Short-term speed predection of wind farm on the back propagation neural network

Based on wind speed sequences, three-layer neural network model of wind speed prediction is analyzed to obtain the selecting method of neural network input, output and hidden layers´ node parameters, and to predict wind speed through rolling wind speed data. In accord with the nonlinear of wind speed sequences, a BP neural network model is established to forecast wind speed. The feasibility and validity of this method is proved by simulation experiments.