刘瑞霖,余洋,刘鋆,李伟
LIU Ruilin,YU Yang,LIU Jun,LI Wei

• 1:新能源电力系统全国重点实验室(华北电力大学(保定))
• 2:河北省分布式储能与微网重点实验室(华北电力大学(保定))
• 3:内蒙古电力集团蒙电经济技术研究院有限责任公司

摘要(Abstract)：

针对传统蒙特卡洛模拟方法在电动汽车充电负荷预测中存在的概率分布拟合误差大及数据随机性强等问题，提出基于GMM（高斯混合模型）与GRA（灰色关联度分析）的改进方法。首先，利用GMM拟合充电行为特征的多峰分布特性，并通过BIC（贝叶斯信息准则）优化高斯分量个数，提升概率分布模型准确性。其次，应用GRA分析蒙特卡洛随机生成的充电行为数据组与原始数据的关联度，筛选最优数据组以降低极端值影响。最后，引入电池容量伽马分布模型表征不同类型电动汽车的异质性充电需求，进行电动汽车负荷预测。仿真结果表明：采用改进方法拟合的起始充电时间与起始充电SOC（荷电状态）相关系数最高达0.999 5，预测的总充电负荷预测峰值更贴合实际用电高峰时段，显著提升了预测精度。该方法通过融合GMM多峰拟合能力与GRA数据筛选机制，解决了传统方法中概率模型单一性和数据随机性问题，为电网负荷规划与动态平衡提供了技术支撑。
To address the limitations of conventional Monte Carlo simulation methods in electric vehicle(EV) charging load forecasting—particularly their large probability distribution fitting errors and strong data randomness—this paper proposes an enhanced approach integrating Gaussian mixture model(GMM) and grey relational analysis(GRA). First, GMM is employed to fit the multimodal distribution characteristics of charging behaviors, with the Bayesian information criterion(BIC) optimizing the number of Gaussian components to enhance probability distribution model accuracy. Second, GRA evaluates the relational degree between Monte Carlo-generated charging behavior datasets and the original data, screening optimal datasets to mitigate extreme value impacts. Finally, a Gamma distribution model of battery capacity is introduced to characterize heterogeneous charging demands across EV types. Simulation results demonstrate that the enhanced method achieves a correlation coefficient of up to 0.999 5 for initial charging time and initial state of charge(SOC) fitting, while the forecasted peak total charging load aligns closely with actual peak demand periods, significantly improving forecasting precision. By combining GMM's multimodal fitting capability with GRA's data screening mechanism, this method resolves the oversimplified probability models and data randomness inherent in traditional approaches, offering robust technical support for grid load planning and dynamic balancing.

关键词(KeyWords)： 电动汽车;充电负荷预测;蒙特卡洛模拟;高斯混合模型;灰色关联度分析
EV;charging load forecasting;Monte Carlo simulation;GMM;GRA

Abstract:

Keywords:

基金项目(Foundation): 国家重点研发计划(2018YFE0122200)

