程颖,董炜,姜震韬,汤奕,冯长有
CHENG Ying,DONG Wei,JIANG Zhentao,TANG Yi,FENG Changyou

• 1:国网浙江省电力有限公司电力科学研究院
• 2:东南大学电气工程学院
• 3:国家电网有限公司国家电力调度控制中心

摘要(Abstract)：

负荷构成的实时精确辨识对电力系统仿真分析具有重要意义。当前基于传统优化方法的辨识过程难以处理多维时序特征且计算耗时高，导致辨识精度不足，无法满足在线应用。对此，融合注意力机制的特征加权能力与CNN（卷积神经网络）的特征提取能力，提出一种适用于有源综合负荷模型的负荷构成在线辨识方法。首先，从机理层面提出了含光伏和储能的有源综合负荷模型；然后，构建融合多尺度卷积与注意力机制的特征提取网络，并行捕捉异构负荷特征并突出关键信息；最后，以负荷节点间参数的全域灵敏度之比作为评价指标，筛选出关键负荷节点并针对目标节点进行辨识。算例结果表明，与现有方法相比，所提方法具有较高的辨识精度和鲁棒性，可满足电力系统在大部分运行场景下的在线安全分析需求。
Real-time and accurate identification of load composition is of great significance for power system simulation and analysis. Current identification processes based on conventional optimization methods struggle to handle multidimensional temporal characteristics and are computationally intensive, leading to insufficient identification accuracy and an inability to meet the demands of online applications. To address this challenge, an online load composition identification method suitable for active integrated load models is proposed, integrating the feature weighting capability of attention mechanisms with the feature extraction capability of convolutional neural network(CNN). Firstly, an active integrated load model incorporating photovoltaics and energy storage is proposed from a mechanistic perspective. Subsequently, a feature extraction network integrating multi-scale convolution and attention mechanisms is constructed to capture heterogeneous load features in parallel and highlight critical information. Finally, key load nodes are screened based on the ratio of global parameter sensitivity among load nodes as an evaluation metric, and target nodes are identified accordingly. Case study results demonstrate that, compared to existing methods, the proposed approach achieves higher identification accuracy and robustness, meeting the requirements for online security analysis in most power system operational scenarios.

关键词(KeyWords)： 综合负荷模型;构成辨识;卷积神经网络;注意力机制;深度学习
integrated load model;composition identification;CNN;attention mechanism;deep learning

Abstract:

Keywords:

基金项目(Foundation): 国家电网有限公司科技项目（5100-202419015A-1-1-ZN）

作者(Author): 程颖,董炜,姜震韬,汤奕,冯长有
CHENG Ying,DONG Wei,JIANG Zhentao,TANG Yi,FENG Changyou

