George Bebis
University of Nevada-Reno, NV·Computer Science
Much of the world's groundwater flows through cracks in underground rock. These fracture networks are hard to map because they are hidden and observations are sparse.
- Fields
Computer Science
Earth & Environmental Science
Engineering
- Topics
deep learning
generative models
conditional generative modeling
- Grants
- 1 known grant, through Sep 2029
What they work on
- Develop a physics-informed conditional generative deep learning framework that infers underground fracture network geometry and hydraulic properties from hydraulic and tracer tomography data.
- Embed numerical simulators of groundwater flow physics into the learning framework to improve robustness and generalization when field observations are sparse, and to quantify uncertainty in the estimated fracture fields.
- Validate the fracture-mapping methods using synthetic datasets, laboratory experiments on 3D-printed fractured rock, and field observations.
More about the research
Better maps of fractures could improve how water and energy resources underground are managed and reduce risks in subsurface engineering.
Why it matters
Water managers and engineers working underground need accurate pictures of rock fractures that control how groundwater and fluids move. AI tools that infer fracture networks and their uncertainty from limited field tests could improve groundwater and subsurface energy resource management. They could also lower risks in projects that depend on understanding the subsurface.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Collin Sutton
Colorado School of Mines, CO·Earth & Environmental Science
Develop a reduced-order, bipartite graph framework that captures the essential flow, solute transport and geochemical reactions in fractured rock without the cost of full-physics simulation.
Liangping Li
South Dakota School of Mines and Technology, SD·Earth & Environmental Science
Model how hydraulic fractures spread during stimulation of deep rock, using Idaho National Laboratory's network flow-discrete element model and experiments at the Sanford Underground Research Facility.
Qi Yu
University of California-San Diego, CA·Computer Science·Early-career
Develop deep learning models that forecast spatiotemporal dynamics while obeying physical laws, with the goal of simulating ocean currents, traffic and epidemic spread faster than numerical solvers.
The school
University of Nevada-Reno: 11 labs on PI Pages, $194M a year on research, 186 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.