【摘要】 风电固有的间歇性和不确定性给准确预测风电出力带来了挑战，也给风电并网带来了棘手的问题。本文提出一种用于风电短期预测的卷积神经网络-双向长短期记忆网络(CNN-BiLSTM)混合深度学习模型。首先利用双向长短期记忆网络挖掘双向时间特征，然后利用卷积神经网络提取空间特征。最后使用本文模型对某风电场风电功率进行预测。对比结果表明，使用CNN-BiLSTM方法能显著改善预测性能，降低风电功率预测误差。更多还原

【Abstract】 The inherent intermittency and uncertainty of wind power pose challenges to accurately forecasting wind power output,as well as thorny issues for wind power grid integration. In this paper, a convolutional neural network-bidirectional long short-term memory network(CNN-BiLSTM) hybrid deep learning model is proposed for wind power short-term prediction. First, bidirectional temporal features are mined using BiLSTM, and then spatial features are extracted using CNN. Finally, the proposed method is used to predict the wind power of a wind farm. The comparison results show that using the CNN-BiLSTM method can significantly improve the prediction performance and reduce the wind power prediction error.更多还原