Geothermal Operational Optimization with Machine Learning

Grant Buster, Nicole Taverna, Michael Rossol, Jay Huggins, Jon Weers, Andy Blair, Paul Siratovich

Research output: NRELPresentation


The Geothermal Operational Optimization with Machine Learning (GOOML) project has developed a generic and extensible component-based system modeling framework to study complex geothermal fields using a data-driven approach. Through building a digital twin of a geothermal steam field with the GOOML modeling framework, operators can analyze historical and forecasted power production, explore possible steam field configurations, and optimize real world operations, all in a cost-effective digital environment. The GOOML modeling software is based on a historical data-assimilation framework that uses first-principal thermodynamics to model steam field components using historical data, and a forecast framework that uses machine-learning-driven models of steam field components to predict future operations. This modeling framework creates countless new opportunities for digital exploration of steam field design and operations. To date, digital twins have been developed for several steam fields in New Zealand and the United States. These digital twins have been validated by comparing hindcast predictions against historical production data. Field design and operations have been explored using genetic optimization and reinforcement learning. Initial results show compelling and often surprising opportunities for improved design and operation of fields with 2 to 5 percent improvements in annual energy production. GOOML is driving a step-change in geothermal operations by applying state-of-the-art machine learning algorithms, comprehensive data analytics, and a first-of-its-kind intelligent geothermal systems model.
Original languageAmerican English
Number of pages19
StatePublished - 2021

Publication series

NamePresented at the 15th International Conference on Energy Sustainability (ES2021), 16-18 June 2021, Denver, Colorado

NREL Publication Number

  • NREL/PR-6A20-79934


  • geothermal machine learning
  • geothermal operations optimization
  • steamfield optimization


Dive into the research topics of 'Geothermal Operational Optimization with Machine Learning'. Together they form a unique fingerprint.

Cite this