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

doi:10.11930/j.issn.1004-9649.202403112

中国电力 ›› 2024, Vol. 57 ›› Issue (12): 60-70.DOI: 10.11930/j.issn.1004-9649.202403112

• 面向新型电力系统的源荷预测技术 • 上一篇下一篇

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

杨鹏伟¹(), 赵丽萍¹, 陈军法², 甄钊³(), 王飞³^,⁴^,⁵, 李利明⁶

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. 北京送变电有限公司，北京　102401
3. 华北电力大学电力工程系，河北保定　071003
4. 新能源电力系统全国重点实验室（华北电力大学），北京　102206
5. 河北省分布式储能与微网重点实验室（华北电力大学），河北保定　071003
6. 北京清电科技有限公司，北京　100190

收稿日期:2024-03-27 出版日期:2024-12-28 发布日期:2024-12-27
作者简介:杨鹏伟（1984—），硕士，高级工程师，从事配网生产研究，E-mail：ypw0508@gmail.com
甄钊（1989—），通信作者，博士（后），讲师，从事光伏发电功率预测、风电功率预测、电价与电量预测研究，E-mail：zhenzhao@ncepu.edu.cn
基金资助:
国网冀北电力有限公司张家口供电公司科技项目（新型有源配电网分布式光伏发电集群功率预测关键技术研究与示范， SGJBZJ00PWJS2310934）；国家自然科学基金资助项目（52007092）。

Distributed Photovoltaic Ultra-short-term Power Forecasting Method Based on Temporal Analog Matching Approach and Transformer Network Modeling

Pengwei YANG¹(), Liping ZHAO¹, Junfa CHEN², Zhao ZHEN³(), Fei WANG³^,⁴^,⁵, Liming LI⁶

1. Zhangjiakou Power Supply Company, State Grid Jibei Electric Power Co., Ltd., Zhangjiakou 075000, China
2. Beijing Power Transmission and Distribution Co., Ltd., Beijing 102401, China
3. Department of Power Engineering, North China Electric Power University, Baoding 071003, China
4. State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources (North China Electric Power University), Beijing 102206, China
5. Hebei Key Laboratory of Distributed Energy Storage and Microgrid (North China Electric Power University), Baoding 071003, China
6. Beijing Tsingdian Technology Co., Ltd., Beijing 100190, China

Received:2024-03-27 Online:2024-12-28 Published:2024-12-27
Supported by:
This work is supported by Science &Technology Project of State Grid Jibei Electric Power Co., Ltd. Zhangjiakou Power Supply Company (Research and Demonstration of Key Technologies for Power Forecasting of Distributed Photovoltaic Power Generation Cluster in New Active Distribution Network, No.SGJBZJ00PWJS2310934), National Natural Science Foundation of China (No.52007092).

摘要/Abstract

摘要：

由于缺乏气象数据，分布式光伏在天气骤变场景下预测精度不高，提出了一种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。首先，将相似时段概念由日扩展至更灵活的时间段，并提出了一种历史功率与卫星遥感信息融合的匹配策略，旨在无须依赖气象数据的情况下，高效识别出对预测最为关键的相似功率时段。在此基础上，融合Transformer网络的强大时序建模能力，动态解析多源相似时段中的隐藏关联，深入挖掘功率关键特征信息，从而为天气骤变条件下的分布式光伏系统提供更为精确的超短期功率预测。最后，通过实际分布式光伏功率数据验证了所提方法的有效性。

关键词: 分布式光伏, 相似时段, Transformer模型, 超短期功率预测, 卫星遥感信息

Abstract:

To address the challenge of low prediction accuracy of distributed photovoltaic (PV) power generation under sudden weather change scenarios due to the lack of meteorological data, this paper proposes a distributed PV ultra-short-term power prediction method based on temporal analog matching approach (TAMA) and Transformer network modeling. Firstly, the concept of similar time periods is extended from days to more flexible time periods, and a matching strategy integrating historical power and satellite remote sensing information is proposed to efficiently identify the most critical time periods of similar power for prediction without relying on meteorological data. Based on this, the powerful temporal modeling capability of the Transformer network is used to dynamically resolve the hidden correlations in multi-source similar time periods, and deeply mine the key features of power, thus providing more accurate ultra-short-term power prediction for distributed PV systems under sudden weather change conditions. Finally, the effectiveness of the proposed method is verified through actual distributed PV power generation data.

Key words: distributed PV power, similar time periods, transformer model, ultra-short-term power forecasting, satellite remote sensing information

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.

