基于谱聚类算法的配电网电压控制研究  

Research on voltage control of power distribution network based on spectral clustering algorithm

作　　者：顾少平 刘乙 郑众 宋杰[2] 陈嘉栋 GU Shaoping;LIU Yi;ZHENG Zhong;SONG Jie;CHEN Jiadong(Suzhou Power Supply Company,State Grid Jiangsu Electric Power Company Limited Suzhou Jiangsu 215000,China;Nari Technology Development Limited Company,Nanjing 211106,China)

机构地区：[1]国网江苏省电力有限公司苏州供电分公司,江苏苏州215000 [2]国电南瑞科技股份有限公司,南京211106

出　　处：《自动化与仪器仪表》2025年第3期274-278,共5页Automation & Instrumentation

基　　金：国家重点研发计划《城区用户与电网供需友好互动系统》(2016YFB0901100)。

摘　　要：针对传统配电网电压控制存在控制效果不佳的问题,提出一种基于粒子群优化(Particle Swarm Optimization,PSO)谱聚类算法(Spectral clustering,SC)的配电网控制策略。首先,采用SC算法对配电网电压进行分级处理;然后基于PSO算法加入二阶振荡环节,以加强PSO算法多样性;最后以全网网损为优化目标,发电机机端电压为优化变量,搭建一个基于改进PSO算法的配电网电压控制模型。结果表明,相同迭代次数下,改进PSO算法在迭代至87次时即可实现收敛,寻优精度高达97.45%,明显高于PSO的算法、遗传算法GA和强化学习RL。相较于传统的DQN-DDPG模型和GA-CNN模型,本模型对区域B电压控制的负荷曲线与发生扰动前的负荷电压曲线基本拟合,未出现明显波动。由此说明,本模型能够实现配电网电压控制和优化,具备有效性和适用性。In view of the problem of poor control effect of voltage control of traditional distribution network,a distribution network control strategy based on particle swarm optimization(Particle Swarm Optimization,PSO) spectral clustering algorithm(Spectral clustering,SC) is proposed.First,the SC algorithm is applied to process the distribution network voltage;then the PSO algorithm,the voltage of the improved PSO algorithm.The results show that with the same iteration times,the improved PSO algorithm converges to 87 iterations,and the optimization accuracy is as high as 97.45%,which is significantly higher than the PSO algorithm,genetic algorithm GA and reinforcement learning RL.Compared with the traditional DQN-DDPG model and GA-CNN model,the load curve of region B voltage control basically fits the load voltage curve before the disturbance without obvious fluctuations.This shows that this model can realize the voltage control and optimization of distribution network,and has effective and applicability.

关 键 词：谱聚类算法 粒子群优化 配电网 电压控制 二阶振荡 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]