Uncertainty quantification for basin-scale geothermal conduction models.

Area of Science:

• Geophysics
• Energy Science
• Computational Science

Background:

• Geothermal energy is a key renewable resource for the energy transition, offering a low CO2 footprint.
• Accurate subsurface characterization is crucial for responsible geothermal resource management, but often involves high uncertainties.
• Traditional deterministic simulations are insufficient for quantifying these uncertainties, and probabilistic approaches are computationally infeasible due to long simulation times.

Purpose of the Study:

• To develop and implement a computationally efficient method for generating full state predictions in geothermal simulations.
• To enable probabilistic simulations and inverse approaches by significantly reducing computation time.
• To provide tools for quantifying and communicating subsurface uncertainties in geothermal energy applications.

Main Methods:

• Utilized a reduced basis method to significantly decrease geothermal simulation runtimes.
• Integrated the reduced basis method into an existing simulation framework.
• Generated 2D and 3D predictive uncertainty maps for geothermal applications.

Main Results:

• The reduced basis method enables the generation of full state predictions, drastically cutting simulation time.
• Probabilistic simulations and inverse approaches become feasible, allowing for uncertainty quantification.
• Predictive uncertainty maps were generated, identifying regions with high temperatures and low uncertainties.

Conclusions:

• The developed method facilitates responsible geothermal resource management by enabling uncertainty quantification.
• The approach significantly reduces simulation times, making probabilistic analyses practical.
• The flexible implementation allows for transferability to other geophysical simulations requiring state and uncertainty information.