Progress in coupling a transport solver to global gyrokinetic simulations

POSTER

Abstract

Predictive modeling with transport solvers has tended to use surrogate models, such as quasilinear transport models, to represent turbulent fluxes. One route to improved predictive modeling is instead to couple with direct numerical simulations (DNS) of gyrokinetic turbulence. We have coupled the transport code Tango to the gyrokinetic code GENE. The first results used global gyrokinetic simulations with adiabatic electrons in a circular magnetic geometry, and the transport solver evolved ion pressure. Here, we report on progress in using a DIII-D-like geometry. We also report on progress using kinetic electrons and evolving ion pressure, electron pressure, and density. An additional complication of coupling a transport solver to turbulence simulations rather than a surrogate model is that the fluctuations hinder convergence, and we analyze this issue.

*Work supported by US DOE under Contract DE-AC52-07NA27344, a SciDAC award (Partnership for Multiscale Gyrokinetic Turbulence), and the Exascale Computing Project (17-SC-20-SC).

Presenters

  • Jeffrey Parker

    • Lawrence Livermore Natl Lab

Authors

  • Jeffrey Parker

    • Lawrence Livermore Natl Lab
  • Lynda L LoDestro

    • Lawrence Livermore Natl Lab
  • Lee Ricketson

    • Lawrence Livermore Natl Lab
  • Alejandro Campos

    • Lawrence Livermore Natl Lab
  • J A Hittinger

    • Lawrence Livermore Natl Lab
  • Daniel Told

    • Max Planck Institute for Plasma Physics - Garching
    • MPI for Plasma Physics - Garching
  • Gabriele Merlo

    • University of Texas - Austin
    • UT Austin
  • Frank Jenko

    • Max Planck Institute for Plasma Physics, 85748 Garching, Germany, Technical University of Munich, 85748 Garching, Germany
    • Max Planck Institute for Plasma Physics - Garching, University of Texas - Austin