Pith. sign in

REVIEW 3 major objections 6 minor 49 references

Semi-Lagrangian implicit Bhatnagar-Gross-Krook collision model for the finite-volume discrete Boltzmann method

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A semi-Lagrangian implicit collision scheme that replaces the next-step equilibrium with the current one improves temporal accuracy by up to a factor of eight, slightly lowers cost, and matches the stability of the variable-transformation…

desk verdict A simple, useful implicit collision rule with credible case-specific evidence; the general explanation overreaches. read the letter →

arxiv 1908.08531 v3 pith:XIYZTB2N submitted 2019-08-21 physics.comp-ph

Leitao Chen , Sauro Succi , Xiaofeng Cai , Laura Schaefer This is my paper · ORCID
classification physics.comp-ph
keywords latticeBoltzmannmethoddiscretefinitevolumeunstructuredmeshBGKcollisionsemi-Lagrangiantemporalaccuracycomputationalcost
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

This paper proposes a semi-Lagrangian implicit collision (SLIC) model for the finite-volume discrete Boltzmann method (FVDBM) and shows that its simplest version, SLIC+INT1, is the best of the non-iterative implicit collision treatments tested. The scheme exploits the fact that particle distribution functions are constant along their characteristic paths: tracking them back one time step and using first-order (piecewise-constant) interpolation turns the unknown next-step equilibrium $f_\alpha^{eq,n+1}$ into the known current-step equilibrium $f_\alpha^{eq,n}$. On the Taylor-Green vortex, this one-line replacement reduces the temporal error by up to a factor of eight (about four on average) against the temporal-extrapolation and variable-transformation schemes, without affecting spatial accuracy, and it does so at slightly lower computational cost. The paper also shows why the simple scheme beats its own second-order-interpolation counterpart: temporal error tracks the size of the deviation $\delta=f_\alpha^{eq,n+1}-f_\alpha^{eq,n}$, and SLIC+INT1 is the only scheme with $\delta=0$.

What carries the argument

The central object is the identity $f_\alpha^{eq,n+1}=f_\alpha^{eq,n}$ (Eq.\,32), produced by applying the semi-Lagrangian idea to the collision term: the PDFs that meet at a cell barycenter at $t^{n+1}$ came from tracked-back locations $\mathbf{X}(P_\alpha)=\mathbf{X}(P)-\boldsymbol{e}_\alpha\Delta t$ at $t^n$, and first-order interpolation makes each such value the piecewise-constant barycenter value. The paper unifies all four schemes into $f_\alpha^{eq,n+1}=f_\alpha^{eq,n}+\delta$ (Eq.\,46), so the machinery reduces to the magnitude of $\delta$; $\delta=0$ for SLIC+INT1, while TE, VT, and SLIC+INT2 carry nonzero $\delta$. The theoretical engine for why zero is best is hypothesis H2: in a correct simulation the global change $f_\alpha^{eq,n+1}-f_\alpha^{eq,n}$ decays to zero over time because streaming and relaxation cancel in the material derivative of the equilibrium (the 'ever-shifting battle'), making the zero-$\delta$ closure the most accurate.

What would settle it

Run SLIC+INT1 against the variable-transformation scheme on a flow whose equilibrium never settles, such as an oscillating-lid cavity or a cylinder wake; if the temporal-error advantage over VT drops well below the eightfold factor reported for the decaying Taylor-Green vortex, hypothesis H2 and the claimed near-order-of-magnitude improvement do not carry over to unsteady regimes.

Watch

Extended reading notes

Core claim

The central claim is that the implicitness of the BGK collision equilibrium can be removed by tracing the particle distribution functions back along their characteristic lines during the collision step; with first-order interpolation, each traced-back distribution equals the cell-barycenter value, and therefore $f_\alpha^{eq,n+1}(P)=f_\alpha^{eq,n}(P)$ (Eq.\,32). The paper argues that this replacement, far from being a crude temporal extrapolation, is the natural semi-Lagrangian closure, and that it improves temporal accuracy by almost an order of magnitude in the tested range while slightly reducing cost. Stability in the $\Delta t/\tau$ limit improves by an order of magnitude relative to the temporal-extrapolation scheme and equals the variable-transformation scheme, with no change in the $\Delta t/\Delta x$ advection limit. The same rule of thumb, 'replace $f_\alpha^{eq,n+1}$ with $f_\alpha^{eq,n}$ anywhere it appears,' also upgrades the variable-transformation scheme.

Load-bearing premise

The whole argument leans on hypothesis H2—that in a correct simulation the equilibrium distribution changes negligibly from one step to the next, globally—which is demonstrated only on a decaying Taylor-Green vortex and a steady lid-driven cavity, not on flows with persistent unsteadiness or moving boundaries.

Editorial extensions

If this is right

  • Users can take larger time steps without sacrificing temporal accuracy: in the SOU-flux configuration, larger $\Delta t$ reduced the error while decreasing the number of steps needed to reach a given physical time.
  • The rule of thumb 'replace $f_\alpha^{eq,n+1}$ with $f_\alpha^{eq,n}$' can be applied to the variable-transformation scheme, producing a faster algorithm that is mathematically identical to SLIC+INT1.
  • The $\Delta t/\tau$ stability limit rises from about 2.6 for temporal extrapolation to at least 100 for SLIC and variable transformation, with the $\Delta t/\Delta x$ advection limit unchanged.
  • The observed convergence of $f_\alpha^{eq,n+1}-f_\alpha^{eq,n}$ to zero is proposed as a design guideline for developing future implicit collision closures.

Reading between the lines

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

  • The authors' own hypothesis H2 implies the scheme's advantage is tied to flows where the equilibrium distribution settles over time; a forced or moving-boundary flow with persistent unsteadiness is the natural stress test, and the eightfold gain may shrink there.
  • The $\delta$-formulation suggests a design family of 'zero-$\delta$' closures: any reconstruction that feeds moments computed only from current PDFs into the next equilibrium avoids the temporal penalty of SLIC+INT2 while potentially recovering spatial accuracy.
  • Because the stability scans stop at $\Delta t/\tau=100$, a scan beyond that point could reveal whether SLIC and variable transformation part ways at extreme collision rates.
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 / 6 minor

Summary. The paper proposes a semi-Lagrangian implicit collision (SLIC) model for the finite volume discrete Boltzmann method (FVDBM). With first-order spatial interpolation of the traced-back distributions, the scheme reduces to replacing the implicit equilibrium f_alpha^{eq,n+1} by f_alpha^{eq,n} (Eq. 32). The authors compare SLIC+INT1, a second-order variant (SLIC+INT2), the temporal extrapolation (TE) scheme, and the variable transformation (VT) scheme on a Taylor-Green vortex (TGV) and a lid-driven square cavity, using two different time-marching/flux combinations (theta=1 with SOU flux; theta=1/2 with PL flux). The central numerical claims are: SLIC+INT1 improves temporal accuracy by up to a factor of eight in the tested range (about four on average), slightly lowers computational cost, and raises the Delta-t/tau stability limit to match the VT scheme (capped at Delta-t/tau = 100). The paper also presents hypotheses H1 and H2 to explain the accuracy advantage and proposes a rule of thumb: replace f_alpha^{eq,n+1} with f_alpha^{eq,n} wherever it appears, including inside the VT scheme.

Significance. If the accuracy and cost claims hold beyond the tested cases, the SLIC+INT1 rule is a simple, parameter-free modification that could be applied directly in existing FVDBM solvers and even inside the VT scheme (Sec. 6.4). The paper's strengths are its transparent derivation, the consistency of the results across two independent numerical configurations, the honest acknowledgment that the stability comparison is capped at Delta-t/tau = 100, and the explicit limitation of the headline accuracy factor to the tested range. The main weakness is that the explanatory framework (H1/H2) is validated only on flows that satisfy H2 by construction, so the generality of the rule of thumb is not established.

