AiION: novel deep learning chemical geothermometer for temperature prediction of deep geothermal reservoirs.

AlGaiar, Mahmoud; Bano, Shahana; Lashin, Aref; Hossain, Mamdud; Faisal, Nadimul Haque; Abu Salem, Hend S.

Abstract

This study introduces AiION, a novel deep learning chemical geothermometer designed to predict deep geothermal reservoir temperatures and address the limitations of traditional geothermometry methods. By integrating classical geothermometry, multi-component geothermometry, and existing machine learning insights, AiION was trained on a comprehensive dataset of 674 water samples from Nevada. Among four evaluated machine learning algorithms, AiION, a deep neural network model, demonstrated superior performance, explaining over 97% of the variance in both training and test data. The global applicability of AiION was validated through successful evaluation on 42 new well samples from diverse geothermal fields worldwide. This research significantly advances solute geothermometry by providing a reliable, data-driven tool for geothermal exploration and development, contributing to sustainable energy efforts. The novelty of AiION lies in its large training dataset, high prediction accuracy, and global applicability, which overcome the limitations of traditional and existing machine learning methods for reliable subsurface temperature prediction in diverse geothermal systems.

Citation

ALGAIAR, M., BANO, S., LASHIN, A., HOSSAIN, M., FAISAL, N.H. and ABU SALEM, H.S. 2025. AiION: novel deep learning chemical geothermometer for temperature prediction of deep geothermal reservoirs. Renewable energy [online], 248, article number 123154. Available from: https://doi.org/10.1016/j.renene.2025.123154

• Journal Article Type: Article
• Acceptance Date: Apr 11, 2025
• Online Publication Date: Apr 12, 2025
• Publication Date: Aug 1, 2025
• Deposit Date: Apr 14, 2025
• Publicly Available Date: Apr 14, 2025
• Journal: Renewable energy
• Print ISSN: 0960-1481
• Electronic ISSN: 1879-0682
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 248
• Article Number: 123154
• DOI: https://doi.org/10.1016/j.renene.2025.123154
• Keywords: Geothermometry; Artificial intelligence; Geothermal energy; Geothermal exploration; Machine learning; Random forest; Gradient boosting; Artificial neural networks; Deep neural networks
• Public URL: https://rgu-repository.worktribe.com/output/2795032