Improved chance constrained random vector functional link network for prediction interval of wind power forecasting

In recent years, the application proportion of renewable energy in the power system has gradually increased, and the proportion of renewable energy access will also steadily increase in the future. To cope with the instability brought about by a large number of renewable energy sources connected to the power system, power prediction research is highly significant. In this paper, an improved chance constrained random vector functional link (ICCRVFL) network interval prediction model is innovatively proposed to maximize the interval sharpness performance under the premise of ensuring the reliability of probabilistic prediction. The method firstly combines chance constrained programming and random vector functional link (RVFL) network, and then improves the network structure of RVFL network to construct the prediction models with new network structures. Finally, real data from wind farms are cited for case studies, and the results show that the proposed interval prediction method has high reliability and sharpness.