A dynamic adaptive algorithm for statistical up-scaling based short-term wind power prediction of regions

Wind Power Prediction (WPP) of regions is beneficial for the department of power grid dispatching to evaluate the wind power output characteristics from a global perspective and to formulate a scientific and reasonable scheduling scheme for power grid. The statistical up-scaling based WPP of regions is one of the most popular methods, which has been widely applied by the industrial companies. The selection of reference wind farms is one of the key factors of the statistical up-scaling method based WPP of regions. However, most of the methods for the selection of reference wind farms are not flexible, which restrict the improvement of the WPP accuracy, as the weather of wind farm sub-regions are always changeable. To overcome the challenge, a dynamic adaptive algorithm based WPP of regions is presented in the paper, in terms of the dynamic selection of reference wind farms, the dynamic selection of the WPP methods based on the classification of the weather types of the wind farms and the dynamic selection of the up-scaling method. The results show that the proposed method is with higher accuracy than the traditional statistical up-scaling based WPP of regions and is applicable for industrial WPP of regional wind farms.