Current-Sensorless Model-Free Predictive Voltage Control with Optimal Switching Sequence for LC-Filtered Grid-Forming Converters

The traditional model-based predictive voltage control with optimal switching sequence (OSS) often exhibits compromised precision in voltage regulation when confronted with parameter uncertainties and sensor measurement disturbances inherent in inductor-capacitor (LC)-filtered gridforming converter (GFC). To address this challenge, this paper proposes a current-sensorless model-free predictive voltage control with OSS (CSL-MFPVC-OSS) that eliminates the adverse impacts of parameter variations and measurement noises on voltage prediction, ensuring enhanced robustness of voltage regulation. First, a parameter-less Luenberger observer is designed to accurately estimate both parameter perturbations and capacitor current. Second, the estimated disturbance along with the estimated current are integrated into the voltage prediction, systematically compensating for the effects of parameter uncertainties and sampling errors on the duty-cycle calculation of each voltage vector. Finally, an experimental platform of LC-filtered GFC is established to validate the effectiveness of the CSL-MFPVC-OSS strategy.