major comments (3)
  1. [Sec. 5.4 (H2, Eq. 46, Fig. 10)] The central explanation and the general rule of thumb in Sec. 6 rest on hypothesis H2: 'If a simulation is correct, globally, f_alpha^{eq,n+1} - f_alpha^{eq,n} -> 0 as time progresses'. The numerical evidence in Fig. 10 is obtained from a decaying TGV flow and a steady lid-driven cavity, both of which satisfy H2 by construction. For flows with persistent unsteadiness, forcing, or moving boundaries, partial f_alpha^{eq}/partial t does not vanish, and the O(Delta-t * partial f_alpha^{eq}/partial t) error introduced by Eq. (32) may offset the accuracy advantage. The theoretical argument around Eq. (48), based on the 'ever-shifting battle', is qualitative and does not demonstrate that D f_alpha^{neq}/Dt and f_alpha^{neq}/tau cancel in such flows. I request either an additional test on a persistently unsteady or forced flow (e.g., periodic vortex shedding or an oscillating lid-driven cavity), or a revised statement that explicitly restricts the rule of thumb and the accuracy claim to flows where the global partial f_alpha^{eq}/partial t decays, with a warning for other flows.
  2. [Sec. 5.4 (H1, Figs. 4 and 10)] The validation of H1 is partly circular: the temporal-error ranking in Fig. 4 and the |delta| ranking in Fig. 10 are measured from the same TGV simulations, so the observed correlation between smaller |delta| and smaller temporal error is a single-problem observation, not a proof that delta=0 is optimal in general. The statement that 'H1 can be proven with ease' (p. 20) overstates what the evidence supports. To make the explanation load-bearing, the authors should provide a separate truncation-error analysis for a generic unsteady flow, or a test on an independent flow class, rather than relying on the same decaying-flow data that produced the accuracy ranking.
  3. [Sec. 5.1 and Sec. 7 (claim 1)] The headline quantitative claim - 'improve the temporal accuracy by almost an order of magnitude in the tested range' - is based on the TGV tests in Figs. 4 and 11, i.e., a single decaying-flow configuration at one Reynolds number (tau = 0.009) and Delta-t/tau up to 0.2. The companion claim that the scheme does 'not affect spatial accuracy' is inferred from the steady LDSC comparison (Fig. 8), not from a transient spatial-accuracy measurement. Because the paper itself shows (Fig. 12) that the choice of collision closure interacts with the flux discretization, the spatial-accuracy invariance should be demonstrated explicitly (e.g., via a grid-convergence study at matched temporal error) or the claim should be softened accordingly.
minor comments (6)
  1. [Eq. (40)] The summation notation in Eq. (40) is garbled; it should read E_t = sum_{n=1}^{S} epsilon_n.
  2. [Sec. 1 (p. 8)] The sentence 'And this why it is "semi"' contains a typo; it should be 'this is why'.
  3. [Sec. 5.2 (p. 17)] The phrase 'the VT scheme requires two times of variable transformations' should be 'two variable transformations' or 'twice'.
  4. [Sec. 5.4 (p. 21)] The phrase 'the famous "ever-shifting battle" originally evoked by Ludwig Boltzmann' should use 'invoked' rather than 'evoked'.
  5. [Fig. 3] The text states that the slope b of each curve is shown in Fig. 3, but the caption does not list the slopes; please ensure the fitted slopes are visible in the figure or are given explicitly in the caption or text.
  6. [Sec. 6.5.2 (Table 2)] The statement that 'the SLIC+INT1 presents a 13% improvement in speed' should state the reference scheme (VT) explicitly, since the percentage depends on the baseline.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity; SLIC+INT1 is a derived scheme with independent numerical comparisons, and the H1/H2 explanation is a post-hoc generality limitation, not a circular reduction.

full rationale

The central derivation of SLIC+INT1 is self-contained: Section 4 derives f_alpha^{eq,n+1}=f_alpha^{eq,n} (Eq. (32)) from characteristic tracing of PDFs (Eq. (25)), moment preservation along characteristics (Eq. (26)), and the first-order spatial interpolation assumption (Eq. (30)); no fitted parameter or target result is used as an input. The claimed accuracy, cost, and stability advantages are reported from direct numerical comparisons (Figs. 4, 6, 9, 11-13) against reference schemes under identical flux and time-marching settings, so the main claim is not equivalent to its inputs by construction. The explanatory hypotheses H1 and H2 in Section 5.4 are post hoc and are supported by Eq. (48) and Fig. 10; although H2 is stated globally ('If a simulation is correct, globally, f_alpha^{eq,n+1}-f_alpha^{eq,n} -> 0 as time progresses') while the evidence is limited to a decaying Taylor-Green vortex and a steady lid-driven cavity, this is a limitation on generality rather than a circular step. Self-citations to the authors' prior flux and interpolation schemes ([27], [28]) provide standard numerical ingredients and are not load-bearing for the new collision model; the reference to [7] for the 'ever-shifting battle' narrative is a textbook explanation, not an imported uniqueness theorem that forces the conclusion. No equation is redefined as its own prediction, and no fitted parameter is renamed as a predicted quantity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No new physical entities or fitted parameters are introduced; the scheme rests on standard kinetic and semi-Lagrangian assumptions plus the paper-specific hypothesis H2.