作者(Author): 刘瑞霖,余洋,刘鋆,李伟
LIU Ruilin,YU Yang,LIU Jun,LI Wei

DOI: 10.19585/j.zjdl.202508002

参考文献(References)：

• [1]新华社.我国新能源汽车产销量连续10年位居全球第一[EB/OL].(2025-01-21)[2025-03-02].https：//www.gov.cn/lianbo/bumen/202501/content_7000306.htmslb=true.
• [2]新华社.我国将加强新能源汽车动力电池回收利用[EB/OL].(2025-02-22)[2025-03-02]. https：//www. gov. cn/zhengce/202502/content_7004966.htmddtab=true.
• [3]宣羿，樊立波，孙智卿，等.考虑低碳交通的电动汽车充电站优化配置方法[J].浙江电力，2024,43(6):69-79.XUAN Yi,FAN Libo,SUN Zhiqing,et al.An optimal allocation method for electric vehicle charging stations considering lowcarbon transportation[J]. Zhejiang Electric Power,2024,43(6):69-79.
• [4]黄健，陈建红，何剑杰，等.基于GCN-LSTM的电动汽车负荷预测方法[J].浙江电力，2024,43(12):59-67.HUANG Jian,CHEN Jianhong,HE Jianjie,et al.An EV load forecasting method for using GCN-LSTM[J].Zhejiang Electric Power,2024,43(12):59-67.
• [5]黄凤娇.规模化电动汽车充电负荷对电网影响的研究[J].电工技术，2024(20):113-116.HUANG Fengjiao.Study on the impact of load of largescale EV charging on power grid[J].Electric Engineering,2024(20):113-116.
• [6]王浩林，张勇军，毛海鹏.基于时刻充电概率的电动汽车充电负荷预测方法[J].电力自动化设备，2019,39(3):207-213.WANG Haolin,ZHANG Yongjun,MAO Haipeng.Charging load forecasting method based on instantaneous charging probability for electric vehicles[J].Electric Power Automation Equipment,2019,39(3):207-213.
• [7] GUO C L,ZHU K J,CHEN C C,et al. Characteristics and effect laws of the large-scale electric Vehicle’s charging load[J].eTransportation,2020,3:100049.
• [8]游磊，金小明，刘云.电动汽车集群充电负荷计算方法研究[J].南方能源建设，2024,11(5):159-167.YOU Lei,JIN Xiaoming,LIU Yun.Research on the charging load calculation method for electric vehicle cluster[J].Southern Energy Construction,2024,11(5):159-167.
• [9] TIAN J,LV Y,ZHAO Q,et al.Electric vehicle charging load prediction considering the orderly charging[J]. Energy Reports,2022,8:124-134.
• [10]郭创新，刘洞宇，朱承治，等.电动汽车居民区充电负荷建模分析[J].电力自动化设备，2020,40(1):1-9.GUO Chuangxin,LIU Dongyu,ZHU Chengzhi,et al.Modeling and analysis of electric vehicle charging load in residential area[J]. Electric Power Automation Equipment,2020,40(1):1-9.
• [11]蒋建东，韩文文，苑子俊，等.居民区电动汽车充电负荷建模研究[J].郑州大学学报（理学版），2019,51(1):73-77.JIANG Jiandong,HAN Wenwen,YUAN Zijun,et al.The modeling of electric vehicle charging load in residential areas[J].Journal of Zhengzhou University(Natural Science Edition),2019,51(1):73-77.
• [12]毛建斌，杨少兵，杨湘彦，等.基于有限混合模型的规模化电动汽车充电负荷在线预测方法[J].电网技术，2025,49(5):1931-1940.MAO Jianbin,YANG Shaobing,YANG Xiangyan,et al.Charging load online forecasting method based on finite mixture model for large-scale electric vehicles[J]. Power System Technology,2025,49(5):1931-1940.
• [13]熊紫维，秦志友，邓志坤，等.基于神经网络和混合高斯模型的光伏出力预测研究[C]//2024年电力行业技术监督专业技术交流研讨会优秀论文集.南宁，2024:1106-1117.Xiong Ziwei,Qin Zhiyou,Deng Zhikun,et al.Research on photovoltaic power output prediction based on neural network and Gaussian mixture model[C]//Proceedings of Electric Power Industry Technical Supervision Professional Technology Exchange Conference 2024. Nanning,2024:1106-1117.
• [14]朱亮，支妍力，梅贱生，等.基于GMM-FHMM的工业产线非介入式负荷辨识[J].浙江电力，2024,43(12):68-76.ZHU Liang,ZHI Yanli,MEI Jiansheng,et al. Nonintrusive load monitoring for industrial production line based on GMM-FHMM[J]. Zhejiang Electric Power,2024,43(12):68-76.
• [15]闫威，李南，沈月秀，等.基于CNN-GAN与半监督回归的电动汽车充电负荷预测[J].浙江电力，2023,42(2):83-89.YAN Wei,LI Nan,SHEN Yuexiu,et al.Electric vehicle charging load forecasting based on CNN-GAN and semisupervised regression[J].Zhejiang Electric Power,2023,42(2):83-89.
• [16]夏飞，袁博，彭道刚，等.基于信息量准则的锂离子电池变阶RC等效电路模型建模及优化方法[J].中国电机工程学报，2018,38(21):6441-6451.XIA Fei,YUAN Bo,PENG Daogang,et al.Modeling and optimization of variable-order RC equivalent circuit model for lithium ion batteries based on information criterion[J].Proceedings of the CSEE,2018,38(21):6441-6451.
• [17]肖碧涛，赖晓路，郭鹏，等.高斯混合模型风电机组异常数据识别方法研究[J].可再生能源，2023,41(8):1051-1056.XIAO Bitao,LAI Xiaolu,GUO Peng,et al. Abnormal data identification for wind turbine based on Gaussian Mixture Model[J]. Renewable Energy Resources,2023,41(8):1051-1056.
• [18]邓俊，张阳，李怡然，等.基于高斯混合模型聚类的双馈风电场动态等值建模方法[J].太阳能学报，2024,45(1):342-350.DENG Jun,ZHANG Yang,LI Yiran,et al. Dynamic equivalence modeling of doubly-fed wind farm based on Gaussian mixture model clustering algorithm[J].Acta Energiae Solaris Sinica,2024,45(1):342-350.
• [19]陈丽丹，聂涌泉，钟庆.基于出行链的电动汽车充电负荷预测模型[J].电工技术学报，2015,30(4):216-225.CHEN Lidan,NIE Yongquan,ZHONG Qing. A model for electric vehicle charging load forecasting based on trip chains[J].Transactions of China Electrotechnical Society,2015,30(4):216-225.
• [20]李磊，赵新，李晓辉，等.基于动态交通信息的电动汽车充电需求预测模型及其对配网的影响分析[J].电网与清洁能源，2020,36(3):107-118.LI Lei,ZHAO Xin,LI Xiaohui,et al. Electric vehicle charging demand prediction model based on dynamic traffic information and its impacts on distribution networks[J].Power System and Clean Energy,2020,36(3):107-118.
• [21] MU Y F,WU J Z,JENKINS N,et al.A Spatial-temporal model for grid impact analysis of plug-in electric vehicles[J].Applied Energy,2014,114:456-465.
• [22]王雁凌，吴梦凯，周子青，等.基于改进灰色关联度的电力负荷影响因素量化分析模型[J].电网技术，2017,41(6):1772-1778.WANG Yanling,WU Mengkai,ZHOU Ziqing,et al.Quantitative analysis model of power load influencing factors based on improved grey relational degree[J]. Power System Technology,2017,41(6):1772-1778.
• [23]方尚尚，王冰，胡庆燚，等.基于蒙特卡洛算法的电动汽车充电需求负荷研究[J].系统仿真技术，2020,16(3):150-155.FANG Shangshang,WANG Bing,HU Qingyi,et al.Research on electric vehicle charging demand load based on Monte Carlo algorithm[J]. System Simulation Technology,2020,16(3):150-155.