Pith. sign in

REVIEW 3 major objections 5 minor 41 references

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

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper presents a stabilized continuous Galerkin mixed scheme for anisotropic heat conduction that cuts spurious tokamak heat loss to about 4%, versus 35% and 32% for standard CG formulations.

desk verdict The SUPG-mixed CG scheme is a genuine, well-tested improvement for anisotropic heat transport in fusion — the 4%-vs-32% result is credible — but the consistency proof does not cover the implemented scheme because tau is cell-wise discontinuous. read the letter →

arxiv 2412.12396 v1 pith:AJ5NFK7L submitted 2024-12-16 math.NA cs.NA

Golo A. Wimmer , Ben S. Southworth , Koki Sagiyama , Xian-Zhu Tang This is my paper · ORCID
classification math.NAcs.NA MSC 65M6065N3065N1276X05
keywords anisotropicheatconductioncontinuousGalerkinSUPGstabilizationmixedformulationmagneticconfinementfusiontokamaksimulationspuriouscross-fielddiffusiondirectionalderivativeauxiliaryvariable
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that a continuous Galerkin (CG) discretization of the anisotropic heat conduction equation can be made nearly as accurate as a discontinuous Galerkin one by adding streamline upwind Petrov-Galerkin (SUPG) stabilization to a mixed formulation. The mixed formulation introduces an auxiliary field for the temperature's directional derivative along the magnetic field, and both the temperature and auxiliary equations receive SUPG stabilization. The authors prove that their discrete scheme is consistent with the strong form of the heat equation, and in a long-run tokamak equilibrium test the stabilized scheme loses only about 4% of the available internal energy to spurious cross-field heat transport, compared with 35% and 32% for the standard primal and mixed CG formulations at matched resolution. The payoff is a practical way to keep the favorable implementation properties of continuous finite elements while controlling the dominant numerical error in strongly anisotropic fusion simulations.

What carries the argument

The workhorse is the SUPG-modified mixed CG scheme (3.8): the temperature $T_h$ and the auxiliary field $\zeta_h$, a scaled directional derivative $\sqrt{\kappa_\Delta}\,b\cdot\nabla T$, both live in continuous Galerkin spaces, and test functions are modified to $\gamma + \tau s\cdot\nabla\gamma$ with $s=\sqrt{\kappa_\Delta}\,b$. The SUPG-modified bilinear forms $M^a$, $M^f$, and $G^f_\parallel$ carry the directional derivative and, after the auxiliary variable is eliminated, the anisotropic term has the form $(G^f_\parallel)^T (M^f)^{-1}(G^f_\parallel + G^f_{\parallel,b})$, mirroring a discrete diffusion operator. The stabilization parameter $\tau = (2/\sqrt{\delta t} + k\sqrt{\kappa_\Delta}/\delta x)^{-1}$ is chosen so the modification is nondimensional and vanishes in the space-time refined limit, in which the scheme reduces to the standard mixed method.

What would settle it

Run the tokamak equilibrium test at the same resolution and time step but with a substantially larger parallel conductivity (say, 10 or 100 times) and check whether the relative available temperature at the final time stays near 0.96 or drops appreciably; a clear drop would indicate that the $\tau$ formula, rather than the structural SUPG improvement, is carrying the result. A second check is to run the same test with $\tau$ forced to a very small value, since the scheme then reduces to the standard mixed method (2.9), which loses about 32%.

Watch

Extended reading notes

Core claim

The central claim is that applying SUPG-type stabilization to both equations of the mixed formulation produces a consistent spatial discretization whose discrete anisotropic operator keeps the same transpose-gradient / mass-inverse / gradient structure as the underlying parabolic operator, up to consistency-related modifications, and that preserving this structure is what keeps the directional derivative along the magnetic field accurately represented. With temperature in a continuous Galerkin space of degree k and the auxiliary variable also in a continuous space, the scheme is proven consistent with the strong solution (Proposition 3.1). In the full-torus tokamak equilibrium sustainment test at second order and matched resolution, the stabilized scheme keeps about 96% of the available temperature, while the primal and standard mixed CG formulations keep about 65% and 68%, respectively; at lower resolution the gap is even larger, with the stabilized scheme retaining roughly 84-86% versus roughly 22-29%.

Load-bearing premise

The scheme's accuracy rests on the heuristic formula for the stabilization parameter $\tau$ in (3.15); the paper proves consistency but not stability or ellipticity, and warns that an uncareful choice of $\tau$ can introduce instabilities, so if (3.15) is not robust across the conductivity and time-step regimes used in fusion simulations, the reported 4% heat loss may not generalize.

