基于改进神经网络的弱电网逆变器谐波自动补偿方法  

Harmonic Automatic Compensation Method for Weak Current Grid Inverter Based on Improved Neural Network

作　　者：韩然 凌霄 李祥新 王英 HAN Ran;LING Xiao;LI Xiangxin;WANG Ying(School of Software Technology,Zhejiang University,Hangzhou 310058,China;State Grid Qingdao Power Supply Company,Qingdao 266000,China)

机构地区：[1]浙江大学软件学院,杭州310058 [2]国网青岛供电公司,青岛266000

出　　处：《自动化与仪表》2025年第4期1-5,共5页Automation & Instrumentation

摘　　要：为抑制逆变器与电网之间的谐振现象,提出基于改进神经网络的弱电网逆变器谐波自动补偿方法。利用粒子群算法对BP神经网络进行改进,结合改进BP神经网络和比例-积分-微分控制对逆变器进行谐波补偿。实验结果表明,该方法在面临电压突变时,电容电压小幅度变化,变化量为9.8 V,持续时间为4.37 ms,而后重新达到稳态。当负载由20Ω突增至30Ω时,相关输出保持稳定,且无锯齿现象。验证了所提方法具有较好的动态响应,有助于提高弱电网的稳定性和电能质量,能够减少背景谐波对电网的影响。To suppress the resonance phenomenon between the inverter and the power grid,a harmonic automatic compensation method for weak current grid inverters based on improved neural networks is proposed.Using particle swarm optimization algorithm to improve the BP neural network,combined with the improved BP neural network and proportional integral derivative control for harmonic compensation of the inverter.The experimental results show that the method proposed in this paper experiences a small change in capacitor voltage when facing a voltage mutation,with a variation of 9.8 V and a duration of 4.37 ms,before reaching steady state again.When the load suddenly increases from 20Ωto 30Ω,the related output remains stable and there is no sawtooth phenomenon.The proposed method has been validated to have good dynamic response,which helps to improve the stability and power quality of weak current networks,and can reduce the impact of background harmonics on the power grid.

关 键 词：逆变器 弱电网 谐波补偿 BP神经网络 粒子群算法 

分 类 号：TM46[电气工程—电器] TP272[自动化与计算机技术—检测技术与自动化装置]