Pengwei YANG, Liping ZHAO, Junfa CHEN, Zhao ZHEN, Fei WANG, Liming LI. Distributed Photovoltaic Ultra-short-term Power Forecasting Method Based on Temporal Analog Matching Approach and Transformer Network Modeling[J]. Electric Power, 2024, 57(12): 60-70.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202403112

https://www.electricpower.com.cn/CN/Y2024/V57/I12/60

图/表 17

图 1 研究方法整体流程

Fig.1 Flowchart of proposed methodology

图 2 相似时段匹配方法流程

Fig.2 Flowchart of TAMA

图 3 单个batch下的特征工程数据流程

Fig.3 Data flow process for a single batch

图 4 功率预测模型结构

Fig.4 Transformer based forecasting model structure

图 5 模型编码器结构

Fig.5 Structure of model encoder

图 6 模型解码器结构

Fig.6 Structure of model decoder

图 7 各时段发电功率K-Means聚类结果

Fig.7 K-Means clustering results for power generation in different time intervals

图 8 相似度评价过程

Fig.8 Process of similarity evaluation

图 9 匹配时段的卫星遥感数据MSSIM和待预测时段的功率皮尔逊相关系数

Fig.9 Satellite remote sensing data MSSIM during matching time period and Pearson correlation coefficient during predicted time period

图 10 同时段的云遮盖率和发电功率对比

Fig.10 Comparison of cloud cover and power generation during the same time intervals

图 11 不同滑动窗口长度下的RMSE预测误差

Fig.11 RMSE error under different sliding window length

表 1 预测模型的RMSE误差对比

Table 1 Comparison of RMSE errors of different forecasting models

方法	多数为晴天	多数为云蓬	多数为云密
LSTM	1.77	1.85	1.89
Transformer	1.66	1.67	1.71
TAMA+LSTM	1.64	1.76	1.78
TAMA +Transformer	1.53	1.58	1.68

表 2 预测模型的MAE误差对比

Table 2 Comparison of MAE errors of different forecasting models

方法	多数为晴天	多数为云蓬	多数为云密
LSTM	1.41	1.48	1.52
Transformer	1.38	1.35	1.48
TAMA+LSTM	1.36	1.41	1.46
TAMA +Transformer	1.26	1.31	1.37

图 12 天气骤变情况下预测结果（大部分为晴天）

Fig.12 Forecasting results under rapid weather changes (mostly clear skies )

图 13 天气骤变情况下预测结果（多数时间云蓬）

Fig.13 Forecasting results under rapid weather changes (mostly clear skies with few cloud)

图 14 天气骤变情况下预测结果（多数时间云密）

Fig.14 Forecasting results under rapid weather changes (mostly clouded weather)

