利用混合深度学习算法的时空风速预测  

SPATIO-TEMPORAL WIND SPEED FORECASTING BASED ON HYBRID DEEP LEARNING ALGORITHM

作　　者：贵向泉[1] 孟攀龙 孙林花 秦三杰[3] 刘靖红 Gui Xiangquan;Meng Panlong;Sun Linhua;Qin Sanjie;Liu Jinghong(School of Computer and Communication,Lanzhou University of Technology,Lanzhou 730050,China;School of Meteorology,Lanzhou Resources Environment Voc-Tech University,Lanzhou 730000,China;Gansu Meteorological Bureau,Lanzhou 730020,China)

机构地区：[1]兰州理工大学计算机与通信学院,兰州730050 [2]兰州资源环境职业技术大学气象学院,兰州730000 [3]甘肃省气象局,兰州730020

出　　处：《太阳能学报》2025年第3期668-678,共11页Acta Energiae Solaris Sinica

基　　金：甘肃省重点研发计划-工业类项目(22YF7GA159);甘肃省基础研究计划-软科学专项(22JR4ZA084);甘肃省教育厅:产业支撑计划项目(2023CYZC-25)。

摘　　要：风速预测的准确性始终不理想,为解决风速复杂的时空相关性和非线性问题,提出一种新颖的混合深度学习模型。首先,采用二次分解法将输入序列分解为具有不同频率振动模式的模态分量(IMF);使用图卷积神经网络(GCN)和双向长短期记忆网络(BiLSTM)来预测高频分量;使用自适应图时空Transformer网络(ASTTN)来预测低频分量,以充分考虑输入序列的时空相关性。最后将高频分量和低频分量合并叠加,得到最终的预测结果。将该模型应用于甘肃省某风电场进行风速预测,实验结果表明,所提出混合深度学习模型能有效提高风速预测的准确性。The accuracy of wind speed forecasting is crucial for the efficient operation and dispatch of power grid systems.To tackle the complex spatio-temporal correlations and nonlinearity of wind speed,a novel hybrid deep learning model is proposed.Firstly,a secondary decomposition method is employed to decompose the input sequence into intrinsic mode functions(IMFs)with different frequency vibration modes.Graph convolutional network(GCN)and a bidirectional long short-term memory(BiLSTM)network are used to predict the high-frequency components.An adaptive graph spatio-temporal transformer network(ASTTN)is employed to predict the low-frequency components,fully considering the spatio-temporal correlation of the input sequence.Finally,the high-frequency and low-frequency components are combined and superimposed to obtain the ultimate prediction results.The model is applied to wind speed forecasting in a wind farm in Gansu Province,and experimental results demonstrate that the proposed hybrid deep learning model effectively enhances the accuracy of wind speed forecasting.

关 键 词：风速 预测 深度学习 图卷积神经网络 双向长短期记忆网络 自适应图时空Transformer 

分 类 号：TM614[电气工程—电力系统及自动化]