考虑温控型负荷特性影响的集群用户超短期负荷预测方法

doi:10.11930/j.issn.1004-9649.202502075

中国电力 ›› 2025, Vol. 58 ›› Issue (12): 63-72, 85.DOI: 10.11930/j.issn.1004-9649.202502075

• 协同海量分布式灵活性资源的韧性城市能源系统关键技术 • 上一篇

考虑温控型负荷特性影响的集群用户超短期负荷预测方法

孟浩¹(), 徐飞²(), 符帅¹, 孙鹏¹, 郝玲², 刘博宇², 刘芷维²

1. 国网内蒙古东部电力有限公司通辽供电公司，内蒙古通辽　028000
2. 新型电力系统运行与控制全国重点实验室（清华大学）北京　100084

收稿日期:2025-02-08 修回日期:2025-11-17 发布日期:2025-12-27 出版日期:2025-12-28
作者简介:
孟浩（1984），硕士，高级工程师，从事配网生产研究工作，E-mail：1490358109@qq.com
徐飞（1974），通信作者，研究员，从事配电网优化运行、可再生能源消纳提升技术研究，E-mail：xufei@tsinghua.edu.cn
基金资助:
国网内蒙古东部电力有限公司通辽供电公司科技项目（蒙东农牧区配电网典型负荷特征分析预测建模及平衡调控技术研究，526620240008）。

Ultra-Short-Term Load Forecasting Method for Aggregated Users Considering the Impact of Temperature-Controlled Load Characteristics

MENG Hao¹(), XU Fei²(), FU Shuai¹, SUN Peng¹, HAO Ling², LIU Boyu², LIU Zhiwei²

1. Tongliao Power Supply Company, State Grid Inner Mongolia East Electric Power Co., Ltd., Tongliao 028000, China
2. State Key Laboratory of Power System Operation and Control (Tsinghua University), Beijing 100084, China

Received:2025-02-08 Revised:2025-11-17 Online:2025-12-27 Published:2025-12-28
Supported by:
This work is supported by Science and Technology Project of State Grid East Inner Mongolia Electric Power Co., Ltd. Tongliao Power Supply Company (Research on Typical Load Characteristic Analysis and Prediction Modeling and Balance Control Technology of Distribution Network in Agricultural And Pastoral Areas of Eastern Inner Mongolia, No.526620240008).

摘要/Abstract

摘要：

含高比例温控型负荷的集群用户负荷易受气温变化等因素影响而发生特性突变，使得历史不同时间的负荷特性存在时序分布偏移，导致已有的集群用户负荷预测建模方法因泛化性不足而效果不佳。借鉴迁移学习提取空间维度域不变特征的思路，提出了一种基于时域不变特征建模的集群用户超短期负荷预测方法。由于负荷时序分布发生偏移的周期和各周期的边界通常是未知的，首先，对负荷的时序分布偏移情况进行了量化，将负荷划分为具有显著分布差异的序列，以支撑后续样本间时域共性特征的提取；然后，提出了基于Transformer的时域不变特征提取算法，通过最小化不同分布数据样本间的时序分布差异，提取时域不变特征，进而优化负荷预测建模，提升负荷特性突变场景下的预测精度。最后，基于真实居民负荷数据验证了所提方法的优越性。

关键词: 温控型负荷, 数据时序分布偏移, Transformer模型, 超短期负荷预测, 深度学习

Abstract:

The load of aggregated users with a high proportion of temperature-controlled loads is prone to abrupt in characteristics due to factors such as temperature variations, leading to temporal distribution shifts in load characteristics across different historical periods. This results in poor performance of existing load forecasting modeling methods for aggregated users due to insufficient generalization capability. Drawing on the concept of transfer learning for extracting domain-invariant features in the spatial dimension, an ultra-short-term load forecasting method for aggregated users based on time-domain invariant feature modeling is proposed. Since the cycles of temporal distribution shifts in load data and the boundaries of these cycles are typically unknown, firstly, the temporal distribution shift is quantified, and the load is segmented into sequences with significant distribution differences to support the subsequent extraction of time-domain common features among samples. Then, a Transformer-based time-domain invariant feature extraction algorithm is proposed, which minimizes the temporal distribution differences among data samples with varying distributions to extract time-domain invariant features, thereby optimizing load forecasting modeling and improving prediction accuracy under scenarios of abrupt load characteristic changes. Finally, the superiority of the proposed method is validated using real residential load data.

Key words: temperature-controlled load, data temporal distribution shift, transformer model, ultra-short-term load forecasting, deep learning

孟浩, 徐飞, 符帅, 孙鹏, 郝玲, 刘博宇, 刘芷维. 考虑温控型负荷特性影响的集群用户超短期负荷预测方法[J]. 中国电力, 2025, 58(12): 63-72, 85.

MENG Hao, XU Fei, FU Shuai, SUN Peng, HAO Ling, LIU Boyu, LIU Zhiwei. Ultra-Short-Term Load Forecasting Method for Aggregated Users Considering the Impact of Temperature-Controlled Load Characteristics[J]. Electric Power, 2025, 58(12): 63-72, 85.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202502075

https://www.electricpower.com.cn/CN/Y2025/V58/I12/63

图/表 10

图 1 数据偏移程度指标

Fig.1 Data shift degree indicator

图 2 研究方法整体流程

Fig.2 Flow chart of research methods

图 3 负荷预测模型结构

Fig.3 Structure of load forecasting model

表 1 不同域数量的精度对比

Table 1 Accuracy comparison of different numbers of domains

域数量	调和精度/%	RMSE精度/%	MAE精度/%
2	95.72	94.91	96.43
3	95.82	95.25	96.91
4	96.08	94.89	96.44
5	96.17	95.47	97.11
6	96.07	95.33	96.90
7	96.10	94.93	96.52
8	95.75	94.32	95.81

图 4 不同序列划分方法预测精度对比

Fig.4 Comparison of forecasting accuracy of different sequence partitioning methods

表 2 测试集精度

Table 2 Test set accuracy

方法	调和精度/%	NRMSE精度/%	NMAE精度/%
1	94.65	94.37	95.61
2	96.02	95.13	96.91
3	95.89	95.27	96.93
4	91.93	90.68	92.50
5	96.17	95.47	97.11

图 5 不同预测步长时RMSE精度

Fig.5 RMSE accuracy for different forecasting steps

图 6 不同方法在不同时间尺度下预测结果对比

Fig.6 Comparison of forecasting results of different methods under different time scales

