摘要
针对分布式光伏并网逆变器独立构网接入非线性负载时易引起交流侧输出电压失真、谐波含量大等问题,提出了一种基于改进神经网络算法的光伏逆变器独立构网控制策略。所提策略具有以下优势:利用遗传算法(GA)优化了所构建的反向传播(BP)神经网络关键参数,使得基于神经网络的逆变控制策略具有更快的收敛速度;当光伏逆变器独立构网为非线性负载供电时,与传统比例积分(PI)控制策略相比具有更低的电压谐波含量并有效提高逆变器抗负载干扰特性。最后通过构建一台额定功率10 kW的原理样机对所提策略的有效性及正确性进行了实验验证。
For the problems of AC output voltage distortion and high harmonic content when the distributed photovoltaic grid-connected inverter is connected to nonlinear load,a photovoltaic inverter independent grid control strategy based on improved neural network algorithm is proposed.The proposed strategy has following advantages:genetic algorithm(GA)is used to optimize the key parameters of the constructed backpropagation(BP)neural network,so that the inverter control strategy based on neural network has a faster convergence speed;compared with the traditional proportional integral(PI)control strategy,the photovoltaic inverter has lower voltage harmonic content and effectively improve the anti-load interference characteristic when powered by the independent network for nonlinear load.Finally,a prototype with rated power of 10 kw is constructed to verify the efectiveness and correctness of the proposed strategy.
作者
潘鸣宇
柴志超
赵贺
孙钦斐
PAN Ming-yu;CHAI Zhi-chao;ZHAO He;SUN Qin-fei(State Grid Beijing Electric Power Company Electric Power Research Institute,Beijing 100075,China)
出处
《电力电子技术》
北大核心
2023年第12期80-82,93,共4页
Power Electronics
基金
国家电网有限公司科技项目(52022322000C)。
关键词
光伏逆变器
独立构网
非线性负载
改进神经网络
photovoltaic inverter
independent grid
nonlinear load
improved neural network
