Oak Ridge National Laboratory


News Release

Media Contact: Ron Walli (wallira@ornl.gov)
Communications and External Relations


ORNL supercomputer helping shed brighter light on fusion

OAK RIDGE, Tenn., Sep. 11, 2000 — Some of the secrets to achieving fusion energy may be unlocked by calculations developed by researchers at the Department of Energy's Oak Ridge National Laboratory (ORNL) and run on the lab's supercomputer.

Fusion energy, evident in the sun and stars, is the ultimate source of power because it provides an environmentally acceptable alternative to energy generated by fossil fuels. The work of Fred Jaeger and Lee Berry of ORNL's Fusion Energy Division is significant because it enables scientists to better understand radio waves in the plasma that would be at the heart of a fusion power plant.

"Our research is allowing us to study the high-power radio waves we use in fusion research experiments," Jaeger said. "This newly gained knowledge should help us get a clearer picture of the physics of the heating system and control for a fusion machine."

One of the techniques used to transform the fuel into the plasma state needed for fusion is to use intense electromagnetic waves, much as a microwave oven heats food. But because instruments cannot be placed inside plasma, which is more than 500 million degrees Fahrenheit, experimental measurements on the waves used for heating must be indirect.

"It's essential to have a good theoretical understanding of the wave behavior and to be able to calculate it accurately," Jaeger said. "But computing these waves is difficult because the particles are so hot they move at almost the speed of light. This motion makes it difficult to calculate how the plasma particles will respond to the wave and how much electric current they will produce."

Until now, researchers wanting to calculate the effects of radio waves in plasma have been forced to either ignore the variation of the plasma in all but one direction or consider just waves having long wavelengths and low frequencies.

"The first choice, treating the plasma as one-dimensional, is akin to adopting tunnel vision," Berry said. "The wave is computed along a single line through the plasma, but we don't get a picture of what is occurring in the whole plasma cross-section."

The second choice, Berry said, eliminates from consideration many of the wave processes that are of most importance in today's fusion experiments that require high frequencies and can have very short wavelengths.

With the technique developed at ORNL, they can compute plasma waves across an entire plasma cross-section. It does not require any restriction on wavelength or frequency.

"With this approach, the limit on attainable resolution comes not from the theory, but from the size and speed of the available computer required to solve the enormous sets of equations," Jaeger said.

Working with Ed D'Azevedo of ORNL's Computer Science and Mathematics Division, Jaeger and Berry have developed a computer program to solve the equations that take advantage of the massively parallel structure of modern supercomputers.

They obtained the first high-definition picture in two dimensions of a donut-shaped fusion device called a tokamak for a process called "mode conversion." In this process, researchers inject radio waves from outside the device. At a certain location, the waves suddenly change character to a different type of wave having very fine scale structure and are absorbed by the plasma.

The solutions were obtained running on ORNL's 1 trillion operation-per-second (teraflop) IBM RS/6000 SP supercomputer, which with Berry and Jaeger's program achieved speeds of 650 billion operations per second.

"These calculations are considered to be a breakthrough for wave studies in fusion machines," said Don Batchelor, section head of the theory group in the Fusion Energy Division.

The new computer program provides high-resolution pictures that clearly detail the formation of the short wavelength structures and how the various waves propagate, reflect and are absorbed in the plasma. Researchers expect the new technique to be useful to much more complex plasma shapes than the tokamak type used in this experiment.

In the quest for fusion, scientists have attained a number of important milestones. They have achieved temperatures as high as 520 million degrees, more than 20 times the temperature at the center of the sun. And more than 16 million watts of fusion power have been produced in the laboratory. Their next tasks are to demonstrate sustained reactions that produce substantial amounts of energy and to build and demonstrate a fusion power plant. The ORNL work will contribute to the application of waves for sustaining and controlling fusion plasmas.

The research is funded by DOE. ORNL is a DOE multiprogram facility operated by UT-Battelle.