betaFeedback

WorksPhysics and Astronomy

Comprehensive full-f drift-kinetic and delta-f gyrokinetic simulations of a linear plasma device based on the gyro-moment approach

Abstract

First of a kind comprehensive full-f drift-kinetic (DK) and δδ-f gyrokinetic (GK) turbulent simulations are carried out in a linear plasma device. We self-consistently derive an electrostatic model including large-scale slowly-varying DK-ordered fields coupled to small-scale rapidly-fluctuating GK-ordered fields. By relying on the critical balance ordering, we show that the electrons are described by a drift-reduced Braginskii model while we rely on a Hermite-Laguerre spectral expansion for describing both the DK and GK parts of the ion distribution function. Global simulations are carried out using the parameters of the linear device LAPD, showing that the DK part of the ion distribution function is approximately a bi-Maxwellian. Fast spectral convergence both for the DK and GK Hermite-Laguerre expansion coefficients is observed, and that the GK fields do not affect the DK fields at the physical LAPD collisionality. Only when the collisionality is reduced and the source term is amplified for the GK fluctuations, an amplification of small-scale turbulent structures is observed. The findings are supported by linear results that show that the simulations are dominated by turbulent fluctuations that are Kelvin-Helmholz driven. Additionally, a GK Kelvin-Helmholz-like mode is observed in the low-GK-collisionality regime which can non-linearly drive small-scale structures.

Introduction

Modeling the boundary region of magnetically confined fusion plasmas remains a challenge for the fusion community. Fluctuations at all scales need to be considered, as well as kinetic effects and collisional processes. Developing a model for simulating the plasma boundary requires overcoming the limitations of existing models. Full-f fluid codes focus on the long-wavelength limit [32, 14], and their validity is limited to the highly collisional regime. Gyrofluid codes apply Padé approximation to relax the long-wavelength approximation [17, 18]. Additionally, they use kinetic closures to avoid considering the highly collisional limit, although the success of these closures is highly case dependent [3, 40, 41]. On the other hand, gyrokinetic (GK) δ\delta-f models, which offer a good description of the core of fusion plasmas [29, 21], present the challenge of including collisions at the level of relevance for the boundary. Furthermore, their application to the boundary is questionable due to large-scale fluctuations present in this region. Full-f GK codes often focus on the long-wavelength limit [5, 7, 16, 26, 39] and thus compromise physical fidelity at the cost of computational performance.

To overcome the limitations of these approaches and accurately describe the boundary, [36, 15, 9], propose a model that couples a DK and GK ordering. Here, we leverage the work of Frei et al. [9] and present a first implementation of a model that couples DK and GK orderings. More precisely, this work relies on splitting the distribution functions into a drift-kinetic (DK) large-scale slowly-varying part and a GK small-scale rapidly-varying part. The DK fields are allowed to be large in amplitude but only vary on long scale, employing a long-wavelength limit, while the GK fields are small in amplitude but allowed to vary on all scales.

The model is based on single particle equations of motion derived from first principles starting from the particle one-form. From the single particle equations of motion, the gyro-averaged Boltzmann equation is derived, split into a GK part and a DK part, and projected onto a Hermite-Laguerre basis giving a DK and a GK hierarchy equation for the projection coefficients, denoted as gyromoments. The Hermite-Laguerre projection allows for an efficient implementation of the plasma dynamics, including multiple different collision operators. Finally, we split the electrostatic potential into a DK part and a GK part, similarly to the distribution functions, and we derive a DK and GK Poisson equation from first-principles by using a variational principle.

Previous δ\delta-f GK gyromoment simulations based on the Hermite-Laguerre projection are shown to efficiently reduce the computational complexity of solving the GK system of equations compared to continuum δ\delta-ff GK simulations Hoffmann et al. [20]; Mandell et al. [24]. Similarly, full-ff gyromoment simulations were performed in a linear plasma device, within a DK ordering Frei et al. [10]; Mencke & Ricci [25]. This work combines the advantages of the two previously developed models allowing for order-one DK fluctuations with small wavenumbers and, at the same time, small-amplitude GK fluctuations with arbitrary wavenumbers. It then enables the study of instabilities on all scales while still removing the single-particle gyromotion from the description.