图 7 不同时序分布偏移程度下精度对比

Fig.7 Accuracy comparison under different temporal distribution shift degrees

图 8 不同时序分布偏移程度下的提升程度

Fig.8 Improvement degree under different temporal distribution shift degrees

参考文献 31

1	李思维, 许中平, 于龙, 等. 计及电动汽车和温控负荷集群的负荷控制用户组合优化方法[J]. 中国电力, 2025, 58 (3): 86- 97. DOI
	LI Siwei, XU Zhongping, YU Long, et al. A Load Control User Combination Optimization Method with Electric Vehicle and Temperature Control Load Cluster[J]. Electric Power, 2025, 58 (3): 86- 97. DOI
2	徐玉婷, 田世明, 陈宋宋, 等. 基于LSTM的居民负荷预测及其可调节潜力分析[J]. 电力信息与通信技术, 2023, 21 (05): 1- 8. DOI
	XU Yuting, TIAN Shiming, CHEN Songsong, et al. Resident load forecasting based on LSTM and its adjustable potential analysis[J]. Electric Power Information and Communication Technology, 2023, 21 (05): 1- 8. DOI
3	史媛, 周浩, 张俊娜, 等. 双碳背景下的用户负荷主动优化方法研究[J]. 电力信息与通信技术, 2023, 21 (5): 25- 32. DOI
	SHI Yuan, ZHOU Hao, ZHANG Junna, et al. Research on user load active optimization method under dual carbon background[J]. Electric Power Information and Communication Technology, 2023, 21 (5): 25- 32. DOI
4	杨胡萍, 余阳, 汪超, 等. 基于VMD-CNN-BIGRU的电力系统短期负荷预测[J]. 中国电力, 2022, 55 (10): 71- 76.
	YANG Huping, YU Yang, WANG Chao, et al. Short-term load forecasting of power system based on VMD-CNN-BIGRU[J]. Electric Power, 2022, 55 (10): 71- 76.
5	葛鑫鑫, 付志扬, 徐飞, 等. 面向新型电力系统的虚拟电厂商业模式与关键技术[J]. 电力系统自动化, 2022, 46 (18): 129- 146. DOI
	GE Xinxin, FU Zhiyang, XU Fei, et al. Business model and key technologies of virtual power plant for new power system[J]. Automation of Electric Power Systems, 2022, 46 (18): 129- 146. DOI
6	周颖, 白雪峰, 王阳, 等. 面向虚拟电厂运营的温度敏感负荷分析与演变趋势研判[J]. 中国电力, 2024, 57 (1): 9- 17.
	ZHOU Ying, BAI Xuefeng, WANG Yang, et al. Analysis and evolution trend of temperature-sensitive loads for virtual power plant operation[J]. Electric Power, 2024, 57 (1): 9- 17.
7	杨燕, 王云, 胡熊伟, 等. 基于5G无线通信的可中断负荷快速响应系统[J]. 电力信息与通信技术, 2024, 22 (10): 83- 89. DOI
	YANG Yan, WANG Yun, HU Xiongwei, et al. Interruptible load fast response system based on 5G wireless communicatio[J]. Electric Power Information and Communication Technology, 2024, 22 (10): 83- 89. DOI
8	朱明, 夏宇栋, 常凯, 等. 基于粒子群优化算法的空调负荷灰箱模型辨识[J]. 电力科学与技术学报, 2023, 38 (4): 214- 221. DOI
	ZHU Ming, XIA Yudong, CHANG Kai, et al. Identification of grey box model for air conditioning load based on particleswarm optimization algorithm[J]. Journal of Electric Power Science and Technology, 2023, 38 (4): 214- 221. DOI
9	马涛, 李津, 曹晓波, 等. 计及用户行为差异性和配电网潮流的电采暖负荷群优化调度策略研究[J]. 电力科学与技术学报, 2023, 38 (1): 77- 87. DOI
	MA Tao, LI Jin, CAO Xiaobo, et al. Research on optimal dispatch strategy of electric heating load groups considering userbehavior difference and distribution network power flow[J]. Journal of Electric Power Science and Technology, 2023, 38 (1): 77- 87. DOI
10	徐宇颂, 邹山花, 卢先领. 基于特征选择和组合模型的短期电力负荷预测[J]. 中国电力, 2022, 55 (7): 121- 127. DOI
	XU Yusong, ZOU Shanhua, LU Xianling. Short-term load forecasting based on feature selection and combination model[J]. Electric Power, 2022, 55 (7): 121- 127. DOI
11	杨珂, 朱洪斌, 李达, 等. 基于压缩感知的纵向联邦学习园区负荷预测方法[J]. 电力信息与通信技术, 2024, 22 (5): 36- 42.
	YANG Ke, ZHU Hongbin, LI Da, et al. Load forecasting method for vertical federated learning park based on compressed sensing[J]. Electric Power Information and Communication Technology, 2024, 22 (5): 36- 42.
12	吴桐, 惠红勋, 张洪财. 商业建筑空调系统参与城市电网负荷调控综述[J]. 中国电力, 2023, 56 (7): 1- 11.
	WU Tong, HUI Hongxun, ZHANG Hongcai. Review of commercial air conditioners for participating in urban grid regulation[J]. Electric Power, 2023, 56 (7): 1- 11.
13	ALKHAYAT G, MEHMOOD R. A review and taxonomy of wind and solar energy forecasting methods based on deep learning[J]. Energy and AI, 2021, 4, 100060. DOI
14	郭艳飞, 程林, 李洪涛, 等. 基于支持向量机和互联网信息修正的空间负荷预测方法[J]. 中国电力, 2019, 52 (4): 80- 88. DOI
	GUO Yanfei, CHENG Lin, LI Hongtao, et al. Spatial load forecasting method based on support vector machine and internet information correction[J]. Electric Power, 2019, 52 (4): 80- 88. DOI
15	吴军英, 路欣, 刘宏, 等. 基于Spearman-GCN-GRU模型的超短期多区域电力负荷预测[J]. 中国电力, 2024, 57 (6): 131- 140.
