High-Fidelity Gyrokinetic Simulation of Tokamak and ITER Edge Physics

PI Choongseok Chang, Princeton Plasma Physics Laboratory
n experimental tokamak fusion plasma

Poloidal cross-sectional view of the perturbed plasma density representing electrostatic turbulence in the edge of a realistic tokamak reactor plasma. The connected streamer structures inside the magnetic separatrix surface (black line) and the isolated blobby structures around and outside the magnetic separatrix can be observed in the locally enlarged box. The nonlinear turbulence structures shown here in the edge plasma form the basis for the plasma confinement physics that determines the fusion yield in a fusion reactor. Image: Seung-Hoe Ku, Princeton Plasma Physics Laboratory; David Pugmire, Oak Ridge National Laboratory

This multi-year INCITE project seeks to advance the understanding of the edge plasma physics critical to fusion reactors, with a focus on ITER, an international collaboration to design, construct, and assemble a burning plasma experiment that can demonstrate the scientific and technological feasibility of fusion.

The team is applying its 5D gyrokinetic particle code, XGC1, on DOE leadership computing resources to address some of the most difficult plasma physics questions facing ITER and future fusion reactors. In particular, they are performing studies on three high-priority challenges: (1) quantifying the narrowness of the heat-flux width on the ITER divertor material plates in the high-confinement mode (H-mode) operation during tenfold energy gain operation; (2) understanding the basic physics behind the low-to-high mode L-H transition and pedestal formation at the edge, which is necessary to achieve a tenfold energy gain in ITER; and (3) studying the control of edge localized instabilities that could damage the material wall in ITER and future magnetic fusion reactors. These large-scale studies are time- urgent for the successful planning of ITER operation and require an intensive, concentrated computing effort using extreme-scale supercomputers.