Although our approach ultimately is intended for the boundary of magnetically confined fusion devices, we consider the simpler case of a linear plasma device in this work. A linear plasma device offers an ideal testbed for new global turbulent plasma simulations. The absence of drifts associated with the curvature of the magnetic field, together with the smaller size of the simulations compared to toroidal devices reduces the complexity and cost of the simulations and allows for a deeper understanding of the physics at play Friedman et al. [11]. We note that, despite their simple geometries, previous simulations of linear plasma devices show complex turbulent behavior Shi et al. [38]; Pan et al. [29, 30]; Popovich et al. [34]; Rogers & Ricci [37]; Frei et al. [10]; Mencke & Ricci [25]. Finally, the perpendicular incidence of the magnetic field with the walls at the end of the plasma column simplify the boundary conditions of the simulations.

This paper is structured as follows. In section § 2 we present the ordering used as well as the equations describing the evolution of the distribution functions. The DK model for the ions and electrons is derived in § 3 from a kinetic description, and similarly, the GK model is derived in § 4. We close the system in § 5, by deriving the DK and GK Poisson equations used to solve for the two electrostatic potentials. The numerical implementation of the physical model is described in § 6. In § 7 we report the results of the comprehensive turbulent simulations comparing it with a full-ff DK simulation where the GK perturbations are suppressed. § 8 provides an analysis of the findings in § 7 by presenting linear investigations of the turbulence-driving instabilities. Finally, the conclusions follow in § 9.

Ordering assumptions and equations of motion

We focus on the case of a fully-ionized plasma with one single-ionized ion species such that the plasma is composed of species a=i,ea=i,e with ee denoting the electrons and ii the only ion species with charge qi=eq_i=e. We split the electrostatic potential

ϕ(x)=ϕDK(x)+ϕGK(x),(1)\phi(\mathbf{x})=\phi_{DK}(\mathbf{x})+\phi_{GK}(\mathbf{x}), \tag*{(1)}

where x\mathbf{x} is the particle position and we impose that the GK part of the electrostatic potential has a small amplitude compared to the DK part, ϕGK/ϕDK∼ϵδ<1\phi_{GK}/\phi_{DK}\sim\epsilon_\delta<1. We order the DK and GK quantities such that they vary on different length scales. More precisely, we order the perpendicular gradient lengths, k⊥DK∼∣∇⊥ln⁡FaDK∣∼∣∇⊥ln⁡ϕDK∣∼ϵ⊥/ρsk_{\perp DK}\sim\lvert\nabla_\perp\ln F_{aDK}\rvert\sim\lvert\nabla_\perp\ln\phi_{DK}\rvert\sim\epsilon_\perp/\rho_s with ϵ⊥<1\epsilon_\perp<1 and k⊥GK∼∣∇⊥ln⁡ϕGK∣∼1/ρsk_{\perp GK}\sim\lvert\nabla_\perp\ln\phi_{GK}\rvert\sim1/\rho_s, being ρs=cs/Ωi\rho_s=c_s/\Omega_i the ion sound Larmor radius with cs=Te/mic_s=\sqrt{T_e/m_i} the sound speed, TeT_e the electron temperature, and Ωi=qiB/mi\Omega_i=q_iB/m_i the ion gyro frequency, with BB the magnetic field strength. Furthermore, we order the parallel length scale as k∥∼ϵk⊥DKk_\parallel\sim\epsilon k_{\perp DK} and the frequency as ω∼Ωiϵϵ⊥\omega\sim\Omega_i\epsilon\epsilon_\perp. This ordering imposes that ϕDK\phi_{DK} gives a contribution to the E×BE\times B velocity on the order of ϵ⊥/ϵδ\epsilon_\perp/\epsilon_\delta with respect to that of ϕGK\phi_{GK} Frei et al. [9]. By further ordering ϵ⊥∼ϵδ\epsilon_\perp\sim\epsilon_\delta, we impose that the E×BE\times B drift from the DK electrostatic potential contributes to the drifts, comparably to the GK counterpart.

