A data sample division method for wind power prediction based on China's 24 solar terms. (12th February 2020)

• Record Type: Journal Article
• Authors: Han, Shuang; Zhang, Luna; Liu, Yongqian; Ma, Yuanchi; Yan, Jie; Li, Li
• Other Names: Wang Yi guestEditor.
• Is Part Of: International transactions on electrical energy systems. Volume 30:Number 7(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-02-12
• Subjects: data samples -- similarity measure -- wind power prediction -- 24 solar terms; Electric power -- Periodicals; Electric power systems -- Periodicals; Electrical engineering -- Periodicals; 621.3
• DOI: 10.1002/2050-7038.12342
• Languages: English
• ISSNs: 2050-7038

Abstract

Summary: Wind power prediction (WPP) is one of the key techniques to eliminate the adverse impact of large proportion of wind power on power grid, while it faces the main problem of low prediction accuracy. The main contribution of this paper is a new data sample division method based on China's 24 solar terms (24STs) for wind power prediction. The 24STs is an astronomical and meteorological calendar summarized by ancient Chinese, and it more accurately reflects the law of the seasonal changes, the phenomenon of phenology, and climate characteristics compared with the calendar month. In the paper, firstly, the reasonableness of the division of 24STs is verified from the point of temperature and wind speed; secondly, four meteorological factors related to wind power output are used for similarity calculation, and the results show that the data samples in the same STs have higher similarity than those in the same semi‐month based on Gregorian; finally, the WPP models are established with the 24STs sample division method and the semi‐month sample division method, and the case proves that the 24STs method has higher prediction accuracy because it can grasp the seasonal changes law of some meteorological elements more accurately.