Electrostatic quasi-neutral formulation of global cross-separatrix particle simulation in field-reversed configuration geometry

Physics of Plasmas ARTICLE
Electrostatic quasi-neutral formulation of global cross-separatrix particle simulation in field-reversed configuration geometry
scitation.org/journal/php
Cite as: Phys. Plasmas 27, 082504 (2020); doi: 10.1063/5.0012439 Submitted: 30 April 2020 . Accepted: 14 July 2020 .
Published Online: 4 August 2020
1
1 1,2
1,3
1,a) 1 2 2
C. K. Lau, D. P. Fulton, J. Bao, Z. Lin, S. Dettrick, AFFILIATIONS
M. Binderbauer, T. Tajima,
and L. Schmitz
1TAE Technologies, Inc., Foothill Ranch, California 92610, USA
2Department of Physics and Astronomy, University of California, Irvine, California 92697, USA 3Department of Physics and Astronomy, University of California, Los Angeles, California 90095, USA
Note: This paper is part of the Special Collection: Building the Bridge to Exascale Computing: Applications and Opportunities for Plasma Science
a)Author to whom correspondence should be addressed: clau@tae.com
ABSTRACT
A quasi-neutral blended drift-Lorentz particle model of the field-reversed configuration (FRC) has been developed and implemented in the particle-in-cell code named ANC. A field-aligned mesh and corresponding mesh operations are constructed for solving self-consistent elec- tric fields in FRC geometry. Particle dynamics are described in cylindrical coordinates to allow for cross-separatrix simulation coupling the core and scrape-off layer regions of the FRC. This new model is successfully verified against analytically derived dispersion relations, and FRC turbulence is studied using the blended model for the first time.
Published under license by AIP Publishing. https://doi.org/10.1063/5.0012439
I. INTRODUCTION
An ideal field-reversed configuration (FRC) is a compact toroid (CT) with purely poloidal magnetic fields. It consists of two regions separated by a separatrix: an inner, closed field-line core region and an outer, open field-line scrape-off layer (SOL) region. The FRC concept as a fusion reactor has several benefits: (1) b (the ratio of plasma pressure to magnetic energy density) near unity suggests more economical magnetic energy costs; (2) compact shape simplifies the device hull and external magnetic field coils; (3) SOL may be connected to the divertor arbitrarily far from the FRC core; and (4) the lack of toroidal magnetic fields radically changes the magnetic topology and the consequential stability of the plasma.
In the C2/C2U FRC experiments at TAE Technologies, Doppler backscattering (DBS)1 measurements of density fluctuations showed distinct qualities between the core and SOL. In the SOL, the fluctua- tion spectrum is highest in amplitude at ion-scale wavenumbers, decreasing exponentially toward electron-scale wavenumbers. In the core, fluctuations are overall lower in amplitude relative to the SOL,
Phys. Plasmas 27, 082504 (2020); doi: 10.1063/5.0012439 Published under license by AIP Publishing
and the highest amplitude fluctuations are located at electron-scale
2
wavenumbers.
Early efforts to understand this phenomenon in the FRC used
linear GTC [the Fortran based gyrokinetic (GK) toroidal code3] simu- 4
lations localized to drift-surfaces using gyrokinetic particles. Within
gyrokinetically valid parameters (hqi =Ln i < 1), these simulations
show the FRC core to be stable while various instabilities were able to
arise at a range of toroidal wavenumbers and pressure profiles in the
5,6
simulations were insufficient for turbulence simulations due to the coupled nature of the core and SOL. In order to address the coupled nature of the core and SOL regions, a new simulation model spanning across the separatrix has been developed. The formulation described in this paper is implemented as the particle-in-cell (PIC) code named “ANC.” Closely related to and based on the parallelism of GTC, this code is written in Fortran and uses OpenMP/message passing interface (MPI) parallelization while updates for graphics processing unit (GPU) acceleration are in progress. Unlike the more “device agnostic”
27, 082504-1
SOL.
Although consistent with experiment, the localized linear GTC