of large variations in inductance and resistance values, validating it as a viable alternative solution.