Based on this ordering, we derive the single-particle dynamics in § 2.1. The behavior of the full distribution function is described in § 2.2 using the Boltzmann equation. Finally, the evolution of the DK and GK parts of the distribution function, is derived from the Boltzmann equation in § 2.3.

Single-particle dynamics

We assume that the collisionless particle dynamics be governed by the gyrocenter single-particle Lagrangian, ΓA\Gamma_A. By using a Lie approach and including all terms up to second order in the GK and DK parts of the E×BE\times B drift (i.e., to order cs2ϵ⊥2c_s^2\epsilon_\perp^2 and cs2ϵδ2c_s^2\epsilon_\delta^2), while assuming a straight and constant magnetic field, we have Frei et al. [9]

Γa(R,μ,v∥,t)=qaAa∗⋅R˙−μBΩiθ˙−mav∥2/2−qaΦa∗+O(mamiTaϵ⊥3,mamiTaϵδ3),(2)\Gamma_a(\mathbf{R},\mu,v_\parallel,t)=q_a\mathbf{A}_a^*\cdot\dot{\mathbf{R}}-\frac{\mu B}{\Omega_i}\dot{\theta}-m_av_\parallel^2/2-q_a\Phi_a^*+O\left(\frac{m_a}{m_i}T_a\epsilon_\perp^3,\frac{m_a}{m_i}T_a\epsilon_\delta^3\right), \tag*{(2)}

where v∥=v⋅bv_\parallel=\mathbf{v}\cdot\mathbf{b} is the velocity component parallel to the magnetic field, b\mathbf{b} being the unit vector in the direction of the magnetic field, μ=mav⊥2/B\mu=m_av_\perp^2/B is the adiabatic invariant, mam_a is the mass of species aa, and v⊥=∣v−bv∥∣v_\perp=\lvert\mathbf{v}-\mathbf{b}v_\parallel\rvert is the velocity component perpendicular to the magnetic field. We also define the fields qaΦa∗=qaϕDK+μB+qaΦE+qa⟨ψ⟩Rq_a\Phi_a^*=q_a\phi_{DK}+\mu B+q_a\Phi_E+q_a\langle\psi\rangle_{\mathbf{R}} and qaAa∗=qaA+mav∥b+mauEq_a\mathbf{A}_a^*=q_a\mathbf{A}+m_av_\parallel\mathbf{b}+m_a\mathbf{u}_E, where qaΦE=ma∣uE∣2/2+μB∇⊥2ϕDK/(2BΩa)q_a\Phi_E=m_a\lvert\mathbf{u}_E\rvert^2/2+\mu B\nabla_\perp^2\phi_{DK}/(2B\Omega_a) and uE=−∇ϕDK×B/B2\mathbf{u}_E=-\nabla\phi_{DK}\times\mathbf{B}/B^2 is the E×BE\times B velocity for the DK potential. The effective GK gyro-averaged potential is defined as

⟨ψ⟩R=⟨ϕGK⟩R+qa22maΩa∂μ(⟨ϕGK2⟩R−⟨ϕGK⟩R2)+qa2maΩa2⟨(b×∇ϕ~GK)⋅∇ϕ~GK⟩R(3)\langle\psi\rangle_R = \langle\phi_{GK} \rangle_R + \frac{q_a^2}{2m_a\Omega_a}\partial_\mu\left(\langle\phi_{GK}^2 \rangle_R-\langle\phi_{GK}\rangle_R^2\right)+\frac{q_a}{2m_a\Omega_a^2}\left\langle\left(\mathbf{b}\times\nabla\widetilde{\phi}_{GK}\right)\cdot\nabla\widetilde{\phi}_{GK}\right\rangle_R \tag*{(3)}

