基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测

doi:10.11930/j.issn.1004-9649.202306104

中国电力 ›› 2024, Vol. 57 ›› Issue (6): 121-130.DOI: 10.11930/j.issn.1004-9649.202306104

基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测

王宇飞¹(), 杜桐¹(), 边伟国¹(), 张钊²(), 刘慧婷²(), 杨丽君²()

1. 国网冀北电力有限公司张家口供电公司，河北张家口　075000
2. 河北省电力电子节能与传动控制重点实验室（燕山大学电气工程学院），河北秦皇岛　066004

收稿日期:2023-06-28 接受日期:2024-01-16 出版日期:2024-06-28 发布日期:2024-06-25
作者简介:王宇飞（1991—），男，工程师，从事配电网规划与建设研究，E-mail：15133330416@139.com
杜桐（1993—），男，经济师，从事电力经济分析研究，E-mail：18332550961@163.com
边伟国（1990—），男，工程师，从事电力线损管理研究，E-mail：weiguo1382183@sohu.com
张钊（1992—），男，博士研究生，从事机器学习在电力系统中的应用研究， E-mail：1311017915@qq.com
刘慧婷（1997—），女，硕士研究生，从事电力系统负荷预测研究，E-mail：1072398868@qq.com
杨丽君（1972—），女，教授，通信作者，从事电力及综合能源系统恢复控制及优化调度研究，E-mail：195320008@qq.com
基金资助:
国网冀北张家口供电公司2022年群众性创新项目（B30107220006）；河北省自然科学基金资助项目（E2021203004）。

Short-term Load Forecasting Based on DTW K-medoids and VMD Multi-branch Neural Network for Multiple Users

Yufei WANG¹(), Tong DU¹(), Weiguo BIAN¹(), Zhao ZHANG²(), Huiting LIU²(), Lijun YANG²()

1. State Grid Jibei Electric Power Co., Ltd. Zhangjiakou Power Supply Company, Zhangjiakou 075000, China
2. Key Laboratory of Power Electronics for Energy Conservation and Drive Control of Heibei Province (School of Electrical Engineering, Yanshan University), Qinhuangdao 066004, China

Received:2023-06-28 Accepted:2024-01-16 Online:2024-06-28 Published:2024-06-25
Supported by:
This work is supported by 2022 Mass Innovation Project of State Grid Zhangjiakou Power Supply Company (No.B30107220006) and Natural Science Foundation of Hebei Province (No.E2021203004).

摘要/Abstract

摘要：

多用户电力负荷预测是指根据历史负荷数据对多个用户或区域的电力负荷进行预测，可使电网企业掌握不同用户或区域的电力需求，以便更好地开展规划和实施调度优化等。然而由于各用户呈现出复杂多样的用电行为，采用传统方法难以进行统一建模并实现快速准确预测。为此，构建了一种基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测模型。首先，采用DTW K-medoids法进行用户负荷数据聚类，利用动态时间弯曲（dynamic time warping，DTW）计算数据间的距离，取代K-medoids算法中传统的欧氏距离度量方式，以改善多用户负荷聚类的效果；在此基础上，为充分表征负荷历史数据的长短期时序依赖特征，建立了一种基于变分模态分解（variational mode decomposition，VMD）-多分支神经网络模型的并行预测方法，用于多用户短期负荷预测；最后，使用某地区20个用户365天的负荷数据进行聚类、训练和测试实验，结果显示该模型结果的平均绝对误差和均方根误差等指标均较对比模型有较大幅度降低，表明该方法可有效表征多类用户的用电行为，提升多用户负荷预测效率和精度。

关键词: 多用户, 负荷预测, DTW K-medoids聚类, 变分模态分解（VMD）, 多分支神经网络

Abstract:

