An improved predictive power control with disturbance rejection for three-phase AC/DC converters at low switching frequency for grid operation

With the rapid advancement of digital signal processors, model predictive power control (MPPC) has found extensive use in power converters due to its simple principle and rapid dynamic response. Although conventional MPPC efficiently minimizes apparent power error in each sampling period through cost function, it overlooks switching frequency, resulting in significant losses. Such large switching frequency has caused a great decline in the service life of power devices. Therefore, a balance between switching frequency and power control effect should be considered in practical use. In addition, since MPPC requires an accurate mathematical model, the system disturbance will significantly affect the control effect, so the robustness of MPPC is not ideal. This paper proposes a predictive power control method with disturbance rejection at low switching frequency. Employing circular boundary-based predictive control achieves a balance between power distortion and switching losses without a precise mathematical model. Additionally, a linear extended state observer mitigates control inaccuracies caused by system disturbances. Results demonstrate that this method effectively reduces switching losses by over half, albeit with minimal DC voltage fluctuations.