An accurate SUPG-stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion

2024

Read An accurate SUPG-stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-12396.txt.

Introduction

Wimmer, Southworth, Sagiyama, and Tang present a new spatial discretization for anisotropic heat conduction in magnetized plasma, in a paper from twenty twenty-four. Such conduction matters for magnetic confinement fusion. Standard continuous Galerkin methods fare worse at the extreme anisotropy found in these plasmas. Their scheme aims to fix that.

The method uses a mixed formulation. It adds an auxiliary variable for the directional derivative along the magnetic field. Both fields sit in a continuous Galerkin space. Streamline upwind Petrov-Galerkin stabilization is applied to both equations. The authors prove the scheme is consistent. In tests tied to fusion it beat past methods clearly. One tokamak case in balance showed only four percent spurious heat loss. Older schemes lost roughly a third of the heat in that case.

Catalog record

YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-12396

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

An accurate SUPG-stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion Abstract We present a novel spatial discretization for the anisotropic heat conduction equation, aimed at improved accuracy at the high levels of anisotropy seen in a magnetized plasma, for example, for magnetic confinement fusion. The new discretization is based on a mixed formulation, introducing a form of the directional derivative along the magnetic field as an auxiliary variable and discretizing both the temperature and auxiliary fields in a continuous Galerkin (CG) space. Both the temperature and auxiliary variable equations are stabilized using the streamline upwind Petrov-Galerkin (SUPG) method, ensuring a better representation of the directional derivatives and therefore an overall more accurate solution. This approach can be seen as the CG-based version of our previous work (Wimmer, Southworth, Gregory, Tang, 2024), where we considered a mixed discontinuous Galerkin (DG) spatial discretization including DG-upwind stabilization. We prove consistency of the novel discretization, and demonstrate its improved accuracy over existing CG-based methods in test cases relevant to magnetic confinement fusion. This includes a long-run tokamak equilibrium sustainment scenario, demonstrating a 35% and 32% spurious heat loss for existing primal and mixed CG-based formulations versus 4% for our novel SUPG-stabilized discretization. Golo A. Wimmer 1 * * * correspondence to: [email protected] , Ben S. Southworth 1 , Koki Sagiyama 2 , Xian-Zhu Tang 1 1 Los Alamos National Laboratory , 2 Imperial College London August 24, 2026 Keywords. Anisotropic heat conduction, anisotropic diffusion, auxiliary operator, continuous Galerkin, SUPG 1 Introduction Plasmas in magnetic confinement fusion exhibit extremely anisotropic heat fluxes [ 24 ] , conducting heat at a rate up to 10 orders of magnitude greater parallel to magnetic field lines than perpendicular [ 18 ] . If the parallel heat flux is not represented accurately in numerical simulations, the resulting increased cross-diffusion perpendicular to magnetic field lines may lead to spuriously short energy confinement times. One way to ensure such an accurate representation is to align the computational mesh with magnetic flux surfaces [ 13 , 23 , 24 ] ; however, this may not be possible in more complex magnetohydrodynamic (MHD) simulations including magnetic islands or even stochastic magnetic field lines [ 14 , 35 ] . In practice, such simulations therefore require numerical methods that accurately represent the parallel heat flux independently of mesh alignment. Due to the importance of anisotropic heat flux in extended MHD and two-fluids simulations for magnetic fusion plasmas, a range of numerical methods have been developed in recent years in order to tackle this challenge. This includes, but is not limited to, improved accuracy through higher polynomial order discretizations [ 16 , 17 ] , mesh refinement at areas of large expected cross-diffusion errors [ 37 ] , asymptotic preserving methods (and extensions thereof), aimed at well-posedness in the limit of infinite anisotropy strength [ 11 , 12 , 25 , 32 , 39 , 40 ] , and first-order hyperbolic system methods including auxiliary variables for the temperature’s gradient components together with pseudo-time advancement terms [ 10 ] (and references therein). Another auxiliary variable-based approach is presented in [ 19 ] , where the auxiliary variable represents the