Multi-user power load forecasting refers to the power load forecasting of multiple users or regions based on historical loads data, which can make the grid companies understand the power demands of different users or regions, so as to better carry out the planning and scheduling optimization of the power system. However, different users have complex and diverse power consumption behaviors, so it is difficult to use traditional methods to universally model different power users' loads and achieve accurate prediction. Therefore, a new multi-user short-term load prediction model based on DTW K-medoids and VMD-multi-branch neural network is established. Firstly, in order to improve the clustering performance of traditional clustering methods, the DTW K-medoids method is used to cluster users' load data, and the distance between loads data is calculated using the dynamic time warping (DTW) instead of the traditional Euclidean distance measurement method in K-medoids to improve the clustering effects of multiple users' load. Secondly, in order to fully characterize the long short-term time series-dependent characteristics of load history data, a parallel load forecasting method based on VMD-multi-branch neutral network model is established for multi-user short-term load forecasting. Finally, the 365-day load data of 20 users in a region is used for clustering, training and experiment, and the results show that the MAE and RMSE indexes of the proposed model significantly decrease compared with that of the comparative models, indicating that the proposed method can effectively characterize the power consumption behaviors of multiple users and improve the prediction efficiency and accuracy of multi-user loads.

Key words: multi-user, load forecasting, DTW K-medoids clustering, variational mode decomposition, multi-branch neural network

王宇飞, 杜桐, 边伟国, 张钊, 刘慧婷, 杨丽君. 基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测[J]. 中国电力, 2024, 57(6): 121-130.

Yufei WANG, Tong DU, Weiguo BIAN, Zhao ZHANG, Huiting LIU, Lijun YANG. Short-term Load Forecasting Based on DTW K-medoids and VMD Multi-branch Neural Network for Multiple Users[J]. Electric Power, 2024, 57(6): 121-130.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202306104

https://www.electricpower.com.cn/CN/Y2024/V57/I6/121

图/表 13

图 1 预测流程

Fig.1 Prediction flowchart

图 2 三维特征图重构模块

Fig.2 The 3 D feature map reconstruction module

图 3 GCNN结构

Fig.3 Structure of GCNN

图 4 多分支神经网络

Fig.4 The multi-branch neural network

表 1 预测模型参数

Table 1 Parameters for prediction models

模型		参数
VMD		分解模态数=3，初始中心频率=1
Kemiod		聚类数4
GCNN		5层
GRU		3层

图 5 肘部法则确定聚类数

Fig.5 Number of clusters determined by elbow method

图 6 聚类中心可视化

Fig.6 Cluster center visualization

表 2 不同聚类算法对预测模型的影响

Table 2 Influence of different clustering algorithms on prediction models

模型	MAE/MW	RMSE/MW	PE/%
K-means	12.82	18.77	25.99
K-medoids	12.38	18.40	25.74
DTW K-means	12.23	18.31	25.49
DTW K-medoids	11.19	17.98	25.10

表 3 VMD分解不同模态数对预测模型的影响

Table 3 Influence of different mode number of VMD on prediction models

类别	模态数IMF	MAE/MW	RMSE/MW	PE/%
1	3	15.86	19.71	26.67
	5	17.26	21.77	27.56
	7	19.53	25.39	28.50
2	3	15.05	25.36	25.93
	5	15.96	26.59	26.37
	7	17.30	27.16	26.71
3	3	7.39	10.92	24.20
	5	8.33	12.06	24.91
	7	9.52	13.07	25.03
4	3	8.51	12.29	24.83
	5	9.74	14.04	24.93
	7	11.13	15.35	25.45

图 7 VMD分解可视化

Fig.7 VMD visualization

图 8 不同类别预测结果

Fig.8 Prediction results of different categories of data

表 4 不同类别数据的预测误差比较

Table 4 Comparison of prediction errors of different categories of data

负荷类别	MAE/MW	RMSE/MW	PE/%
1	15.86	19.71	28.02
2	15.05	25.36	26.61
3	7.39	10.92	24.10
4	8.51	12.29	24.42

表 5 不同的预测模型误差对比

Table 5 Error comparison of different prediction models

模型	MAE/MW	RMSE/MW	PE/%
LSTM	17.41	25.05	28.83
CNN	19.96	28.16	30.02
GCNN	16.92	22.31	26.48
LSTM-GCNN	15.17	18.42	25.41
本文	11.19	17.98	25.10

