betaFeedback

WorksMathematics

A Realizability-Based Kinetic-Fluid Domain Decomposition Method for Single-Temperature Gas Mixtures

Abstract

This paper introduces a domain-adaptive numerical method for multi-species kinetic models that relax toward a single-temperature and single-velocity hydrodynamic limit. To this end, we derive moment realizability matrices for this regime. In this way the ideas of Levermore, Morokoff, and Nadiga are extended to non-equilibrium mixtures in which inter-species and intra-species collisions occur at comparable rates, so that all species relax to a common velocity and a common temperature. The realizability conditions are used as an indicator to divide the computational domain, at each time step, into macroscopic and kinetic regions. In the kinetic regions, where the continuum description fails (e.g., across strong shocks, non-equilibrium interfaces, or kinetic boundary layers), the method solves the multi-species kinetic equations. Close to local thermodynamic equilibrium, the Euler equations of the mixture are solved instead. Confining the kinetic computations to the non-equilibrium regions reduces the simulation time, while the results remain close to those of the full kinetic solver. The method is discretized on Cartesian grids with high-order schemes in space and time. The kinetic equations are integrated with an Asymptotic-Preserving Implicit-Explicit (IMEX) scheme, so that a consistent discretization of the single-velocity, single-temperature Euler system of the mixture is recovered when the Knudsen number goes to zero. To the best of our knowledge, this is the first realizability-based adaptive spatial domain decomposition method for gas mixtures in this single-velocity, single-temperature setting.

Introduction

Gas mixtures appear in many problems, from aerospace engineering to evaporation–condensation [38] and plasma physics [10]. In transitional regimes, as in the re-entry of spacecraft at high altitude [25] or in micro-electro-mechanical systems (MEMS), the continuum hypothesis is no longer valid. In these conditions macroscopic models such as the Navier–Stokes equations do not correctly describe localized structures like shock profiles and kinetic boundary layers, and kinetic models, such as the multi-species Boltzmann equation, are needed.

The theory of multi-species kinetic models is well developed [11, 12], and more recent works also consider granular mixtures and their long-time behaviour [48]. These equations couple a linear transport operator with nonlinear collision terms, which become stiff in the fluid regime. Their numerical solution is expensive, because the distribution functions depend on up to seven variables [26]. When the mean free path goes to zero, the solution of the kinetic system approaches that of the Euler equations of the mixture, which are much cheaper to solve. For this reason, using a kinetic solver in the whole domain is wasteful when the flow is close to equilibrium in most of it.

Hybrid methods, or multi-physics domain decomposition strategies, reduce this cost [26, 13, 29, 49, 15, 17, 19]. They divide the computational domain into regions where different models are solved, according to the local distance from thermodynamic equilibrium.

Kinetic–fluid coupling for gas mixtures has also been studied in other works. A particle-based micro–macro method was proposed in [22]. Domain decomposition methods for multi-species flows were proposed in [23], where a multi-species kinetic model is coupled with a multi-fluid model, and, closer to the present setting, in [20], where a multi-species Vlasov–BGK model is coupled with its single-velocity, single-temperature hydrodynamic limit through buffer regions. In both domain decomposition methods the kinetic and fluid regions are fixed before the computation and do not change in time. Here, instead, the partition is dynamic: it is updated at every time step with a realizability-based indicator, so that the kinetic and fluid regions evolve in time with the flow.

The stiffness of the collision terms is handled with Asymptotic-Preserving (AP) schemes [21, 35, 14, 40, 36, 39, 41, 32, 24, 28, 27, 38, 34, 6]. AP schemes maintain numerical stability across highly collisional regimes without requiring spatial or temporal step sizes to resolve the microscopic scales. They guarantee that the appropriate macroscopic continuum equations are recovered at the discrete level, allowing the time step to be selected independently of the local Knudsen number [37].