assumptions (5)
  • domain assumption BGK collision model with single relaxation time τ (Eq. 1) adequately represents the collision process.
    Standard kinetic model; the paper studies implicit treatments of this collision operator, not the physical fidelity of BGK.
  • domain assumption D2Q9 discrete velocity set with second-order truncated equilibrium (Eq. 13) is used throughout.
    Standard LBM D2Q9 model; generalization to other lattices is not demonstrated.
  • domain assumption PDFs preserve their values along characteristics during the advection substep, so f_α^{n+1}(P)=f_α^n(P_α) (Eq. 25).
    This is the semi-Lagrangian split approximation; collision is handled separately in the implicit term.
  • domain assumption First-order interpolation: f_α^n(P_α)=f_α^n(P) within a control volume (Eq. 30).
    This defines SLIC+INT1; it is accurate only when PDFs are nearly constant over a cell.
  • ad hoc to paper H2: globally, f_eq^{n+1}-f_eq^n tends to zero (∂f_eq/∂t tends to 0) for correct simulations.
    Introduced in Section 5.4 to justify δ=0 as the optimal collision closure; validated only on TGV and steady cavity flows.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semi-Lagrangian implicit Bhatnagar-Gross-Krook collision model for the finite-volume discrete Boltzmann method." pith.science (2026). https://pith.science/paper/XIYZTB2N

@misc{pith2026190808531,
  author       = {Pith},
  title        = {Pith review of: Semi-Lagrangian implicit Bhatnagar-Gross-Krook collision model for the finite-volume discrete Boltzmann method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XIYZTB2N}},
  note         = {Machine review of arXiv:1908.08531}
}
abstract

A new implicit BGK collision model using a semi-Lagrangian approach is proposed in this paper. Unlike existing models, in which the implicit BGK collision is resolved either by a temporal extrapolation or by a variable transformation, the new model removes the implicitness by tracing the particle distribution functions (PDFs) back in time along their characteristic paths during the collision process. An interpolation scheme is needed to evaluate the PDFs at the traced-back locations. By using the first-order interpolation, the resulting model allows for the straightforward replacement of ${f_{\alpha}}^{eq,n+1}$ by ${f_{\alpha}}^{eq,n}$ no matter where it appears. After comparing the new model with the existing models under different numerical conditions (e.g. different flux schemes and time marching schemes) and using the new model to successfully modify the variable transformation technique, three conclusions can be drawn. First, the new model can improve the accuracy by almost an order of magnitude. Second, it can slightly reduce the computational cost. Therefore, the new scheme improves accuracy without extra cost. Finally, the new model can significantly improve the ${\Delta}t/{\tau}$ limit compared to the temporal interpolation model while having the same ${\Delta}t/{\tau}$ limit as the variable transformation approach. The new scheme with a second-order interpolation is also developed and tested; however, that technique displays no advantage over the simple first-order interpolation approach. Both numerical and theoretical analyses are also provided to explain why the new implicit scheme with simple first-order interpolation can outperform the same scheme with second-order interpolation, as well as the existing temporal extrapolation and variable transformation schemes.

Figures

Figures reproduced from arXiv: 1908.08531 by the authors.