DOI: 10.19585/j.zjdl.202603008

参考文献(References)：

• [1]陈国平，李明节，许涛，等.关于新能源发展的技术瓶颈研究[J].中国电机工程学报，2017,37(1):20-27.CHEN Guoping,LI Mingjie,XU Tao,et al.Study on technical bottleneck of new energy development[J].Proceedings of the CSEE,2017,37(1):20-27.
• [2]孙华东，李佳豪，李文锋，等.大规模电力系统仿真用新能源场站模型结构及建模方法研究（二）：机电暂态模型[J].中国电机工程学报，2023,43(6):2190-2202.SUN Huadong,LI Jiahao,LI Wenfeng,et al.Research on model structures and modeling methods of renewable energy station for large-scale power system simulation（Ⅱ）：electromechanical transient model[J]. Proceedings of the CSEE,2023,43(6):2190-2202.
• [3]陈忠玉，徐晋，汪可友，等.计及跟随型和支撑型分布式光伏的广义综合负荷模型及两阶段参数聚合等效方法[J].电力自动化设备，2023,43(3):86-93.CHEN Zhongyu,XU Jin,WANG Keyou,et al. Generalized composite load model and two-stage parameter aggregation and equivalent method considering grid-following and grid-forming distributed photovoltaic[J]. Electric Power Automation Equipment,2023,43(3):86-93.
• [4]鞠平，郭德正，曹路，等.含主动负荷的综合电力负荷建模研究综述与展望[J].河海大学学报（自然科学版），2020,48(4):367-376.JU Ping,GUO Dezheng,CAO Lu,et al.Review and prospect of modeling on generalized synthesis electric load containing active loads[J].Journal of Hohai University(Natural Sciences),2020,48(4):367-376.
• [5]徐重酉，王明月，刘天元，等.非理想通信环境下主动配电网柔性负荷建模与优化控制方法[J].浙江电力，2023,42(11):104-113.XU Chongyou,WANG Mingyue,LIU Tianyuan,et al.A modeling and optimal control method of flexible load in the active distribution networks in non-ideal communication environments[J]. Zhejiang Electric Power,2023,42(11):104-113.
• [6]王国春，许洪强，冯长有，等.新一代在线安全分析技术架构及未来展望[J].电力系统自动化，2023,47(24):110-120.WANG Guochun,XU Hongqiang,FENG Changyou,et al.Technical architecture and future prospect for new generation of online security analysis[J].Automation of Electric Power Systems,2023,47(24):110-120.
• [7]舒印彪，张智刚，郭剑波，等.新能源消纳关键因素分析及解决措施研究[J].中国电机工程学报，2017,37(1):1-9.SHU Yinbiao,ZHANG Zhigang,GUO Jianbo,et al.Study on key factors and solution of renewable energy accommodation[J].Proceedings of the CSEE,2017,37(1):1-9.
• [8]郭剑波，王铁柱，罗魁，等.新型电力系统面临的挑战及应对思考[J].新型电力系统，2023(1):32-43.GUO Jianbo,WANG Tiezhu,LUO Kui,et al. Development of new power systems:challenges and solutions[J].New Type Power Systems,2023(1):32-43.
• [9]LAN T K,SUN H D,WANG Q,et al. Synthesis load model with renewable energy sources for transient stability studies[J].IEEE Transactions on Power Systems,2024,39(1):1647-1663.
• [10]周佩朋，李光范，孙华东，等.基于频域阻抗分析的直驱风电场等值建模方法[J].中国电机工程学报，2020,40（增刊1）:84-90.ZHOU Peipeng,LI Guangfan,SUN Huadong,et al.Equivalent modeling method of PMSG wind farm based on frequency domain impedance analysis[J]. Proceedings of the CSEE,2020,40(S1):84-90.
• [11]ARIF A,WANG Z Y,WANG J H,et al.Load modeling:a review[J]. IEEE Transactions on Smart Grid,2018,9(6):5986-5999.
• [12]HISKENS I A.Nonlinear dynamic model evaluation from disturbance measurements[J]. IEEE Transactions on Power Systems,2001,16(4):702-710.
• [13]HUANG Q H,HUANG R K,PALMER B J,et al.A generic modeling and development approach for WECC composite load model[J].Electric Power Systems Research,2019,172:1-10.
• [14]ZHAO J B,WANG Z Y,WANG J H. Robust timevarying load modeling for conservation voltage reduction assessment[J].IEEE Transactions on Smart Grid,2018,9(4):3304-3312.
• [15]胡心远.考虑多扰动场景的广义综合负荷模型参数辨识方法研究[D].天津：天津大学，2022.HU Xinyuan.Study on parameter identification method of generalized composite load model considering multiscenario disturbance[D].Tianjin:Tianjin University,2022.