The main objective of this paper is the design of a spatially adaptive, automated numerical method for gas mixtures operating within a single-temperature, single-velocity hydrodynamic limit. At the microscopic level, this physical regime is characterized by intra- and inter-species collisions acting on the same asymptotic time scale. Thanks to these interactions, the distinct components of the mixture rapidly converge toward a common macroscopic velocity and temperature. Since all species of the mixture converge toward a single local equilibrium state, our domain decomposition algorithm treats the whole mixture as a single entity (i.e. all species switch together between the macroscopic solver and the kinetic solver).

From a mathematical point of view, the main new contribution of this work is the derivation of a moment realizability matrix for this single-velocity and single-temperature limit. This matrix is used to detect the regions out of local equilibrium, and therefore to decompose the domain. In this paper the criterion is derived and tested for a multi-species BGK operator. The collision model enters the indicator only through the Chapman–Enskog corrections, so the same construction can be used for the multi-species Boltzmann operator once these corrections are known. The coupling between the two solvers does not depend on the collision operator, so the proposed strategy can be extended to the full multi-species Boltzmann collision integral.

To reduce the cost, the kinetic solver is used only in the cells selected by the indicator. The coupling is non-intrusive. The fluid and kinetic solvers only exchange cell values near the interfaces, and they only need to use the same Runge–Kutta stages.

The opposite regime, in which collisions between particles of the same species dominate, is considered in the companion paper [18]. There the hydrodynamic limit has a velocity and a temperature for each species, the indicator is evaluated species by species, and the method is second order on unstructured triangular meshes. Here all collisions occur on the same time scale, so the indicator is derived for the mixture as a whole and the same partition of the domain is used for all species. The kinetic model, the implicit relaxation step and the coupling between the solvers follow [18], with the present scaling, and are briefly recalled to keep the paper self-contained. Besides the indicator, the main difference is the order of accuracy: both solvers use a third-order CWENO reconstruction and a third-order IMEX scheme [7, 8], with a penalization of the collision term, on Cartesian meshes.

The article is organized as follows. § 2 presents the kinetic model and its hydrodynamic limit. In § 3 we derive the realizability conditions and the switching criterion. § 4 describes the numerical schemes, in particular the AP IMEX time integration and the coupling between the solvers. § 6 presents the numerical tests, and § 7 discusses the computational cost. Conclusions and future work are given in § 8.

Kinetic model for single-temperature, single-velocity gas mixtures

Following [3, 4], we consider a mixture of NN monatomic gases with elastic collisions. The distribution fi(t,x,v)f_i(t,x,v) of species ii, with x,v∈R3x,v \in\mathbb{R}^3 and t∈R+t \in\mathbb{R}_+, satisfies

{∂fi∂t+v⋅∇xfi=Qi[f],Qi[f]=∑k=1NQi,k[fi,fk],fi(0,x,v)=fi,0(x,v).(1)\begin{cases} \frac{\partial f_i}{\partial t}+v\cdot\nabla_x f_i=Q_i[f],\\ Q_i[f]=\sum_{k=1}^{N}Q_{i,k}[f_i,f_k],\\ f_i(0,x,v)=f_{i,0}(x,v). \tag*{(1)} \end{cases}

Here, the term Qi[f]Q_i[f] represents the total collisional source for the ii-th species. The specific binary interaction integral Qi,k[fi,fk]Q_{i,k}[f_i,f_k] between species ii and kk is expressed as

Qi,k[fi,fk]=∫R3∫S2gi,k(∣y∣,y^⋅Ω)[fi(v′)fk(w′)−fi(v)fk(w)] dΩ dw.(2)Q_{i,k}[f_i,f_k]=\int_{\mathbb{R}^3}\int_{\mathbb{S}^2}g_{i,k}(|y|,\hat{y}\cdot\Omega)\left[f_i(v')f_k(w')-f_i(v)f_k(w)\right]\,\mathrm{d}\Omega\,\mathrm{d}w. \tag*{(2)}

In this integral, y=v−wy=v-w denotes the relative velocity vector of the interacting particle pair vv and ww, while v′v' and w′w' refer to their respective velocities prior to the collision. The intermolecular forces enter through the collision kernel gi,kg_{i,k}, which depends on the relative speed ∣y∣|y| and on μ=y^⋅Ω\mu=\hat{y}\cdot\Omega, with the unit vector defined as y^=y/∣y∣\hat{y}=y/|y|.