Figure 1
Figure 1. Advection of PDFs along characteristics Unlike the TE and VT schemes, the SLIC method tracks the PDFs along their characteristics back in time in a Lagrangian way and was initially proposed by Groppi et al. [50]. The SLIC method consists of two steps, which will be explained as follows. First, according to Eq. (23), at the barycenter P of any CV, as shown in [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 3
Figure 3. First, the TE and VT schemes are almost identical in terms of error; and second, the SLIC+INT2 approach produces larger errors than SLIC+INT1. The analysis behind these observations will be given in subsequent subsections [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 7
Figure 7. The effect of viscosity on lid-driven square cavity flow [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figures from the paper (4 more)
Figure 8
Figure 8. Figure 8: The numerical viscosities of different implicit collision schemes 5.2 Computational cost The measured computational cost is the update time, 𝑡𝑈, in this paper, which is defined as: 𝑡𝑈 = 𝑡𝑇 𝑊𝑆 (43) where 𝑡𝑇 is the total runtime for the simulation, W is the total number …
Figure 10
Figure 10. Figure 10 [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 10
Figure 10. Figure 10: The transient behaviors of 𝜕𝑓𝛼 𝑒𝑞/𝜕𝑡 for different models on the Taylor-Green vortex flow for (a) direction 0; (b) direction 1; (c) direction 2; (d) direction 3; (e) direction 4; (f) direction 5; (g) direction 6; (h) direction 7; (i) direction 8 6. The application of …
Figure 4
Figure 4. Figure 4: This is because the PL flux scheme can produce much less numerical viscosity than the SOU [PITH_FULL_IMAGE:figures/full_fig_p025_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 49 canonical work pages

  1. [1]

    G. R. McNamara and G. Zanetti, Use of the Boltzmann Equation to Simulation Lattice-Gas Automata, Phys. Rev. Lett. 61, 2332 (1988)

    work page 1988

  2. [2]

    F. J. Higuera, S. Succi, and R. Benzi, Lattice Gas Dynamics with Enhanced Collisions, Europhys. Lett. 9, 4 (1989)

    work page 1989

  3. [3]

    F. J. Higuera and S. Succi, Simulating the Flow Around A Circular Cylinder with A Lattice Boltzmann Equation, Europhys. Lett. 8, 6 (1989)

    work page 1989

  4. [4]

    F. J. Higuera and J. Jiménez, Boltzmann Approach to Lattice Gas Simulations, Europhys. Lett. 9, 7 (1989)

    work page 1989

  5. [5]

    Succi, Lattice Boltzmann 2038, Europhys

    S. Succi, Lattice Boltzmann 2038, Europhys. Lett. 109, 5 (2015)

    work page 2015

  6. [6]

    G. A. Bird, Molecular Gas Dynamics and the Direct Simulation of Gas Flows, A Clarendon Press, 1994

    work page 1994

  7. [7]

    Succi, The Lattice Boltzmann Equation: For Complex States of Flowing Matter, Oxford University Press, 2018

    S. Succi, The Lattice Boltzmann Equation: For Complex States of Flowing Matter, Oxford University Press, 2018

    work page 2018

  8. [8]

    Chen and G

    S. Chen and G. D. Doolen, Lattice Boltzmann Method for Fluid Flows, Annu. Rev. Fluid Mech. 30 (1998)

    work page 1998

Show all 49 references
  1. [9]

    Succi, G

    S. Succi, G. Amati, and R. Benzi, Challenges in Lattice Boltzmann Computing, J. Stat. Phys. 81, 5 (1995)

    1995

  2. [10]

    Nannelli and S

    F. Nannelli and S. Succi, The Lattice Boltzmann Equation on Irregular Lattices, J. Stat. Phys. 68, (1992)

    1992

  3. [11]

    Peng and H

    G. Peng and H. Xi, On Boundary Conditions in the Finite Volume Lattice Boltzmann Method on Unstructured Meshes, Int. J. Modern Phys. C 106 (1999)

    1999

  4. [12]

    G. Peng, H. Xi, C. Duncan, and S. Chou, Lattice Boltzmann Method on Irregular Meshes, Phys. Rev. E 58, 4 (1998)

    1998

  5. [13]

    G. Peng, H. Xi, C. Duncan, and S. Chou. Finite Volume Scheme for the Lattice Boltzmann Method on Unstructured Meshes, Phys. Rev. E 59, 4 (1999)

    1999

  6. [14]

    H. Xi, G. Peng, and S. Chou, Finite-Volume Lattice Boltzmann Method, Phys. Rev. E 59, 6202 (1999)

    1999

  7. [15]

    Ubertini, G

    S. Ubertini, G. Bella, and S. Succi, Lattice Boltzmann Method on Unstructured Grids: Further Developments, Phys. Rev. E 68, 016701 (2003)

    2003

  8. [16]

