aiION: a machine learning approach to geothermal exploration.

AlGaiar, Mahmoud M.

Abstract

Geothermal energy, a renewable resource derived from the Earth's subsurface, holds great promise for sustainable power generation. This research aimed to develop a data-driven geothermometer using supervised machine learning (ML) to predict subsurface temperatures in deep geothermal fields based on geochemical data. The study involved managing large geochemical and thermal databases, using data pre-processing techniques to manage missing values, and clustering analysis to identify patterns within the data. Geographic Information System (GIS) mapping was also used to spatially represent the results. Various ML algorithms, including regression modelling and artificial neural networks (ANN), were used to train the model, followed by rigorous validation against actual temperature measurements from well logs. The results showed a strong correlation between geochemical parameters and subsurface temperatures, demonstrating that the developed geothermometer can effectively estimate temperatures even in areas without direct measurements. This innovative approach highlights the role of artificial intelligence in improving the exploration and management of geothermal resources, ultimately contributing to renewable energy sustainability efforts. The research highlights the integration of geochemistry with advanced data science as a transformative step towards optimizing geothermal exploration practices for a cleaner energy future.

Citation

ALGAIAR, M.M. 2025. aiION: a machine learning approach to geothermal exploration. Presented at the 2025 Geothermal seminar (GEOTHERMAL 2025): gaining momentum, 26-27 February 2025, [virtual event].

• Presentation Conference Type: Presentation / Talk
• Conference Name: 2025 Geothermal seminar (GEOTHERMAL 2025): gaining momentum
• Start Date: Feb 26, 2025
• End Date: Feb 27, 2025
• Deposit Date: Feb 28, 2025
• Publicly Available Date: Feb 28, 2025
• Peer Reviewed: Peer Reviewed
• Keywords: Geothermal energy; Renewable energy; Sustainable power; Machine learning (ML)
• Public URL: https://rgu-repository.worktribe.com/output/2720807