The number density ni(t,x)n_i(t,x), the mean velocity ui(t,x)u_i(t,x) and the temperature Ti(t,x)T_i(t,x) of each species are

{ni=∫R3fi(v) dv,ui=1ni∫R3vfi(v) dv,Ti=mi3ni∫R3∣v−ui∣2fi(v) dv.(3)\begin{cases} n_i=\int_{\mathbb{R}^3}f_i(v)\,\mathrm{d}v,\\ u_i=\frac{1}{n_i}\int_{\mathbb{R}^3}v f_i(v)\,\mathrm{d}v,\\ T_i=\frac{m_i}{3n_i}\int_{\mathbb{R}^3}|v-u_i|^2f_i(v)\,\mathrm{d}v. \tag*{(3)} \end{cases}

The equilibrium states of (1) are local Maxwellians with a common velocity uu and a common temperature TT:

fieq=ni(mi2πT)32exp⁡(−mi2T∣v−u∣2),(4)f_i^{\mathrm{eq}}=n_i\left(\frac{m_i}{2\pi T}\right)^{\frac{3}{2}}\exp\left(-\frac{m_i}{2T}|v-u|^2\right), \tag*{(4)}

where the total number density nn, the total mass density ρ\rho, the velocity uu and the temperature TT of the mixture are

{n=∑i=1Nni,ρ=∑i=1Nρi=∑i=1Nmini,u=1ρ∑i=1Nminiui,T=13n∑i=1Nmi∫R3∣v−u∣2fi(v) dv=1n∑i=1NniTi+13n∑i=1Nmini∣ui−u∣2.(5)\begin{cases} n=\sum_{i=1}^{N}n_i,\\ \rho=\sum_{i=1}^{N}\rho_i=\sum_{i=1}^{N}m_i n_i,\\ u=\frac{1}{\rho}\sum_{i=1}^{N}m_i n_i u_i,\\ T=\frac{1}{3n}\sum_{i=1}^{N}m_i\int_{\mathbb{R}^3}|v-u|^2f_i(v)\,\mathrm{d}v =\frac{1}{n}\sum_{i=1}^{N}n_iT_i+\frac{1}{3n}\sum_{i=1}^{N}m_i n_i|u_i-u|^2. \tag*{(5)} \end{cases}

As in [3, 4], we replace the Boltzmann collision operator with a BGK-type model, in which the collision term is a sum of relaxation terms:

Qi[f]=∑k=1NQ~i,k[fi,fk],Q~i,k[fi,fk]=νik(niMik−fi).(6)\begin{aligned} Q_i[f] &= \sum_{k=1}^{N} \widetilde{Q}_{i,k}[f_i,f_k],\\ \widetilde{Q}_{i,k}[f_i,f_k] &= \nu_{ik}(n_iM_{ik}-f_i). \tag*{(6)} \end{aligned}

Here νik\nu_{ik} is the collision frequency, independent of vv, between species ii and kk, and MikM_{ik} is a Maxwellian with the mass mim_i of species ii and with auxiliary velocity uiku_{ik} and temperature TikT_{ik}:

