Learning from ChatGPT: A Transformer-Based Model for Wind Power Forecasting

Xiaoran Dai1; Guo-Ping Liu2; Wenshan Hu1; Zhongcheng Lei1; Hong Zhou1

10.1109/EEEIC/ICPSEurope57605.2023.10194807

2023

979-8-3503-4744-9

2023 IEEE International Conference on Environment and Electrical Engineering and 2023 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe)

1-6

6-9 June 2023

Madrid, Spain

Wind power forecasting is a crucial aspect of re-newable energy production, as it helps to optimize energy output and ensure grid stability. In recent years, Transformer-based language models such as ChatGPT have been successfully used in natural language processing tasks, but their application in wind power forecasting remains largely unexplored. In this article, we propose using a Transformer model, the core of ChatGPT, to improve the accuracy of wind power forecasting. Using the self-attention mechanism, the developed model can capture the complex temporal relationships in large-scale time series data. Furthermore, the proposed method is evaluated on a test set using various performance metrics. Results show that our model outperforms traditional forecasting models, achieving higher accuracy. Our findings suggest that Transformer-based models have significant potential for improving wind power forecasting accuracy and ultimately contributing to a more sustainable energy future.

Wind power forecasting ChatGPT Transformer-based model Time series prediction

其他

EI

20233514630588

Natural language processing systems ; Weather forecasting ; Wind power

Meteorology:443 ; Wind Power (Before 1993, use code 611 ):615.8 ; Data Processing and Image Processing:723.2 ; Mathematical Statistics:922.2

IEEE

1.School of Electrical Engineering and Automation, Wuhan University, Wuhan, China
2.Center for Control Science and Technology, Southern University of Science and Technology, Shenzhen, China

Xiaoran Dai,Guo-Ping Liu,Wenshan Hu,et al. Learning from ChatGPT: A Transformer-Based Model for Wind Power Forecasting[C],2023:1-6.