	WU Junying, LU Xin, LIU Hong, et al. Ultra-short-term multi-region power load forecasting based on Spearman-GCN-GRU model[J]. Electric Power, 2024, 57 (6): 131- 140.
16	史佳琪, 马丽雅, 李晨晨, 等. 基于串行–并行集成学习的高峰负荷预测方法[J]. 中国电机工程学报, 2020, 40 (14): 4463- 4472. DOI
	SHI Jiaqi, MA Liya, LI Chenchen, et al. Daily peak load forecasting based on sequential-parallel ensemble learning[J]. Proceedings of the CSEE, 2020, 40 (14): 4463- 4472. DOI
17	吴潇雨, 和敬涵, 张沛, 等. 基于灰色投影改进随机森林算法的电力系统短期负荷预测[J]. 电力系统自动化, 2015, 39 (12): 50- 55.
	WU Xiaoyu, HE Jinghan, ZHANG Pei, et al. Power system short-term load forecasting based on improved random forest with grey relation projection[J]. Automation of Electric Power Systems, 2015, 39 (12): 50- 55.
18	李瑾, 刘金朋, 王建军. 采用支持向量机和模拟退火算法的中长期负荷预测方法[J]. 中国电机工程学报, 2011, 31 (16): 63- 66. DOI
	LI Jin, LIU Jinpeng, WANG Jianjun. Mid-long term load forecasting based on simulated annealing and SVM algorithm[J]. Proceedings of the CSEE, 2011, 31 (16): 63- 66. DOI
19	孔祥玉, 郑锋, 鄂志君, 等. 基于深度信念网络的短期负荷预测方法[J]. 电力系统自动化, 2018, 42 (5): 133- 139. DOI
	KONG Xiangyu, ZHENG Feng, E Zhijun, et al. Short-term load forecasting based on deep belief network[J]. Automation of Electric Power Systems, 2018, 42 (5): 133- 139. DOI
20	罗澍忻, 麻敏华, 蒋林, 等. 考虑多时间尺度数据的中长期负荷预测方法[J]. 中国电机工程学报, 2020, 40 (S1): 11- 19. DOI
	LUO Shuxin, MA Minhua, JIANG Lin, et al. Medium and long-term load forecasting method considering multi-time scale data[J]. Proceedings of the CSEE, 2020, 40 (S1): 11- 19. DOI
21	YIN Lifei, XIE Jiaxing. Multi-temporal-spatial-scale temporal convolution network for short-term load forecasting of power systems[J]. Applied Energy, 2021, 283, 116328. DOI
22	WANG J Q, DU Y, WANG J. LSTM based long-term energy consumption prediction with periodicity[J]. Energy, 2020, 197, 117197. DOI
23	魏晓川, 王新刚. 基于气象大数据的城市电力负荷预测[J]. 电测与仪表, 2021, 58 (2): 90- 95. DOI
	WEI Xiaochuan, WANG Xingang. Prediction of urban electricity load based on meteorological big data[J]. Electrical Measurement & Instrumentation, 2021, 58 (2): 90- 95. DOI
24	郭占伍, 张泽亚, 周兴华, 等. 考虑气象因素的电采暖负荷预测研究[J]. 电测与仪表, 2022, 59 (2): 154- 158. DOI
	GUO Zhanwu, ZHANG Zeya, ZHOU Xinghua, et al. Study on forecasting method of electric heating load considering meteorological factors[J]. Electrical Measurement & Instrumentation, 2022, 59 (2): 154- 158. DOI
25	王溪. 基于大数据分析方法的温控负荷可调度容量预测及应用[D]. 合肥工业大学, 2018.
	WANG Xi. Prediction and application of dispatchable capacity of temperature control load based on big data analysis method [D]. Hefei: Hefei University of Technology, 2018.
26	SCHNEIDER K P, SORTOMME E, VENKATA S S, et al. Evaluating the magnitude and duration of cold load pick-up on residential distribution using multi-state load models[J]. IEEE Transactions on Power Systems, 2015, 31 (5): 3765- 3774.
27	李香龙, 马龙飞, 赵向阳, 等. 基于LSTM网络的时间多尺度电采暖负荷预测[J]. 电力系统及其自动化学报, 2021, 33 (4): 71- 75. DOI
	LI Xianglong, MA Longfei, ZHAO Xiangyang, et al. Multi-time scale electric heating load forecasting based on long short-term memory network[J]. Proceedings of the CSU-EPSA, 2021, 33 (4): 71- 75. DOI
28	WANG X, XU Y, FU Z, et al. A dynamic interactive optimization model of CCHP system involving demand-side and supply-side impacts of climate change. Part Ⅰ: Methodology development[J]. Energy Conversion and Management, 2022, 252, 115112. DOI
29	胡可心, 李康平, 刘春阳, 等. 动态分时电价下居民用户需求响应基线负荷预测方法[J]. 电测与仪表, 2025, 62 (9): 62- 72. DOI
	HU Kexin, LI Kangping, Liu Chunyang, et al. A baseline load forecasting method for residential demand response under dynamic time-of-use electricity price[J]. Electrical Measurement & Instrumentation, 2025, 62 (9): 62- 72. DOI
30	GONG Z, TANG Y, LIANG J. Patchmixer: A patch-mixing architecture for long-term time series forecasting[J]. arXiv preprint: 2310.00655:2023.
31	ALI M, LI B, YING Z, et al. Day-ahead demand response potential prediction in residential buildings with HITSKAN: a fusion of kolmogorov-arnold networks and N-HiTS[J]. Energy and Buildings, 2025: 115455.