References

  1. [1]Bailo, R., and Rey, T. Projective and telescopic projective integration for non-linear kinetic mixtures. Journal of Computational Physics 458 (2022), 111082.DOI
  2. [2]Bisi, M., Bosch eri, W., Dimarco, G., Groppi, M., and Martalò, G. A new mixed Boltzmann-BGK model for mixtures solved with an IMEX finite volume scheme on unstructured meshes. Applied Mathematics and Computation 433 (2022), 127416.DOI
  3. [3]Bisi, M., Groppi, M., and Martalò, G. Macroscopic equations for inert gas mixtures in different hydrodynamic regimes. Journal of Physics A: Mathematical and Theoretical 54, 8 (2021), 085201.DOI
  4. [4]Bobylev, A., Bisi, M., Groppi, M., Spiga, G., and Potapenko, I. A general consistent BGK model for gas mixtures. Kinetic & Related Models 11 (2018), 1377–1393.DOI
  5. [5]Boscarino, S., Cho, S. Y., and Russo, G. A local velocity grid conservative semi-Lagrangian schemes for BGK model. Journal of Computational Physics 460 (2022), 111178.DOI
  6. [6]Boscarino, S., Cho, S.-Y., Russo, G., and Yun, S.-B. High Order Conservative Semi-Lagrangian Scheme for the BGK Model of the Boltzmann Equation. Commun. Comput. Phys. 29 (2021), 1–56.arxiv.org/abs/1905.03660
  7. [7]Boscarino, S., Pareschi, L., and Russo, G. Implicit-explicit Runge–Kutta schemes for hyperbolic systems and kinetic equations in the diffusion limit. SIAM J. Sci. Comput. 35, 1 (2013), A22–A51.DOI
  8. [8]Boscarino, S., Pareschi, L., and Russo, G. Implicit-Explicit Methods for Evolutionary Partial Differential Equations, vol. 24 of Mathematical Modeling and Computation. SIAM, 2024.DOI
  9. [9]Bosch eri, W., and Dimarco, G. High order finite volume schemes with IMEX time stepping for the Boltzmann model on unstructured meshes. Computer Methods in Applied Mechanics and Engineering 387 (2021), 114180.
  10. [10]Bott, A. F. A., Cowley, S. C., and Schekochihin, A. A. Kinetic stability of Chapman–Enskog plasmas. J. Plasma Phys. 90, 2 (2024), 975900207.arxiv.org/abs/2310.17754
  11. [11]Briant, M. Stability of global equilibrium for the multi-species Boltzmann equation in L∞ settings. Discrete Contin. Dyn. Syst. 36, 12 (2016), 6669–6688.DOI
  12. [12]Briant, M., and Daus, E. S. The Boltzmann Equation for a Multi-species Mixture Close to Global Equilibrium. Arch. Ration. Mech. Anal. 222, 3 (2016), 1367–1443.DOI
  13. [13]Burt, J. M., and Boyd, I. D. A hybrid particle approach for continuum and rarefied flow simulation. J. Comput. Phys. 228, 2 (2009), 460–475.DOI
  14. [14]Caflisch, R. E., Jin, S., and Russo, G. Uniformly Accurate Schemes for Hyperbolic Systems with Relaxation. SIAM J. Numer. Anal. 34, 1 (1997), 246–281.DOI
  15. [15]Caparello, D., Pareschi, L., and Rey, T. Hierarchical dynamic domain decomposition for the multiscale Boltzmann equation. arXiv e-prints (2025), arXiv:2505.03360.arxiv.org/abs/2505.03360
  16. [16]Caparello, D., Pareschi, L., and Rey, T. High-Order Asymptotic-Preserving IMEX schemes for an ES-BGK model for Gas Mixtures. arXiv preprint arXiv:2511.22304 (2025).arxiv.org/abs/2511.22304
  17. [17]Caparello, D., Pareschi, L., and Rey, T. An IMEX multi-level domain-decomposition method for the Boltzmann equation on unstructured meshes. arXiv preprint arXiv:2610.21740 (2026).arxiv.org/abs/2610.21740
