刘志远,许江蛟
LIU Zhiyuan,XU Jiangjiao

• 1:上海电力大学电气工程学院

摘要(Abstract)：

针对高比例分布式能源的接入背景下跨区域电网对高效可靠无功优化的需求，以及区域间敏感运行数据无法直接共享等问题，提出一种基于自适应联邦学习与ADMM（交替方向乘子法）的分布式无功优化方法。首先，通过联邦学习在各节点开展分布式训练，以模型参数交换代替原始数据传输，在保障数据隐私的同时完成边界量估计；其次，引入自适应学习率和早停策略，动态调整学习率并终止冗余训练，从而降低计算与通信开销；再次，将联邦学习的预测结果嵌入ADMM迭代，利用预测量替代部分跨区域传输的边界变量，提升收敛效率与稳定性。最后，在IEEE 9节点、IEEE 33节点和IEEE 118节点系统中进行仿真分析，验证了所提方法的有效性。
To address the need for efficient and reliable reactive power optimization in interconnected power grids with high penetration of distributed energy resources, as well as the constraint that sensitive operational data cannot be directly shared across regions, a distributed reactive power optimization method based on adaptive federated learning(AFL) and the alternating direction method of multipliers(ADMM) is proposed. First, federated learning is employed to perform distributed training across multiple nodes, where model parameter exchange replaces raw data transmission, enabling boundary variable estimation while preserving data privacy. Second, an adaptive learning rate and an early stopping strategy are introduced to dynamically adjust the learning rate and terminate redundant training, thereby reducing computational and communication overhead. Third, the predictions obtained from federated learning are embedded into the ADMM iterative process, where the predicted values are used to replace part of the boundary variables that would otherwise require inter-regional exchange, thus improving convergence efficiency and stability. Finally, simulation studies on the IEEE 9-bus, IEEE 33-bus, and IEEE 118-bus systems verify the effectiveness of the proposed method.

关键词(KeyWords)： 无功优化;联邦学习;交替方向乘子法;隐私保护;自适应算法
reactive power optimization;federated learning;ADMM;privacy preservation;adaptive algorithm

