Abstract Details
Abstracts
Author: David Green
Requested Type: Poster
Submitted: 2017-03-17 16:08:59
Co-authors:
Contact Info:
Oak Ridge National Laboratory
P.O. Box 2008
Oak Ridge, TN 37831-6
USA
Abstract Text:
Recent efforts to move toward quantitative prediction of high power RF antenna designs for fusion applications has meant dual requirements of resolving the geometric details of the antenna and other plasma facing surfaces, together with the hot plasma kinetics of the core plasma response. The typical approach to this problem is to couple a spectral solver for the core plasma, and a finite-difference or finite-element cold-plasma representation (no wavenumber dependence in the conductivity). While these approaches have seen some success, they are still limited by several fundamental restrictions of the Fourier spectral method for core kinetics. These include the numerical limitations of uniform resolution in any spectral direction, matrices approaching dense (or indeed fully dense), and N^3 work scaling to invert such dense matrices, as well as physics limitations including the stationary-phase approximation. An alternative approach was presented by [Green et al., Comm. Phys. Comm., 185, 736 (2014)] where the computational power of modern computing architectures was leveraged in an operator-split, iterative approach to the inclusion of kinetic effects to existing cold-plasma full-wave solvers. Such an approach allows a single-domain solution where kinetic effects can be included only where they are required, as with mesh resolution. Here we examine the stability and convergence properties of the iterative component of this approach on a fusion relevant benchmark case containing mode-conversion, kinetic ion waves, and electron Landau damping.
This research used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.
Comments: