Scientists at the U.S. Department of Energy’s (DOE) Princeton Plasma Physics Laboratory (PPPL) who have developed a novel computer simulation framework and code predicting plasma behavior throughout the entire volume of a fusion machine were part of a group that won a DOE Secretary’s Honor Award for Achievement. 

This prestigious award is given to teams of federal employees, contractors or military members who “together accomplished significant achievements on behalf of the department.” Those teams should also have demonstrated “cooperation and teamwork attaining their goals.” Winners receive a commemorative plaque and a certificate signed by the U.S. Secretary of Energy.

The computational project was part of a larger DOE project focused on adapting existing fusion simulation programs so that they could make scientific discoveries through simulations that would include richer science and run more quickly on faster computers. This Exascale Computing Project (ECP) operated from 2016 to 2023 and focused on creating both exascale computing hardware and software that can run on the new hardware, as well as ways that researchers can use exascale computing to advance science. These new machines can perform 1 quintillion, or 1 million million million, operations per second and could power discoveries in a range of disciplines. 

The computer program, known as Whole Device Model Application (WDMApp), is the first code that simulates both the outer and inner regions of fusion plasma simultaneously using a technique known as “tight code coupling.” This technique combines two or more codes using the fundamental equations of plasma physics, making them depend highly on one another. This capability means that scientists can get a more detailed understanding of plasma behavior than before and design fusion facilities to operate more efficiently.

Specifically, WDMApp combines the modeling capabilities of three codes. Two of them simulate the plasma core: the Gyrokinetic Electromagnetic Numerical Experiment (GENE) code, developed at Germany’s Max Planck Institute for Plasma Physics, and the Gyrokinetic ElectroMagnetic (GEM) code, developed at the University of Colorado Boulder. The third code is the X-Point Included Gyrokinetic Code (XGC), primarily developed by PPPL Managing Principal Research Physicist Choongseok (CS) Chang and his team, which simulates the plasma edge.

A combination of simulations 

Scientists want a code that models how the core and edge interact in fusion devices called tokamaks because the plasma behavior in the core is affected by what happens at the edge and vice versa, including disruptions that can damage internal components and the flow of waste heat exiting the device. 

Increasing understanding of these phenomena can help scientists design future fusion facilities that can operate for long periods without needing maintenance, increasing efficiency and lowering costs. “Unless we model the entire plasma in a fusion device, we lose the connection between core and edge and then our simulations lose accuracy,” said Chang, who was the co-lead principal investigator of the project.

“It was a great privilege to lead the WDMApp team,” said Amitava Bhattacharjee, a professor of astrophysical sciences at Princeton University and former head of the Theory Department at PPPL. “Participating in the ECP was an exciting experience for all of us helping usher in the exascale era. I thank every member of the WDMApp team and the ECP community at large for their contributions.”

“Historically, more powerful computers have allowed scientists to simulate fusion processes with more detail than before, allowing predictions of plasma behavior that can improve the operations of fusion facilities,” Bhattacharjee said. “The development of WDMApp — the first time that scientists have stitched together core and edge fusion codes — will continue this tradition.”

Bhattacharjee and Chang would like acknowledge the valuable contributions of many individuals, including Julien Dominski, Stéphane Ethier, Robert Hager, Greg Hammett, Seung-Hoe Ku, Albert Mollen, Aaron Scheinberg, Benjamin Sturdevant, Pallavi Trivedi, and George Wilkie. Bhattacharjee and Chang said that the work would not have been possible without key collaborations between PPPL and researchers from Argonne National Laboratory, the University of Texas at Austin, the University of Colorado Boulder, Lawrence Livermore National Laboratory, Oak Ridge National Laboratory, Rutgers University, Jubilee Development, Rensselaer Polytechnic Institute, the University of New Hampshire and the University of Utah.