基于深度强化学习的新型电力系统无功电压优化控制  

Reactive voltage optimization control of novel power system based on deep reinforcement learning

作　　者：周良才 周毅 沈维健 黄志龙 李雷 李艺丰 ZHOU Liangcai;ZHOU Yi;SHEN Weijian;HUANG Zhilong;LI Lei;LI Yifeng(East China Branch,State Grid Corporation of China,Shanghai 200002,China;NARI Group Corporation(State Grid Electric Power Research Institute),Nanjing 211106,China;State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210000,China)

机构地区：[1]国家电网有限公司华东分部,上海200002 [2]南瑞集团有限公司(国网电力科学研究院有限公司),南京211106 [3]国网江苏省电力有限公司,南京210000

出　　处：《电测与仪表》2024年第9期182-189,共8页Electrical Measurement & Instrumentation

基　　金：国家电网有限公司科技项目(5108-20233058A-1-1-ZN)。

摘　　要：新型电力系统框架下可再生能源的随机性和波动性、负荷的主动性以及电网的电力电子化使得电网的运行方式和实时控制面临着新的挑战。无功电压优化控制是保证电网安全稳定运行的基础,针对可再生能源系统的大规模接入、储能系统的灵活配置带来的无功电压控制问题,提出了基于深度强化学习的电网无功电压优化控制方法,综合考虑运行效率、经济性和安全性建立电网无功优化模型,利用马尔可夫决策过程将无功优化问题转化为强化学习序贯决策优化,充分考虑无功调压设备时间、空间耦合特性,利用深度确定性梯度算法(deep deterministic policy gradient,DDPG)进行模型求解。最后,在改进的IEEE 33算例上进行仿真验证,通过对比验证了文中所提方法在无功优化决策过程中的有效性和可靠性。Under the framework of the novel power system,the randomness and volatility of renewable energy,along with load initiative and electronic power grid integration,pose new challenges to the operation mode and real-time control of the power grid.Reactive voltage optimization control serve as the foundation for ensuring a safe and stable operation of the power grid.This paper proposes a reactive voltage optimization control method based on deep reinforcement learning to address issues caused by large-scale integration of renewable energy system and flexible configuration of energy storage system.The proposed method establishes a comprehensive reactive power optimization model considering operational efficiency,economy,and security.By utilizing Markov decision-making process,we transform the reactive power optimization problem into sequential decision optimization through reinforcement learning while fully considering time-space coupling characteristics of reactive power regulation.Deep deterministic policy gradient(DDPG) is employed to solve this model.Finally,an improved IEEE 33 example is simulated to verify both effectiveness and reliability in optimizing reactive power decision-making process.

关 键 词：深度强化学习 马尔可夫决策过程 新型电力系统 无功优化 

分 类 号：TM761[电气工程—电力系统及自动化]