Wind speed prediction in China with fully-convolutional deep neural network

Zongwei Zhang, Lianlei Lin, Sheng Gao, Junkai Wang and Hanqing Zhao

Renewable and Sustainable Energy Reviews, 2024, vol. 201, issue C

Abstract: Accurate and efficient short-term wind speed forecasts are critical for maintaining safe and stable operation of the wind power system. Consequently, a UNet-based fully convolutional neural network, STD-UNet, was proposed to predict the refined grid wind speed and direction in China. STD-UNet extracts and reconstructs the spatiotemporal characteristics with embedded spatiotemporal coupling and fits the wind data from wind speed, high-frequency information and wind direction using a hybrid loss function. The experiments showed the accuracy of STD-UNet in predicting the wind speed with mean absolute error of 0.74 m/s in 3 h, 1.06 m/s in 6 h, and 1.43 m/s in 12 h. The accuracy of the wind direction forecast in eight directions reached 97.66 % in 1 h and 93.1 % in 3 h. The transferability of STD-UNet was tested by accurately predicting vector wind speed in Northwest China and North America using Northeast China model. Finally, the efficiency experiments showed that STD-UNet took only 26 ms to predict the wind speed and direction in the next 24 h. Therefore, STD-UNet can be used as an effective tool for wind power centers to predict the vector wind speed and aid in wind power's the ultra-short-term and short-term deployment planning owing to its accuracy and efficiency.

Keywords: Fully convolutional network; Hybrid loss; Nonparametric attention; Refined grid wind field; Spatiotemporal forecasting; Vector wind speed prediction (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S1364032124003496
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:rensus:v:201:y:2024:i:c:s1364032124003496

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/600126/bibliographic
http://www.elsevier. ... 600126/bibliographic

DOI: 10.1016/j.rser.2024.114623

Access Statistics for this article

Renewable and Sustainable Energy Reviews is currently edited by L. Kazmerski

More articles in Renewable and Sustainable Energy Reviews from Elsevier
Bibliographic data for series maintained by Catherine Liu ().