temperature’s directional derivative along magnetic field lines. The same idea is expanded on in our previous work [ 38 ] , where a mixed, discontinuous Galerkin (DG) finite element-based method is derived based on classical DG-upwinding as frequently employed for advection terms in hydrodynamic problems. In particular, we showed the directional gradient’s improved representation resulting from transport stabilization leads to greatly improved accuracy when compared to its non-stabilized counterpart from [ 19 ] in test cases with extreme anisotropy. Many of the aforementioned approaches are based on the finite element method, which is often considered in magnetic confinement fusion simulations due to, for instance, their flexibility in meshing complex geometries. Next to the aforementioned DG-based work in [ 16 , 17 , 38 ] , this includes additional DG-based approaches such as hybrid DG [ 15 ] and local DG [ 22 ] . However, DG-based methods may lead to challenges such as an increased number of degree of freedoms (DOFs), as well as an increased complexity in terms of implementation when compared to continuous Galerkin (CG) methods. Further, DG-based implementations often require interior penalty formulations, leading to a potential loss in the order of accuracy. For these reasons, and since magnetic confinement fusion scenarios generally involve a shock-free low Mach number regime, many finite element-based MHD codes used to simulate such regimes, such as [ 5 , 7 , 23 , 35 ] , deploy CG (or higher order regularity Galerkin) rather than DG methods. In this paper, we present an extension of our previous DG-based upwind-stabilized work in [ 38 ] to CG-based discretizations. As in [ 38 ] , we consider a mixed formulation of the anisotropic heat flux, using a form of the temperature’s directional gradient along magnetic field lines as the auxiliary variable. Analogously to upwinding in the DG case, we apply a form of the Streamline Upwind Petrov-Galerkin (SUPG) method [ 8 ] for the two directional derivatives comprising the anisotropic diffusion operator. We motivate our particular choice of SUPG formulation and further demonstrate consistency of the resulting spatial discretization for the anisotropic heat flux equation. Additionally, in a series of numerical tests, we validate its order of accuracy as well as superior performance in realistic test cases when compared to a primal weak formulation and the mixed formulation of [ 19 ] . In particular, we study the spread of a temperature perturbation in a 2D magnetic flux surface, an MHD equilibrium in a full-torus tokamak domain including a temperature-dependent parallel conductivity coefficient, as well as the spread of a temperature perturbation through a flux tube in the latter domain. This includes a long run full-torus tokamak equilibrium scenario, demonstrating a 35% and 32% spurious heat loss for existing primal and mixed CG-based formulations versus 4% for our novel SUPG-stabilized discretization. Further, the tests contain parameter scales suitable for MHD models, including the choice of time step and parallel conductivity. The remainder of this paper is structured as follows: In Section 2 , we review the anisotropic heat conduction equation and include a brief description of standard CG-based primal and mixed space discretizations. In Section 3 , we introduce our novel SUPG-based scheme, including a discussion on its derivation as well as consistency and the choice of SUPG stabilization parameter. In Section 4 , we present and discuss numerical results. Finally, in Section 5 , we review our results and discuss possible future work. 2 Background In this section, we briefly review the anisotropic heat flux equation. Further, we recall existing continuous Galerkin-based methods for the equation’s spatial discretization, with which we will compare our novel method in the numerical results section. Anisotropic heat flux equation. We consider the evolution of temperature T within a plasma domain \Omega as governed by a heat flux \mathbf{q} and forcing S according to \displaystyle\frac{\partial T}{\partial t}-\nabla\cdot\mathbf{q} \displaystyle=S, (2.1a) \displaystyle\mathbf{q} \displaystyle=\kappa_{\parallel}\nabla_{\parallel}T+\kappa_{\perp}\nabla_{\perp}T, (2.1b) for parallel and perpendicular heat conductivities \kappa_{\parallel} , \kappa_{\perp} , respectively. The directional gradients are given by \nabla_{\parallel}T=\mathbf{b}\big(\mathbf{b}\cdot\nabla T\big),\hskip 28.45274pt\nabla_{\perp}T=\nabla T-\nabla_{\parallel}T, (2.2) where \mathbf{b}=\mathbf{B}/|\mathbf{B}| denotes the normalized

Read the complete text