    Zarghami, M

    A. Zarghami, M. Ma ghrebi, J. Ghasemi, and S. Ubertini, Lattice Boltzmann Finite Volume Formulation with Improved Stability, Commun. Comput. Phys. 12, 42 (2012)

    2012

  9. [17]

    Stiebler, J

    M. Stiebler, J. Tölke, and M. Krafczyk, An Upwind Discretization Scheme for the Finite Volume Lattice Boltzmann Method, Comput. & Fluids 35, 814 (2006)

    2006

  10. [18]

    D. V. Patil and K. N. Lakshmisha, Finite Volume TVD Formulation of Lattice Boltzmann Simulation on Unstructured Mesh, J. Comput. Phys. 228, 5262 (2009)

    2009

  11. [19]

    D. V. Patil and K. N. Lakshmisha, Two-Dimensional Flow Past Circular Cylinders Using Finite Volume Lattice Boltzmann Formulations, Int. J. Num. Methods Fluids 69, 1149 (2012)

    2012

  12. [20]

    S. K. Choi and C. L. Lin, A Simple Finite-Volume Formulation of the Lattice Boltzmann Method for Laminar and Turbulent Flows, Num. Heat Transfer Part B 58, 242 (2010)

    2010

  13. [21]

    Chen and L

    L. Chen and L. Schaefer, A Unified and Preserved Dirichlet Boundary Treatment for the Cell-Centered Finite Volume Discrete Boltzmann Method, Phys. Fluids 27, 027104 (2015)

    2015

  14. [22]

    Lee and C

    T. Lee and C. Lin, A Characteristic Galerkin Method for Discrete Boltzmann Equation, J. Comput. Phys. 171, 336 (2001)

    2001

  15. [23]

    Lee and C

    T. Lee and C. Lin, An Eulerian Description of the Streaming Process in the Lattice Boltzmann Equation, J. Comput. Phys. 185, 445 (2003)

    2003

  16. [24]

    K. E. Wardle and T. Lee, Finite Element Lattice Boltzmann Simulations of Free Surface Flow in A Concentric Cylinder, Comput. Math. Appl. 65, 230 (2013)

    2013

  17. [26]

    D. Yu, R. Mei, L. Luo, and W. Shyy, Viscous Flow Computations with the Method of Lattice Boltzmann Equation, Prog. Aero. Sci. 39, 329 (2003)

    2003

  18. [27]

    Chen and L

    L. Chen and L. Schaefer, Godunov-type Upwind Flux Schemes of the Two-Dimensional Finite Volume Discrete Boltzmann Method, Comp. Math. App. 75, 3105 (2018)

    2018

  19. [28]

    L. Chen, L. Schaefer , and X. Cai. An Accurate Unstructured Finite Volume Discrete Boltzmann Method , ASME 2018 Int. Mech. Eng. Con. Expo. IMECE2018-87136, V007T09A010 (2018)

    2018

  20. [29]

    P. L. Bhatnagar, E. P. Gross, and M. Krook. A Model for Collision Processes in Gases. I. Small Amplitude Processes in Charged and Neutral One-Component System, Phys. Rev. 94, 3 (1954)

    1954

  21. [30]

    Mei, and W

    R. Mei, and W. Shyy, On the Finite Difference-Based Lattice Boltzmann Method in Curvilinear Coordinates, J. Comput. Phys. 143, 426 (1998)

    1998

  22. [31]

    Bardow, I

    A. Bardow, I. V. Karlin, and A. A. Gusev , General Characteristic-Based Algorithm for Off-Lattice Boltzmann Simulations, Europhys. Lett. 75, 3 (2006)

    2006

  23. [32]

    X. He, S. Chen, and G. D. Doolen, A Novel Thermal Model for the Lattice Boltzmann Method in Incompressible limit, J. Comput. Phys. 146, 1 (1998)

    1998

  24. [33]

    Bösch and I

    F. Bösch and I. V. Karlin, Exact Lattice Boltzmann Equation, Phys. Rev. Lett. 111, 9 (2013)

    2013

  25. [34]

    Ubertini, P

    S. Ubertini, P. Asinari, and S. Succi, Three Ways to Lattice Boltzmann: A Unified Time-Marching Picture, Phys. Rev. E 81, 016311 (2010)

