Wind Energy Resources Prediction Based on EOF Method and RBF Neural Network

In this paper, research starts from the data captured from several wind measuring stations. Firstly, the main spatial Patterns are extracted by EOF (empirical orthogonal function) method, and then the time coefficient series corresponding to principal spatial patterns are processed and predicted by RBF (radial basis function) neural network. Furthermore, according to the EOF decomposition method, inversely the new prediction time coefficient series are used to calculate the wind speed values in the future. Finally, the validity and advantages of this prediction approach are tested by the simulation results.