There has been increasing global attention to the growing demand for a cost-efficient and carbon free energy resource to power development of artificial intelligence (AI), as well as the application of AI for scientific challenges. AI has the potential to both transform geothermal energy exploration and greatly increase the supply of renewable energy. The development of exploration-focused AI techniques, however, requires a large dataset of geologic models of geothermal systems for training algorithms. This dataset needs to be both geologically realistic and sufficiently comprehensive in order for geoscientists to apply these algorithms effectively to real-world exploration campaigns. Unfortunately, such a dataset for geothermal exploration is currently unavailable. The scarcity of geothermal sites, compared to conventional oil and gas fields, as well as the lack of digitized and systematically formatted data from earlier exploration efforts, has contributed to this gap. In this study, we address this issue by developing a dataset of geologic models of geothermal systems that combines real-world data with advanced physical simulations. This dataset, named “GeoGym”, represents a significant step forward in AI training and algorithm development for geothermal exploration. It enables a quantitative assessment of algorithm-assisted decision-making and data collection strategies.

Press the Back button in your browser, or search again.

Copyright 2025, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.