温子健,李伟风,刘泽晖,易永海,李幸聪,曾君
WEN Zijian,LI Weifeng,LIU Zehui,YI Yonghai,LI Xingcong,ZENG Jun

• 1:广东电网有限责任公司江门供电局
• 2:华南理工大学自动化学院
• 3:华南理工大学电力学院

摘要(Abstract)：

电动汽车数量的快速增长，对充电负荷预测精度提出了更高要求。为此，提出了一种基于时间特征分类、并行预测与误差修正的充电负荷预测方法。首先，基于K-means分簇方法对负荷序列进行分类标记，以分离不同负荷时间特性的时间点。其次，先利用LSTM（长短期记忆网络）和GRU（门控循环单元）对各类时间点并行独立预测，再利用CatBoost（类别提升）根据时间点类别进行预测输出的选择，并将预测结果与真实结果相减，获得历史误差序列。最后，将历史误差序列按半天、一天、一周、两周、一月以及一季度周期进行序列分解，分别利用Transformer模型进行误差成分预测，实现对预测误差的跟踪与修正，获得精确的负荷预测结果。结果表明，所提方法相较传统基准预测方法在预测精度上有所提升。
The rapid growth in the number of electric vehicles(EVs) has put forward higher requirements for the accuracg of charging load forecasting. Accordingly, this paper proposes a charging load forecasting method based on temporal feature classification, parallel forecasting, and error correction. First, load series are classified and labeled using the K-means clustering method to distinguish time points with different load temporal characteristics. Second, long short-term memory(LSTM) and gated recurrent unit(GRU) networks are employed to perform parallel and independent forecasting for each category of time points. CatBoost(categorical boosting) is then applied to select the forecast output according to the time point category, and the forecast results are subtracted from the actual values to obtain a historical error series. Finally, the historical error series is decomposed into components with periods of half a day, one day, one week, two weeks, one month, and one quarter. A Transformer model is utilized to forecast each error component separately, enabling tracking and correction of the forecast error and yielding a refined load forecast. Results show that the proposed method improves forecast accuracy compared with conventional benchmark forecasting approaches.

关键词(KeyWords)： 电动汽车充电;负荷预测;时间特征分类;并行预测;误差修正
electric vehicle charging;load forecasting;temporal feature classification;parallel forecasting;error correction

Abstract:

Keywords:

基金项目(Foundation): 广东省自然科学基金（2024A1515012428）;; 广东电网有限责任公司科技项目（GDKJXM20240065）

作者(Author): 温子健,李伟风,刘泽晖,易永海,李幸聪,曾君
WEN Zijian,LI Weifeng,LIU Zehui,YI Yonghai,LI Xingcong,ZENG Jun

