Wind power prediction model based on CNN + LSTM

The invention discloses a wind power prediction model based on CNN + LSTM. The wind power prediction model can be divided into two parts: a data preprocessing stage and a model training use stage. Inthe data preprocessing stage, weather forecast data (NWP) and historical observation data of a wind field are utilized to extract characteristics of wind speed, wind direction, atmospheric pressure, temperature, air humidity and the like, and normalization processing is performed on the data. In a model training use stage, the processed data is put into a CNN + LSTM model for prediction, and the CNN network comprises a Conv1D layer, a Pooling layer and a Dropout layer. The LSTM network comprises a basic LSMT layer and a final full connection layer. According to the method, a deep learning method is used, and the CNN and the LSTM network are combined to predict the wind power. 本发明公开了一种基于CNN+LSTM的风电功率预测模型，可分为两部分：数据预处理阶段与模型训练使用阶段；在数据预处理阶段，利用风场的天气预报数据(NWP)与历史观测数据，提取了风速、风向、大气压力、温度、空气湿度等特征，对数据归一化处理；在模型训练使用阶段，将处理后的数据放入CN