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金（52377111）

作者(Author): 刘志远,许江蛟
LIU Zhiyuan,XU Jiangjiao

DOI: 10.19585/j.zjdl.202606007

参考文献(References)：

• [1]邵海明，王家福，李传生，等.面向新型电力系统的高电压计量技术研究进展及展望[J].高电压技术，2026,52(2):493-508.SHAO Haiming,WANG Jiafu,LI Chuansheng,et al.Review and prospect on the research status of high-voltage metrology technology for new-type power systems[J].High Voltage Engineering,2026,52(2):493-508.
• [2]PESCHOL J,PIERCY D,TINNEY W,et al.Optimum control of reactive power flow[J].IEEE Transactions on Power Apparatus and Systems,1968,PAS-87(1):40-48.
• [3]ZHAO B,GUO C X,CAO Y J.A multiagent-based particle swarm optimization approach for optimal reactive power dispatch[J].IEEE Transactions on Power Systems,2005,20(2):1070-1078.
• [4]TAKAHASHI N,HAYASHI Y.Centralized voltage control method using plural D-STATCOM with controllable dead band in distribution system with renewable energy[C]//2012 3rd IEEE PES Innovative Smart Grid Technologies Europe(ISGT Europe). October 14-17,2012,Berlin,Germany.IEEE,2012:1-5.
• [5]金煜，焦康威.智能用电互动模式及安全性研究[J].电工技术，2025(16):174-183.JIN Yu,JIAO Kangwei.Research on intelligent electricity interaction modeand securityam[J].Electric Engineering,2025(16):174-183.
• [6]李忠凡，陈曦，黄海涛.基于多智能体分层强化学习的多园区综合能源系统优化运行[J].浙江电力，2025,44(9):46-57.LI Zhongfan,CHEN Xi,HUANG Haitao.Optimal operation of MPIESs based on hierarchical multi-agent reinforcement learning[J].Zhejiang Electric Power,2025,44(9):46-57.
• [7]刘海天，郭烨，石剑涛，等.分布式优化在互联电网中的应用分析及展望[J].电力系统自动化，2024,48(22):22-42.LIU Haitian,GUO Ye,SHI Jiantao,et al. Application analysis and prospect of distributed optimization in interconnected power grid[J]. Automation of Electric Power Systems,2024,48(22):22-42.
• [8]王江，杨军，骆亚帅，等.考虑分区规模平衡性的主动配电网无功电压优化方法[J].浙江电力，2025,44(9):81-91.WANG Jiang,YANG Jun,Luo Yashuai,et al.A reactive voltage optimization method for active distribution networks considering partition size balancing[J]. Zhejiang Electric Power,2025,44(9):81-91.
• [9]WANG Y M,WU L,WANG S X.A fully-decentralized consensus-based ADMM approach for DC-OPF with demand response[J]. IEEE Transactions on Smart Grid,2017,8(6):2637-2647.
• [10]SUN K Z,SUN X A.A two-level ADMM algorithm for AC OPF with global convergence guarantees[J]. IEEE Transactions on Power Systems,2021,36(6):5271-5281.
• [11]OH B,LEE D H,JEONG W C,et al.Distributed optimal power flow for distribution system using second order cone programming and consensus alternating direction method of multipliers[J].Journal of Electrical Engineering&Technology,2022,17(2):999-1008.
• [12]陈池瑶，苗世洪，姚福星，等.基于多智能体算法的多微电网-配电网分层协同调度策略[J].电力系统自动化，2023,47(10):57-65.CHEN Chiyao,MIAO Shihong,YAO Fuxing,et al.Hierarchical cooperative dispatching strategy of multimicrogrid and distribution networks based on multi-agent algorithm[J]. Automation of Electric Power Systems,2023,47(10):57-65.
• [13]郭庆来，田年丰，孙宏斌.支撑能源互联网协同优化的隐私计算关键技术[J].电力系统自动化，2023,47(8):2-14.Guo Qinglai,Tian Nianfeng,Sun Hongbin.Key technologies of privacy computation supporting collaborative optimization of energy Internet[J]. Automation of Electric Power Systems,2023,47(8):2-14.
• [14]HUANG Z H,HU R,GUO Y X,et al. DP-ADMM:ADMM-based distributed learning with differential privacy[J].IEEE Transactions on Information Forensics and Security,2020,15:1002-1012.
• [15]LV Y W,YANG G H,SHI C X.Differentially private distributed optimization for multi-agent systems via the augmented Lagrangian algorithm[J]. Information Sciences,2020,538:39-53.
• [16]LIU Y Y,GENG J C,SHANG F H,et al. Laplacian smoothing stochastic ADMMs with differential privacy guarantees[J].IEEE Transactions on Information Forensics and Security,2022,17:1814-1826.
• [17]WU T,ZHAO C H,ZHANG Y A.Privacy-preserving distributed optimal power flow with partially homomorphic encryption[J].IEEE Transactions on Smart Grid,2021,12(5):4506-4521.
• [18]YUAN Z P,LI P,LI Z L,et al.A fully distributed privacypreserving energy management system for networked microgrid cluster based on homomorphic encryption[J].IEEE Transactions on Smart Grid,2024,15(2):1735-1748.
• [19]管桂林，陶政坪，支婷，等.面向医疗场景的去中心化联邦学习隐私保护方法[J].计算机应用，2024,44（增刊2）:112-117.GUAN Guilin,TAO Zhengping,ZHI Ting,et al.Decentralized federated learning privacy protection method for medical scenarios[J]. Journal of Computer Applications,2024,44(S2):112-117.
• [20]郑润达，张维，张永杰，等.金融领域中的联邦学习研究[J].数智技术研究与应用，2025,1(1):1-17.ZHENG Runda,ZHANG Wei,ZHANG Yongjie,et al.Research on federated learning in the field of finance[J].SmartTech Innovations,2025,1(1):1-17.
• [21]智慧，段苗苗，杨利霞，等.一种基于区块链和联邦学习融合的交通流预测方法[J].电子与信息学报，2024,46(9):3777-3787.ZHI Hui,DUAN Miaomiao,YANG Lixia,et al.A traffic flow prediction method based on the fusion of blockchain and federated learning[J].Journal of Electronics&Information Technology,2024,46(9):3777-3787.
• [22]王蓓蓓，朱竞，王嘉乐，等.电表数据隐私保护下的联邦学习行业电力负荷预测框架[J].电力系统自动化，2023,47(13):86-93.WANG Beibei,ZHU Jing,WANG Jiale,et al.Federatedlearning based industry load forecasting framework under privacy protection of meter data[J].Automation of Electric Power Systems,2023,47(13):86-93.
• [23]郭光华，亓新宏，王瑞琪，等.基于联邦学习的综合能源系统集成需求响应机制[J].电力自动化设备，2023,43(8):33-39.GUO Guanghua,QI Xinhong,WANG Ruiqi,et al. Integrated demand response mechanism for integrated energy system based on federated learning[J].Electric Power Automation Equipment,2023,43(8):33-39.
• [24]郭方洪，刘师硕，吴祥，等.基于联邦学习的含不平衡样本数据电力变压器故障诊断[J].电力系统自动化，2023,47(10):145-152.GUO Fanghong,LIU Shishuo,WU Xiang,et al. Federated learning based fault diagnosis of power transformer with unbalanced sample data[J]. Automation of Electric Power Systems,2023,47(10):145-152.
• [25]TURITSYN K,SULC P,BACKHAUS S,et al. Local control of reactive power by distributed photovoltaic generators[C]//2010 First IEEE International Conference on Smart Grid Communications.October 4-6,2010,Gaithersburg,MD,USA.IEEE,2010:79-84.