主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2025, Vol. 33 ›› Issue (11): 41-53.doi: 10.16381/j.cnki.issn1003-207x.2023.1947

• • 上一篇    下一篇

基于深度强化学习算法的跨区域多微电网系统扩展规划研究

周剑1, 聂孝婷1, 庞可欣1, 王小越2(), 马义中1   

  1. 1.南京理工大学经济管理学院,江苏 南京 210094
    2.北京工商大学商学院,北京 100048
  • 收稿日期:2023-01-23 修回日期:2024-05-01 出版日期:2025-11-25 发布日期:2025-11-28
  • 通讯作者: 王小越 E-mail:xiaoyue@btbu.edu.cn
  • 基金资助:
    国家自然科学基金项目(72101116);国家自然科学基金项目(72371003);国家自然科学基金项目(72001006);江苏高校哲学社会科学研究重大项目(2025SJZD125)

Multi-area Multi-Microgrid System Expansion Planning Based on Deep Reinforcement Learning Algorithm

Jian Zhou1, Xiaoting Nie1, Kexin Pang1, Xiaoyue Wang2(), Yizhong Ma1   

  1. 1.School of Economic and Management,Nanjing university of science and technology,Nanjing 210094,China
    2.Business School,Beijing Technology and Business University,Beijing 100048,China
  • Received:2023-01-23 Revised:2024-05-01 Online:2025-11-25 Published:2025-11-28
  • Contact: Xiaoyue Wang E-mail:xiaoyue@btbu.edu.cn

摘要:

微电网在提升电力供应韧性和减少温室气体排放等方面展现出了巨大潜力,孤岛型微电网通过互联成为跨区域的多微电网系统,有利于实现微电网的经济性和供电韧性。针对跨区域多微电网系统的扩展规划问题,考虑相邻微电网间的能量互济,将供电韧性和环境效益作为约束,提出以最小化多微电网系统总成本为目标的长期扩展规划框架。基于深度强化学习算法,对此动态、随机决策优化问题给出了求解方法,结合真实数据构造了包含三个区域的多微电网系统,并以此作为算例验证模型的有效性。算例仿真结果表明,针对跨区域多微电网系统的规划框架不仅可提升微电网的供电韧性,而且能够考虑跨区域的微电网结构的影响,适时调整投资规划,选取电力依赖性更高、用途更广泛的区域进行微电网设施的投资,有效解决了跨区域多微电网系统的规划问题。

关键词: 微电网, 韧性, 扩展规划, 跨区域, 强化学习

Abstract:

Microgrids have shown great potential in improving the resilience of power supply and reducing greenhouse gas (GHG) emissions. The interconnection of island microgrids into a multi-area multi-microgrids (MMGs) system will help improve the economy and power supply resilience of microgrids. Aiming at the expansion planning problem for MMGs, a long-term expansion planning framework is proposed with the goal of minimizing the total cost of the MMGs while taking power supply resilience and environmental benefits as constraints. The energy sharing between adjacent microgrids is also considered. Based on the real data, a MMGs system containing three microgrids is constructed as a case study to demonstrate the effectiveness of the proposed model. The dynamic and stochastic optimization problem is solved by deep reinforcement learning algorithm. The results show that the planning framework for MMGs can improve the resilience of the microgrid power supply and reduce GHG emissions. The proposed framework also considers the impact of the interconnection structure of MMGs and appropriately adjusts strategies based on the frequency of outages and outage losses of individual microgrid. This research has important practical significance for the expansion planning of MMGs.

Key words: microgrid, resilience, expansion planning, multi-area, reinforcement learning

中图分类号: