Skip to Main Content U.S. Department of Energy
Energy and Environment Directorate
Page 17 of 27

Research Highlights

October 2011

Supercomputer simulation tool reaches major milestone


This graphic shows a simulation that predicts the extent and direction of uranium contamination from a waste site 10 years into the future. The simulation was conducted on a high-performance computing system, simultaneously using 256 processors-the average personal computer uses just one processor. The red and yellow tones are uranium concentrations; the blue is a tracer. The tracer shows the spread of uranium-contaminated groundwater, taking into account that some of the uranium binds with the sediments and that the tracer travels at the same speed as water. ASCEM research is enabling this type of detailed and accurate simulation.

Results:

Researchers at PNNL and four other national laboratories recently demonstrated an early prototype of an advanced simulation capability that will better predict the long-term behavior of subsurface contaminant plumes. The Advanced Simulation Capability for Environmental Management (ASCEM) marries petaflop supercomputing—a thousand trillion operations per second—with new capabilities enabling users to integrate data, software and the latest understanding of subsurface science. ASCEM employs a transformational approach that will help better evaluate cleanup approaches and estimate how long cleanup will take, thereby reducing uncertainties, risks, and costs.

Why it matters:

DOE's Office of Environmental Management (EM) is addressing one of the largest environmental challenges in the world—the safe cleanup of legacy waste from our nation's five decades of nuclear weapons production and nuclear energy research. ASCEM is being developed to enhance cleanup decisions and reduce lifecycle costs on some of our country's most challenging nuclear waste contamination issues. The modular and open- source approach being used by ASCEM will result in faster development times, reduced lifecycle costs, and broad community involvement in the cleanup process. ASCEM is being designed to support DOE-EM's cleanup efforts, but also may end up helping DOE Fossil Energy's goal of sequestering carbon dioxide emissions, and DOE Nuclear Energy's goal of evaluating high-level nuclear waste repositories. ASCEM was recently demonstrated on a plume containing radioactive and hazardous contaminants, including uranium, Strontium-90, and Iodine 129, at the Savannah River Site in Aiken, South Carolina.

Methods:

Unlike previous simulators, ASCEM is being designed to incorporate new and emerging parallel supercomputer architectures-systems that break a problem into parts and use many processors at the same time to reduce simulation times. Most other subsurface simulators have been modified to work on parallel architectures and don't have the power that a simulator specifically designed for parallel architectures has. Additionally, ASCEM is being developed in a modular format with open-source codes so the subsurface science community can use the simulation capability for a variety of subsurface flow and transport problems. By making ASCEM consistent and easy to use, it can be employed throughout the DOE-EM cleanup effort to improve efficacy and reduce overall costs.

What's Next:

The ASCEM team is developing the first user release prototype, a code version for researchers to begin working with, and is planning a second series of demonstrations that include modeling and visualization of deep vadose zone contamination at the Hanford Site in eastern Washington.

Acknowledgments:

This research is funded by the U.S. Department of Energy, Office of Environmental Management. A team of scientists from five national laboratories including PNNL, Los Alamos National Laboratory, Lawrence Berkeley National Laboratory, Oak Ridge National Laboratory and Savannah River National Laboratory are collaborating on the development of ASCEM. The DOE Office of Science community provides both scientific data for the models and tools for accelerated development of ASCEM.

Reference:

Williamson, M, J Meza, D Moulton, I Gorton, M Freshley, P Dixon, R Seitz, C Steefel, S Finsterle, S Hubbard, M Zhu, K Gerdes, R Patterson, and YT Collazo. 2011. "Advanced simulation capability for environmental management (ASCEM): An overview of initial results." Technology and Innovation 13:1-100.

PNNL-SA-83193


Page 17 of 27

Our Business

About Energy and Environment