Unlocking hidden geothermal potential: leveraging artificial intelligence for subsurface exploration.

AlGaiar, Mahmoud M.

Abstract

Geothermal energy shows potential as a renewable and sustainable energy source. However, geothermal exploration is challenging and costly due to the subsurface complexities involved in locating potential reservoirs. Recent advancements in artificial intelligence and machine learning provide opportunities to enhance exploration efficiency, reduce costs, and accelerate discovery. Machine learning algorithms can analyze large, multidimensional datasets from various sources, including geophysical, geological, geochemical, thermal, and geospatial datasets, to identify intricate patterns and relationships that guide the exploration of hidden geothermal resources. For instance, supervised machine learning algorithms, such as random forests and neural networks, can be trained on labeled datasets to identify patterns that correlate with geothermal productivity, including geospatial features, subsurface structures, and temperature gradients. Alternatively, utilizing unsupervised learning techniques such as clustering analysis can identify outliers and reveal distinct sets of characteristics associated with hidden geothermal potential. Simultaneously, the utilization of deep neural networks can automate analysis of raw geophysical imagery and seismic survey data to identify key subsurface features and stratification. Artificial Intelligence approaches for geothermal exploration are limited by factors such as data availability and quality. The uncovering of hidden geothermal systems requires large datasets that are often costly to obtain, the relationships between subsurface features and geothermal activity are highly complex, and it is difficult to generalize models to new geological environments because geothermal activity can vary significantly in different regions. Expert human insight is also required in geothermal exploration to improve the interpretability of data-driven models and to fully capture the dynamic complexities of subsurface reservoirs. Artificial intelligence can integrate diverse data sources to provide comprehensive geothermal predictive modeling and decision support. By utilizing data-driven techniques with further research and development, exploration can be accelerated, productivity can be improved, and hidden geothermal resources can be developed in a strategic, economical, and sustainable manner.

Citation

ALGAIAR, M.M. 2024. Unlocking hidden geothermal potential: leveraging artificial intelligence for subsurface exploration. Presented at the 3rd Annual geothermal seminar (Geothermal 2024): heating up the market, 21-22 February 2024, [virtual event].

• Presentation Conference Type: Lecture
• Conference Name: 3rd Annual geothermal seminar (Geothermal 2024): heating up the market
• Conference Location: [virtual event]
• Start Date: Feb 21, 2024
• End Date: Feb 22, 2024
• Deposit Date: Feb 27, 2024
• Publicly Available Date: Apr 18, 2024
• Keywords: Geothermal energy; Renewable energy; Sustainable energy source; Artificial intelligence; Machine learning
• Public URL: https://rgu-repository.worktribe.com/output/2255420