计及相似日与ECM的ICEEMDAN-BiGRU-XGBoost-CrossAttention超短期光伏功率预测

李练兵; 高一波; 陈业; 代亮亮; 景睿雄; 高国强

doi:10.19912/j.0254-0096.tynxb.2024-1944

PDF(1316 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (3) : 656-667. DOI: 10.19912/j.0254-0096.tynxb.2024-1944

计及相似日与ECM的ICEEMDAN-BiGRU-XGBoost-CrossAttention超短期光伏功率预测

李练兵¹, 高一波¹, 陈业¹, 代亮亮², 景睿雄¹, 高国强¹

作者信息 +

ICEEMDAN-BiGRU-XGBoost-CrossAttention ULTRA-SHORT-TERM PV POWER PREDICTION TAKING INTO ACCOUNT SIMILARITY DAY AND ECM

Li Lianbing¹, Gao Yibo¹, Chen Ye¹, Dai Liangliang², Jing Ruixiong¹, Gao Guoqiang¹

Author information +

文章历史 +

摘要

为提高光伏功率的预测精度,提出一种考虑相似日选取与误差修正模型（error correction model,ECM）的超短期光伏功率预测方法。首先,利用改进自适应噪声完备集合经验分解（improved complete ensemble empirical mode decomposition with adaptive noise,ICEEMDAN）方法将数据分解并重构为高频与低频分量,输入基于交叉注意力机制（CrossAttention）的双向门控循环单元（bidirectional gated recurrent unit,BiGRU）与优化的分布式梯度提升库（extreme gradient boosting,XGBoost）组合的特征提取与预测模型;其次,利用灰色关联分析方法计算预测日与历史日间的综合相似因子,选取预测日的气象相似日,作为基于BiGRU的相似日信息增强模块的输入,并在初始预测序列基础上构造残差预测序列,构建基于BiGRU的误差修正模型;最后,融合相似日信息后的ICEEMDAN-BiGRU-XGBoost-CrossAttention模型预测结果,叠加误差修正模型的预测误差,得出最后的日内光伏功率预测结果。基于实际光伏场站气象以及光伏发电功率数据,对比不同光伏发电功率模型,验证了所提方法提高了日内超短期光伏发电功率预测精度,具有一定应用价值。

Abstract

To enhance the accuracy of photovoltaic power forecasting, this study proposes an ultra-short-term photovoltaic power prediction method incorporating similar-day selection and an Error Correction Model（ECM）. First, data is decomposed and reconstructed into high-frequency and low-frequency components using the Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise （ICEEMDAN） method. which feeds into a feature extraction and prediction model combining CrossAttention-based Bidirectional Gated Recurrent Units（BiGRU） and eXtreme Gradient Boosting（XGBoost）. Second, the comprehensive similarity factor between the forecast day and historical days is calculated using grey correlation analysis. Meteorologically similar days for the forecast day are selected as input to the BiGRU-based similar-day information enhancement module. A residual forecast sequence is constructed based on the initial forecast sequence to build an error correction model using BiGRU. Finally, the prediction results from the ICEEMDAN-BiGRU-XGBoost-CrossAttention model, integrated with similar-day information, are combined with the prediction errors from the error correction model to derive the final intraday PV power prediction. Using actual meteorological and PV power generation data from photovoltaic stations, comparisons with different PV power generation models validate that the proposed method enhances the accuracy of intraday ultra-short-term PV power prediction and demonstrates practical application value.

导出引用

李练兵, 高一波, 陈业, 代亮亮, 景睿雄, 高国强. 计及相似日与ECM的ICEEMDAN-BiGRU-XGBoost-CrossAttention超短期光伏功率预测[J]. 太阳能学报. 2026, 47(3): 656-667 https://doi.org/10.19912/j.0254-0096.tynxb.2024-1944

Li Lianbing, Gao Yibo, Chen Ye, Dai Liangliang, Jing Ruixiong, Gao Guoqiang. ICEEMDAN-BiGRU-XGBoost-CrossAttention ULTRA-SHORT-TERM PV POWER PREDICTION TAKING INTO ACCOUNT SIMILARITY DAY AND ECM[J]. Acta Energiae Solaris Sinica. 2026, 47(3): 656-667 https://doi.org/10.19912/j.0254-0096.tynxb.2024-1944

