Skip to main content
Research Highlight

Building better geothermal models

Topics: Clean Energy Fossil Energy Renewable Energy

from ORNL Review, vol 49, no 3, 2016

The ingredients that make geothermal energy suitable for electricity production are simple: heat, permeability (usually fractured rock) and water. Until recently the trick has been finding all in the same place and in the right amounts.

Enormous heat energy is available between 3 and 10 kilometers below the surface of the U.S., notes Yarom Polsky, an engineer in ORNL’s Energy and Environmental Sciences Directorate, but researchers haven’t yet determined how to extract it. Current thinking suggests we will need to develop enhanced or engineered geothermal systems. These are systems that lack at least one of the ingredients—usually water or fractured rock. 

To create an enhanced geothermal system—or EGS—an energy producer would drill a well down to where the rocks are hot enough (in the range of 250–300 degrees Celsius or roughly 500–550 degrees Fahrenheit) and then hydraulically fracture the rock to create permeability. 

In the oil and gas industry this permeability allows oil and gas to be recovered. In EGS wells, permeability allows water pumped down the well to flow through the fractures and absorb heat from the surrounding rock. The heated water will then be pumped out of the ground at a second well that intersects the fracture network and be used to generate electricity by powering a steam turbine.

Polsky and his colleagues hope to improve energy producers’ ability to analyze proposed EGSs before drilling begins by producing more accurate computer models of how water flows through these systems under a variety of conditions.

Traditionally, energy producers wanting to estimate the rate of flow through a proposed EGS have had to rely on idealized models of fluid flow—models that are often inaccurate. To address this shortcoming, Polsky and his colleagues are using neutrons generated by ORNL’s High Flux Isotope Reactor to develop a more sophisticated way to visualize the flow of fluid through the complex fractures that occur in real geological samples.

“We're developing a way to visualize the flow through more complicated systems so we can use that as a basis for calibrating flow models,” Polsky said.

Typically, researchers experimentally measuring flow through complex geometries have had to use materials they could see through, usually machined plastics, which limited both the complexity of the systems and the pressures they could use. For Polsky and his colleagues, the compelling thing about visualizing flow with neutrons is that rock samples are transparent to neutrons, while water is not, enabling researchers to see the water inside actual rock samples rather than through plastic surrogates. 

Because Polsky and his colleagues were using neutrons to image continuously moving flows, they had to develop methods that used contrast agents that could be tracked. In this case they injected small bubbles of a fluid called Fluorinert that looks different from water to the neutrons. Then they captured neutron radiographs of the rock sample about once every 10 milliseconds. By doing this they could visualize the motion of the fluid.

“If we can calculate the velocity fields within actual complex fracture geometries, then we can validate computational fluid dynamics models that help predict how fluid will flow through the system,” Polsky said. “This becomes the basis for estimating the system’s heat production.”

Not only are rocks transparent to neutrons, but so is the pressure vessel that holds the rock, meaning Polsky and his colleagues can match the conditions of the sample being studied to those of the rocks in the proposed EGS. This is true for pressures up to 10,000 pounds per square inch and temperatures up to 350 degrees Celsius, or about 660 degrees Fahrenheit. 

“The problem with studying subsurface systems is that we never really know what we have down there,” Polsky said. “However, we can come up with a range of possibilities. I’ve done a lot of modeling and simulation work, and we always get an answer. Whether or not the answer reflects reality is another question. 

“We need a way to validate the results of our simulations—either with field experiments or with laboratory experiments. That’s what we’re providing with this research.”