Wind power prediction with LSSVM model based on data stratification pretreatment

Abstract

Wind speed and wind power prediction are the keys to solve the wind power with grid problems. The invalid sample data affects the wind power model. To get the relationships of wind speed and wind power, layered statistics method is used to modify the wind power curve. This paper uses least square support vector machine model to predict the modified data. In order to verify the predicted effect, experienced power curve method is used for comparison. The results show that layered statistics method can eliminate the invalid data effectively and improve the accuracy of the prediction.