  18. [18]Caparello, D., Pareschi, L., and Rey, T. Realizability-Based Kinetic-Fluid Domain Decomposition for Multi-temperature Gas Mixtures on Unstructured Meshes. arXiv preprint arXiv:2611.17131 (2026).arxiv.org/abs/2611.17131
  19. [19]Caparello, D., and Tenna, T. A Coupled IMEX Domain Decomposition Method for High-Order Time Integration of the ES-BGK Model of the Boltzmann Equation. arXiv preprint arXiv:2512.03586 (2025).arxiv.org/abs/2512.03586
  20. [20]Chunaga, T., Sagert, I., Murillo, M. S., and Haack, J. R. Multi-species kinetic-fluid coupling for high-energy density simulations. Journal of Computational Physics 505 (2024), 112908.
  21. [21]Coron, F., and Perthame, B. Numerical passage from kinetic to fluid equations. SIAM J. Numer. Anal. 28 (1991), 26–42.DOI
  22. [22]Crestetto, A., Klingenberg, C., and Pirner, M. Kinetic/Fluid Micro-Macro Numerical Scheme for a Two Component Gas Mixture. Multiscale Model. Simul. 18, 3 (2020), 970–998.arxiv.org/abs/1806.11485
  23. [23]Datta, I. A. M., and Shumlak, U. Computationally efficient high-fidelity plasma simulations by coupling multi-species kinetic and multi-fluid models on decomposed domains. Journal of Computational Physics 483 (2023), 112073.DOI
  24. [24]Diaz, M. A., Chen, M.-H., and Yang, J.-Y. High-Order Conservative Asymptotic-Preserving Schemes for Modeling Rarefied Gas Dynamical Flows with Boltzmann-BGK Equation. Commun. Comput. Phys. 18, 4 (2015), 1012–1049.DOI
  25. [25]Dillman, R., and Corliss, J. Overview of the Mars Sample Return Earth Entry Vehicle. In Proceedings of the 6th International Planetary Probe Workshop (Atlanta, GA, 2008). NASA Technical Reports Server, Document ID 20080023907.
  26. [26]Dimarco, G., and Pareschi, L. Numerical methods for kinetic equations. Acta Numer. 23 (2014), 369–520.DOI
  27. [27]Filbet, F., and Jin, S. A class of asymptotic-preserving schemes for kinetic equations and related problems with stiff sources. J. Comput. Phys. 229, 20 (2010), 7625–7648.DOI
  28. [28]Filbet, F., and Jin, S. An Asymptotic Preserving Scheme for the ES-BGK Model of the Boltzmann Equation. J. Sci. Comput. 46 (2011), 204–224.DOI
  29. [29]Filbet, F., and Rey, T. A Hierarchy of Hybrid Numerical Methods for Multiscale Kinetic Equations. SIAM J. Sci. Comput. 37, 3 (2015), A1218–A1247.arxiv.org/abs/1402.6304
  30. [30]Filbet, F., and Xiong, T. A hybrid discontinuous Galerkin scheme for multi-scale kinetic equations. J. Comput. Phys. 372 (2018), 841–863.DOI
  31. [31]Gamba, I., and Tharkabhushanam, S. Spectral-Lagrangian methods for collisional models of non-equilibrium statistical states. Journal of Computational Physics 228 (2009), 2012–2036.arxiv.org/abs/0710.5308
  32. [32]Gosse, L., and Toscani, G. An asymptotic-preserving well-balanced scheme for the hyperbolic heat equations. Comptes Rendus Mathematique 334, 4 (2002), 337–342.DOI
  33. [33]Horn, R. A., and Johnson, C. R. Matrix Analysis, 2 ed. Cambridge University Press, 2012.
  34. [34]Hu, J., and Wang, L. An asymptotic-preserving scheme for the semiconductor Boltzmann equation toward the energy-transport limit. J. Comput. Phys. 281 (2015), 806–824.arxiv.org/abs/1309.1330