参考文献 26

1	孔祥玉, 马玉莹, 艾芊, 等. 新型电力系统多元用户的用电特征建模与用电负荷预测综述[J]. 电力系统自动化, 2023, 47 (13): 2- 17. DOI
	KONG Xiangyu, MA Yuying, AI Qian, et al. Review on electricity consumption characteristic modeling and load forecasting for diverse users in new power system[J]. Automation of Electric Power Systems, 2023, 47 (13): 2- 17. DOI
2	白星振, 赵康, 葛磊蛟, 等. 基于EWT-GRU-RR的配电网短期电力负荷预测模型[J]. 山东科技大学学报(自然科学版), 2023, 42 (5): 77- 87.
	BAI Xingzhen, ZHAO Kang, GE Leijiao, et al. Short-term power load forecasting model based on EWT-GRU-RR[J]. Journal of Shandong University of Science and Technology (Natural Science), 2023, 42 (5): 77- 87.
3	董雷, 陈振平, 韩富佳, 等. 基于图卷积神经网络与K-means聚类的居民用户集群短期负荷预测[J]. 电网技术, 2023, 47 (10): 4291- 4301.
	DONG Lei, CHEN Zhenping, HAN Fujia, et al. Short-term load forecasting of residential user groups based on graph convolutional neural network and K-means clustering[J]. Power System Technology, 2023, 47 (10): 4291- 4301.
4	朱子意, 孙晓燕, 柳先彪, 等. 基于相似用电单元及图卷积神经网络的电力负荷预测[J]. 电力科学与工程, 2023, 39 (7): 9- 23. DOI
	ZHU Ziyi, SUN Xiaoyan, LIU Xianbiao, et al. Power load forecasting based on similar power units and graph convolution neural network[J]. Electric Power Science and Engineering, 2023, 39 (7): 9- 23. DOI
5	梁京章, 黄星舒, 吴丽娟, 等. 基于KPCA和改进K-means的电力负荷曲线聚类方法[J]. 华南理工大学学报(自然科学版), 2020, 48 (6): 143- 150.
	LIANG Jingzhang, HUANG Xingshu, WU Lijuan, et al. Clustering method of power load profiles based on KPCA and improved K-means[J]. Journal of South China University of Technology (Natural Science Edition), 2020, 48 (6): 143- 150.
6	李康宇, 吴青娥, 刘磊, 等. 聚类算法在用户电力负荷分类中的应用[J]. 智能电网, 2018, 8 (2): 189- 203. DOI
	LI Kangyu, WU Qing’e, LIU Lei, et al. Application of clustering algorithm in user power load classification[J]. Smart Grid, 2018, 8 (2): 189- 203. DOI
7	荀超, 陈伯建, 吴翔宇, 等. 基于改进K-means算法的电力短期负荷预测方法研究[J]. 电力科学与技术学报, 2022, 37 (1): 90- 95.
	XUN Chao, CHEN Bojian, WU Xiangyu, et al. Research on short-term power load forecasting method based on improved K-means algorithm[J]. Journal of Electric Power Science and Technology, 2022, 37 (1): 90- 95.
8	KOIVISTO M, HEINE P, MELLIN I, et al. Clustering of connection points and load modeling in distribution systems[J]. IEEE Transactions on Power Systems, 2013, 28 (2): 1255- 1265. DOI
9	SYED D, ABU-RUB H, GHRAYEB A, et al. Deep learning-based short-term load forecasting approach in smart grid with clustering and consumption pattern recognition[J]. IEEE Access, 2021, 9, 54992- 55008. DOI
10	宗文泽, 吴永明, 徐计, 等. 基于DTW-kmedoids算法的时间序列数据异常检测[J]. 组合机床与自动化加工技术, 2022, (5): 120- 124, 128.
	ZONG Wenze, WU Yongming, XU Ji, et al. Anomaly detection of time series data based on DTW-kmedois clustering algorithm[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2022, (5): 120- 124, 128.
11	丛玮, 谢道仪, 严仪炅. 基于DTW的K-Medoids扇区交通特征聚类研究[J]. 交通技术, 2021, 10 (1): 59- 69. DOI
	CONG Wei, XIE Daoyi, YAN Yijiong. Clustering of sector traffic characteristics based on DTW K-Medoids[J]. Open Journal of Transportation Technologies, 2021, 10 (1): 59- 69. DOI
12	SEMERO Y K, ZHANG J H, ZHENG D H. EMD–PSO–ANFIS-based hybrid approach for short-term load forecasting in microgrids[J]. IET Generation, Transmission & Distribution, 2020, 14 (3): 470- 475.
13	雷炳银, 王子驰, 苏雨晴, 等. 基于EEMD-CS-LSSVM的短期负荷预测方法研究[J]. 电力系统及其自动化学报, 2021, 33 (9): 117- 122.
