华南理工大学学报(自然科学版) ›› 2026, Vol. 54 ›› Issue (3): 10-20.doi: 10.12141/j.issn.1000-565X.250222

• 能源、动力与电气工程 • 上一篇    下一篇

基于数据驱动的配电网-用户能源协同管理策略

边瑞恩1,2  刘亚东1   

  1. 1.上海交通大学  电子信息与电气工程学院,上海 200241;

    2.南方电网供应链集团有限公司,广东 广州 510630

  • 出版日期:2026-03-25 发布日期:2025-10-24

Data-Driven Collaborative Management Strategy for Distribution Networks and Home Energy

BIAN Ruien 1,2  LIU Yadong1   

  1. 1.School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200241, China;

    2. Southern Power Grid Supply Chain Group Co., Ltd., Guangzhou 510630, Guangdong Province, China

  • Online:2026-03-25 Published:2025-10-24

摘要:

传统的家庭能源管理系统主要关注用户家庭自身的能源管理,通常忽视了用户舒适度及其与配电网的双向互动潜力。为了进一步提升家庭能源管理的智能化水平,本文提出了一种双层架构的家庭能源管理策略,其中上层考虑了配电网的系统运行约束和成本,下层以尽可能满足用户的用电偏好及用电成本要求,在考虑与配电网互动的前提下,实现家庭能源最优管理。针对常规配电网潮流计算和室内热响应算法在精确建模方面的不足和局限性,引入数据驱动方法(LSTM用于潮流估计,OS-ELM用于热响应估计)以更准确地评估配电网状态和室内热动态,并利用高效的自然聚合算法(NAA)来提升优化过程的求解速度和准确率。为评估所提算法的性能,设计了详尽的对比实验进行分析,基于IEEE 33节点系统的实验验证表明,所提策略在降低用户用电成本、保障配电网安全运行及提升用户舒适度方面均显著优于传统方法,验证了其有效性和优越性。该方法为解决高渗透率分布式能源社区中源荷协同优化问题提供了有效的技术支撑,具有较好的工程应用前景。

关键词: 用户偏好, 数据驱动, 家庭能源管理, 热效应估计

Abstract:

Traditional home energy management methods mainly focus on the energy management within individual households, overlooking considerations of comfort and the interaction with the grid. To enhance the capabilities of home energy management, a two-layered home energy management strategy is proposed. The upper layer considers the operational constraints and costs of the distribution network and the lower layer aims to fulfill users' electricity consumption preferences and cost requirements, achieving optimal household energy management under the premise of interacting with the grid. Acknowledging the limitations of conventional power flow calculations for distribution networks and indoor thermal response algorithms, data-driven methods are introduced to improve the accuracy in assessing both the grid's state and indoor thermal conditions. Furthermore, a natural aggregation algorithm is employed to increase the precision of the optimization process. Extensive experiments are conducted to evaluate the accuracy of the proposed algorithm, and the results convincingly demonstrate the effectiveness and scientific rigor of the proposed strategy.


Key words: user preference;data-driven, home energy management, thermal effect estimation