[1]	李科, 潘庭龙, 许德智. 基于MSCNN-BiGRU-Attention的短期电力负荷预测[J]. 中国电力, 2025, 58(6): 10-18.
[2]	姜通海, 王峰, 刘子琪, 单帅杰. 基于改进生成对抗网络的风光气象资源联合场景生成方法[J]. 中国电力, 2025, 58(3): 183-192.
[3]	武同心, 纪鑫, 杨成月, 陈屹婷, 杨智伟. 基于双层注意力机制的电力文本分类模型[J]. 中国电力, 2025, 58(11): 156-163.
[4]	杨鹏伟, 赵丽萍, 陈军法, 甄钊, 王飞, 李利明. 基于相似时段匹配与Transformer网络建模的分布式光伏超短期功率预测方法[J]. 中国电力, 2024, 57(12): 60-70.
[5]	刘东, 张越, 皮俊波, 单连飞, 刘赫, 宋鹏程, 姜涛. 面向电网调控信息智能检索的知识图谱构建及应用[J]. 中国电力, 2023, 56(7): 78-84.
[6]	张鑫, 叶俊杰, 崔瑶, 黄鑫, 仲林林. 基于语义信息距离解耦的变电运维多类别缺陷图像检测[J]. 中国电力, 2023, 56(6): 209-218.
[7]	周震震, 宋云海, 何宇浩, 王黎伟, 黄和燕, 何珏, 朱志航, 闫云凤. 基于分组查询注意力的可扩展电力人员行为分类方法[J]. 中国电力, 2023, 56(11): 77-85.
[8]	李丰君, 王磊, 赵健, 张建宾, 张世尧, 田杨阳. 基于天气融合和LSTM网络的分布式光伏短期功率预测方法[J]. 中国电力, 2022, 55(11): 149-154.
[9]	赵永良, 付鑫, 郭阳, 边迎迎, 王思宁. 基于深度学习和图像识别的电力配件智能出入库[J]. 中国电力, 2021, 54(3): 55-60.
[10]	马静怡, 崔昊杨, 张明达, 孙益辉, 许永鹏. 基于改进Faster RCNN的小尺度入侵目标识别及定位[J]. 中国电力, 2021, 54(3): 38-44.
[11]	周俊煌, 黄廷城, 谢小瑜, 范纹郡, 易婷婷, 张勇军. 视频图像智能识别技术在输变电系统中的应用研究综述[J]. 中国电力, 2021, 54(1): 124-134,166.
[12]	吴云亮, 张建新, 李豹, 李鹏, 李智勇, 周鑫, 杨燕, 赖晓文. 深度学习辅助约束辨识的电力市场快速出清方法[J]. 中国电力, 2020, 53(9): 90-97,207.
[13]	赵会茹, 赵一航, 郭森. 基于互补集合经验模态分解和长短期记忆神经网络的短期电力负荷预测[J]. 中国电力, 2020, 53(6): 48-55.
[14]	杨晓楠, 孙博, 郎燕生. 基于深度学习的特高压直流闭锁故障智能调度决策[J]. 中国电力, 2020, 53(6): 8-17.
[15]	周艳真, 查显煜, 兰健, 郭庆来, 孙宏斌, 薛峰, 王胜明. 基于数据增强和深度残差网络的电力系统暂态稳定预测[J]. 中国电力, 2020, 53(1): 22-31.

考虑温控型负荷特性影响的集群用户超短期负荷预测方法

Ultra-Short-Term Load Forecasting Method for Aggregated Users Considering the Impact of Temperature-Controlled Load Characteristics

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 15

编辑推荐

Metrics

模态框（Modal）标题

考虑温控型负荷特性影响的集群用户超短期负荷预测方法

Ultra-Short-Term Load Forecasting Method for Aggregated Users Considering the Impact of Temperature-Controlled Load Characteristics

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 15

编辑推荐

Metrics