Editorial extensions

If this is right

  • Existing CG-based MHD codes can adopt the scheme with only a test-space modification and one extra continuous auxiliary field, without switching to discontinuous Galerkin degrees of freedom or interior penalty terms.
  • At the coarse resolutions typical of resistive MHD simulations run on dissipative time scales, the scheme keeps spurious thermal energy loss at a few percent rather than tens of percent.
  • The accuracy advantage persists under temperature-dependent parallel conductivity, where standard formulations are amplified by a self-reinforcing cross-diffusion cycle; with the conductivity limiter removed, the standard formulations deteriorate far more than the SUPG one.
  • A fourth-order primal CG run at the same degree-of-freedom budget still loses about 10% of the available temperature, five times more than the lowest-order SUPG scheme in the same tokamak test.
  • In the flux-tube perturbation test, the stabilized scheme keeps the temperature perturbation between the expected bounding flux surfaces, while the standard mixed formulation leaks heat across them.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the same SUPG-modified mixed structure could be applied to other strongly anisotropic elliptic operators in plasma models, such as resistive diffusion or separate ion and electron energy equations, wherever a directional derivative must be represented accurately on meshes not aligned with the magnetic field.
  • Editorial inference: the stabilization parameter $\tau$ is the main tuning knob, and a natural testable extension is the paper's own suggestion to replace the local cell length $\delta x$ by the average length of the magnetic field line segment through each cell; one could check whether that makes the scheme less sensitive to mesh anisotropy.
  • Editorial inference: because the method loses symmetry and ellipticity, robust use in production codes will likely require solvers specialized for non-symmetric transport-dominated systems rather than standard symmetric multigrid, which the paper flags as future work.
  • Editorial inference: since the method reduces to the standard mixed scheme as $\tau\to 0$, an adaptive choice of $\tau$ driven by a local error indicator could be tested to see whether the reported gains survive at even coarser grids or larger time steps.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes a continuous Galerkin (CG) discretization for the anisotropic heat-conduction equation, using a mixed formulation in which an auxiliary variable approximates the directional derivative s·∇T and SUPG stabilization is applied to both the temperature and auxiliary-variable equations. The authors prove consistency of the bilinear form (Proposition 3.1), discuss the loss of symmetry and ellipticity of the resulting Petrov-Galerkin operator, and validate the method on a manufactured convergence test, a 2D flux-surface perturbation, a full-torus tokamak equilibrium sustainment case with κ⊥=0, and a tokamak flux-tube perturbation. The headline numerical claim is that the new method loses about 4% of available thermal energy in the long-run equilibrium test, versus about 35% and 32% for the primal and standard mixed CG formulations at matched resolution.

Significance. If the claims hold, this is a practically relevant contribution: it offers a CG-based alternative to the authors' earlier DG-upwind method, using the same auxiliary-variable idea but with lower implementation complexity for existing CG-based MHD codes. The test design is a strength: the κ⊥=0 equilibrium has an exact stationary solution, so the measured energy loss is directly attributable to spurious perpendicular transport; the comparisons are at matched degrees of freedom; and the consistency calculation in Proposition 3.1 is clean under its stated regularity assumptions. The authors also candidly acknowledge the loss of symmetry and ellipticity and the heuristic nature of τ. However, the consistency proof does not cover the quadrature-evaluated, cell-wise discontinuous τ actually used in the tokamak runs, and the definition of τ in Eq. (3.15) contains a dimensional inconsistency; these issues must be resolved before the central accuracy claim is fully supported.

