控制理论(社会学)
模型预测控制
涟漪
电压
逆变器
光伏系统
计算机科学
电容器
电压基准
共模信号
加权
工程类
电气工程
控制(管理)
物理
电信
人工智能
传输(电信)
声学
模拟信号
作者
Chang Liu,Chunshui Du,Xiangyang Xing,Ying Jiang,Rui Zhang,Xiaoyan Li,Chenghui Zhang
出处
期刊:IEEE Transactions on Industry Applications
[Institute of Electrical and Electronics Engineers]
日期:2022-09-01
卷期号:58 (5): 6297-6309
标识
DOI:10.1109/tia.2022.3178818
摘要
The five-level active neutral-point-clamped (5L-ANPC) inverter has been applied to increase the power capacity of photovoltaic (PV) systems. However, this type of configuration requires to balance neutral point (NP) voltage and flying capacitor (FC) voltage properly. For the operation of the transformerless five-level inverter, another requirement is to suppress the leakage current associated with the common-mode voltage (CMV) vectors. Due to the mutually coupled issues, the conventional model predictive control (MPC) methods reported so far cannot deal with them properly at the same time. To address these problems, a novel double-vector MPC (DV-MPC) strategy with zero CMV vector is proposed in this article. First, 19 voltage vectors with zero CMV are implemented as candidate vectors. Then, for the purpose to reduce output current ripples, two voltage vectors are penalized by the cost function without weighting factors to track the reference voltage. In addition, the dwell times are optimized to further mitigate current ripples. Finally, to avoid the sharply increasing computational burden of the online iteration caused by the introduction of the second vector and its corresponding redundant switching states, a voltage-balance strategy is presented for the DV-MPC method, which can determine the suitable switching combinations directly and adapt them to the selected vectors to balance the voltages of the NP and FCs together. The effectiveness of the proposed method is verified by the simulation and experiment at the end of this article.
科研通智能强力驱动
Strongly Powered by AbleSci AI