where the gyroaverage operator is introduced,

⟨f⟩R=12π∫02πdθ∫dx δ(R+ρ−x)f(x,μ,v∥,t)(4)\langle f\rangle_R=\frac{1}{2\pi}\int_0^{2\pi}d\theta\int d\mathbf{x}\,\delta(\mathbf{R}+\boldsymbol{\rho}-\mathbf{x})f(\mathbf{x},\mu,v_\parallel,t) \tag*{(4)}
=12π∫02πdθ∑n=0∞(ρa⋅∇)nn!f(R,μ,v∥,t)(5)=\frac{1}{2\pi}\int_0^{2\pi}d\theta\sum_{n=0}^{\infty}\frac{(\boldsymbol{\rho}_a\cdot\nabla)^n}{n!}f(\mathbf{R},\mu,v_\parallel,t) \tag*{(5)}
=∑n=0∞(2μqaΩa)n∇2n2nn!n!f(R,μ,v∥,t),(6)=\sum_{n=0}^{\infty}\left(\frac{2\mu}{q_a\Omega_a}\right)^n\frac{\nabla^{2n}}{2^n n!n!}f(\mathbf{R},\mu,v_\parallel,t), \tag*{(6)}

with δ(⋅)\delta(\cdot) the Dirac delta function and ρa=av⊥/Ωa\boldsymbol{\rho}_a=a\mathbf{v}_\perp/\Omega_a, the Larmor radius, being a a unit vector perpendicular to both the magnetic field and the gyromotion of the particle Frei et al. [9]. Finally, we define

ϕ~GK=ϕGK−⟨ϕGK⟩R,(7)\widetilde{\phi}_{GK}=\phi_{GK}-\langle\phi_{GK}\rangle_R, \tag*{(7)}

and

ϕ‾GK=∫0θdθ′ ϕ~GK(x(R,θ′),μ,v∥,t),(8)\overline{\phi}_{GK}=\int_0^\theta d\theta'\,\widetilde{\phi}_{GK}\left(\mathbf{x}(\mathbf{R},\theta'),\mu,v_\parallel,t\right), \tag*{(8)}

in Eq. (3).

Collective behavior

The collective behavior of the distribution function is governed by the gyroaveraged Boltzmann equation Frei et al. [9],

