并网逆变器最优开关序列模型预测电流控制
Optimal switching sequence model predictive current control for grid connected inverter
胡存刚 1孙晓磊 1张悦 2芮涛 3尹政 1冯壮壮 1王尧1
作者信息
- 1. 安徽大学 电气工程与自动化学院,安徽 合肥230601
- 2. 中国电力科学研究院有限公司 新能源与储能运行控制国家重点实验室,北京100192
- 3. 安徽大学 互联网学院,安徽 合肥230601
- 折叠
摘要
传统的并网逆变器模型预测电流控制方法仅在控制周期内对电流波形进行单点预测,忽略了逆变器开关状态变化引起的输出电流纹波,为此提出一种改进的最优开关序列并网逆变器模型预测电流控制策略.首先,为减小开关切换次数、优化频谱分布,对空间电压矢量组合进行重新排列,设计6组开关序列,并根据开关序列提出基于无差拍的矢量持续时间计算方法.其次,在一个控制周期内设置8个电流预测点,以计算开关序列在控制周期内作用所产生的电流预测误差,并通过累加电流预测误差构造预测控制的价值函数,选择使价值函数最低的开关序列作为最优序列组合,提高并网电流波形质量.最后,通过实验验证所提策略在降低并网逆变器输出电流纹波方面的有效性.
Abstract
The traditional model predictive current control method of grid connected inverter only predicts the current waveform at a single point during the control cycle,ignoring the output current ripple caused by the change of inverter switch state.To solve this problem,an improved model predictive current con-trol strategy of optimal switching sequence grid connected inverter was proposed.Firstly,in order to re-duce the switching times and optimize the spectrum distribution,the space voltage vector combination was rearranged,and six groups of switch sequences were designed.According to the switch sequences,a vec-tor duration calculation method based on deadbeat was proposed;Secondly,eight current prediction points were set in a control cycle to calculate the current prediction error generated by the action of the switch sequence in the control cycle,and the value function of predictive control was constructed by ac-cumulating the current prediction error. The switch sequence with the lowest value function was selected as the optimal sequence combination,which improves the waveform quality of grid connected current;Fi-nally,effectiveness of the proposed strategy in reducing the output current ripple of grid connected invert-er was verified by experiments.
关键词
并网逆变器/模型预测电流控制/最优开关序列/无差拍控制/优化价值函数Key words
grid connected inverter/model predictive current control/optimal switching sequence/dead-beat control/optimize value function引用本文复制引用
基金项目
新能源与储能运行控制国家重点实验室开放基金(NYB51202201697)
国家自然科学基金(52207184)
出版年
2024
NETL
NSTL
万方数据