February 9, 2017
Web Feature

Two Brains are Better Than One: How the Geothermal Community Upped the Game for Computer Codes

Modeling an underground world of uncertainty requires collaboration

The GTO-CCS team brought 11 research institutions together.

Heat from the Earth’s core contains an immeasurable amount of renewable energy; problem is it’s not very accessible. According to DOE, enhanced geothermal systems (EGS) could generate more than 100 GW of clean, renewable energy. That’s enough electricity to meet the nation’s power needs many times over.

EGS works by fracturing the rock beneath the Earth’s surface and connecting the fracture network to geothermal wells. Geothermal wells are located in a fracture network’s path so geothermal fluid can be pumped underground and then flowed through the fractures. The rock heats the fluid, turning it to steam at the ground surface, which is then used to turn turbines and generate electricity.

Codes Can Cut Costs

High upfront installation costs create a barrier to fast, widespread deployment of geothermal power plants. The initial cost in U.S. for a large field and power plant is around $2,500 per installed kW and upwards of $5,000/kWe for a small (<1Mwe) power plant. Most of the cost comes from drilling the geothermal wells and creating a successful geothermal reservoir. Since it’s difficult to “see” below the earth’s surface, geothermal developers rely on computer codes called numerical simulators. These codes reduce the uncertainty around where production geothermal wells should be drilled and the flow and thermal recovery characteristics of the reservoir. They also prove to be extremely valuable because of their ability to simulate how a reservoir can be created and sustained, how flow will change over time, the impacts of stimulation and circulation, and ultimately may predict a reservoir’s power generation capability.

Test, Compare, and Refine

Several computer codes exist, but not all codes are created equal. To better understand discrepancies between the codes and improve geothermal models, the DOE’s Geothermal Technologies Office(Offsite link) (GTO) initiated a Code Comparison Study. Led by PNNL engineer Mark White, the GTO Code Comparison Study brought together 11 research institutions to test, compare, improve, and discuss their codes.

The team regularly collaborated on what they called benchmark problems—strictly defined problems that were relatively straightforward to solve using numerical simulators. The numerical simulators generally accounted for coupled thermal, hydrologic, geomechanical, and geochemical processes—all processes relevant to EGS. Because of the team’s various locations, they collaborated using a PNNL-developed software, called GTO-Velo. Each team submitted their unique numerical simulators and tested them against the seven benchmark problems. Using GTO-Velo’s visual comparison tool, participants compared their results against the other simulators. After seeing how their simulation results compared to others, participants had the opportunity to resubmit as many times as they wanted. Through GTO-Velo, the participants created a forum to collaborate and improve their codes.

A Real World “Challenge Problem”

In addition to the benchmark problems, the team tackled a significant challenge facing geothermal development: EGS stimulation. Perhaps the most challenging aspect of EGS is understanding how stimulation occurs and predicting where and when it will result in a geothermal fracture network. Stimulation creates small fractures in the rock beneath the Earth’s surface. Currently, EGS stimulation is thought to occur by opening natural, pre-existing fractures in the rock and occasional hydraulic fractures through pressurized fluid injection.

One example of how the concept of stimulation stumped scientists can be seen at Fenton Hill near Los Alamos, New Mexico. Between 1974 and 1995, Los Alamos National Laboratory researchers created one shallow and one deep reservoir via hydraulic and thermal stimulation at Fenton Hill and conducted extensive EGS testing to understand stimulation mechanisms. Years after their successful experiments, the Fenton Hill researchers were left with a few unresolved questions.

The Code Comparison Study team wanted to see if they could solve some of the scientific mysteries of Fenton Hill; so, the team took on the site as their challenge problem and applied their modern simulation codes to historical Fenton Hill data. Through their analysis, the team investigated several conventional theories about stimulation mechanisms, such as shear dilation of existing fractures. But they also developed or tested several new concepts, such as relationships between joint stiffness and fracture reactivation size, and mixed-mechanism conceptual models for fracturing.

Mixed-mechanism conceptual modeling contradicts conventional theories for EGS creation. Mixed-mechanism hypothesizes fracture networks form through a combination of natural and induced fractures. After natural fractures open, new underground fractures build upon the natural factures as high pressure fluid injection occurs. If correct, the mixed-mechanism conceptual model could provide new understanding to geothermal reservoir simulation, which could help facilitate widespread EGS development.

For more information check out the Benchmark Problems of the GTO Code Comparison Study.

Research Resulting from Code Comparison Study Efforts

Key Capabilities

Published: February 9, 2017

PNNL Research Team

Mark White and Signe White