Abstract Details
Abstracts
Author: Andreas Kleiner
Requested Type: Poster
Submitted: 2024-04-12 10:56:30
Co-authors: F. Ebrahimi, A. Pankin, N. Ferraro, J. King, C. Zhao, Y. Liu
Contact Info:
Princeton Plasma Physics Laboratory
100 Stellarator Road
Princeton, New Jersey 08543
United States
Abstract Text:
We provide a benchmark of different extended-MHD codes for simulations of peeling-ballooning (PB) modes in ELMy and non-ELMing H-mode discharges on NSTX. PB modes are typically associated with ELMs and are destabilized by the strong pressure gradient in the H-mode pedestal and resulting strong current density. Extended-MHD effects can crucially affect ELM stability thresholds as was recently discovered on NSTX and JET-ILW. M3D-C1 and NIMROD are two leading nonlinear initial value stability codes that are widely employed for simulations of the macroscopic dynamics of fusion plasmas. Both codes implement two-fluid models that are derived from the Braginskii equations and use advanced numerical solvers. These codes have been benchmarked for simulations of disruptions before. However, for the purpose of code verification and to ensure ELM-stability results are of high fidelity, it is important to benchmark these codes for edge stability simulations, as the edge region can be very challenging for numerical simulation. In addition, MARS-K and MARS-Q implement reduced, linear models that can be used for PB simulations and will be included for a benchmark of linear stability results. M3D-C1 uses streamfunctions and C1 continuous (reduced) quintic finite elements on an unstructured (and adapted) triangular mesh to solve its time-dependent set of MHD equations. NIMROD similarly uses high-order finite elements in the (R-Z) plane and a pseudospectral FFT representation in toroidal direction. The first set of benchmarks is performed with and without equilibrium rotation, at different resistivity values including a comparison with the ideal-MHD limit, and finally with and without finite-Larmor radius effects. Our benchmark comprises linear and nonlinear simulations. We find good agreement in terms of the linear growth rates in between the initial value extended-MHD codes. The nonlinear simulations are ongoing, but during the early phase of the simulation we find similar results.
Comments:
This abstract is for a poster on collaborative work within a SciDAC project I would like to present on behalf of the team in addition to a possible invited talk for which I have submitted an abstract.