基于混合卷积神经网络算法的风场预测研究

Wind Field Prediction Based on Hybrid Convolutional Neural Network Algorithm

在农业生产中,准确的风速预报对农作物安全防范有着至关重要的作用。针对云南地区的高海拔和多山,基于卷积神经网络框架,提出了卷积长短时序分析神经网络-卷积门控循环单元神经网络(ConvLSTM-ConvGRU)混合风速预测模型。通过神经网络框架的改进,有效的提高了模型对风场空间特征的提取。利用美国国家环境预报中心(NCEP)提供的再分析风速数据集,使用ConvLSTM、ConvGRU、ConvLSTM-ConvGRU混合模型分别对云南地区的风速进行。实验结果表明:ConvLSTM-ConvGRU混合风速预测模型能够有效对云南地区风场进行预测,相较于另外两个模型提高了预测准确度。

In agricultural production,accurate wind speed forecasts are crucial to crop safety precautions.Based on the convolutional neural network framework,a ConvLSTM-ConvGRU hybrid wind speed prediction model is proposed for the high altitude and mountainous nature of Yunnan.Through the improvement of the neural network framework,the extraction of the spatial features of the wind farm by the model is improved.Using the reanalyzed wind speed dataset provided by the National Environmental Forecasting Center of the United States,the wind speed prediction of Yunnan region was carried out using the ConvLSTM,ConvGRU,and ConvLSTM-ConvGRU hybrid models.Experimental results show that the ConvLSTM-ConvGRU hybrid wind speed prediction model can effectively predict the wind field in Yunnan,and improve the prediction accuracy compared with the other two models.