Multivariable sequential model predictive control of LCL‐type grid connected inverter

Abstract

The conventional single variable indirect model predictive control (SIMPC) of LCL-type three-level grid- connected inverter (GCI) is to indirectly control the grid current by controlling the output current of the inverter, which usually produces large errors and reduces the system anti-disturbance performance. Although the previously proposed multivariable weighted model predictive control (MWMPC) method of LCL-type GCI can directly control many variables, such as grid current, it requires setting the weight coefficient for each variable and involves a substantial amount of calculation. To due with these problems, this paper proposes a multivariable sequential model predictive control (MSMPC) method for LCL-type three-level GCI. The inverter output current, the filter capacitor voltage, and the grid current are gradually predicted and optimized under the variable control sequence of the LCL-type GCI mathematical model. Under the premise of achieving direct control of the variables above, this method can reduce the number of weighting coefficients and gradually contract the optimization space to reduce the computation required for multivariable prediction and optimization. This paper compares the proposed MSMPC to the conventional SIMPC and MWMPC in steady state, transient state performance, and anti-disturbance performance. The effectiveness of the proposed MSMPC is demonstrated through simulation and experimentation.