Wind Speed Prediction Based on Time series Neural Network Algorithm

This paper choose short-term prediction as the research content of the wind speed, the time series and neural network theory are uesd into to the wind speed prediction model. On the basis of BP model, ARMA-BP model is proposed. Then the ARMA-RBF model was established, makes the prediction error smaller. Simulation experimental results shows that using neural network to establish the sequence of the network, training time significantly shortened, network between predicted values and the real observation values of output doesn't appear too big deviation, the new training sample set fitting is good, and prediction accuracy improved.