References

  1. [1]Arakawa, A. 1997 Computational design for long-term numerical integration of the equations of fluid motion: Two-dimensional incompressible flow. Part I. Journal of computational physics 135 (2), 103.
  2. [2]Balay, Satish, Abhyankar, Shrirang, Adams, Mark F., Benson, Steven, Brown, Jed, Brune, Peter, Buschelman, Kris, Constantinescu, Emil M., Dalcin, Lisandro, Dener, Alp, Eijkhout, Victor, Faibusowitsch, Jacob, Gropp, William D., Hapla, Václav, Isaac, Tobin, Jolivet, Pierre, Karpeev, Dmitry, Kaushik, Dinesh, Knepley, Matthew G., Kong, Fande, Kruger, Scott, May, Dave A., McInnes, Lois Curfman, Mills, Richard Tran, Mitchell, Lawrence, Munson, Todd, Roman, Jose E., Rupp, Karl, Sanan, Patrick, Sarich, Jason, Smith, Barry F., Zampini, Stefano, Zhang, Hong, Zhang, Hong & Zhang, Junchao 2024 PETSc Web page. https://petsc.org/.
  3. [3]Beer, M. A. & Hammett, G. W. 1996 Toroidal gyrofluid equations for simulations of tokamak turbulence. Physics of Plasmas 3 (11), 4046.
  4. [4]Braginskii, SI 1965 Transport processes in a plasma. Reviews of plasma physics 1, 205.
  5. [5]Chang, C. S., Ku, S., Loarte, A., Parail, V., Koechl, F., Romanelli, M., Maingi, R., Ahn, J.-W., Gray, T., Hughes, J. & others 2017 Gyrokinetic projection of the divertor heat-flux width from present tokamaks to ITER. Nuclear Fusion 57 (11), 116023.DOI
  6. [6]D’Angelo, Nicola 1965 Kelvin—helmholtz instability in a fully ionized plasma in a magnetic field. The Physics of Fluids 8 (9), 1748–1750.DOI
  7. [7]Dorf, M. & Dorr, M. 2020 Progress with the 5D full-F continuum gyrokinetic code COGENT. Contributions to Plasma Physics 60 (5), e201900113.DOI
  8. [8]Dougherty, J. P. 1964 Model Fokker-Planck equation for a plasma and its solution. The Physics of Fluids 7 (11), 1788.DOI
  9. [9]Frei, B. J., Jorge, R. & Ricci, P. 2020 A gyrokinetic model for the plasma periphery of tokamak devices. Journal of Plasma Physics 86 (2), 905860205.DOI
  10. [10]Frei, B. J., Mencke, J & Ricci, P 2024 Full-f turbulent simulation in a linear plasma device using a gyro-moment approach. Physics of Plasmas 31 (1).DOI
  11. [11]Friedman, B., Carter, T. A., Umansky, M. V., Schaffner, D. & Dudson, B. 2012 Energy dynamics in a simulation of lapd turbulence. Physics of Plasmas 19 (10), 102307.DOI
  12. [12]Frigo, Matteo & Johnson, Steven G. 2005 The design and implementation of FFTW3. Proceedings of the IEEE 93 (2), 216–231, special issue on “Program Generation, Optimization, and Platform Adaptation”.DOI
  13. [13]Gekelman, W., Pfister, H., Lucky, Z., Bamber, J., Leneman, D. & Maggs, J. 1991 Design, construction, and properties of the large plasma research device- The LAPD at UCLA. Review of scientific instruments 62 (12), 2875.DOI
  14. [14]Giacomin, M., Ricci, P., Corrado, A., Fourestey, G., Galassi, D., Lanti, E., Mancini, D., Richart, N., Stenger, L. N. & Varini, N. 2022 The gbs code for the self-consistent simulation of plasma turbulence and kinetic neutral dynamics in the tokamak boundary. Journal of Computational Physics p. 111294.DOI
  15. [15]Hahm, TS, Wang, Lu & Madsen, Jens 2009 Fully electromagnetic nonlinear gyrokinetic equations for tokamak edge turbulence. Physics of Plasmas 16 (2).
  16. [16]Hakim, A. H., Mandell, N. R., Bernard, T. N., Francisquez, M., Hammett, G. W. & Shi, E. L. 2020 Continuum electromagnetic gyrokinetic simulations of turbulence in the tokamak scrape-off layer and laboratory devices. Physics of Plasmas 27 (4), 042304.
  17. [17]Held, Markus & Wiesenberger, Matthias 2022 Beyond the oberbeck–boussinesq and long wavelength approximation. Nuclear Fusion 63 (2), 026008.DOI
  18. [18]Held, Markus, Wiesenberger, Matthias & Kendl, Alexander 2020 Padé-based arbitrary wavelength polarization closures for full-f gyro-kinetic and-fluid models. Nuclear Fusion 60 (6), 066014.DOI
  19. [20]Hoffmann, A. C. D., Frei, B. J. & Ricci, P. 2023b Gyrokinetic simulations of plasma turbulence in a z-pinch using a moment-based approach and advanced collision operators. Journal of Plasma Physics 89 (2), 905890214.DOI
  20. [21]Jenko, Frank, Dorland, W, Kotschenreuther, M & Rogers, BN 2000 Electron temperature gradient driven turbulence. Physics of plasmas 7 (5), 1904–1910.pubmed.ncbi.nlm.nih.gov/11136051
