Wind Speed Prediction Model Based on Radial Basis Functional Neural Network

Wind speed forecasting is important to operation of wind power plants and power systems. To solve short term wind speed prediction problem, a radial basis functional neural network prediction model for wind speed time series based on cross iterative fuzzy clustering algorithm and regularized orthogonal least squares algorithm is proposed. First, the optimal fuzzy clustering centers of samples are computed by cross iterative fuzzy clustering algorithm. Then radial basis functional centers are optimized by regularized orthogonal least squares algorithm, and the generalized cross-validation is regarded as criteria to halt center selection. The proposed model centralizes advantages of both algorithms, and it can decrease network scale, improve generalization performance, accelerate network training speed and avoid ill-conditioning of learning problems. A case of practical wind speed time series from wind power plants verifies validity of the proposed model.