网络首发:2026-04-07,
纸质出版:2026
移动端阅览
顾婷婷, 黄亦露, 王亚男, 等. 基于多头注意力机制的ResNet-UNet短期风电功率预测[J]. 太阳能学报, 2026,47(3):474-480.
顾婷婷, 黄亦露, 王亚男, et al. 基于多头注意力机制的ResNet-UNet短期风电功率预测[J]. 2026, 47(3): 474-480.
顾婷婷, 黄亦露, 王亚男, 等. 基于多头注意力机制的ResNet-UNet短期风电功率预测[J]. 太阳能学报, 2026,47(3):474-480. DOI: doi:10.19912/j.0254-0096.tynxb.2024-2351.
顾婷婷, 黄亦露, 王亚男, et al. 基于多头注意力机制的ResNet-UNet短期风电功率预测[J]. 2026, 47(3): 474-480. DOI: doi:10.19912/j.0254-0096.tynxb.2024-2351.
为深入挖掘精细化的数值天气预报信息与风电功率之间的关联特征
提出一种基于多头注意力机制的ResNet-UNet短期风电功率预测方法。首先
考虑不同高度层的风向、风速、气压、相对湿度等气象因子
以网格为单元对数值预报进行特征提取并形成高维特征场。然后
融合UNet模型和ResNet模型
引入多头注意力机制捕获数值预报空间格点的相关特性
搭建风电功率预测模型。最后
采用浙江省某风电场的真实数据进行验证
并与UNet、ResNet、LSTM、BP模型进行对比分析
结果表明
所提出的基于多头注意力机制的ResNet-UNet预测方法能够有效提高预测精度。
Numerical weather prediction has an important impact on the accuracy of short-term wind power prediction models. In order to fully mine the deep mapping relationship between the information of numerical weather prediction and actual wind power
this paper proposes a short- term wind power forecasting model based on ResNet-UNet model with incorporation of multi-head attention mechanism. Firstly
considering the meteorological factors such as wind direction
wind speed
air pressure
temperature
relative humidity at different altitude levels
the features of numerical weather prediction information are extracted using the grid as a unit and then form the high-dimensional feature vector. Secondly
a wind power prediction model is constructed by fusing the UNet model and ResNet model
in which a multi-head attention mechanism is introduced to capture the spatial correlation characteristics of numerical weather prediction. Finally
the actual data of a wind farm in Zhejiang province is used to verify the model and compared with the prediction accuracy of the UNet
the ResNet
the LSTM
the BP models. The results indicate that the proposed method can effectively impove prediction accuracy.
SOLÉ J, GARCÍA-OLIVARES A, TURIEL A, et al. Renewable transitions and the net energy from oil liquids: a scenarios study[J]. Renewable energy, 2018, 116: 258-271.
邵桂萍, 许洪华. 可再生能源综合系统现状与未来发展趋势研究[J]. 太阳能, 2024(7): 129-134.
唐新姿, 顾能伟, 黄轩晴, 等. 风电功率短期预测技术研究进展[J]. 机械工程学报, 2022, 58: 1-24.
李伟, 王冰, 曹智杰, 等. 基于混沌理论的鸡群改进算法及其在风电功率区间预测中的应用[J]. 太阳能学报, 2021, 42(7): 350-358.
DU P W.Ensemble machine learning-based wind forecasting to combine NWP output with data from weather station[J]. IEEE transactions on sustainable energy, 2019, 10(4): 2133-2141.
邬永, 王冰, 陈玉全, 等. 融合精细化气象因素与物理约束的深度学习模型在短期风电功率预测中的应用[J]. 电网技术, 2021, 24(4): 1455-1465.
苗长新, 王霞, 李昊, 等. 基于数值天气预报风速误差修正的风电功率日前预测[J]. 电网技术, 2022, 46(9): 455-3462.
杨国清, 刘世林, 王德意, 等. 基于Attention-GRU风速修正和Stacking的短期风电功率预测[J]. 太阳能学报, 2022, 43(12): 273-281.
王勃, 冯双磊, 刘纯. 基于天气分型的风电功率预测方法[J]. 电网技术, 2014, 38(1): 93-98.
宋家康, 彭勇刚, 蔡宏达, 等. 考虑多位置NWP和非典型特征的短期风电功率预测研究[J]. 电网技术, 2018, 42(10): 3234-3242.
王丽婕, 刘田梦, 王勃, 等. 基于奇异值分解与卡尔曼滤波修正多位置NWP的短期风电功率预测[J]. 太阳能学报, 2022, 43(12): 392-398.
ANDRADE J R, BESSA R J.Improving renewable energy forecasting with a grid of numerical weather predictions[J]. IEEE transactions on sustainable energy, 2017, 8(4): 1571-1580.
邓韦斯, 车建峰, 汪明清, 等. 基于网格型数值天气预报的风电集群日前功率预测方法[J]. 南方电网技术, 2024, 18(6): 51-57.
赵永宁, 叶林. 区域风电场短期风电功率预测的最大相关-最小冗余数值天气预报特征选取策略[J]. 中国电机工程学报, 2015, 35(23): 5985-5994.
方巍, 齐媚涵. 基于深度学习的高时空分辨率降水临近预报方法[J]. 地球科学与环境学报, 2023, 45(3): 706-718.
YANG Y M, MEHRKANOON S.AA-TransUNet: attention augmented TransUNet for nowcasting tasks[C]//2022 International Joint Conference on Neural Networks (IJCNN). Padua, Italy, 2022: 01-08.
RONNEBERGER O, FISCHER P, BROX T.U-Net: convolutional networks for biomedical image segmentation[M]//Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Cham: Springer International Publishing, 2015: 234-241.
关吉平, 高延波, 黄从雷, 等. 基于Attention-Unet网络的FY-4A卫星降水估计[J]. 气象学报, 2024, 82(3): 398-410.
HE K M, ZHANG X Y, REN S Q, et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA, 2016: 770-778.
RAO R M, LIU J, VERKUIL R, etal. MSA transformer[C]//Proceedingsof the 38th International Conference on Machine Learning. Singapore: PMLR, 2021: 8844-8856.
0
浏览量
2
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621