图 15 不同方法预测精度

Fig.15 Forecasting accuracy for different methods

参考文献 23

1	国家能源局. 2023年上半年光伏发电建设运行情况[EB/OL]. (2023-07-27) [2023-12-15]. http://www.nea.gov.cn/2023-07/27/c_1310734298.htm.
2	高岩, 吴汉斌, 张纪欣, 等. 基于组合深度学习的光伏功率日前概率预测模型[J]. 中国电力, 2024, 57 (4): 100- 110.
	GAO Yan, WU Hanbin, ZHANG Jixin, et al. Probabilistic forecasting model for day-ahead power generation of centralized photovoltaic stations based on combine deep learning[J]. Electric Power, 2024, 57 (4): 100- 110.
3	顼佳宇, 李伟, 宋宝同, 等. 分布式光伏接入配电网适应性评估 [J/OL]. 电测与仪表, 1–8[2023-11-13] http://kns.cnki.net/kcms/detail/23.1202.TH.20220602.1656.006.html.
	XIANG Jiayu, LI Wei, SONG Baotong, et al. Adaptability evaluation of distributed photovoltaic access distribution network[J/OL]. Electrical Measurement & Instrumentation, 1–8[2023-11-13] http://kns.cnki.net/kcms/detail/23.1202.TH.20220602.1656.006.html.
4	阮呈隆, 李康平, 李正辉, 等. 分布式光伏集群功率短期预测的空间互补特性初探[J]. 电力系统自动化, 2024, 48 (3): 42- 50.
	RUAN Chenglong, LI Kangping, LI Zhenghui, et al. Preliminary study on spatial complementarity characteristics of short-term power forecasting for distributed photovoltaic clusters[J]. Automation of Electric Power Systems, 2024, 48 (3): 42- 50.
5	LI K P, YAN J C, HU L, et al. Two-stage decoupled estimation approach of aggregated baseline load under high penetration of behind-the-meter PV system[J]. IEEE Transactions on Smart Grid, 2021, 12 (6): 4876- 4885. DOI
6	WANG F, GE X X, DONG Z B, et al. Joint energy disaggregation of behind-the-meter PV and battery storage: a contextually supervised source separation approach[J]. IEEE Transactions on Industry Applications, 2022, 58 (2): 1490- 1501. DOI
7	MAYER M J, GRÓF G. Extensive comparison of physical models for photovoltaic power forecasting[J]. Applied Energy, 2021, 283, 116239. DOI
8	黄莉, 甘恒玉, 刘兴举, 等. 基于Transformer编码器的超短期光伏发电功率预测[J]. 智慧电力, 2024, 52 (5): 16- 22, 59.
	HUANG Li, GAN Hengyu, LIU Xingju, et al. Ultra-short-term photovoltaic power generation prediction based on transformer encoder[J]. Smart Power, 2024, 52 (5): 16- 22, 59.
9	张海涛, 李文娟, 李雪峰, 等. 基于变分模态分解和时间注意力机制TCN网络的光伏发电功率预测[J/OL]. 电测与仪表: 1–8 [2024-05-25]. http://kns.cnki.net/kcms/detail/23.1202.th.20240510.1542.003.html.
	ZHANG Haitao, LI Wenjuan, LI Xuefeng, et al. Photovoltaic power forecasting based on TPA-TCN model and variational modal decomposition[J/OL]. Electrical Measurement & Instrumentation: 1–8 [2024-05-25]. http://kns.cnki.net/kcms/detail/23.1202.th.20240510.1542.003.html.
10	张家安, 王琨玥, 陈建, 等. 基于空间相关性的分布式光伏出力预测[J]. 电力建设, 2020, 41 (3): 47- 53.
	ZHANG Jiaan, WANG Kunyue, CHEN Jian, et al. Research on prediction of distributed photovoltaic output considering spatial relevance[J]. Electric Power Construction, 2020, 41 (3): 47- 53.
11	王彪, 吕洋, 陈中, 等. 考虑信息时移的分布式光伏机理-数据混合驱动短期功率预测[J]. 电力系统自动化, 2022, 46 (11): 67- 74.
	WANG Biao, LYU Yang, CHEN Zhong, et al. Hybrid mechanism-data-driven short-term power forecasting of distributed photovoltaic considering information time shift[J]. Automation of Electric Power Systems, 2022, 46 (11): 67- 74.
12	ZHANG M, ZHEN Z, LIU N, et al. Optimal graph structure based short-term solar PV power forecasting method considering surrounding spatio-temporal correlations[J]. IEEE Transactions on Industry Applications, 2023, 59 (1): 345- 357. DOI
13	王玉庆, 徐飞, 刘志坚, 等. 基于动态关联表征与图网络建模的分布式光伏超短期功率预测[J]. 电力系统自动化, 2023, 47 (20): 72- 82.
	WANG Yuqing, XU Fei, LIU Zhijian, et al. Ultra-short-term power forecasting of distributed photovoltaic based on dynamic correlation characterization and graph network modeling[J]. Automation of Electric Power Systems, 2023, 47 (20): 72- 82.
14	赵洪山, 孙承妍, 温开云, 等. 无气象信息条件下基于AGCRN的分布式光伏出力超短期预测方法[J]. 高电压技术, 2024, 50 (1): 65- 73.
	ZHAO Hongshan, SUN Chengyan, WEN Kaiyun, et al. Ultra-short-term prediction of distributed photovoltaic power method based on AGCRN in the absence of meteorological information[J]. High Voltage Engineering, 2024, 50 (1): 65- 73.
15	WANG Y Q, FU W J, ZHANG X D, et al. Dynamic directed graph convolution network based ultra-short-term forecasting method of distributed photovoltaic power to enhance the resilience and flexibility of distribution network[J]. IET Generation, Transmission & Distribution, 2024, 18 (2): 337- 352.
16	臧海祥, 程礼临, 刘玲, 等. 基于数据驱动的太阳辐射估计和预测研究与展望[J]. 电力系统自动化, 2021, 45 (11): 170- 183.
	ZANG Haixiang, CHENG Lilin, LIU Ling, et al. Research and prospect for data-driven estimation and prediction of solar radiation[J]. Automation of Electric Power Systems, 2021, 45 (11): 170- 183.
17	王涛, 王旭, 许野, 等. 计及相似日的LSTM光伏出力预测模型研究[J]. 太阳能学报, 2023, 44 (8): 316- 323.
	WANG Tao, WANG Xu, XU Ye, et al. Study on LSTM photovoltaic output prediction model considering similar days[J]. Acta Energiae Solaris Sinica, 2023, 44 (8): 316- 323.
18	李超然, 潘鹏程, 杨伟荣, 等. 基于改进相似日优化HBA-BiLSTM-KELM的光伏发电功率预测[J]. 太阳能学报, 2024, 45 (5): 508- 516.
	LI Chaoran, PAN Pengcheng, YANG Weirong, et al. Research on PV system power forecasting based on improved similar day and HBA-BiLSTM-KELM neural network[J]. Acta Energiae Solaris Sinica, 2024, 45 (5): 508- 516.
19	LIU X J, LIU Y Y, KONG X B, et al. Deep neural network for forecasting of photovoltaic power based on wavelet packet decomposition with similar day analysis[J]. Energy, 2023, 271, 126963. DOI
20	张青山, 王丽婕, 郝颖, 等. 基于卫星云图和晴空模型的分布式光伏电站太阳辐照度超短期预测[J]. 高电压技术, 2022, 48 (8): 3271- 3281.
	ZHANG Qingshan, WANG Lijie, HAO Ying, et al. Ultra-short-term solar irradiance forecasting of distributed photovoltaic power stations based on satellite cloud images and clear sky model[J]. High Voltage Engineering, 2022, 48 (8): 3271- 3281.
21	YAN J C, HU L, ZHEN Z, et al. Frequency-domain decomposition and deep learning based solar PV power ultra-short-term forecasting model[J]. IEEE Transactions on Industry Applications, 2021, 57 (4): 3282- 3295. DOI
22	HE K M, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]//2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy. IEEE, 2017: 2980–2988.
23	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//NIPS 2017, CA, USA, 2017.

[1]	孙通, 张沈习, 曹毅, 曹佳晨, 程浩忠. 计及5G基站可调特性的配电网分布式光伏准入容量鲁棒优化[J]. 中国电力, 2025, 58(2): 140-146.
[2]	周才期, 刘静利, 孙鹏凯, 张玉敏. 分布式光伏波动事件多级区间滚动预警方法[J]. 中国电力, 2024, 57(8): 96-107.
[3]	王海军, 居蓉蓉, 董颖华. 基于时空关联特征与B-LSTM模型的分布式光伏功率区间预测[J]. 中国电力, 2024, 57(7): 74-80.
[4]	刘韵艺, 汤渊, 苏盛, 吴裕宙, 王晓倩. 基于STL-Bayesian时空模型的分布式光伏系统异常检测[J]. 中国电力, 2024, 57(5): 222-231.
[5]	张宇, 魏新迟, 冯凯辉, 时珊珊, 孟子涵, 梁媛. 计及分布式光伏承载力的配电网侧独立储能充放电策略[J]. 中国电力, 2024, 57(4): 111-117.
[6]	陈璨, 苏紫诺, 马原, 刘佳林, 王玉庆, 王飞. 基于分层关联性建模的分布式光伏功率超短期概率预测[J]. 中国电力, 2024, 57(12): 50-59.
[7]	陈奇芳, 李若凡, 夏明超, 吴迪, 邵尹池. 计及多维性能评估的新型配电网光伏选址定容方法[J]. 中国电力, 2024, 57(10): 172-178, 207.
[8]	吴晓刚, 阎洁, 葛畅, 唐雅洁, 倪筹帷, 季青锋. 基于改进GRU-CNN的风光水一体化超短期功率预测方法[J]. 中国电力, 2023, 56(9): 178-186,205.
[9]	白晶, 金广厚, 孙鹤林, 胡楚樊, 赵晋泉, 陈天华. 高渗透光伏配电网第三方主体调压辅助服务补偿与获取[J]. 中国电力, 2023, 56(4): 95-103,111.
[10]	王晶, 袁懿, 邵尹池, 张晋奇, 丁然, 巩彦江. 考虑资源储备度的主动配电网多目标集群划分与电压控制方法[J]. 中国电力, 2023, 56(12): 69-79.
[11]	段瑶, 高崇, 程苒, 吴亚雄, 黄烨, 刘志文. 考虑5G基站可调度潜力的配电网分布式光伏最大准入容量评估[J]. 中国电力, 2023, 56(12): 80-85, 99.
[12]	李娟, 汪家铭, 许苏迪, 姜云龙, 杨雄. 基于量子萤火虫算法的配电网故障恢复策略[J]. 中国电力, 2023, 56(12): 191-198.
[13]	李世辉, 王琪, 贾晓卜, 吴文传, 梁华洋, 朱琳. 考虑热泵负荷和分布式光伏的配微网协调调度[J]. 中国电力, 2022, 55(9): 29-37.
[14]	陈璨, 白明辉, 张婉明, 席海阔, 尹兆磊, 邵尹池, 王枭枭, 马原. 分布式光伏边界渗透率快速定位及消纳方案择优[J]. 中国电力, 2022, 55(8): 40-50.
[15]	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.

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

Distributed Photovoltaic Ultra-short-term Power Forecasting Method Based on Temporal Analog Matching Approach and Transformer Network Modeling

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 17

参考文献 23

相关文章 15

编辑推荐

Metrics

模态框（Modal）标题

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

Distributed Photovoltaic Ultra-short-term Power Forecasting Method Based on Temporal Analog Matching Approach and Transformer Network Modeling

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 17

参考文献 23

相关文章 15

编辑推荐

Metrics