Pith. sign in

REVIEW 4 major objections 5 minor 41 references

Adjoint shape optimization from the continuum to free-molecular gas flows

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read This paper develops an adjoint shape-optimization method for gas flows that computes accurate drag sensitivities and converges to optimal airfoils in roughly a dozen iterations across Knudsen numbers from 0.001 to 10.

desk verdict First adjoint shape-optimization for diffuse-boundary BGK, with real FDM validation, but leaning heavily on [18] and worth refereeing with a request for self-contained derivation. read the letter →

arxiv 2501.00923 v1 pith:33J3MZHT submitted 2025-01-01 physics.comp-ph

Ruifeng Yuan , Lei Wu This is my paper · ORCID
classification physics.comp-ph
keywords shapeoptimizationadjointmethodrarefiedgasflowdiscretevelocityBGKequationdiffuseboundaryconditionCSTparameterizationKnudsennumber
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 tries to establish that shape optimization of solid bodies in gas flows can be done uniformly from the continuum regime to the free-molecular regime by deriving shape sensitivities from the BGK kinetic equation with diffuse-reflection boundary conditions. Previous adjoint optimization for rarefied gases used topology optimization, which requires one design variable per mesh cell and a non-body-fitted boundary; the proposed method uses a body-fitted mesh and parameterizes the boundary with a small number of design variables. The paper claims that the calculated sensitivities match finite-difference sensitivities at Knudsen numbers 0.001, 0.1, and 10, that the CST parameterization removes the sensitivity oscillations caused by velocity-space discretization, and that airfoil drag optimizations converge in roughly a dozen steps in 5 to 20 minutes on 40 to 160 cores. In a Mach-2 case the drag is reduced by 13.36%. If these claims hold, gradient-based design becomes practical for rarefied and multiscale gas flows without solving expensive surrogate problems.

What carries the argument

The load-bearing object is the adjoint BGK system, specifically the adjoint distribution function $\varphi$ and its wall flux $\phi_w$; the adjoint equation propagates sensitivity information backwards through the flow, and the wall boundary condition converts the objective into a constant flux source $m_J$. Around this, the method combines continuous adjoint analysis (to derive that equation) with discrete-adjoint-style differentiation (to convert derivatives with respect to surface element area, centroid, and normal into derivatives with respect to boundary nodes), and then a chain rule maps those node derivatives onto CST parameters. The CST parameterization is more than a geometry representation: it acts as a low-pass filter that removes the high-frequency sensitivity oscillations induced by discretizing the molecular velocity space, making the gradient usable by the optimizer.

What would settle it

Take one of the optimized airfoils (e.g., the Kn = 0.01 supersonic case) and recompute its drag with an independent, high-accuracy kinetic solver such as direct simulation Monte Carlo under the same Mach number, wall temperature, and Knudsen number; if the recomputed drag is not lower than that of the initial NACA0012 airfoil, or if significant differences appear when the same comparison is done for the other optimized shapes, the adjoint gradient that drove the optimization is wrong.

Watch

Extended reading notes

Core claim

The central discovery is that the adjoint of the BGK equation, together with the diffuse-reflection boundary condition, provides shape sensitivities for gas flows at every Knudsen number, and that the velocity-space discretization noise in those sensitivities is eliminated by parameterizing the boundary with CST shape functions. The paper derives a Lagrangian for the objective (drag or heat) constrained by the steady BGK equation, obtains a continuous adjoint equation whose wall boundary condition contains the objective moment $m_J$ as a flux source, and then computes the final sensitivity by discrete-like differentiation of the discretized objective and adjoint boundary terms with respect to the boundary element's area, centroid position, and outward normal. This sensitivity is transferred through a chain rule to CST design variables and fed to an SLSQP quasi-Newton optimizer. Numerically, the paper shows that the adjoint sensitivity agrees with central finite differences under the same velocity-space discretization, that the CST parameterization yields smooth and mesh-converged sensitivity profiles, and that the optimized airfoil shapes match those from a previous topology-optimization method.

Load-bearing premise

The governing assumption is that the implicit multiscale gas-kinetic scheme adopted from Reference [18] works correctly for the adjoint BGK system; the paper states that the detailed computation procedures are omitted because they are 'generally the same' as in [18], so if that solver is inaccurate for the adjoint equation, the sensitivities and therefore all optimized shapes would be wrong.

Editorial extensions

If this is right

  • Drag-reduction optimizations of airfoils in a channel converge in 9 to 18 optimization steps at Kn = 0.001, 0.1, and 10, with total runtimes of 6 to 20 minutes on 40 to 160 cores.
  • Under Mach 2 flow, the method reduces drag by 13.36% at Kn = 0.01 and 4.22% at Kn = 0.5, with the optimized airfoil converting a detached bow shock into two oblique shocks.
  • The method reproduces the optimal shapes obtained by the earlier topology-optimization approach of Reference [18] for several flow conditions, despite using a completely different geometry representation.
  • Because the boundary is body-fitted and the design space is low-dimensional (18 CST parameters in the tested setup), quasi-Newton optimization is feasible without the hundreds of design variables that hinder topology optimization.
  • The sensitivity validation against finite differences across Kn = 0.001 to 10 indicates that the adjoint derivation, including the diffuse boundary condition, is algebraically correct in all flow regimes.

Reading between the lines

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

  • Beyond the paper: the same hybrid continuous/discrete adjoint structure could be applied to other kinetic models, such as the Shakhov or ellipsoidal BGK equations, with extra collision terms appearing in the adjoint equation; nothing in the derivation is specific to the relaxation-time model.
  • Beyond the paper: if CST smoothing is understood as a low-pass filter on a noisy gradient, then coarser velocity grids might be usable in optimization without losing accuracy, at the price of reparameterizing the boundary; this could cut cost further in three dimensions.
  • Beyond the paper: the paper's claim that extension to 3D only requires the 3D analogue of the geometric derivative (equation 37) is plausible, but the velocity-space oscillations that had to be smoothed in 2D may be more severe in 3D because the velocity grid is much larger.
  • Beyond the paper: the non-monotonic optimal-thickness trend with Knudsen number (maximum thickness at Kn = 0.1) is a prediction about rarefied aerodynamics that could be tested experimentally or with high-fidelity direct simulation Monte Carlo in a controlled flow.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes an adjoint-based shape optimization method for gas flows described by the BGK equation with diffuse-reflection boundary conditions. A continuous adjoint equation is stated and solved with a multiscale discrete-velocity finite-volume scheme taken from the authors' prior work [18], and the sensitivity of boundary functionals is obtained through a discrete-adjoint-like differentiation of the discretized boundary integrals. The sensitivities are validated against finite differences for an elliptic cylinder at Kn = 10, 0.1, and 0.001, and the method is applied to drag reduction of NACA0012-type airfoils in subsonic channel flow and supersonic flow, reporting drag reductions up to 13.36% and convergence in roughly a dozen optimization steps.

Significance. The main contribution is a shape-optimization framework that avoids the high design-variable count and lower boundary accuracy of topology optimization while covering the whole Knudsen regime. The finite-difference validation of sensitivities at three Knudsen numbers is a genuine strength and provides direct evidence that the adjoint implementation is discretely consistent. The observation that CST parameterization suppresses velocity-discretization-induced sensitivity oscillations is practically useful. However, the paper defers the core numerical scheme to a prior publication, and the derivation of the adjoint equation is omitted; these gaps currently limit the reproducibility and self-containedness of the central method. If those details are supplied and the topology-optimization comparison is appropriately qualified, the method would be a valuable contribution to rarefied-flow design.

major comments (4)
  1. [Section 3.1, Eqs. (26)-(27)] The discretization and solution procedure for both the primal and adjoint equations are deferred to the authors' prior work [18], with the statement that the procedures are 'generally the same' as in that reference. Since the accuracy of the adjoint sensitivities and all downstream optimization results depends directly on this multiscale scheme, and since the adjoint equation (22) involves terms (negative advection, boundary integrals coupling to phi_w, the phi_tau term) that are specific to this problem, the paper should include either the full derivation and discretization of the adjoint system or a substantial appendix describing the numerical scheme. Without this, the central claim is not reproducible from the manuscript alone.
  2. [Section 2.5, Eq. (22)] The adjoint governing equation is stated without derivation. The treatment of the diffuse boundary condition, especially the coupling through the wall density rho_w and the associated integral boundary condition, is nontrivial; the finite-difference comparisons in Section 4.1 test the final sensitivity but do not demonstrate that the derivation is correct for other objectives or flow conditions. A derivation, or at least a detailed outline of the variational steps leading to Eq. (22), should be added.
  3. [Section 4.2 and 4.3, Figs. 8 and 13] The agreement with the topology-optimization results of [18] is presented as validation, but it is not independent: both methods use the same multiscale gas-kinetic solver, and the shape optimization is initialized with the chord length set to the [18] optimum, as the paper itself states in Section 4.2. The Re = 200 topology result in [18] is also admitted to be unconverged. This comparison should be reframed as a consistency check between two geometry representations rather than as independent verification of physical accuracy, and the text should be adjusted accordingly.
  4. [Section 4.1, Fig. 2] The node-coordinate sensitivities show persistent oscillations that do not disappear even at a 300x300 velocity-space discretization, and the paper asserts that CST parameterization removes these oscillations as a 'low-pass filter'. It is not demonstrated, however, whether this filtering removes numerical noise or also removes physically meaningful high-frequency sensitivity that could bias the optimizer. A quantitative convergence study (for example, CST sensitivities at increasing Bernstein-polynomial order, or a comparison against explicitly filtered node sensitivities) would support the claim that the parameterization does not alter the converged optimum.
minor comments (5)
  1. [Section 4.1, Eq. (46)] The finite-difference step size epsilon in Eq. (46) is never specified; please provide the value used and, ideally, a short step-size sensitivity check.
  2. [Section 4.1, Fig. 2 caption] The FDM results are reported only for the 60x60 velocity discretization; please clarify whether finer FDM runs were attempted and, if not, why the 60x60 comparison was considered sufficient.
  3. [Section 4.1, text near Fig. 4] The sentence 'It is shown that the our adjoint results agree well with the FDM results' contains a grammatical error and should read 'our adjoint results'.
  4. [Section 5] The claim that extension to 3D is 'straightforward' is plausible but not substantiated; at minimum, the 3D analogue of Eq. (37) should be outlined or explicitly referenced.
  5. [Section 3.4, Eq. (44)] The convergence criterion (44) depends only on the change in the objective; reporting the constraint feasibility at each iteration would be useful to demonstrate that the volume constraint is actually satisfied at convergence.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the adjoint sensitivity and optimization are self-contained, with a minor, non-load-bearing reliance on the authors' prior multiscale solver.

full rationale

The adjoint sensitivity derivation in Section 2.5 is self-contained: the Lagrangian (18), adjoint equation (22), and sensitivity formulas (33)-(36) are derived from the stated BGK model and diffuse boundary condition, with no fitted constants. The finite-difference checks in Section 4.1 verify the discrete consistency of the adjoint against the same solver; this is a standard numerical consistency test, not an equivalence-by-construction. The optimization results are compared with the authors' previous topology optimization [18], which shares the same gas-kinetic solver, so this comparison is a consistency check rather than an independent physical benchmark; the paper itself notes that the Re=200 case in [18] did not fully converge (Section 4.2). The main load-bearing component not re-derived here is the multiscale gas-kinetic scheme from [18] (Section 3.1), but this is a normal reuse of prior published work, not a prediction that reduces to its own input. No parameter is fitted to the target drag values, and no uniqueness theorem or prior claim by the same authors is invoked to forbid alternative designs. Hence no significant circularity.

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

The central claim rests on standard kinetic theory assumptions (BGK model, diffuse reflection, hard-sphere viscosity) and on the prior multiscale solver from the same authors. No free parameters are fitted to data; the only numerical choices are discretization resolutions. The paper introduces no new physical entities.

assumptions (5)
  • domain assumption BGK model equation with relaxation time tau = mu/p adequately represents gas flows from continuum to free-molecular regimes.
    The paper uses BGK instead of the full Boltzmann equation, citing standard references; this is a common approximation but limits quantitative accuracy.
  • domain assumption Diffuse reflection with zero wall velocity and specified wall temperature describes gas-surface interaction.
    The method assumes all molecules thermalize at the wall; real surfaces may have accommodation coefficients other than 1.
  • domain assumption Hard-sphere viscosity model mu is proportional to sqrt(T) is valid.
    The paper uses this to set tau; other models would change the results.
  • ad hoc to paper The multiscale gas-kinetic scheme from [18] is accurate for both primal and adjoint equations.
    The paper does not describe the scheme, relying on the authors' prior work; this is a self-referential assumption.
  • standard math Velocity-space integration limits are insensitive to boundary normal perturbations because integrands vanish at the bounds.
    Section 3.2 point 1; this is a stated mathematical claim used to simplify the boundary variation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adjoint shape optimization from the continuum to free-molecular gas flows." pith.science (2026). https://pith.science/paper/33J3MZHT

@misc{pith2026250100923,
  author       = {Pith},
  title        = {Pith review of: Adjoint shape optimization from the continuum to free-molecular gas flows},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/33J3MZHT}},
  note         = {Machine review of arXiv:2501.00923}
}
read the original abstract

An adjoint-based shape optimization method for solid bodies subjected to both rarefied and continuum gas flows is proposed. The gas-kinetic BGK equation with the diffuse-reflection boundary condition is used to describe the multiscale gas flows. In the vicinity of the gas-solid interface, a body-fitted mesh is utilized, and the sensitivity with respect to the boundary geometry is analyzed through a combined continuous and discrete adjoint methods. The primal and adjoint governing equations are resolved using efficient multiscale numerical schemes, ensuring the precision of the sensitivity analysis in all flow regimes. The sensitivity data is subsequently integrated into a quasi-Newton optimization algorithm to facilitate rapid convergence towards the optimal solution. Numerical experiments reveal that the discretization of the molecular velocity space can induce sensitivity oscillations; however, these can be effectively eliminated by employing appropriate parameterization of the boundary geometry. In optimizing 2D airfoils for drag reduction under varying degrees of gas rarefaction, our method achieves the optimal solution in just a dozen optimization iterations and within a time frame of 5 to 20 minutes (utilizing parallel computation with 40 to 160 cores), thereby underscoring its exceptional performance and efficiency.

Figures

Figures reproduced from arXiv: 2501.00923 by the authors.

Figure 1
Figure 1. The setups of mesh and boundary condition for the flow pa [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Flow over an elliptic cylinder inside a channel at Kn = 10: sensit [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
Figure 3
Figure 3. Flow over an elliptic cylinder inside a channel at Kn = 10: sensit [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Flow over an elliptic cylinder inside a channel: sensitivity of the [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]
Figure 5
Figure 5. Figure 5: Optimization of the airfoil inside a channel: the mesh for the [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]
Figure 7
Figure 7. Figure 7: The most [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 6
Figure 6. Figure 6: Optimization of the airfoil inside a channel: streamlines and p [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]
Figure 7
Figure 7. Figure 7: Optimization of the airfoil inside a channel: comparison of th [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]
Figure 8
Figure 8. Figure 8: Optimization of the airfoil inside a channel: the comparison o [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: Optimization of the airfoil under supersonic flow: the boun [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Optimization of the airfoil under supersonic flow: the mes [PITH_FULL_IMAGE:figures/full_fig_p023_10.png]
Figure 11
Figure 11. Figure 11: Optimization of the airfoil under supersonic flow: stream [PITH_FULL_IMAGE:figures/full_fig_p023_11.png]
Figure 12
Figure 12. Figure 12: Optimization of the airfoil under supersonic flow: compar [PITH_FULL_IMAGE:figures/full_fig_p024_12.png]
Figure 13
Figure 13. Figure 13: Optimization of the airfoil under supersonic flow: the com [PITH_FULL_IMAGE:figures/full_fig_p026_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [18]

    R. Yuan, L. Wu, A design optimization method for rarefied and co ntinuum gas flows, Journal of Computational Physics 517 (2024) 113366

    work page 2024

  2. [1]

    A. I. Forrester, A. J. Keane, Recent advances in surrogate- based optimization, Progress in aerospace sciences 45 (1-3) (2009) 50–79

    work page 2009

  3. [2]

    Han, Kriging surrogate model and its application to design optim ization: a review of recent progress, Acta Aeronautica et Astronautica Sinica 37 (11) (2016) 3197–322 5

    Z. Han, Kriging surrogate model and its application to design optim ization: a review of recent progress, Acta Aeronautica et Astronautica Sinica 37 (11) (2016) 3197–322 5

    work page 2016

  4. [3]

    Bhosekar, M

    A. Bhosekar, M. Ierapetritou, Advances in surrogate based m odeling, feasibility analysis, and optimiza- tion: A review, Computers & Chemical Engineering 108 (2018) 250–2 67

    work page 2018

  5. [4]

    Jameson, Aerodynamic design via control theory, Journal o f scientific computing 3 (1988) 233–260

    A. Jameson, Aerodynamic design via control theory, Journal o f scientific computing 3 (1988) 233–260

    work page 1988

  6. [5]

    Jameson, Aerodynamic shape optimization using the adjoint me thod, Lectures at the Von Karman Institute, Brussels (2003)

    A. Jameson, Aerodynamic shape optimization using the adjoint me thod, Lectures at the Von Karman Institute, Brussels (2003)

    work page 2003

  7. [6]

    E. Reed, H. Alkandry, J. Codoni, J. McDaniel, I. Boyd, Investiga tion of the interactions of reaction control systems with mars science laboratory aeroshell, in: 48th A IAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition, 2010, p. 1558

    work page 2010

  8. [7]

    J. Li, D. Jiang, X. Geng, J. Chen, Kinetic comparative study on ae rodynamic characteristics of hy- personic reentry vehicle from near-continuous flow to free molecu lar flow, Advances in Aerodynamics 3 (2021) 1–10

    work page 2021

Show all 41 references
  1. [8]

    M. H. Hablanian, High-vacuum technology: a practical guide, CRC Press, 1997

    1997

  2. [9]

    Sharipov, P

    F. Sharipov, P. Fahrenbach, A. Zipp, Numerical modeling of the H olweck pump, Journal of Vacuum Science & Technology A 23 (5) (2005) 1331–1339

    2005

  3. [10]

    Bakshi, EUV lithography, SPIE press, 2009

    V. Bakshi, EUV lithography, SPIE press, 2009

    2009

  4. [11]

    Tantos, S

    C. Tantos, S. Varoutis, C. Day, Deterministic and stochastic m odeling of rarefied gas flows in fusion particle exhaust systems, Journal of Vacuum Science & Technolog y B 38 (6) (2020)

    2020

  5. [12]

    K. Lin, S. Zhang, C. Liu, H. Yang, B. Zhang, Aerodynamic optimiz ation of naca 0012 airfoils with attached gurney flap in the rarefied gas flow, AIP Advances 13 (12 ) (2023)

    2023

  6. [13]

    G. A. Bird, Molecular gas dynamics and the direct simulation of gas flows, Clarendon Press, 1994

    1994

  7. [14]

    J. Y. Yang, J. C. Huang, Rarefied flow computations using nonlin ear model Boltzmann equations, Journal of Computational Physics 120 (2) (1995) 323–339

    1995

  8. [15]

    Mieussens, Discrete velocity model and implicit scheme for the BGK equation of rarefied gas dy- namics, Mathematical Models and Methods in Applied Sciences 10 (08) (2000) 1121–1149

    L. Mieussens, Discrete velocity model and implicit scheme for the BGK equation of rarefied gas dy- namics, Mathematical Models and Methods in Applied Sciences 10 (08) (2000) 1121–1149

    2000

  9. [16]

    A. Sato, T. Yamada, K. Izui, S. Nishiwaki, S. Takata, A topology optimization method in rarefied gas flow problems using the Boltzmann equation, Journal of Computatio nal Physics 395 (2019) 60–84

    2019

  10. [17]

    K. Guan, K. Matsushima, Y. Noguchi, T. Yamada, Topology optim ization for rarefied gas flow problems using density method and adjoint IP-DSMC, Journal of Computatio nal Physics 474 (2023) 111788

    2023

  11. [19]

    Caflisch, D

    R. Caflisch, D. Silantyev, Y. Yang, Adjoint DSMC for nonlinear Bo ltzmann equation constrained optimization, Journal of Computational Physics 439 (2021) 11040 4. 27