	LEI Bingyin, WANG Zichi, SU Yuqing, et al. Research on short-term load forecasting method based on EEMD-CS-LSSVM[J]. Proceedings of the CSU-EPSA, 2021, 33 (9): 117- 122.
14	盛四清, 金航, 刘长荣. 基于VMD-WSGRU的风电场发电功率中短期及短期预测[J]. 电网技术, 2022, 46 (3): 897- 904.
	SHENG Siqing, JIN Hang, LIU Changrong. Short-term and mid-short-term wind power forecasting based on VMD-WSGRU[J]. Power System Technology, 2022, 46 (3): 897- 904.
15	李文武, 石强, 李丹, 等. 基于VMD和PSO-SVR的短期电力负荷多阶段优化预测[J]. 中国电力, 2022, 55 (8): 171- 177.
	LI Wenwu, SHI Qiang, LI Dan, et al. Multi-stage optimization forecast of short-term power load based on VMD and PSO-SVR[J]. Electric Power, 2022, 55 (8): 171- 177.
16	叶剑华, 曹旌, 杨理, 等. 基于变分模态分解和多模型融合的用户级综合能源系统超短期负荷预测[J]. 电网技术, 2022, 46 (7): 2610- 2622.
	YE Jianhua, CAO Jing, YANG Li, et al. Ultra short-term load forecasting of user level integrated energy system based on variational mode decomposition and multi-model fusion[J]. Power System Technology, 2022, 46 (7): 2610- 2622.
17	GENG G C, HE Y, ZHANG J, et al. Short-term power load forecasting based on PSO-optimized VMD-TCN-attention mechanism[J]. Energies, 2023, 16 (12): 4616. DOI
18	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.
19	吕志星, 张虓, 王沈征, 等. 基于K-Means和CNN的用户短期电力负荷预测[J]. 计算机系统应用, 2020, 29 (3): 161- 166.
	LYU Zhixing, ZHANG Xiao, WANG Shenzheng, et al. Hybrid method for short-term load forecasting based on K-means and convolutional neural network[J]. Computer Systems & Applications, 2020, 29 (3): 161- 166.
20	季玉琦, 严亚帮, 和萍, 等. 基于K-Medoids聚类与栅格法提取负荷曲线特征的CNN-LSTM短期负荷预测[J]. 电力系统保护与控制, 2023, 51 (18): 81- 93.
	JI Yuqi, YAN Yabang, HE Ping, et al. CNN-LSTM short-term load forecasting based on the K-Medoids clustering and grid method to extract load curve features[J]. Power System Protection and Control, 2023, 51 (18): 81- 93.
21	姚程文, 杨苹, 刘泽健. 基于CNN-GRU混合神经网络的负荷预测方法[J]. 电网技术, 2020, 44 (9): 3416- 3424.
	YAO Chengwen, YANG Ping, LIU Zejian. Load forecasting method based on CNN-GRU hybrid neural network[J]. Power System Technology, 2020, 44 (9): 3416- 3424.
22	姚芳, 汤俊豪, 陈盛华, 等. 基于ISSA-CNN-GRU模型的电动汽车充电负荷预测方法[J]. 电力系统保护与控制, 2023, 51 (16): 158- 167.
	YAO Fang, TANG Junhao, CHEN Shenghua, et al. Charging load prediction method for electric vehicles based on an ISSA-CNN-GRU model[J]. Power System Protection and Control, 2023, 51 (16): 158- 167.
23	李坤奇, 孟润泉, 李凤莲. 融合聚类-卷积-门循环的居民用电短期负荷预测方法[J]. 太原理工大学学报, 2023, 54 (1): 203- 210.
	LI Kunqi, MENG Runquan, LI Fenglian. Short-term load forecasting method based on BIRCH-CNN-GRU model[J]. Journal of Taiyuan University of Technology, 2023, 54 (1): 203- 210.
24	YANAGIMTO H, HASHIMOTO K, OKADA M. Attention visualization of gated convolutional neural networks with self attention in sentiment analysis[C]//2018 International Conference on Machine Learning and Data Engineering (iCMLDE). Sydney, NSW, Australia. IEEE, 2018: 77–82.
25	HEINRICH K, ZSCHECH P, JANIESCH C, et al. Process data properties matter: introducing gated convolutional neural networks (GCNN) and key-value-predict attention networks (KVP) for next event prediction with deep learning[J]. Decision Support Systems, 2021, 143, 113494. DOI
26	樊江川, 于昊正, 刘慧婷, 等. 基于多分支门控残差卷积神经网络的短期电力负荷预测[J]. 中国电力, 2022, 55 (11): 155- 162, 174.
	FAN Jiangchuan, YU Haozheng, LIU Huiting, et al. Short-term load forecasting based on multi-branch residual gated convolution neural network[J]. Electric Power, 2022, 55 (11): 155- 162, 174.

