Abstract

This paper presents a model predictive control (MPC) strategy to solve an optimization problem based on system state and control variables, subject to constraints imposed by the control objective. The cost function exhibits convex characteristics, ensuring a unique control law. Additionally, the proposed control technique includes a formal analysis that ensures stability in the sense of Lyapunov. The control strategy is applied to a shunt active power filter (SAPF) connected to a voltage source and a nonlinear load, aiming to suppress the harmonics demanded by the load while ensuring that the source current remains sinusoidal and balanced, with a total harmonic distortion (THD) within regulatory standards. Simulation and hardware-in-the-loop (HIL) tests are conducted to evaluate the proposed approach. In both tests, two cases are considered: the first assumes balanced voltage conditions, while the second includes unbalanced voltage conditions and harmonic distortion. The simulation test, carried out in MATLAB/SIMULINK, demonstrates that our MPC reduces the THD from 26.71% to 1.59% in Case 1 and to 1.95% in Case 2. The HIL test, implemented using two RT boxes, shows a THD reduction from 24.57% to 2.77% in Case 1 and to 2.96% in Case 2. These results highlight the effectiveness of the proposed strategy, achieving values below regulatory standards for distribution networks. Furthermore, the MPC demonstrates superior performance in comparison with a passivity-based control strategy that employs interconnection and damping assignment (IDA-PBC), further emphasizing its efficiency and practicality. © 2025 The Authors