Short Term Wind Power Prediction Based On Support Vector Machine

The arithmetic of wind power prediction plays an important part in the development of wind power prediction.In this paper,based on the principles of support vector machine(SVM),SVM model for wind power prediction is built up.Piecewise Support Vector Machine(PSVM) model is proposed to improve the precision by analyzing the characteristics of power curves of wind turbine generator systems.Parameters of this model are optimized by using two methods to research the relationship between precision and parameters.Optimal wind turbine prediction method is used for prediction on wind farm.Two types of training sample are set up for the comparison of precision as well.The operation data from a wind farm in North China are used to test the proposed model,the mean relative error of PSVM model is 4.76%less than that of SVM model. Results also indicate that this model is robust.For the time frame of one to six hours,the average error of optimal wind turbine prediction method is 12.55%.