电器与能效管理技术 ›› 2022, Vol. 0 ›› Issue (2): 12-20.doi: 10.16628/j.cnki.2095-8188.2022.02.003

• 研究与分析 • 上一篇    下一篇

共享储能模式下多微电网博弈优化方法

郑海林1, 温步瀛1,2, 朱振山1,3, 翁智敏1   

  1. 1.福州大学 电气工程与自动化学院, 福建 福州 350108
    2.福建省新能源发电与电能变换重点实验室, 福建 福州 350108
    3.智能配电网装备福建省高校工程研究中心, 福建 福州 350108
  • 收稿日期:2021-10-09 出版日期:2022-02-28 发布日期:2022-03-31
  • 作者简介:郑海林(1997—),男,硕士研究生,研究方向为电力系统运行与控制。|温步瀛(1967—),男,教授,博士,研究方向为电力系统优化运行与电力市场、风电并网运行技术。|朱振山(1989—),男,讲师,博士,研究方向为电力系统运行与控制。
  • 基金资助:
    福建省教育厅中青年教师教育科研项目(JAT190043);福州大学科研启动项目(510901)

Optimization of Multi Microgrid Game Method Under Shared Energy Storage Mode

ZHENG Hailin1, WEN Buying1,2, ZHU Zhenshan1,3, WENG Zhimin1   

  1. 1. College of Electrical Engineering and Automation,Fuzhou University,Fuzhou 350108,China
    2. Fujian Key Laboratory of New Energy Generation and Power Conversion,Fuzhou 350108,China
    3. Fujian Province University Engineering Research Center of Smart Distribution Grid Equipment,Fuzhou 350108,China
  • Received:2021-10-09 Online:2022-02-28 Published:2022-03-31

摘要:

储能的高投资成本是限制其商业化发展的主要障碍,通过储能聚合商协调储能设备运行,提高储能的利用率并降低成本。首先,综合考虑了微电网中的火电机组、充电站、可中断负荷等可调节灵活性资源的成本以及共享储能的费用分摊,以各方效益最大化为目标,构建了各微电网与共享储能聚合商的博弈优化运行模型。其次,采用了多智能体强化学习方法求解多主体下博弈问题,引入KL散度优化智能体的学习率,提高算法的收敛性。最后,以3个相邻微电网的算例分析,共享储能模式下提升了各主体的经济效益,验证了共享储能模式的优越性与算法改进的有效性。

关键词: 共享储能, 多主体博弈, 强化学习, KL散度, 自适应学习率

Abstract:

The high investment cost of energy storage is the main obstacle to its commercial development.Through the energy storage aggregators to coordinate the operation of energy storage equipment,the utilization rate of energy storage is improved and the cost is reduced.Firstly,the cost of adjustable flexible resources such as thermal power units,charging stations and interruptible loads in microgrid and the cost sharing of shared energy storage are comprehensively considered.Aiming at maximizing the benefits of all parties,the game optimization operation model between microgrid and shared energy storage aggregator is constructed.Secondly,the Multi-Agent Reinforcement learning method is used to solve the multi-agent game problem,and KL divergence is introduced to optimize the agent learning rate and improve the convergence of the algorithm.Finally,taking three adjacent microgrids as examples,the economic benefits of each subject are improved under the shared energy storage mode,which verifies the superiority of the mode and the effectiveness of the algorithm improvement.

Key words: shared energy storage, multi agent game, reinforcement learning, Kullback-Leibler divergence, adaptive learning rate

中图分类号: