Na-yeon An, Sang-uk Park, Journal of Internet Computing and Services, Vol. 25, No. 6, pp. 23-36, Dec. 2024
10.7472/jksii.2024.25.6.23, Full Text:  HTML
Keywords: Transformer, Deep Learning, Machine Learning, AI, Hydrogen Energy, Solar energy

Abstract

South Korea is focusing on the potential of green hydrogen as a key energy source to achieve its 2050 carbon neutrality goal. Green hydrogen is produced by water electrolysis using renewable energy sources such as solar and wind power, and is considered to be a completely clean energy source that does not emit any greenhouse gases and fine dust. In this study, a Transformer-based forecasting model was implemented to predict the demand and supply of hydrogen energy based on solar power generation data and local weather data in Samcheok, Gangwon Special Self-Governing Province. The correlation between solar power generation and meteorological data was analyzed and used as input variables for the model, and the performance of the prediction model was evaluated using the mean absolute value error (MAE), mean square error (MSE), and root mean square error (RMSE). We also analyzed the difference in prediction performance according to the ratio of training and test data to suggest the optimal data partitioning strategy. The prediction model proposed in this study can easily grasp the predicted hydrogen production through visualization, which can contribute to the efficient management and operation of green hydrogen energy.

Statistics
Cite this article