基于多智能体深度强化学习的配电网双时间尺度电压控制策略  

Distribution Network Dual Time Scale Voltage Control Strategy Based on Multi-Agent Deep Reinforcement Learning

作　　者：赵晶晶 张超立 王涵 盛杰 ZHAO Jingjing;ZHANG Chaoli;WANG Han;SHENG Jie(College of Electrical Engineering,Shanghai University of Electric Power,Shanghai 200090,China)

机构地区：[1]上海电力大学电气工程学院,上海200090

出　　处：《南方电网技术》2025年第2期68-79,共12页Southern Power System Technology

基　　金：国家自然科学基金资助项目(52007112)。

摘　　要：风电、光伏(photovoltaics,PV)在新型电力系统中的渗透率日益增加,使得配电网电压波动加剧,而储能(energy storage,ES)、电动汽车(electric vehicles,EV)对降低配电网电压波动有重要作用。与此同时,智能电表、智能传感器以及改进的通信网络广泛部署,可获取的数据量越来越大,数据驱动技术兴起。提出了一种基于多智能体深度强化学习(multi-agent deep reinforcement learning,MADRL)的配电网双时间尺度有功-无功功率协调的电压控制策略。慢时间尺度下用双深度Q网络算法(double deep Q-network algorithm,DDQN)求解电容器组(capacitor banks,CBs)、有载调压变压器(on-line tap changer,OLTC)与ES有功-无功功率优化问题。快时间尺度下用具有注意力机制的经验增强多智能体柔性参与者-评论家算法(experience augmentation-multi-agent soft actor critic,EA-MASAC)调节PV、风机(wind turbine,WT)、静止无功补偿装置(static var compensator,SVC)的无功功率与EV的有功功率。最后,在IEEE-33节点系统上验证了所提方法的有效性。The increasing penetration rate of wind power and photovoltaics(PV)in new power systems exacerbates voltage fluctua-tions in distribution networks,while energy storage(ES)and electric vehicles(EV)play important roles in reducing voltage fluctua-tions in distribution networks.At the same time,smart meters,smart sensors and improved communication networks are widely deployed,the amount of data available is increasing,and data-driven technology is emerging.This paper proposes a multi-agent deep reinforcement learning(MADRL)based dual time scale active and reactive power coordinated voltage control strategy for distribu-tion networks.Using the double deep Q-network algorithm(DDQN)to solve the optimization problems of capacitor banks(CBs),on line tap transformers(OLTC),and ES active and reactive power at a slow time scale.At a fast time scale,the EA-MASAC algorithm with attention mechanism is used to enhance the reactive power of PV,wind turbines(WT),and static var compensators(SVCs),as well as the active power of EVs.Finally,the effectiveness of the proposed method is verified on an IEEE-33 node system.

关 键 词：数据驱动 多智能体深度强化学习 双时间尺度 电压控制 功率优化 

分 类 号：TM732[电气工程—电力系统及自动化]