中图分类号： TM615

参考文献

[1] 贾凌云, 云斯宁, 赵泽妮, 等. 神经网络短期光伏发电预测的应用研究进展[J]. 太阳能学报, 2022, 43(12): 88-97.
JIA L Y, YUN S N, ZHAO Z N,et al.Progress of applied research on short-term photovoltaic power generation forecasting by neural networks[J]. Acta energiae solaris sinica, 2022, 43(12): 88-97.
[2] 朱琼锋, 李家腾, 乔骥, 等. 人工智能技术在新能源功率预测的应用及展望[J]. 中国电机工程学报, 2023, 43(8): 3027-3047.
ZHU Q F, LI J T, QIAO J,et al.Application and outlook of artificial intelligence technology in new energy power prediction[J]. Proceedings of the CSEE, 2023, 43(8): 3027-3047.
[3] 姜建国, 杨效岩, 毕洪波. 基于VMD-FE-CNN-BiLSTM的短期光伏发电功率预测[J]. 太阳能学报, 2024, 45(7): 462-473.
JIANG J G, YANG X Y, BI H B.Short-term photovoltaic power prediction based on VMD-FE-CNN-BiLSTM[J]. Acta energiae solaris sinica, 2024, 45(7): 462-473.
[4] 崔杨, 陈正洪, 许沛华. 基于机器学习的集群式风光一体短期功率预测技术[J]. 中国电力, 2020, 53(3): 1-7.
CUI Y, CHEN Z H, XU P H.Short-term power prediction for wind farm and solar plant clusters based on machine learning method[J]. Electric power, 2020, 53(3): 1-7.
[5] 张倩, 马愿, 李国丽, 等. 频域分解和深度学习算法在短期负荷及光伏功率预测中的应用[J]. 中国电机工程学报, 2019, 39(8): 2221-2230.
ZHANG Q, MA Y, LI G L, et al.Application of frequency domain decomposition and deep learning algorithm in short-term load and photovoltaic power forecasting[J]. Proceedings of the CSEE, 2019, 39(8): 2221-2230.
[6] 曾亮, 狄飞超, 兰欣, 等. 基于CEEMD-CNN-BiGRU-RF模型的短期风电功率预测[J]. 可再生能源, 2022, 40(2): 190-195.
ZENG L, DI F C, LAN X, et al.Short-term wind power prediction based on CEEMD-CNN-BiGRU-RF model[J]. Renewable energy resources, 2022, 40(2): 190-195.
[7] LIU Y W, FENG H, LI H Y, et al.An improved whale algorithm for support vector machine prediction of photovoltaic power generation[J]. Symmetry, 2021, 13(2): 212.
[8] 杨锡运, 李艳军, 柏永华, 等. 基于聚类集成的LSSVM-Adaboost模型的短期光伏功率预测[J]. 华北电力大学学报(自然科学版), 1-12. [2026-03-25]. https://link.cnki.net/urlid/13.1212.TM.20240709.1736.002.
YANG X Y, LI Y J, BER Y H, et al. Short-term photovoltaic power prediction based on clustering integrated LSSVM-Adaboost model[J]. Journal of North China Electric Power University(Natural Science Edition), 1-12. [2026-03-25]. https://link.cnki.net/urlid/13.1212.TM.20240709.1736.002.
[9] ZHOU S Y, ZHOU L, MAO M X, et al.Transfer learning for photovoltaic power forecasting with long short-term memory neural network[C]//2020 IEEE International Conference on Big Data and Smart Computing (BigComp). Busan, Korea, 2020: 125-132.
[10] 熊图, 赵宏伟, 蔡智洋, 等. 动态组合深度学习模型在短期负荷及光伏功率预测中的应用[J]. 可再生能源, 2020, 38(4): 458-463.
XIONG T, ZHAO H W, CAI Z Y, et al.Application of dynamic combinatorial deep learning model in short-term load and photovoltaic power forecasting[J]. Renewable energy resources, 2020, 38(4): 458-463.
[11] IBRAHIM I A, HOSSAIN M J, DUCK B C.An optimized offline random forests-based model for ultra-short-term prediction of PV characteristics[J]. IEEE transactions on industrial informatics, 2020, 16(1): 202-214.