major comments (3)
  1. [Definition 3.1; Stabilization parameter; Proposition 3.1] The consistency proof assumes τ is continuous: Definition 3.1 states that M^f is well-defined 'provided that the stabilization parameter τ is continuous', and the integration by parts leading to Eq. (3.9) uses a global differentiable τ with no inter-element jumps. In the implementation, however, τ is prescribed by Eq. (3.15) with the local cell length δx and is 'evaluated as an expression ... on quadrature points'; on the unstructured tokamak mesh δx (and hence τ) is discontinuous across cell faces, so the distributional derivative s·∇(τ υ) contains interface contributions that the quadrature evaluation omits. Consequently Proposition 3.1 does not establish consistency of the scheme used for the headline 4% result in Section 4.3. The convergence study of Section 4.1 uses a uniform mesh, where δx (and τ) is constant, so it does not exercise this regime. Please either define τ through a continuous finite-element projection of Eq. (3.15), prove consistency for the discontinuous-τ form including jump terms, or add a numerical consistency test on an unstructured mesh.
  2. [Eq. (3.15)] The stabilization parameter in Eq. (3.15) is dimensionally inconsistent with the unit analysis given just above it. The authors state that τ must have units s^{1/2}, requiring the denominator to have units s^{-1/2}. The term k√κΔ/δx has units s^{-1/2}, but 2√δt has units s^{1/2} (if δt is a time), so the two terms cannot be added. If the intended formula is (2/√δt + k√κΔ/δx)^{-1}, as the dimensional argument and the remark that τ→0 under temporal refinement suggest, please correct it; as printed the formula is not reproducible and changes the time-step scaling of the stabilization.
  3. [Section 3, 'Ellipticity'] The method is presented without a stability or coercivity analysis. The 'Ellipticity' paragraph acknowledges that the Petrov-Galerkin operator (3.13) is non-symmetric and possibly non-elliptic, and that 'potential instabilities' may arise if τ is not chosen carefully; the only guidance for τ is the heuristic formula (3.15), motivated by a dimensional argument and a CFD reference. Because the central accuracy claim (including the 4% loss in Section 4.3) depends on this parameter, the paper should provide either a coercivity/inf-sup analysis under explicit conditions on τ or a systematic numerical study (for example, varying τ around (3.15) and varying κ∥/κ⊥ and δt) to demonstrate robustness.
minor comments (5)
  1. [Remark 3.2] The text refers to 'our SUPG-modified spatial discretization (3.2)'; this should be (3.8) or Definition 3.2, since Eq. (3.2) is the transport-equation example.
  2. [Section 4.1] The mesh description '10 × 2kx cells' should be typeset as 10 × 2^{k_x} cells for clarity.
  3. [Appendix A and Section 2] There are minor typos: 'Branginskii' should be 'Braginskii' in Appendix A, and 'discontinous' should be 'discontinuous' in Section 2.
  4. [Remark 3.1] In Eq. (3.16b), 'were Trhs corresponds to the strong form' should read 'where Trhs corresponds to the strong form'.
  5. [Eq. (4.6)] The limiter parameters Tl and σl are introduced without discussion of how their values were chosen; a sentence explaining the choice would aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the consistency proof is a standard strong-form consistency check, the headline 4% result is measured against an external equilibrium invariant with an a priori stabilization parameter, and no fitted quantity is repackaged as a prediction.

full rationale

No circular step is present. Proposition 3.1 is a standard consistency proof: it shows that a strong solution of (2.1) satisfies the discrete weak form (3.8) by substituting the strong PDE residual into the SUPG-modified test functional; this is the usual meaning of consistency, not a reduction of the method's claims to its inputs. The headline tokamak result in Section 4.3 compares against the exact equilibrium solution T(x,t)=T(x,0), so spurious heat loss is measured against an external invariant, and the SUPG parameter (3.15) is set a priori from δt, δx, k and κ∆ rather than fitted to the benchmark. The convergence test in Section 4.1 uses an independent Fourier/analytic solution, and the tokamak comparisons match resolution or degrees of freedom across methods. Self-citations to the authors' prior DG work [38] describe lineage and solver context but do not supply the accuracy claim, which is established by the paper's own proofs and benchmarks. The manuscript does explicitly flag a real gap in Section 3 ('Stabilization parameter'): τ is required to be continuous for M^f to be well-defined and for the integration-by-parts step in Proposition 3.1, yet τ from (3.15) is evaluated cell-wise at quadrature points, making it generally discontinuous on unstructured meshes. This is a rigor/correctness limitation between the proof and the implemented scheme, not a circularity, because the numerical claims do not reduce by construction to the assumptions that generate them.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The method introduces no new physical entities; the auxiliary variable ζ and the SUPG stabilization are existing techniques. The central claim depends on a heuristic stabilization parameter and on the unproved practical stability of the resulting non-symmetric operator.