[1]	孙庆超, 李嘉靓, 江万里, 王若愚, 李植鹏, 胡亚荣, 朱健斌. 基于数据驱动时空网络的城市中长期电力负荷预测[J]. 中国电力, 2025, 58(3): 168-174.
[2]	吴军英, 路欣, 刘宏, 张彬, 柴守亮, 刘蕴春, 王佳楠. 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.
[3]	陈中飞, 赵越, 蔡秋娜, 张乔榆, 王泽林, 戴晓娟, 陈雨果. 基于净负荷预测误差统计的电力系统爬坡能力充裕度评估[J]. 中国电力, 2024, 57(5): 50-60.
[4]	刘航, 申皓, 杨勇, 纪陵, 余洋. 基于高阶马尔可夫链的纯电重卡集群负荷预测[J]. 中国电力, 2024, 57(5): 61-69.
[5]	李丹, 贺帅, 颜伟, 胡越, 方泽仁, 梁云嫣. 考虑动态时间锚点和典型特征约束的年日均负荷曲线预测[J]. 中国电力, 2024, 57(11): 36-47.
[6]	周颖, 白雪峰, 王阳, 邱敏, 孙冲, 武亚杰, 李彬. 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.
[7]	唐旭辰, 潮铸, 段秦尉, 苏炳洪, 陈卉灿. 基于分层测量数据的高压变电站概率负荷预测方法[J]. 中国电力, 2023, 56(8): 143-150.
[8]	贾巍, 黄裕春. 基于小样本数据差分扩容的微电网负荷预测方法[J]. 中国电力, 2023, 56(8): 151-156,165.
[9]	彭生江, 杨德州, 孙传帅, 袁铁江, 刘永成. 基于氢负荷需求的氢能系统容量规划[J]. 中国电力, 2023, 56(7): 13-20,32.
[10]	杨喜行, 黄纯, 申亚涛, 胡念恩, 万子恒. 基于损失功率匹配的配电线路故障定位方法[J]. 中国电力, 2022, 55(8): 113-120.
[11]	李文武, 石强, 李丹, 胡群勇, 唐芸, 梅锦超. 基于VMD和PSO-SVR的短期电力负荷多阶段优化预测[J]. 中国电力, 2022, 55(8): 171-177.
[12]	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.
[13]	李丹, 张远航, 李黄强, 童华敏, 王凌云. 基于精准气象数据的地区电网短期负荷智能预测系统设计[J]. 中国电力, 2022, 55(7): 128-133.
[14]	郑豪丰, 杨国华, 康文军, 刘志远, 刘世涛, 伍弘, 张鸿皓. 基于多负荷特征和TCN-GRU神经网络的负荷预测[J]. 中国电力, 2022, 55(11): 142-148.
[15]	樊江川, 于昊正, 刘慧婷, 杨丽君, 安佳坤. 基于多分支门控残差卷积神经网络的短期电力负荷预测[J]. 中国电力, 2022, 55(11): 155-162,174.

基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测

Short-term Load Forecasting Based on DTW K-medoids and VMD Multi-branch Neural Network for Multiple Users

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 26

相关文章 15

编辑推荐

Metrics

模态框（Modal）标题

基于DTW K-medoids与VMD-多分支神经网络的多用户短期负荷预测

Short-term Load Forecasting Based on DTW K-medoids and VMD Multi-branch Neural Network for Multiple Users

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 26

相关文章 15

编辑推荐

Metrics