The U.S. Department of Energy Office of Science selected a geothermal energy project led by the University of California, Irvine to advance into a three-year Phase 2 program focused on using artificial intelligence to improve geothermal energy development.
The initiative, called MAESTRO, or Multi-Agent AI Expert for Subsurface Reasoning and Optimization, brings together researchers across multiple universities, national laboratories, and private companies.
The project is designed to use AI to better understand complex underground environments and help identify locations where enhanced geothermal systems can be developed safely and economically.
MAESTRO is part of the Department of Energy’s Genesis Mission Phase 2 awards, which are intended to apply AI to major scientific and engineering challenges.
The research team includes participants from UC Berkeley, UC Irvine, UC Riverside, UC San Diego, and UC Santa Cruz.
Four national laboratories are also involved: Los Alamos National Laboratory, Lawrence Livermore National Laboratory, Lawrence Berkeley National Laboratory, and Pacific Northwest National Laboratory.
Three private-sector partners are participating as well.
The project combines expertise in geophysics, geomechanics, geochemistry, artificial intelligence, machine learning, applied mathematics, and geothermal engineering.
Enhanced geothermal systems typically work by circulating fluid through low-permeability rock at depths of up to approximately 2.5 miles to extract heat and generate electricity.
A major challenge is identifying locations where underground conditions, including stress state, rock properties, water availability, drilling accessibility, existing fractures, and seismic characteristics, are suitable for development.
Many of those properties cannot be measured directly.
MAESTRO will use AI-based simulations that combine large volumes of geophysical observations with previous learning to infer subsurface conditions and predict how fractures may develop during geothermal operations.
The team aims to create models that can estimate the likelihood of successful enhanced geothermal development before operators commit significant capital.
The AI system will also use a self-improving feedback loop.
Areas of uncertainty will be identified, additional simulations will be run to fill knowledge gaps, and the resulting information will be incorporated into updated models.
The project also includes safeguards intended to ensure that AI-generated recommendations remain physically consistent.
Another major focus will be induced seismicity associated with geothermal fracturing.
Researchers plan to improve understanding and control of earthquakes that can be triggered by subsurface operations, while exploring whether the resulting tools could also contribute to early-warning capabilities for naturally occurring earthquakes.
The MAESTRO project will be open source.
The team plans to package its tools into a free browser-based platform that researchers, policymakers, industry professionals, and members of the public can use to explore geothermal potential, evaluate risk, and review data across the United States.
KEY QUOTES:
“There is more than enough potential geothermal energy underground to power the entire United States, but unlocking it safely and economically is still a major challenge. The MAESTRO project is designed to resolve the gaps in our knowledge of complex underground environments by using an integrated AI framework to attack the problem from multiple angles.”
Mohammad Javad Abdolhosseini Qomi, Lead Principal Investigator and Professor of Civil and Environmental Engineering at UC Irvine
“This is, of course, an honor but also a major responsibility to meet the grand vision set by the DOE for the Genesis Mission.”
Mohammad Javad Abdolhosseini Qomi, Lead Principal Investigator and Professor of Civil and Environmental Engineering at UC Irvine
“Increasingly, addressing critical challenges that our society faces requires collaboration across disciplines and institutions. Doing great science is a requirement, of course, but we need people who have the talent to put great teams together and lead them. We are fortunate to have in our school colleagues who succeed in this difficult task.”
Faryar Jabbari, Interim Dean of the Samueli School of Engineering at UC Irvine
“A major strength of this project is the unique set of partnerships it creates, which brings together for the first time geophysics, geochemistry, applied math and computation to address this singular national challenge for geothermal energy.”
Kevin Rosso, Battelle Fellow and Geochemist at Pacific Northwest National Laboratory
“The next frontier is AI that learns by doing: simulation in the loop. This is exactly what’s needed for guiding complex engineering operations like geothermal energy.”
Maziar Raissi, Associate Professor of Mathematics at UC Riverside
“Geothermal power could be a major source of energy in the future, and we are excited to grapple with some of the key hurdles to scaling it up, such as induced seismicity and generalizing behavior between fields.”
Emily Brodsky, Distinguished Professor of Earth and Planetary Sciences at UC Santa Cruz
“Professionals in the sustainable energy development industry will ultimately have access to real-time guidance that can reduce the financial risks and time costs of bringing geothermal projects to fruition. For communities, continuous underground monitoring will greatly enhance safety in earthquake-prone regions. Our tools will be packaged into a free, browser-based tool that anyone can use to explore geothermal potential, assess risk and review data across the country.”
Russ Detwiler, Co-Principal Investigator and Professor of Civil and Environmental Engineering at UC Irvine