free parameters (2)
  • SUPG stabilization parameter τ = τ = (2/√δt + k√κΔ/δx)^{-1} (Eq. 3.15)
    Standard SUPG scaling chosen by hand from the CFD literature, not derived from the anisotropic heat equation. The central numerical results depend on this choice, and the paper notes possible further adjustments for mesh anisotropy.
  • Limiter parameters T_l, σ_l in f(T) (Eq. 4.6) = T_l = 0.1, σ_l = 0.04
    Admitted by the authors to be 'somewhat arbitrary' and set roughly to the separatrix temperature. They cap the Braginskii parallel conductivity in the tokamak test; the comparison with standard methods is conservative (worse without the limiter).
assumptions (4)
  • standard math The strong solution T and fields B, τ are sufficiently regular for integration by parts and for the substitution η = s·∇γ in the consistency proof.
    Used in Proposition 3.1 to reduce the SUPG forms to the strong equation; the discrete method only needs test functions in VCG_k.
  • domain assumption The known-source assumption S for the heat equation (split-time MHD) is adequate for the tests.
    Section 2 states 'we will for simplicity assume S to be a known field', which is standard in split-time MHD but excludes full coupling.
  • ad hoc to paper The auxiliary variable ζ can be represented in the continuous space VCG_k, despite the directional derivative of a CG field generally being discontinuous.
    This is the central design choice enabling SUPG; the paper shows in §4.1 that CG P2 for ζ gives slightly larger errors than DG dP1, so the CG choice is a compromise.
  • ad hoc to paper The discrete Petrov-Galerkin operator (3.13) is stable in practice despite the admitted loss of symmetry and ellipticity.
    The paper has no stability or coercivity proof and warns of 'potential instabilities if the SUPG stabilization parameter τ is not chosen carefully'; numerical success is shown only for the tested parameter range.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An accurate SUPG-stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion." pith.science (2026). https://pith.science/paper/AJ5NFK7L

@misc{pith2026241212396,
  author       = {Pith},
  title        = {Pith review of: An accurate SUPG-stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AJ5NFK7L}},
  note         = {Machine review of arXiv:2412.12396}
}
read the original 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.

Figures

Figures reproduced from arXiv: 2412.12396 by the authors.

Figure 1
Figure 1. Convergence plot for relative L 2 errors (4.2) for analytic Gaussian profile diffusion test case (4.1), for primal CG (solid cyan), standard mixed CG (dashed orange), and SUPG-based (dotted purple) formulations (2.7), (2.9), and (3.8), respectively. 2D magnetic flux surface (see [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Images depicting temperature field T for test case (4.4). Top left: T at initial time, together with straight white lines tracing a single (periodic) magnetic field line and a section of the quadrilateral mesh used for the simulation runs. Top right: primal CG formulation (2.7) at t ≈ 14. Bottom row, left to right: standard mixed and SUPG-modified mixed formulations (2.9) and (3.8), respectively, at t ≈ 14. near the… view at source ↗
Figure 3
Figure 3. Images depicting error of temperature field [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Relative L 2 error development over time for magnetic flux surface test case, for primal CG (solid cyan), standard mixed CG (dashed orange), and SUPG-based (dotted purple) formulations (2.7), (2.9), and (3.8), respectively [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Images depicting toroidally periodic domain for tokamak test case setup. Left: non [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Evolution of the relative total available temperature (4.7) over time for tokamak equi [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Images for perturbation test (4.9), at t ≈ 120 Alfven times. Left: forcing S(t). Center and right: temperature field for standard mixed and SUPG discretizations (2.9), (3.8), respectively. White lines denote two magnetic flux surfaces forming a magnetic flux tube that …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    P. R. Amestoy, I. S. Duff, and J.-Y. L’Excellent. Multifrontal parallel distributed symmetric and unsymmetric solvers. Computer Methods in Applied Mechanics and Engineering , 184(2- 4):501–520, 2000

    work page 2000

  2. [2]

    Balay, S

    S. Balay, S. Abhyankar, M. Adams, J. Brown, P. Brune, K. Buschelman, L. Dalcin, A. Dener, V. Eijkhout, W. Gropp, et al. PETSc users manual. 2019

    work page 2019

  3. [3]

    A. R. Bell. Non-Spitzer heat flow in a steadily ablating laser-produced plasma. The Physics of Fluids, 28(6):2007–2014, 1985

    work page 2007

  4. [4]

    D. Biskamp. Nonlinear Magnetohydrodynamics. Cambridge Monographs on Plasma Physics. Cambridge University Press, 1993

    work page 1993

  5. [5]

    Bonilla, J

    J. Bonilla, J. N. Shadid, X.-Z. Tang, M. M. Crockatt, P. Ohm, E. G. Phillips, R. P. Pawlowski, S. Conde, and O. Beznosov. On a fully-implicit VMS-stabilized FE formulation for low Mach number compressible resistive MHD with application to MCF. Computer Methods in Applied Mechanics and Engineering , 417:116359, 2023

    work page 2023

  6. [6]

