Multi-objective Model Predictive Control of Grid-connected Inverters in Unbalanced Grids

Under unbalance grid voltage, grid-tie inverter have the problem of excessive grid current distortion, large power ripple, and DC side voltage instability. In order to solve these problems, a mulit-objective model predictive power control(MOMPPC) is proposed to address the issue of the distorting grid current in grid-tied inverter. Firstly, the instantaneous power when the grid is unbalanced is detailed analyzed to obtain instantaneous active/reactive power; then adjust the active/reactive power reference and change the reference coefficient matrix to achieve three flexible control goals. In order to avoid frequent inverter switching, optimize the design of the cost function, increase the number of switching times, and the DC side midpoint voltage weighting function to effectively reduce the DC voltage ripple. The comparison of Matlab/Simulink simulation and 3.8 kW prototype experiment show that MOMPPC can achieve flexible control of three goals. The traditional MPPC and MOMPPC control current harmonics THD of the grid side are 4.8% and 3.7%, respectively. The results show the effectiveness of MOMPPC strategy.