• [16]徐岩，靳伟佳，朱晓荣.基于遗传粒子群算法的光伏并网逆变器参数辨识[J].太阳能学报，2021,42(7):103-109.XU Yan,JIN Weijia,ZHU Xiaorong.Parameter identification of photovoltaic grid-connected inverter based on gapso[J].Acta Energiae Solaris Sinica,2021,42(7):103-109.
• [17]盛四清，关皓闻，雷业涛，等.基于混沌海鸥优化算法的含光伏发电系统负荷模型参数辨识[J].太阳能学报，2022,43(7):64-72.SHENG Siqing,GUAN Haowen,LEI Yetao,et al.Parameter identification of load model of photovoltaic power generation system based on chaotic seagull optimization algorithm[J].Acta Energiae Solaris Sinica,2022,43(7):64-72.
• [18]MA J,DONG Z Y,ZHANG P. Using a support vector machine(SVM)to improve generalization ability of load model parameters[C]//2009 IEEE/PES Power Systems Conference and Exposition. March 15-18,2009,Seattle,WA,USA:IEEE,2009:1-8.
• [19]LEE S H,SON S E,LEE S M,et al.Kalman-filter based static load modeling of real power system using K-EMS data[J].Journal of Electrical Engineering and Technology,2012,7(3):304-311.
• [20]虞殷树，陈东海，朱耿，等.基于改进关联分析的行业短期电力负荷预测[J].浙江电力，2023,42(11):29-38.YU Yinshu,CHEN Donghai,ZHU Geng,et al.A shortterm power load forecasting method for industrial sectors based on an improved correlation analysis[J]. Zhejiang Electric Power,2023,42(11):29-38.
• [21]杨莹，刘启航，赵为光，等.基于深度强化学习的光氢储新能源汽车充能站优化调度[J].电力需求侧管理，2025,27(4):92-97.YANG Ying,LIU Qihang,ZHAO Weiguang,et al.Optimization scheduling of photovoltaic-hydrogen-storage for new energy vehicle charging station based on deep reinforcement learning[J].Power Demand Side Management,2025,27(4):92-97.
• [22]刘斯亮，郑泽南，张勇军，等.计及灰数据的知识-数据驱动低压有源配电网潮流计算[J].电测与仪表，2025,62(6):2-10.LIU Siliang,ZHENG Zenan,ZHANG Yongjun,et al.Knowledge-data-driven power flow calculation for lowvoltage active distribution network considering gray data[J].Electrical Measurement&Instrumentation,2025,62(6):2-10.
• [23]李钟平，向月.深度强化学习驱动的风储系统参与能量-调频市场竞价策略[J].电力工程技术，2025,44(3):30-42.LI Zhongping,XIANG Yue.Deep reinforcement learningdriven bidding strategy for wind-storage systems in energy and frequency regulation markets[J].Electric Power Engineering Technology,2025,44(3):30-42.
• [24]王建军，潘佳音，赵珍珠，等.基于特征时间双注意力机制的短期光伏发电预测深度学习模型研究[J].智慧电力，2025,53(4):81-87.WANG Jianjun,PAN Jiayin,ZHAO Zhenzhu,et al. A deep learning model for short-term photovoltaic power generation forecasting based on dual feature-temporal attention mechanism[J].Smart Power,2025,53(4):81-87.
• [25]戴瑞成，董翔，赵璧，等.基于混合深度学习的变电站巡检机器人目标识别算法研究[J].智慧电力，2025,53(3):117-123.DAI Ruicheng,DONG Xiang,ZHAO Bi,et al.Research on target recognition algorithm for substation inspection robot based on hybrid deep learning[J]. Smart Power,2025,53(3):117-123.
• [26]翟乐庆，刘益青，魏元健，等.基于同步提取变换和卷积神经网络的有源配电网单相接地故障选线方法[J].山东电力技术，2025,52(2):65-77.ZHAI Leqing,LIU Yiqing,WEI Yuanjian,et al.A singlephase grounding fault line selection method for active distribution network based on synchroextracting transform and convolutional neural network[J]. Shandong Electric Power,2025,52(2):65-77.
• [27]焦昊，赵佳伟，韦磊，等.基于深度迁移学习的电力系统暂态状态估计[J].电力建设，2025,46(1):97-106.JIAO Hao,ZHAO Jiawei,WEI Lei,et al.Transient state estimation for power system based on deep transfer learning[J].Electric Power Construction,2025,46(1):97-106.
• [28]白云鹏，张志艳，许才，等.基于多头注意力机制改进图神经网络的新能源电力系统风险评估[J].电力建设，2025,46(1):147-157.BAI Yunpeng,ZHANG Zhiyan,XU Cai,et al.Risk assessment of renewable energy power systems via graph multiattention networks[J].Electric Power Construction,2025,46(1):147-157.