Machine Learning and Artificial Intelligence

NREL's geothermal and machine learning experts have teamed up to develop a suite of algorithms and tools that improve reservoir characterization, economize drilling, and optimize geothermal steam field operations.

Advancements in machine learning and artificial intelligence are creating opportunities for innovation and optimization in the geothermal industry. NREL is using the latest machine learning techniques—including genetic algorithms, deep neural networks, reinforcement learning, and generative adversarial networks—in combination with our high-performance computing capabilities and geothermal expertise to create physics-based tools that can solve complex problems in a matter of minutes. These innovative tools improve the accuracy of geothermal exploration, reduce the cost of geothermal exploration, and optimize geothermal operations—driving down the overall cost of geothermal energy.

Capabilities

Our machine learning expertise encompasses a range of artificial intelligence and machine learning techniques, including:

• Deep learning
• Convolutional neural networks
• Genetic algorithms
• Reinforcement learning experiments
• Generative adversarial networks.

These techniques are used in combination with our expertise in geophysics, geology, drilling, power conversion, electricity generation, and thermodynamics to generate artificially intelligent, physics-based tools capable of improving:

• Subterranean resource characterization
• Discrete microseismic event identification and characterization
• Advanced drilling fault detection and fault prediction
• Optimization of steam field operations.

Partnerships

NREL works with a variety of industry partners, domestically and internationally, to accelerate the adoption of machine learning and artificial intelligence technologies and to ground-truth machine learning findings on steam fields and in wells around the world.

Publications

GOOML: Geothermal Operational Optimization with Machine Learning, Transactions (2020)

GOOML: Geothermal Operational Optimization with Machine Learning, World Geothermal Congress (2020)

Using Machine Learning To Predict Future Temperature Outputs in Geothermal Systems, Transactions (2020)

Which Geologic Characteristics Control Porosity and Permeability in Hydrothermal Reservoirs?, Transactions (2020)

View all NREL publications about geothermal research.

Contact