  35. [35]Jin, S. Runge-Kutta Methods for Hyperbolic Conservation Laws with Stiff Relaxation Terms. J. Comput. Phys. 122, 1 (1995), 51–67.DOI
  36. [36]Jin, S. Efficient Asymptotic-Preserving (AP) Schemes For Some Multiscale Kinetic Equations. SIAM J. Sci. Comput. 21, 2 (1999), 441–454.DOI
  37. [37]Jin, S. Asymptotic preserving (AP) schemes for multiscale kinetic and hyperbolic equations: a review. Riv. Mat. Univ. Parma (N.S.) 3, 2 (2012), 177–216.
  38. [38]Jin, S., and Li, Q. A BGK-penalization-based asymptotic-preserving scheme for the multispecies Boltzmann equation. Numer. Methods Partial Differential Equations 29, 3 (2013), 1056–1080.DOI
  39. [39]Jin, S., and Pareschi, L. Asymptotic-Preserving (AP) Schemes for Multiscale Kinetic Equations: a Unified Approach. In Hyperbolic Problems: Theory, Numerics, Applications (Basel, 2001), H. Freistühler and G. Warnecke, Eds., Birkhäuser Basel, pp. 573–582.DOI
  40. [40]Jin, S., Pareschi, L., and Toscani, G. Diffusive Relaxation Schemes For Multiscale Discrete-Velocity Kinetic Equations. SIAM J. Numer. Anal. 35 (1998), 2405–2439.DOI
  41. [41]Klar, A. An Asymptotic-Induced Scheme for Nonstationary Transport Equations in the Diffusive Limit. SIAM J. Numer. Anal. 35, 3 (1998), 1073–1094.DOI
  42. [42]Levermore, C. D., Morokoff, W. J., and Nadiga, B. T. Moment realizability and the validity of the Navier-Stokes equations for rarefied gas dynamics. Phys. Fluids 10, 12 (1998), 3214–3226.DOI
  43. [43]Levy, D., Puppo, G., and Russo, G. Compact Central WENO Schemes for Multidimensional Conservation Laws. SIAM J. Sci. Comput. 22, 2 (2000), 656–672.
  44. [44]Li, W., Song, P., and Wang, Y. A Hybrid Moment Method for Multi-scale Kinetic Equations Based on Maximum Entropy Principle. J. Sci. Comput. 89, 2 (2021), 35.arxiv.org/abs/2004.05893
  45. [45]Mieussens, L. Discrete Velocity Model and Implicit Scheme for the BGK Equation of Rarefied Gas Dynamics. Math. Models Methods Appl. Sci. 10, 8 (2000), 1121–1149.DOI
  46. [46]Mitrinovic, D. S. Analytic Inequalities, vol. 1. Springer, 1970.
  47. [47]Puppo, G., Semplice, M., and Visconti, G. Quinpi: Integrating Conservation Laws with CWENO Implicit Methods. Commun. Appl. Math. Comput. 5, 1 (2023), 343–369.arxiv.org/abs/2102.00741
  48. [48]Rey, T., and Tenna, T. The Boltzmann Equation for a Multi-Species Inelastic Mixture. J. Stat. Phys. 192, 11 (2025), 151.arxiv.org/abs/2410.17027
  49. [49]Tiwari, S. Coupling of the Boltzmann and Euler equations with automatic domain decomposition. J. Comput. Phys. 144, 2 (1998), 710–726.DOI
  50. [50]Tiwari, S. Application of Moment Realizability Criteria for the Coupling of the Boltzmann and Euler Equations. Transport Theory and Statistical Physics 29, 7 (2000), 759–783.DOI
  51. [51]Tyrtshnikov, E. E. A Brief Introduction to Numerical Analysis. Birkhäuser, Boston, MA, 1997. Lecture 4, pp. 31–37.
  52. [52]Xiong, T., and Qiu, J.-M. A hierarchical uniformly high order DG-IMEX scheme for the 1D BGK equation. J. Comput. Phys. 336 (2017), 164–191.DOI

Paper details

vlnc:W_v54dffc2
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 2

downloads · 0

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

  • Mathematics

a generative AI tool

View statement
Looking for related papers…
Contents