A Prediction Method for Ultra Short-Term Wind Power Prediction Basing on Long Short -Term Memory Network and Extreme Learning Machine

• Published in: 2020 Chinese Automation Congress (CAC) pp. 7608 - 7612
• Main Authors: Guangxu, Pan, Haijing, Zhang, Wenjie, Ju, Weijin, Yang, Chenglong, Qin, Liwei, Pei, Yuan, Sun, Ruiqi, Wang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 06.11.2020
• Subjects: Data models; extreme learning machine; Extreme learning machines; Logic gates; long short-term memory network; Mathematical model; Predictive models; ultra short-term wind power prediction; Wind forecasting; Wind power generation
• ISSN: 2688-0938
• DOI: 10.1109/CAC51589.2020.9327895

Thigh degree of accuracy towards the prediction of wind power contributes a lot to planning, economic performance and security maintenance. The viewpoint existing in this paper includes a kind of methods for Ultra short-term Wind Power Prediction basing on long short-term memory network and extreme learning machine. In the data preprocessing stage, considering the coupling between wind power and weather, ensemble empirical mode decomposition (EEMD) is used to decompose the wind power sequence, and principal component analysis (PCA) is used to remove features that are poorly correlated with wind power prediction. In the prediction stage, the low-frequency component uses the long short-term memory network prediction model. High-frequency feature points is used for extreme learning machine prediction model. Finally, we reconstruct the prediction results