Author(s)

Zhe Yan, Liang Xu, Mengfang Yuan, Junfeng Guo

Corresponding Author

Zhe Yan

Abstract

The role of renewable energy in the world is becoming more and more significant.Wind power, as a key component, the accuracy of its power prediction is crucial for the stable operation of the power grid. This article is based on short-term data of wind power and utilizes the powerful nonlinear mapping ability of BPNN to construct a special wind power prediction model. Firstly, the model topology and main parameters are determined based on the characteristics of the data samples, and the model is trained and adjusted. Then, historical power data and related meteorological data of a certain wind farm were collected and processed, and a model was trained and predicted. Experiments show that this method has a high precision in the short term wind power forecast, and the error is much lower than traditional one.Finally, the article discusses the potential advantages and limitations of this method in practical applications, and proposes future research directions.

Keywords

Wind power prediction, BPNN, grid stability, machine learning