基于子空间预测器的预测电压控制方法

Finite control-set model predictive control (FCS-MPC) is regarded as a promising control method in power converters due to its excellent performance, simple implementation, and fast dynamic response. However, standard model predictive control suffers from a high dependency on model parameters. To solve this issue, a novel finite set subspace predictor-based predictive voltage control strategy is proposed in this article, aiming to improve the robustness of the controlled system while maintaining the attractive features of standard FCS-MPC. Specifically, by replacing the original physical model with a subspace predictor at each operating point, the optimal control input can be directly obtained according to the reference output trajectory, and the control action can be executed without knowing the system structure and load parameters. This method uses only historical input–output data in the prediction process, thereby avoiding performance degradation caused by variations in load parameters. Finally, experiments on a three-level neutral-point clamped inverter validate the proposed design.