Short-term prediction of wind power based on deep Long Short-Term Memory

This paper proposes a wind power prediction model based on the Long Short-Term Memory model, one of the deep learning method. Deep learning conforms to the trend of big data and has powerful capability of learning and generalization for mass data. Principal component analysis (PCA) is used to choose input samples and reduce the dimensions of the input variables of the LSTM prediction model based on numerical weather prediction(NWP) data. Simulation results show that, compared with BP neural network and support vector machine(SVM) model, the LSTM prediction model has higher prediction accuracy and greater potential for engineering applications. It is effective to apply the Long Short-Term Memory model to the field of wind power prediction.