Model predictive DC-component power control for grid-connected inverters under unbalanced network

Abstract

Under an unbalanced network, model predictive control (MPPC) with a new definition of the instantaneous reactive power has attracted a great deal of attention due to its simple control structure and outstanding steady-state performance. However, the reactive power cannot precisely trace the nonzero reference. In this paper, a model predictive DC-component power control (MPDCPC) is proposed to tackle the above-mentioned problem. Additionally, the MPDCPC can eliminate oscillations on the reactive power and the negative-sequence current. Then, the corresponding mathematical formulas are derived to extract the DC-component powers, to calculate the DC-component power derivative and to modify the power reference. By regulating the DC-component power to trace the modified reference, the MPDCPC can reduce the current harmonic and remove the oscillations on the active power, reactive power or negative-sequence current. Simulation and experimental platforms are established to demonstrate the validity of MPDCPC. Results demonstrate that the MPDCPC can effectively suppress current distortion and remove oscillations on active power, reactive power or negative-sequence current. In addition, the MPDCPC can successfully break through the restrictions of the MPPC with a new definition of instantaneous reactive power.