Controle preditivo aplicado a um conversor estático de potência trifásico conectado à rede elétrica

Abstract

This work presents a comparative analysis of three control methods – PID, Finite Control Set-Model Predictive Control (FCS-MPC) and Continuous Control Set-Model Predictive Control (CCS-MPC) – applied to a three-phase grid-connected converter, considering aspects such as steady-state behavior, dynamic response to disturbances and the quality of the injected current. For all control methods, Space Vector PWM modulation was used to generate the inverter switching signals. Initially, an ideal scenario was considered, without the inclusion of deadtime. Under these conditions, all controllers exhibited satisfactory disturbance rejection, while the PID and CCS-MPC methods also showed low steady-state error. In contrast, the FCS-MPC method resulted in a significant deviation from the current reference due to the high presence of harmonics components. The harmonic analysis showed that the PID and CCS-MPC controllers, when operating with 3% deadtime, produced currents with low total harmonic distortion – 2,17% and 0,93%, respectively – and individual harmonics that comply with the limits established by NBR 16149. On the other hand, the classical predictive control method, FCS-MPC, does not meet the individual harmonic requirements of the standard, despite maintaining total harmonic distortion below the 5% limit (achieving 3,64%). Based on the analysis, the CCS-MPC method demonstrated the best balance among dynamic response, total harmonic distortions and robustness to deadtime.