改进神经网络的光伏逆变器参数自动优化控制研究  

Research on photovoltaic inverter parameter automatic optimization control based on improved neural network

作　　者：赵新宽 叶庭乔 张镇强 ZHAO Xinkuan;YE Tingqiao;ZHANG Zhenqiang(Zhuhai Collaborative Innovation Center of Smart Grid Technology,The Open University of Zhuhai,Zhuhai 519090,China;Zhuhai Singyes Renewable Energy Technology Co.,Ltd.,Zhuhai 519090,China;Zhuhai Maini Electric Control Technology Co.,Ltd.,Zhuhai 519000,China)

机构地区：[1]珠海开放大学珠海市智能电网技术协同创新中心,广东珠海519090 [2]珠海兴业新能源科技有限公司,广东珠海519090 [3]珠海麦尼电控技术有限公司,广东珠海519000

出　　处：《现代电子技术》2021年第1期93-97,共5页Modern Electronics Technique

基　　金：广东省普通高校特色创新项目(2019GKTSCX097);广东省自然科学基金项目(S2012010010400)。

摘　　要：针对当前光伏逆变器参数优化控制不准确,误差大等不足,以获得理想的光伏逆变器参数优化控制结果为目标,提出改进神经网络的光伏逆变器参数自动优化控制方法。首先,分析当前光伏逆变器参数优化控制的研究进展,并设计了光伏逆变器参数优化控制模型;然后,采用神经网络对光伏逆变器参数变化特点进行拟合,并针对神经网络的不足进行改进;最后,在Matlab 2017平台上对光伏逆变器参数优化控制方法性能进行仿真测试。结果表明,改进神经网络不仅提高了光伏逆变器参数优化控制精度,而且光伏逆变器参数优化控制误差小于当前其他方法,具有较好的实际应用价值。In view of the inaccuracy and big error in the current parameter optimization control,a method of photovoltaic inverter parameter automatic optimization control based on improved neural network is proposed to obtain an ideal control result.The research progress of photovoltaic inverter parameter optimization control is analyzed,and its parameter optimization control model is designed.And then,the neural network is used to fit the parameter change characteristics and improve the shortcomings of neural network.The performance of the parameter optimization control method was simulated and tested on platform Matlab 2017.The results show the method based on the improved neural network improves the accuracy of photovoltaic inverter parameter optimization control,and its control error is less than that of other current methods.Therefore,it has better a practical application value.

关 键 词：光伏逆变器 参数优化 神经网络 连接权值 控制精度 自动优化控制 

分 类 号：TN876-34[电子电信—信息与通信工程] TP391[自动化与计算机技术—计算机应用技术]