Wind power prediction based on the chaos theory and the GABP neural network

Abstract

The wind power prediction plays an important role in the steady run of the power system after the wind power grid. Because of the large amount of data used in the prediction process and the large number of data dimensions, the traditional prediction methods are prone to fall into local convergence and affect the accuracy of prediction. In this paper, a prediction method based on chaos theory and GABP neural network is designed. Firstly, the evolution trajectory of wind power time series in high dimensional space is recovered by phase space reconstruction. Then genetic algorithm is applied to optimize BP neural network (GABP), to make up for the randomness of BP neural network in the choice value and threshold value. Then the trajectory recovered by GABP neural network training is applied to predict the evolution trajectory in the future. Finally, taking wind farm data as an example, the results show that the proposed algorithm improves the prediction accuracy of ultra-short-term prediction.