基于云边协同的含分布式新能源配电网电压控制方法  

Voltage control method of distribution network with distributed new energy based on cloud-edge collaboration

作　　者：李江成 徐晓春[1] 胡健雄 Li Jiangcheng;Xu Xiaochun;Hu Jianxiong(State Grid Huai'an Power Supply Company,Huai'an 223000,China;School of Electrical Engineering,Southeast University,Nanjing 210096,China)

机构地区：[1]国网淮安供电公司,江苏淮安223000 [2]东南大学电气工程学院,江苏南京210096

出　　处：《能源与环保》2025年第1期199-206,共8页CHINA ENERGY AND ENVIRONMENTAL PROTECTION

基　　金：国网江苏省电力有限公司科技项目(J2022049)。

摘　　要：针对大量分布式新能源集群并网给配电网带来的电压频繁波动、越限以及现有调控方法难以实现分布式新能源的实时高效管控等问题,提出了一种基于云边协同的含分布式新能源配电网电压控制方法。首先,将边缘计算技术与基于数据驱动的强化学习结合,建立云边协同的主动配电网无功—电压控制框架;然后,以配电网运行网损最低为优化目标,构建电压控制的物理模型并将其转化为分区马尔可夫决策问题,将配电网划分的各区域描述为边缘智能体,并设计了智能体与环境交互的动作、状态空间和奖励函数,采用多智能体柔性行动—评论家算法求解;最后,在IEEE33节点系统上进行的算例仿真表明所提出的电压控制方法在降低配电网运行损耗方面具有显著效果,有效保障系统运行的安全性与经济性。A voltage control method of distribution network with distributed new energy based on cloud-edge collaboration was proposed to address the frequent voltage fluctuations,exceeding limits,and difficulties in achieving real-time and efficient control of distributed new energy caused by the integration of a large number of distributed new energy clusters into the grid.First,combing the edge computing with data-driven reinforcement learning,a cloud-edge collaboration reactive voltage control framework for active distribution networks was established;then,with the optimization objective of minimizing network losses in the operation of the distribution network,a physical model of voltage control was constructed and transformed into a partitioned Markov decision problem.The various regions divided by the distribution network were described as edge agents,and the action,state space,and reward functions for the interaction between the agents and the environment were designed,the multi-agent soft actor-critic algorithm was used to solve the problem;finally,the effectiveness of the proposed voltage control method was confirmed through simulation on an IEEE 33 node system,showing a substantial reduction in network losses during distribution network operation,and effectively ensuring the safety and economy of system operation.

关 键 词：深度强化学习 无功电压控制 分布式新能源 边缘智能 数据驱动 

分 类 号：TM711[电气工程—电力系统及自动化]