Applying a CPU/GPU Hybrid Code to Investigate Beam Ion Losses in the DIII-D Tokamak

POSTER

Abstract

A new CPU/GPU parallelized hybrid code provides for fast calculations of particle orbits in the magnetic equilibrium of the DIII-D tokamak, thereby extending the ability to model the effects of energetic ion transport in experiments. A generic distribution input, i.e., initial particle position and velocity vector, allows the code to determine the orbits of millions of particles within minutes of calculation time. Case studies examine neutral beam prompt loss effects, including limiter heating and power density impacting a new helicon antenna. For issues of limiter heating, a protocol is developed to enable between-shot identification of the specific beam responsible for over-temperature observations in select experiments. This capability can also be used to identify at-risk components along the outer wall, including diagnostics that may be impacted by the new counter-current off-axis beam. The helicon simulations identify power densities up to 7 MW/m2, which requires careful design of protective graphite tile shielding. Including the effects of neutralization and non-axisymmetric fields increases accuracy in the wall striking locations of the ions.

Presenters

  • James Tyler Carbin

    University of Maryland, College Park

Authors

  • James Tyler Carbin

    University of Maryland, College Park

  • David Carl Pace

    General Atomics - San Diego, General Atomics

  • Mark Kostuk

    General Atomics, General Atomics - San Diego

  • Igor Sfiligoi

    General Atomics, General Atomics - San Diego