A short-term wind power prediction approach based on the dynamic classification of the weather types of wind farms

Abstract

Wind Power Prediction (WPP) is an effective way to reduce the uncertainty and randomness of large volume of wind power, and to improve the wind power integration capacity of power grid. Recently, WPP based on the classification of weather types is one of the most popular methods to improve the WPP accuracy, but most of the research focused on the static classification of the weather types, and the dynamic classification of weather types is not reported. Numerical Weather Prediction (NWP) data of a certain time is applied for the static classification of weather type, which is not effective for describing the dynamic changing process of the weather of wind farms. To overcome the challenge, a short-term WPP approach based on the dynamic classification of the weather types, dividing the weather types into stable type, trending type and fluctuating type, on the basis of wind speed change in a period of time, is presented in the paper. The dynamic classification of the weather types based WPP is with significant advantage to describe the wind power characteristics of different weather types, which is an effective method to improve the short-term WPP accuracy, and is recommended for the industrial applications.