    2010

  26. [35]

    Guo and T

    Z. Guo and T. S. Zhao, Explicit Finite-Difference Lattice Boltzmann Method for Curvilinear Coordinates, Phys. Rev. E 67, 066709 (2003)

    2003

  27. [36]

    Z. Guo, R. Wang, and K. Xu, Discrete Unified Gas Kinetic Scheme for All Knudsen Number Flows: Low Speed Isothermal Case, Phys. Rev. E 88, 033305 (2013)

    2013

  28. [37]

    Z. Guo, R. Wang, and K. Xu, Discrete Unified Gas Kinetic Scheme for All Knudsen Number Flows. II. Thermal Compressible Case, Phys. Rev. E 91, 3 (2015)

    2015

  29. [38]

    Zhang, L

    Y. Zhang, L. Zhu, R. Wang, and Z. Guo, Discrete Unified Gas Kinetic Scheme for All Knudsen Number Flows. III. Binary Gas Mixtures of Maxwell Molecules, Phys. Rev. E 97, 5 (2018)

    2018

  30. [39]

    B. Shan, P. Wang, Y. Zhang, and Z. Guo, Discrete Unified Gas Kinetic Scheme for All Knudsen Number Flows. IV. Strongly Inhomogeneous Fluids, Phys. Rev. E 101, 043303 (2020)

    2020

  31. [40]

    P. R. Rao and L. A. Schaefer, Numer ical Stability of Explicit Off-Lattice Boltzmann Schemes: A Comparative Study, J. Comput. Phys. 285, (2015)

    2015

  32. [41]

    Begue, A

    M. Begue, A. Ghizzo, P. Bertrand, E. Sonnendrucker, and O. Coulaud, Two -Dimensional Semi-Lagrangian Vlasov Simulations of Laser– Plasma Interaction in the Relativistic Regime, J. Plasma Phys. 62, 4 (1999)

    1999

  33. [42]

    Besse and E

    N. Besse and E. Sonnendrucker, Semi-Lagrangian Schemes for the Vlasov Equation on An Unstructured Mesh of Phase Space, J. Comput. Phys. 191, (2003)

    2003

  34. [43]

    J. M. Qiu and C. W. Shu, Conservative High Order Semi-Lagrangian Finite Difference WENO Methods for Advection in Incompressible Flow, J. Comput. Phys. 230, 4 (2011)

    2011

  35. [44]

    X. Cai, W. Guo, and J. M. Qiu, A High Order Conservative Semi-Lagrangian Discontinuous Galerkin Method for Two-Dimensional Transport Simulations, J. Sci. Comput. 73, (2017)

    2017

  36. [45]

    C. Shu, X. D. Niu, and Y. T. Chew , Taylor-Series Expansion and Least-Squares-Based Lattice Boltzmann Method: Two -Dimensional Formulation and Its Applications, Phys. Rev. E 65, 3 (2002)

    2002

  37. [46]

    Cheng and K

    M. Cheng and K. C. Hung, Lattice Boltzmann Method on Nonuniform Mesh, Int. J. Comput. Eng. Sci. 05, 02 (2004)

    2004

  38. [47]

    Krämer, K

    A. Krämer, K. Küllmer, D. Reith, W. Joppich, and H. Foysi, Semi-Lagrangian Off-lattice Boltzmann Method for Weakly Compressible Flows, Phys. Rev. E 95, 2 (2017). DOI: https://doi.org/10.1103/PhysRevE.101.063301 Page 32 of 32

    2017 doi

  39. [48]

    Di Ilio, B

    G. Di Ilio, B. Dorschner, D . Bella, and S. Succi. Simulation of Turbulent Flows with the Entropic Multirelaxation Time Lattice Boltzmann Method on Body-Fitted Meshes, J. Fluid Mech. 849, (2018)

    2018

  40. [49]

    Dorschner, F

    B. Dorschner, F. Bösch, and I. V. Karlin, Particles on Demand for Kinetic Theory, Phys. Rev. Lett. 121, 130602 (2018)

    2018

  41. [50]

    Groppi, G

    M. Groppi, G. Russo, and G. Stracquadanio, High Order Semilagrangian Methods for the BGK Equation, Commun. Math. Sci. 14, 2 (2016)

    2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.