    S. I. Braginskii. Transport processes in a plasma. Reviews of plasma physics , 1:205, 1965

    work page 1965

  7. [7]

    Some properties of the M3D-C1 form of the three- dimensional magnetohydrodynamics equations

    J Breslau, N Ferraro, and S Jardin. Some properties of the M3D-C1 form of the three- dimensional magnetohydrodynamics equations. Physics of Plasmas , 16(9), 2009

    work page 2009

  8. [8]

    A. N. Brooks and T. J. R. Hughes. Streamline upwind/Petrov-Galerkin formulations for convec- tion dominated flows with particular emphasis on the incompressible Navier-Stokes equations. Computer methods in applied mechanics and engineering , 32(1-3):199–259, 1982

    work page 1982

Show all 41 references
  1. [9]

    Chac´ on, D

    L. Chac´ on, D. A. Knoll, and J. M. Finn. An implicit, nonlinear reduced resistive MHD solver. Journal of Computational Physics , 178(1):15–36, 2002. 20

    2002

  2. [10]

    A. S. Chamarthi, H. Nishikawa, and K. Komurasaki. First order hyperbolic approach for anisotropic diffusion equation. Journal of Computational Physics , 396:243–263, 2019

    2019

  3. [11]

    Degond, A

    P. Degond, A. Lozinski, J. Narski, and C. Negulescu. An asymptotic-preserving method for highly anisotropic elliptic equations based on a micro–macro decomposition. Journal of Com- putational Physics , 231(7):2724–2740, 2012

    2012

  4. [12]

    Deluzet and J

    F. Deluzet and J. Narski. A two field iterated asymptotic-preserving method for highly anisotropic elliptic equations. Multiscale Modeling & Simulation , 17(1):434–459, 2019

    2019

  5. [13]

    B. D. Dudson et al. BOUT++: Recent and current developments. Journal of Plasma Physics , 81(1), 2015

    2015

  6. [14]

    J. P. Freidberg. Ideal Magnetohydrodynamics. Plenum Press, New York, NY, 1987

    1987

  7. [15]

    Giorgiani, H

    G. Giorgiani, H. Bufferand, F. Schwander, E. Serre, and P. Tamain. A high-order non field- aligned approach for the discretization of strongly anisotropic diffusion operators in magnetic fusion. Computer Physics Communications , 254:107375, 2020

    2020

  8. [16]

    Green, X

    D. Green, X. Hu, J. Lore, L. Mu, and M. L. Stowell. An efficient high-order numerical solver for diffusion equations with strong anisotropy. Computer Physics Communications , 276:108333, 2022

    2022

  9. [17]

    Green, X

    D. Green, X. Hu, J. Lore, L. Mu, and M. L. Stowell. An efficient high-order solver for dif- fusion equations with strong anisotropy on non-anisotropy-aligned meshes. SIAM Journal on Scientific Computing , 46(2):S199–S222, 2024

    2024

  10. [18]

    G¨ unter, K

    S. G¨ unter, K. Lackner, and C. Tichmann. Finite element and higher order difference formula- tions for modelling heat transport in magnetised plasmas. Journal of Computational Physics , 226(2):2306–2316, 2007

    2007

  11. [19]

    G¨ unter, Q

    S. G¨ unter, Q. Yu, J. Kr¨ uger, and K. Lackner. Modelling of heat transport in magnetised plasmas using non-aligned coordinates. J. Comput. Phys. , 209(1):354–370, oct 2005

    2005

  12. [20]

    Guo and X.-Z

    Z. Guo and X.-Z. Tang. Parallel heat flux from low to high parallel temperature along a magnetic field line. Phys. Rev. Lett. , 108:165005, Apr 2012

    2012

  13. [21]

    D. A. Ham, P. H. J. Kelly, L. Mitchell, C. J. Cotter, R. C. Kirby, K. Sagiyama, N. Bouziani, S. Vorderwuelbecke, T. J. Gregory, J. Betteridge, D. R. Shapero, R. W. Nixon-Hill, C. J. Ward, P. E. Farrell, P. D. Brubeck, I. Marsden, T. H. Gibson, M. Homolya, T. Sun, A. T. T. McRa...