  21. [22]Kaufman, Allan N 1986 The electric dipole of a guiding center and the plasma momentum density. The Physics of fluids 29 (5), 1736–1737.DOI
  22. [23]Loizu, J., Ricci, P., Halpern, F. D. & Jolliet, S. 2012 Boundary conditions for plasma fluid models at the magnetic presheath entrance. Physics of Plasmas 19 (12), 122307.
  23. [24]Mandell, N., Dorland, W., Abel, I., Gaur, R., Kim, P., Martin, M. & Qian, T. 2022 GX: a GPU-native gyrokinetic turbulence code for tokamaks and stellarators. arXiv:2209.06731 .arxiv.org/abs/2209.06731
  24. [25]Mencke, JE & Ricci, P 2025 Extended drift-kinetic full-f turbulent simulation of a linear plasma device using the gyro-moment approach. Plasma Physics and Controlled Fusion 67 (3), 035029.DOI
  25. [26]Michels, D., Ulbl, P., Zholobenko, W., Body, T., Stegmeir, A., Eich, T., Griener, M., Conway, G. D., Jenko, F. & Team, ASDEX Upgrade 2022 Full-f electromagnetic gyrokinetic turbulence simulations of the edge and scrape-off layer of ASDEX Upgrade with GENE-X. Physics of Plasmas 29 (3), 032307.DOI
  26. [27]Mosetto, A., Halpern, F. D., Jolliet, S., Loizu, J. & Ricci, P. 2015 Finite ion temperature effects on scrape-off layer turbulence. Physics of Plasmas 22 (1), 012308.DOI
  27. [28]Orszag, Steven A 1971 On the elimination of aliasing in finite-difference schemes by filtering high-wavenumber components. Journal of Atmospheric Sciences 28 (6), 1074–1074.doi.org/10.1175/1520-0469(1971)028<1074:oteoai>2.0.co;2
  28. [29]Pan, Q., Told, D. & Jenko, F. 2016 Fully nonlinear δf gyrokinetics for scrape-off layer parallel transport. Physics of Plasmas 23 (10), 102302.DOI
  29. [30]Pan, Q., Told, D., Shi, E., Hammett, G. W. & Jenko, F. 2018 Full-f version of GENE for turbulence in open-field-line systems. Physics of Plasmas 062303 (25), 1.DOI
  30. [31]Parra, Felix 2019 Lecture notes on Braginskii fluid equations.
  31. [32]Paruta, P., Ricci, P., Riva, F., Wersal, C., Beadle, C. & Frei, B. 2018 Simulation of plasma turbulence in the periphery of diverted tokamak by using the GBS code. Physics of Plasmas 25 (11), 112301.DOI
  32. [33]Pfirsch, Dieter & Morrison, Philip J 1985 Local conservation laws for the maxwell-vlasov and collisionless kinetic guiding-center theories. Physical Review A 32 (3), 1714.pubmed.ncbi.nlm.nih.gov/9896258
  33. [34]Popovich, P., Umansky, M. V., Carter, T. A. & Friedman, B. 2010 Analysis of plasma instabilities and verification of the bout code for the large plasma device. Physics of Plasmas 17 (10), 102107.DOI
  34. [35]Press, WH, Teukolsky, SA, Vettering, WT & Flannery, BP 1992 Numerical Recipes in Fortran, The Art of Scientific Computing. Cambridge University Press.
  35. [36]Qin, H., Cohen, RH, Nevins, WM & Xu, XQ 2006 General gyrokinetic equations for edge plasmas. Contributions to Plasma Physics 46 (7-9), 477–489.DOI
  36. [37]Rogers, B. N. & Ricci, P. 2010 Low-frequency turbulence in a linear magnetized plasma. Physical Review Letters 104 (22), 225002.pubmed.ncbi.nlm.nih.gov/20867177
  37. [38]Shi, E. L., Hammett, G. W., Stoltzfus-Dueck, T. & Hakim, A. 2017 Gyrokinetic continuum simulation of turbulence in a straight open-field-line plasma. Journal of Plasma Physics 83 (3), 905830304.DOI
  38. [39]Ulbl, P., Body, T., Zholobenko, W., Stegmeir, A., Pfennig, J. & Jenko, F. 2023 Influence of collisions on the validation of global gyrokinetic simulations in the edge and scrape-off layer of TCV. Physics of Plasmas 30 (5).DOI
  39. [40]Wiesenberger, Matthias & Held, Markus 2022 Long-wavelength closures for collisional and neutral interaction terms in gyro-fluid models. In Journal of Physics: Conference Series, , vol. 2397, p. 012015. IOP Publishing.DOI
  40. [41]Wiesenberger, M & Held, M 2023 Effects of plasma resistivity in three-dimensional full-f gyro-fluid turbulence simulations. arXiv preprint arXiv:2309.10414 .arxiv.org/abs/2309.10414
  41. [42]Zeiler, A., Drake, J. F. & Rogers, B. 1997 Nonlinear reduced braginskii equations with ion thermal dynamics in toroidal plasma. Physics of Plasmas 4 (6), 2134.DOI

Paper details

vlnc:W_ys2zutrd
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents