Short-term wind power forecasting based on support vector machine

Wind power prediction, especially short-term forecasting is very significant for the security, stability and economy of power grid. Besides, it plays an important role in a micro-grid for load balancing and capacity planning. Precise prediction of wind power of micro-grid is a complex problem due to its strong randomness and little training data. Compared with traditional methods, Support Vector Machine (SVM) based on Structure Risk Minimization principle, plays a better performance on nonlinear and small sample problems. The method of SVM for short-term wind power prediction is proposed in this paper. An improved pattern search algorithm, which takes use of Lagrange interpolation to obtain the initial points, is used to optimize the parameters of SVM prediction model. The simulation results indicate that the method proposed in this paper can realize short-term wind speed prediction effectively. This paper presents some promising patents on prediction of wind power.