National Center for Computational Sciences

The Department of Energy will dedicate 1 million hours of supercomputer processing time to better understand the movement of the uranium plume in the ground water at the Hanford nuclear reservation just north of Richland.

Monday, DOE announced 45 projects that have been awarded time at supercomputers at four sites, including Pacific Northwest National Laboratory in Richland.

About 96 acres of ground water beneath the 300 Area of Hanford near the Columbia River are polluted with uranium. The area once was used to fabricate uranium fuel for the production of plutonium for the nation’s nuclear weapons program.

The supercomputer project, which will be done on a computer at the Oak Ridge National Laboratory in Tennessee, will provide a model of how the uranium contamination is moving and changing.

It also will be an opportunity for researchers to advance the techniques used to simulate ground water flow and contaminant transport, said Glenn Hammond, a PNNL senior research scientist on the project. Researchers from four national laboratories and the University of Illinois will work on the project.

By using a larger computer, researchers will be able to consider increased detail in the model. That will include layering of different soil types in the subsurface of the 300 area, different chemical reactions such as radioactive decay and how much of the uranium is likely to attach to soil.

The 1 million hours of supercomputing time assigned to the project would take 114 years on a single-processor desktop computer. Completing the project will require an additional 7 million hours. Researchers also have been working on creating the computer code for the project with a DOE five-year grant of about $4 million.

“I’m confident we can get a more accurate picture” of the behavior of the uranium contamination, Hammond said.

That should be useful as DOE considers how to clean up the contamination to protect the Columbia River. DOE had expected the plume to naturally dissipate, but that has not happened. Scientists may be able to determine what characteristics of the contaminants and contaminated area make a difference, leading to better information about what characterization data is important and what is not.

PNNL also is involved in two other supercomputing time awards announced Monday.

Steven Ghan of PNNL, in collaboration with researchers from several other institutions, is working on future climate predictions based on changes in energy policies. The project was awarded 5.5 million hours on Oak Ridge supercomputers.

The PNNL supercomputer will dedicate 750,000 processor hours to a Corning Inc. project that will investigate the flow and deformation of dense suspensions.

The project has applications in many industrial processes, ranging from ceramics to polymers and from the food industry to pharmaceuticals.

Source: Tri-City Herald
By: Annette Cary