National Center for Computational Sciences

Understanding and controlling the turbulent transport of hot plasmas is critical for the success of a fusion reactor such as the international ITER project. ITER’s performance will depend on its ability to retain heat, which depends in turn on the balance between heat loss from the turbulence of the heat transport, and the heating caused by the fusion reactions in the plasma. The design relies on the formation of a “transport barrier” in the burning plasma, which the plasma can’t cross easily; this allows for a separation between the low confinement plasma near the toroidal wall, where temperature and pressure is not optimal for fusion conditions, and the high temperature and pressure confinement of the plasma at the inner core of the tokamak. The physical mechanisms for forming the barrier have been studied with some success, but so far scientists lack a predictive model for the barrier formation conditions they observe at ITER and for other future experiments. To achieve a deeper understanding of barrier formation in ITER, they must implement a global kinetic simulation with sufficient resolution.

This study will be carried out by the principal investigators from three SciDAC fusion projects: Zhihong Lin of the Center for Gyrokinetic Simulation of Energetic Particle Turbulence and Transport (GSEP), Patrick Diamond of Gyrokinetic Particle Simulation of Turbulent Transport in Burning Plasmas (GPS-TTBP), and Choong-Seock Chang of the Center for Plasma Edge Simulation (CPES). The team will use 30 million hours on the ORNL’s petascale Jaguar supercomputer to address key physics and computational issues identified by the three SciDAC projects. Using the Gyrokinetic Toroidal Code, these studies will provide important insight into the formation and dynamics of the transport barrier in ITER and will optimize subsequent experiments. This work will also advance the frontier of computational science in the areas of data management, statistical analysis, and advanced visualization.