TS_XGB:Ultra-Short-Term Wind Power Forecasting Method Based on Fusion of Time-Spatial Data and XGBoost Algorithm

• Published in: Procedia computer science Vol. 199; pp. 1103 - 1111
• Main Authors: Jiading, Jiang, Feng, Wang, Rui, Tang, Lingling, Zhang, Xin, Xu
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2022
• Subjects: Feature fusion; Time-Space data; Wind power prediction; XGBoost algorithm; XGBoost algorithm; Feature fusion; Wind power prediction; Time-Space data
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2022.01.140

Ultra-short-term wind power forecasting is of great significance to the safety, stableness, and optimal allocation of the power system. Aiming at the actual needs of increasing accuracy of wind power forecasting, this paper proposes an integrated learning algorithm based on time and spatial data mining according to the change rate of wind speed and the influence of wind direction, which is used in the ultra-short-term wind power forecasting model: TS_XGB. After calculating with the time series data of Xinjiang Datang Ruoqiang Wind Farm in 2018, the TS_XGB algorithm proposed in this paper has better RMSE and MAE performance than Linear Regression, Ridge Regression, SVM and Decision Tree in the 10-fold cross-validation test. The prediction effect shows that the TS_XGB model proposed in this paper can be effectively used for ultra-short-term wind power forecasting.