Wind Farm Short-Term Power Prediction Based on Multiple Intelligent Algorithms

Abstract

This paper illustrates several intelligent algorithms to build a short-term wind power forecasting model using data of a particular Wind Farm in Inner Mongolia. Advantages and backwards of the four different neural network models have been carefully discussed. Calculation methods and formulas are provided to prove the result. A combination prediction method is proposed in order to make a more accurate distribution power prediction by optimizing the information of multiple single models. Therefore, after analyzing the actual wind power data, the most accurate forecasting model is selected to provide an effective reference for power dispatching, operation and equipment maintenance. In order to integrate theory with practice, a wind farm in Inner Mongolia is chosen to make the short-term power prediction using the different intelligent method discussed in this paper. The most accurate prediction algorithm has been proved by real-time data.