Multi-objective optimal model predictive control for active neutral-point-clamped three-level inverter
LI Qianqian1, XIA Ronghua2, LIU Zhan3, DENG Bin3, XIA Zhenglong3
1. Jiangsu Normal University Kewen College, Xuzhou, Jiangsu 221000; 2. Jiangsu College of Safety Technology, Xuzhou, Jiangsu 221000; 3. School of Electrical Engineering & Automation, Jiangsu Normal University, Xuzhou, Jiangsu 221000
Abstract：Active neutral-point-clamped (ANPC) three-level converter, as a grid-connected inverter, can realize the heat loss balance of power devices and is suitable for photovoltaic power generation system. Aiming at the unbalanced heating of switches in traditional neutral-point-clamped (NPC) three-level grid connected inverter, a multi-objective optimal model predictive control (MO2- MPC) method for ANPC three-level grid connected inverter is proposed, using the principle that three-level output switch state needs to follow the unit level jump to exclude the switch state which doesn’t follow the principle and reduce the number of model optimization. According to the structural characteristics of ANPC three-level grid-connected inverter, the power loss of each switch is calculated in real time by linear fitting of the on and off losses of switches, and the unbalanced loss of each switch is added to the objective value function to minimize the unbalanced loss. Finally, the correctness and effectiveness of the proposed control method are verified by the experimental results of direct power control of ANPC three-level grid-connected inverter.
李倩倩, 夏蓉花, 刘战, 邓斌, 夏正龙. 有源中点钳位型三电平并网逆变器多目标优化预测控制[J]. 电气技术, 2021, 22(7): 13-18.
LI Qianqian, XIA Ronghua, LIU Zhan, DENG Bin, XIA Zhenglong. Multi-objective optimal model predictive control for active neutral-point-clamped three-level inverter. Electrical Engineering, 2021, 22(7): 13-18.
2021, Vol. 22 
