Abstract
Here we present first results simulating plasma filaments in non-axisymmetric geometries, using a fluid turbulence extension of the BOUT++ framework. This is made possible by the implementation of the flux coordinate independent (FCI) scheme for parallel derivatives, an extension of the metric tensor components which allows them to vary in three dimensions, and development of grid generation. Tests have been performed to confirm that the extension to three dimensional metric tensors does not compromise the accuracy and stability of the associated numerical operators. Recent changes to the FCI grid generator in BOUT++, including a curvilinear grid system which allows for potentially more efficient computation, are also presented. Initial simulations of seeded plasma filaments in a non-axisymmetric geometry are reported. We characterize filaments propagating in the closed-field-line region of a low-field-period, rotating ellipse equilibrium as inertially-limited by examining the velocity scaling and currents associated with the filament propagation. Finally, it is shown that filaments in a non-axisymmetric rotating ellipse equilibrium propagate in a toroidally nonuniform fashion, and it is determined that the long connection lengths in the scrape-off-layer enable parallel gradients to establish, which has consequences for interpretation of experimental data.
Highlights
Neoclassical transport is the dominant loss mechanism in sufficiently hot stellarator plasmas and can dominate in the plasma core [1]
Here we present first results simulating plasma filaments in non-axisymmetric geometries, using a fluid turbulence extension of the BOUT++ framework
It is shown that filaments in a nonaxisymmetric rotating ellipse equilibrium propagate in a toroidally nonuniform fashion, and it is determined that the long connection lengths in the scrape-off-layer enable parallel gradients to establish, which has consequences for interpretation of experimental data
Summary
Neoclassical transport is the dominant loss mechanism in sufficiently hot stellarator plasmas and can dominate in the plasma core [1]. The recent implementation of the flux coordinate independent (FCI) [10] method for parallel derivatives in BOUT+ + has allowed for simulations in non-axisymmetric geometries [11, 12]. We present the first results simulating plasma fluid turbulence in non-axisymmetric geometries, made possible by extensive modifications to the BOUT++ framework [13, 14]. For an accurate simulation of plasma dynamics in stellarators, BSTING must include metric components which are fully three dimensional This extension to three dimensions is simple in principle (and was mentioned in the introduction of the original BOUT+ + paper [13]), but the geometrical components are integral to many different parts of the code, and the work presented here has required extensive modifications to the framework. The following section provides initial tests for the implementation of these methods
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have