    2023

  14. [22]

    M. Held, M. Wiesenberger, and A. Stegmeir. Three discontinuous Galerkin schemes for the anisotropic heat conduction equation on non-aligned grids.Computer Physics Communications, 199:29–39, 2016

    2016

  15. [23]

    Hoelzl et al

    M. Hoelzl et al. The JOREK non-linear extended MHD code and applications to large- scale instabilities and their control in magnetically confined fusion plasmas. Nuclear Fusion, 61(6):065001, 2021

    2021

  16. [24]

    S. Jardin. Computational methods in plasma physics . CRC press, 2010

    2010

  17. [25]

    S. Jin. Efficient asymptotic-preserving (AP) schemes for some multiscale kinetic equations. SIAM Journal on Scientific Computing , 21(2):441–454, 1999

    1999

  18. [26]

    D. Kuzmin. A guide to numerical methods for transport equations . 2010. 21

    2010

  19. [27]

    J. Li, Y. Zhang, and X.-Z. Tang. Staged cooling of a fusion-grade plasma in a tokamak thermal quench. Nuclear Fusion, 63(6):066030, may 2023

    2023

  20. [28]

    S. Liu, Q. Tang, and X.-Z. Tang. A parallel cut-cell algorithm for the free-boundary Grad– Shafranov problem. SIAM Journal on Scientific Computing , 43(6):B1198–B1225, 2021

    2021

  21. [29]

    T. A. Manteuffel, S. M¨ unzenmaier, J. Ruge, and B. S. Southworth. Nonsymmetric reduction- based algebraic multigrid. SIAM Journal on Scientific Computing , 41(5):S242–S268, 2019

    2019

  22. [30]

    T. A. Manteuffel, J. Ruge, and B. S. Southworth. Nonsymmetric algebraic multigrid based on local approximate ideal restriction (lAIR). SIAM Journal on Scientific Computing , 40(6):A4105–A4130, 2018

    2018

  23. [31]

    A. T. T. McRae, G.-T. Bercea, L. Mitchell, D. A. Ham, and C. J. Cotter. Automated generation and symbolic manipulation of tensor product finite elements. SIAM Journal on Scientific Computing, 38(5):S25–S47, 2016

    2016

  24. [32]

    Narski and M

    J. Narski and M. Ottaviani. Asymptotic preserving scheme for strongly anisotropic parabolic equations for arbitrary anisotropy direction.Computer Physics Communications, 185(12):3189– 3203, 2014

    2014

  25. [33]

    W. Park, E. V. Belova, G. Y. Fu, X. Z. Tang, H. R. Strauss, and L. E. Sugiyama. Plasma simulation studies using multilevel physics models. Physics of Plasmas , 6(5):1796–1803, May 1999

    1999

  26. [34]

    D. A. Serino, Q. Tang, X.-Z. Tang, T. V. Kolev, and K. Lipnikov. An adaptive Newton-based free-boundary Grad-Shafranov solver. arXiv preprint arXiv:2407.03499 , 2024

    2024 arXiv

  27. [35]

    C. R. Sovinec et al. Nonlinear magnetohydrodynamics simulation using high-order finite ele- ments. Journal of Computational Physics , 195(1):355–386, 2004

    2004

  28. [36]

    Tezduyar

    T. Tezduyar. Stabilization parameters and local length scales in SUPG and PSPG formulations. In Proceedings of the Fifth World Congress on Computational Mechanics , volume 81508, 2002

    2002

  29. [37]

    C. J. Vogl, I. Joseph, and M. Holec. Mesh refinement for anisotropic diffusion in magnetized plasmas. arXiv preprint arXiv:2210.16442 , 2022

    2022 arXiv

  30. [38]

    G. A. Wimmer, B. S. Southworth, T. J. Gregory, and X.-Z. Tang. A fast algebraic multigrid solver and accurate discretization for highly anisotropic heat flux I: open field lines. SIAM Journal on Scientific Computing , 46(3):A1821–A1849, 2024

    2024

  31. [39]

    C. Yang, F. Deluzet, and J. Narski. Preserving the accuracy of numerical methods discretizing anisotropic elliptic problems. arXiv preprint arXiv:1911.11482 , 2019

    1911 arXiv

  32. [40]

    C. Yang, F. Deluzet, and Jacek. Narski. On the accuracy of numerical methods for the dis- cretization of anisotropic elliptic problems. Journal of Computational Physics , page 113568, 2024

    2024

  33. [41]

    Zhang, J

    Y. Zhang, J. Li, and X.-Z. Tang. Cooling flow regime of a plasma thermal quench. Europhysics Letters, 141(5):54002, feb 2023. 22

    2023

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.