[12] VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[J]. Advances in neural information processing systems, 2017, 30: 5998-6008.
[13] JU Y, LI J, SUN G Y.Ultra-short-term photovoltaic power prediction based on self-attention mechanism and multi-task learning[J]. IEEE access, 2020, 8: 44821-44829.
[14] 雷柯松,吐松江·卡日,伊力哈木·亚尔买买提,等. 基于WGAN-GP和CNN-LSTM-Attention的短期光伏功率预测[J]. 电力系统保护与控制, 2023, 51(9): 108-118.
LEI K S, TUSONGJIANG KAJI, ILHAMU YARBAYATI, et al.Short-term photovoltaic power prediction based on WGAN-GP and CNN-LSTM-Attention[J]. Power system protection and control, 2023, 51(9): 108-118.
[15] 滕陈源, 丁逸超, 张有兵, 等. 基于VMD-Informer-BiLSTM模型的超短期光伏功率预测[J]. 高电压技术, 2023, 49(7): 2961-2971.
TENG C Y, DING Y C, ZHANG Y B, et al.Ultra-short-term photovoltaic power prediction based on VMD-Informer-BiLSTM model[J]. High voltage engineering, 2023, 49(7): 2961-2971.
[16] 杨丽薇, 高晓清, 蒋俊霞, 等. 基于小波变换与神经网络的光伏电站短期功率预测[J]. 太阳能学报, 2020, 41(7): 152-157.
YANG L W, GAO X Q, JIANG J X, et al.Short-term photovoltaic output power prediction based on wavelet transform and neural network[J]. Acta energiae solaris sinica, 2020, 41(7): 152-157.
[17] 张雲钦, 程起泽, 蒋文杰, 等. 基于EMD-PCA-LSTM 的光伏功率预测模型[J]. 太阳能学报, 2021, 42(9): 62-69.
ZHANG Y Q, CHENG Q Z, JIANG W J, et al.A photovoltaic power prediction model based on EMD-PCA-LSTM[J]. Acta energiae solaris sinica, 2021, 42(9): 62-69.
[18] 杨晶显, 张帅, 刘继春, 等.基于VMD和双重注意力机制ISTM的短期光伏功率预测[J].电力系统自动化2021, 45(3): 174-182.
YANG J X, ZHANG S, LIU J C, et al.Short-term photovoltaic power prediction based on VMD and dual attention mechanism ISTM[J]. Automation of electric power systems, 2021, 45(3): 174-182.
[19] 吴汉斌, 时珉, 郑焕坤, 等. 基于EEMD-ALOCO-SVM模型的光伏功率短期区间预测[J]. 太阳能学报, 2023, 44(11): 64-71.
WU H B, SHI M, ZHENG H K, et al.Short-term interval prediction of photovoltaic power based on EEMD-ALOCO-SVM model[J]. Acta energiae solaris sinica, 2023, 44(11): 64-71.
[20] 王瑞, 高强, 逯静. 基于CEEMDAN-LSSVM-ARIMA模型的短期光伏功率预测[J]. 传感器与微系统, 2022, 41(5): 118-122.
WANG R, GAO Q, LU J.Short-term photovoltaic power prediction based on CEEMDAN-LSSVM-ARIMA model[J]. Transducer and microsystem technologies, 2022, 41(5): 118-122.
[21] ZHOU Y, ZHOU N R, GONG L H, et al.Prediction of photovoltaic power output based on similar day analysis, genetic algorithm and extreme learning machine[J]. Energy, 2020, 204: 117894.
[22] 王瑞, 闫方, 逯静, 等. 运用相似日和LSTM的短期负荷双向组合预测[J]. 电力系统及其自动化学报, 2022, 34(1): 93-99.
WANG R, YAN F, LU J, et al.Short-term load bi-directional combination forecasting using similar day and LSTM[J]. Proceedings of the CSU-EPSA, 2022, 34(1): 93-99.
[23] 高明, 郝妍. 基于BiLSTM网络与误差修正的超短期负荷预测[J]. 综合智慧能源, 2023, 45(1): 31-40.
GAO M, HAO Y.Ultra-short-term load forecasting based on BiLSTM network with error correction[J]. Integrated intelligent energy, 2023, 45(1): 31-40