DOI: 10.19585/j.zjdl.202605007

参考文献(References)：

• [1]曹佳晨，张沈习，张璐，等.考虑交通流的柔性互联配电网电动汽车承载能力计算方法[J].电力系统自动化，2025,49(5):24-37.CAO Jiachen,ZHANG Shenxi,ZHANG Lu,et al.Calculation method for electric vehicle hosting capacity of flexibly interconnected distribution network considering traffic flow[J].Automation of Electric Power Systems,2025,49(5):24-37.
• [2]李振坤，肖天宇，宋治儒，等.基于节假日因素影响的高速服务区充电负荷建模与充电桩优化规划[J].电力建设，2025,46(12):57-69.LI Zhenkun,XIAO Tianyu,SONG Zhiru,et al.Chargingload modeling and optimal charging-pile planning in highway service areas under the influence of holiday factors[J].Electric Power Construction,2025,46(12):57-69.
• [3]王子龙，黄莉.基于改进图时间卷积网络的农村地区电动汽车充电负荷预测及其对农网的影响[J].电力需求侧管理，2024,26(5):88-93.WANG Zilong,HUANG Li. Electric vehicle charging load prediction in rural areas and its impact on rural power grids based on modified graph temporal convolutional network[J].Power Demand Side Management,2024,26(5):88-93.
• [4]郭鑫喆，王业琴，王超，等.基于多目标模态分解与NAHL神经网络的电动汽车充电负荷预测方法[J].电测与仪表，2025,62(3):20-29.GUO Xinzhe,WANG Yeqin,WANG Chao,et al.Electric vehicle charging load prediction method based on multiobjective modal decomposition and NAHL neural network[J].Electrical Measurement&Instrumentation,2025,62(3):20-29.
• [5]胡枭，张泽朕，杨家全，等.融合路网-气象-日期多特征信息的电动汽车充电负荷预测[J].电力建设，2025,46(9):57-70.HU Xiao,ZHANG Zezhen,YANG Jiaquan,et al.Electric vehicle charging load forecasting by integrating multifeature information of road network,meteorology,and date[J].Electric Power Construction,2025,46(9):57-70.
• [6]陈晓华，吴杰康，张勋祥，等.基于鹈鹕优化算法优化广义回归神经网络的电动汽车充电负荷短期预测[J].山东电力技术，2024,51(7):1-9.CHEN Xiaohua,WU Jiekang,ZHANG Xunxiang,et al.Electric vehicle charging load short-term prediction based on generalized regression neural network optimized by pelican optimization algorithm[J].Shandong Electric Power,2024,51(7):1-9.
• [7]丁乐言，柯松，张帆，等.考虑出行需求和引导策略的电动汽车充电负荷预测[J].电力建设，2024,45(6):10-26.DING Leyan,KE Song,ZHANG Fan,et al.Forecasting of electric-vehicle charging load considering travel demand and guidance strategy[J]. Electric Power Construction,2024,45(6):10-26.
• [8]张夏韦，梁军，王要强，等.电动汽车充电负荷时空分布预测研究综述[J].电力建设，2023,44(12):161-173.ZHANG Xiawei,LIANG Jun,WANG Yaoqiang,et al.Overview of research on spatiotemporal distribution prediction of electric vehicle charging[J].Electric Power Construction,2023,44(12):161-173.
• [9]锁军，李龙，贺瀚青，等.考虑交通路况的电动汽车充电负荷预测[J].电网与清洁能源，2022,38(10):141-147.SUO Jun,LI Long,HE Hanqing,et al.Load forecasting of electric vehicle charging considering traffic conditions[J].Power System and Clean Energy,2022,38(10):141-147.
• [10]牛牧童，廖凯，杨健维，等.考虑季节特性的多时间尺度电动汽车负荷预测模型[J].电力系统保护与控制，2022,50(5):74-85.NIU Mutong,LIAO Kai,YANG Jianwei,et al. Multitime-scale electric vehicle load forecasting model considering seasonal characteristics[J]. Power System Protection and Control,2022,50(5):74-85.
• [11]陈勇，江颖达，徐刚，等.规模化电动汽车充电负荷预测[J].电力需求侧管理，2022,24(5):71-77.CHEN Yong,JIANG Yingda,XU Gang,et al. Charging load forecasting for large-scale electric vehicle[J]. Power Demand Side Management,2022,24(5):71-77.
• [12]黄南天，刘德宝，蔡国伟，等.基于多相关日场景生成的电动汽车充电负荷区间预测[J].中国电机工程学报，2021,41(23):7980-7990.HUANG Nantian,LIU Debao,CAI Guowei,et al.Interval prediction of electric vehicle charging load based on scene generation with multiple correlation days[J]. Proceedings of the CSEE,2021,41(23):7980-7990.
• [13]BUZNA L,DE FALCO P,FERRUZZI G,et al.An ensemble methodology for hierarchical probabilistic electric vehicle load forecasting at regular charging stations[J].Applied Energy,2021,283:116337.
• [14]刘敦楠，张悦，彭晓峰，等.计及相似日与气象因素的电动汽车充电负荷聚类预测[J].电力建设，2021,42(2):43-49.LIU Dunnan,ZHANG Yue,PENG Xiaofeng,et al.Clustering prediction of electric vehicle charging load considering similar days and meteorological factors[J]. Electric Power Construction,2021,42(2):43-49.
• [15]袁小溪，潘鸣宇，段大鹏，等.基于网格划分的电动汽车充电负荷预测方法[J].电力科学与技术学报，2021,36(3):19-26.YUAN Xiaoxi,PAN Mingyu,DUAN Dapeng,et al.Prediction method of electric vehicle charging load based on grid division[J]. Journal of Electric Power Science and Technology,2021,36(3):19-26.
• [16]DABBAGHJAMANESH M,MOEINI A,KAVOUSIFARD A. Reinforcement learning-based load forecasting of electric vehicle charging station using Q-learning technique[J]. IEEE Transactions on Industrial Informatics,2021,17(6):4229-4237.
• [17]ZHANG X,CHAN K W,LI H R,et al. Deep-learningbased probabilistic forecasting of electric vehicle charging load with a novel queuing model[J]. IEEE Transactions on Cybernetics,2021,51(6):3157-3170.
• [18]KHAN W,SOMERS W,WALKER S,et al.Comparison of electric vehicle load forecasting across different spatial levels with incorporated uncertainty estimation[J]. Energy,2023,283:129213.
• [19]刘志强，张谦，朱熠，等.计及车-路-站-网融合的电动汽车充电负荷时空分布预测[J].电力系统自动化，2022,46(12):36-45.LIU Zhiqiang,ZHANG Qian,ZHU Yi,et al. Spatialtemporal distribution prediction of charging loads for electric vehicles considering vehicle-road-station-grid integration[J].Automation of Electric Power Systems,2022,46(12):36-45.
• [20]张延宇，张智铭，刘春阳，等.基于动态自适应图神经网络的电动汽车充电负荷预测[J].电力系统自动化，2024,48(7):86-93.ZHANG Yanyu,ZHANG Zhiming,LIU Chunyang,et al.Electric vehicle charging load prediction based on dynamic adaptive graph neural network[J].Automation of Electric Power Systems,2024,48(7):86-93.
• [21]赵子鋆，彭清文，邓铭，等.考虑多因素影响与误差修正的充电站负荷预测[J].浙江电力，2024,43(4):21-28.ZHAO Zijun,PENG Qingwen,DENG Ming,et al.Load forecasting for charging stations considering multiple influencing factors and error correction[J]. Zhejiang Electric Power,2024,43(4):21-28.
• [22]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.
• [23]ZHANG Y S,XU T X,CHEN T,et al.A high-resolution electric vehicle charging transaction dataset with multidimensional features in China[J]. Scientific Data,2025,12(1):643.
• [24]WU H,XU J,WANG J,et al.Autoformer:Decomposition transformers with auto-correlation for long-term series forecasting[J].Advances in neural information processing systems,2021,34:22419-22430.