孙改平,赵德麟,王旭,李芬,刘蓉晖,林顺富
SUN Gaiping,ZHAO Delin,WANG Xu,LI Fen,LIU Ronghui,LIN Shunfu

• 1:上海电力大学电气工程学部
• 2:电力传输与功率变换控制教育部重点实验室(上海交通大学)

摘要(Abstract)：

DPV（分布式光伏）空间分布分散，现有粗粒度气象数据无法解决光伏电站地理位置偏移带来的局部气象差异问题，导致光伏电站功率预测精度不高。针对上述问题，提出一种基于辐照数据精细化辨识与动态图注意力网络的区域DPV功率预测模型。首先，基于部分薄盘样条法，构建了考虑云层厚度影响的DPV空间太阳辐照度精细辨识方法，为DPV电站提供1 km×1 km分辨率的辐照数据。然后，利用图注意力网络单元捕获气象动态变化下区域DPV出力的时空相关性，聚合各电站的邻域特征，并利用编码器-解码器网络完成所述空间特征的时序建模。最后，通过上海某区域实测数据的对比实验验证了所提方法的有效性。
The dispersed spatial distribution of distributed photovoltaic(DPV) systems makes existing coarsegrained meteorological data inadequate for capturing local weather variations caused by the geographical offsets of PV stations, thereby limiting forecasting accuracy. To address this issue, this paper proposes a regional DPV power forecasting model based on fine-grained solar irradiance identification and a dynamic graph attention network. First, a fine-grained spatial solar irradiance identification method considering the influence of cloud thickness is developed using the partial thin-plate spline method, providing high-resolution solar irradiance data of 1 km×1 km for DPV stations. Subsequently, a graph attention network cell is employed to capture the spatiotemporal correlations of regional DPV power outputs under dynamically changing meteorological conditions by aggregating the neighborhood features of individual stations, while an encoder–decoder architecture is adopted to model the temporal evolution of the extracted spatial features. Finally, comparative experiments based on actual measured data from a region in Shanghai verify the effectiveness of the proposed method.

关键词(KeyWords)： 分布式光伏;功率预测;空间气象;图注意力网络;编码器-解码器
DPV;power forecasting;spatial meteorology;graph attention network;encoder-decoder

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金（52277110）

作者(Author): 孙改平,赵德麟,王旭,李芬,刘蓉晖,林顺富
SUN Gaiping,ZHAO Delin,WANG Xu,LI Fen,LIU Ronghui,LIN Shunfu

DOI: 10.19585/j.zjdl.202609004

参考文献(References)：

• [1]辛保安，单葆国，李琼慧，等.“双碳”目标下“能源三要素”再思考[J].中国电机工程学报，2022,42(9):3117-3125.Xin Baoan,Shan Baoguo,Li Qionghui,et al.Rethinking of the“three elements of energy” toward carbon peak and carbon neutrality[J].Proceedings of the Chinese Society for Electrical Engineering,2022,42(9):3117-3125.
• [2]国家能源局.2024年光伏发电建设情况[EB/OL].(2025-01-27)[2025-08-17].https：//www.nea.gov.cn/20250221/f04452701c914d51a89d0c0ea6f4acd1/c.html.
• [3]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.
• [4]Yu S W,Han R L,Zhang J J.Reassessment of the potential for centralized and distributed photovoltaic power generation in China:On a prefecture-level city scale[J]. Energy,2023,262:125436.
• [5]Sanjari M J,Gooi H B.Probabilistic forecast of PV power generation based on higher order Markov chain[J].IEEE Transactions on Power Systems,2017,32(4):2942-2952.
• [6]Sheng H M,Xiao J,Cheng Y H,et al. Short-term solar power forecasting based on weighted Gaussian process regression[J].IEEE Transactions on Industrial Electronics,2018,65(1):300-308.
• [7]邹小龙，温步瀛，王怀远.融合物理信息的光伏发电功率短期预测方法[J].智慧电力，2025,53(12):59-67.Zou Xiaolong,Wen Buying,Wang Huaiyuan. A physicsinformed short-term forecasting method for photovoltaic power generation[J].Smart Power,2025,53(12):59-67.
• [8]王瑞，闫豪，逯静.基于ICEEMDAN-MBWO-WKELM短期光伏功率预测[J].电力电子技术，2024,58(2):59-63.Wang Rui,Yan Hao,Lu Jing. Short-term photovoltaic power prediction based on ICEEMDAN-MBWOWKELM[J].Power Electronics,2024,58(2):59-63.
• [9]张禄晞，石研，曹志远，等.基于深层感知交互Transformer的光伏功率预测[J].智慧电力，2025(8):70-78.Zhang Luxi,Shi Yan,Cao Zhiyuan,et al. Photovoltaic power forecasting based on deep perception-interaction transformer[J].Smart Power,2025(8):70-78.
• [10]高任龙，李峰，马天东，等.融合多元时间序列与TCNBiLSTM的分布式光伏功率超短期预测研究[J].山东电力技术，2025,52(12):38-49.Gao Renlong,Li Feng,Ma Tiandong,et al.The research on ultra-short-term prediction of distributed photovoltaic power based on multivariate time series fusion with TCNBiLSTM network[J].Shandong Electric Power,2025,52(12):38-49.
• [11]段宏，郭成，孙海东，等.基于DDTW聚类和SK TCNGC BiGRU的分布式光伏短期功率预测[J].智慧电力，2025,53(4):71-80.Duan Hong,Guo Cheng,Sun Haidong,et al. Short-term power prediction for distributed photovoltaic systems based on DDTW clustering and SK TCN-GC BiGRU network[J].Smart Power,2025,53(4):71-80.
• [12]李芬，周尔畅，孙改平，等.一种新型天气分型方法及其在光伏功率预测中的应用[J].上海交通大学学报，2021,55(12):1510-1519.Li Fen,Zhou Erchang,Sun Gaiping,et al.A novel weather classification method and its application in photovoltaic power prediction[J].Journal of Shanghai Jiao Tong University,2021,55(12):1510-1519.
• [13]Liu C Y,Zhang X M,Mei S W,et al.Numerical weather prediction enhanced wind power forecasting:Rank ensemble and probabilistic fluctuation awareness[J].Applied Energy,2022,313:118769.
• [14]Brester C,Kallio-myers V,Lindfors A V,et al.Evaluating neural network models in site-specific solar PV forecasting using numerical weather prediction data and weather observations[J].Renewable Energy,2023,207:266-274.
• [15]张静，熊国江.考虑季节特性与数据窗口的短期光伏功率预测组合模型[J].电力工程技术，2025,44(1):183-192.Zhang Jing,Xiong Guojiang. Short-term photovoltaic power prediction combination model considering seasonal characteristic and data window[J]. Electric Power Engineering Technology,2025,44(1):183-192.
• [16]乔颖，孙荣富，丁然，等.基于数据增强的分布式光伏电站群短期功率预测（一）：方法框架与数据增强[J].电网技术，2021,45(5):1799-1808.Qiao Ying,Sun Rongfu,Ding Ran,et al.Distributed photovoltaic station cluster gridding short-term power forecasting part Ⅰ:methodology and data augmentation[J].Power System Technology,2021,45(5):1799-1808.
• [17]张童彦，廖清芬，唐飞，等.基于气象资源插值与迁移学习的广域分布式光伏功率预测方法[J].中国电机工程学报，2023,43(20):7929-7939.Zhang Tongyan,Liao Qingfen,Tang Fei,et al.Wide-area distributed photovoltaic power forecast method based on meteorological resource interpolation and transfer learning[J].Proceedings of the CSEE,2023,43(20):7929-7939.
• [18]盛裴轩，毛节泰，李建国，等.大气物理学[M].北京：北京大学出版社，2003.
• [19]王于波，郝玲，徐飞，等.分布式光伏集群发电功率波动模式识别与超短期概率预测[J].上海交通大学学报，2024,58(9):1334-1343.Wang Yubo,Hao Ling,Xu Fei,et al.Pattern recognition and ultra-short-term probabilistic forecasting of power fluctuating in aggregated distributed photovoltaics clusters[J].Journal of Shanghai Jiao Tong University,2024,58(9):1334-1343.
• [20]欧阳静，秦龙，王坚锋，等.基于PCA-ShapeDTWQWGRU的分布式光伏集群短期功率预测[J].太阳能学报，2024,45(5):458-467.Ouyang Jing,Qin Long,Wang Jianfeng,et al.Short-term power prediction for distributed PV clusters based on pcashapedtw-qwgru[J].Acta Energiae Solaris Sinica,2024,45(5):458-467.
• [21]Liao W L,Bak-jensen B,Pillai J R,et al. Short-term power prediction for renewable energy using hybrid graph convolutional network and long short-term memory approach[J].Electric Power Systems Research,2022,211:108614.
• [22]Sun D Y,Cao X,Liang Z F,et al.Distributed photovoltaic power prediction technology based on spatio-temporal graph neural networks[J].Energy Engineering,2025,122(8):3329-3346.
• [23]Simeunovi??J,Schubnel B,Alet P J,et al.Spatio-temporal graph neural networks for multi-site PV power forecasting[J]. IEEE Transactions on Sustainable Energy,2022,13(2):1210-1220.
• [24]Hutchinson M F.The application of thin plate smoothing splines to continent-wide data assimilation[C]//BMRC Research Report No.27:Data Assimilation Systems.Melbourne:Bureau of Meteorology,1991:104-113.
• [25]Kipf T N,Welling M.Semi-supervised classification with graph convolutional networks[EB/OL].[2026-01-12]. 2016:arXiv:1609.02907. https：//arxiv. org/abs/1609.02907.
• [26]Veli??kovi??P,Cucurull G,Casanova A,et al. Graph attention networks[C]//Proceedings of the 6th International Conference on Learning Representations(ICLR2018).Vancouver,Canada:[s.n.],2018:1-12.
• [27]Cho K,Van Merri??nboer B,Gulcehre C,et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]//Proceedings of the 2014Conference on Empirical Methods in Natural Language Processing(EMNLP).Doha,Qatar:Association for Computational Linguistics,2014:1724-1734.
• [28]Global Modeling and Assimilation Office. MERRA-2tavg1＿2d＿rad＿Nx:2d,1-Hourly,Time-Averaged,SingleLevel,Assimilation,Radiation Diagnostics V5.12.4[DB/OL].(2015-06-30)[2025-06-26].https：//doi.org/10.5067/Q9QMY5PBNV1T.
• [29]邓芳明，刘涛，王锦波，等.基于地基云图与气象因素多模态融合的光伏功率预测方法[J].中国电机工程学报，2025,45(18):7193-7205.Deng Fangming,Liu Tao,Wang Jinbo,et al.Research on photovoltaic power prediction based on multimodal fusion of ground cloud map and meteorological factors[J]. Proceedings of the CSEE,2025,45(18):7193-7205.
• [30]李芬，姚添添，王亚维，等.相邻阵列遮挡影响下的短期光伏功率预测[J/OL].上海交通大学学报，2025-12-26.https：//doi.org/10.16183/j.cnki.jsjtu.2025.239.Li Fen,Yao Tiantian,Wang Yawei,et al.Short-term photovoltaic power prediction under the influence of adjacent array occlusion[J/OL].China Industrial Economics,2025-12-26.https：//doi.org/10.16183/j.cnki.jsjtu.2025.239.
• [31]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.