    2021

  12. [20]

    Cercignani, The Boltzmann Equation and its Applications, Sprin ger-Verlag, New York, 1988

    C. Cercignani, The Boltzmann Equation and its Applications, Sprin ger-Verlag, New York, 1988

    1988

  13. [21]

    Wu, Rarefied Gas Dynamics: Kinetic Modeling and Multi-Scale Simu lation, Springer, 2022

    L. Wu, Rarefied Gas Dynamics: Kinetic Modeling and Multi-Scale Simu lation, Springer, 2022

    2022

  14. [22]

    P. L. Bhatnagar, E. P. Gross, M. Krook, A model for collision pr ocesses in gases. I. Small amplitude processes in charged and neutral one-component systems, Phy sical Review 94 (3) (1954) 511

    1954

  15. [23]

    Tsien, Superaerodynamics, mechanics of rarefied gases , Journal of the Aeronautical Sciences 13 (12) (1946) 653–664

    H.-S. Tsien, Superaerodynamics, mechanics of rarefied gases , Journal of the Aeronautical Sciences 13 (12) (1946) 653–664

    1946

  16. [24]

    Chapman, T

    S. Chapman, T. G. Cowling, The mathematical theory of non-un iform gases: an account of the kinetic theory of viscosity, thermal conduction and diffusion in gases, Cam bridge university press, 1990

    1990

  17. [25]

    Chu, Kinetic-theoretic description of the formation of a sho ck wave, The Physics of Fluids 8 (1) (1965) 12–22

    C. Chu, Kinetic-theoretic description of the formation of a sho ck wave, The Physics of Fluids 8 (1) (1965) 12–22

    1965

  18. [26]

    Nadarajah, A

    S. Nadarajah, A. Jameson, A comparison of the continuous an d discrete adjoint approach to automatic aerodynamic optimization, in: 38th Aerospace sciences meeting and exhibit, 2000, p. 667

    2000

  19. [27]

    J. E. Peter, R. P. Dwight, Numerical sensitivity analysis for aer odynamic optimization: A survey of approaches, Computers & Fluids 39 (3) (2010) 373–391

    2010

  20. [28]

    Z. Guo, K. Xu, R. Wang, Discrete unified gas kinetic scheme for a ll knudsen number flows: Low-speed isothermal case, Physical Review E 88 (3) (2013) 033305

    2013

  21. [29]

    L. Zhu, P. Wang, Z. Guo, Performance evaluation of the gener al characteristics based off-lattice Boltz- mann scheme and DUGKS for low speed continuum flows, Journal of C omputational Physics 333 (2017) 227–246

    2017

  22. [30]

    R. Yuan, C. Zhong, A conservative implicit scheme for steady st ate solutions of diatomic gas flow in all flow regimes, Computer Physics Communications 247 (2020) 1069 72

    2020

  23. [31]

    R. Yuan, S. Liu, C. Zhong, A novel multiscale discrete velocity me thod for model kinetic equations, Communications in Nonlinear Science and Numerical Simulation 92 (2021 ) 105473

    2021

  24. [32]

    B. M. Kulfan, Universal parametric geometry representation method, Journal of aircraft 45 (1) (2008) 142–158

    2008

  25. [33]

    Y. Bu, W. Song, Z. Han, J. Xu, Aerodynamic optimization design o f airfoil based on CST parameteri- zation method, Journal of Northwestern Polytechnical Universit y 31 (5) (2013) 829–836

    2013

  26. [34]

    T. C. Rendall, C. B. Allen, Efficient mesh motion using radial basis fu nctions with data reduction algorithms, Journal of Computational Physics 228 (17) (2009) 62 31–6249

    2009

  27. [35]

    S. G. Johnson, The NLopt nonlinear-optimization package, https://github.com/stevengj/nlopt (2007)

    2007

  28. [36]

    Kraft, A software package for sequential quadratic prog ramming, Forschungsbericht- Deutsche Forschungs- und Versuchsanstalt fur Luft- und Raumfahrt (19 88)

    D. Kraft, A software package for sequential quadratic prog ramming, Forschungsbericht- Deutsche Forschungs- und Versuchsanstalt fur Luft- und Raumfahrt (19 88)

  29. [37]

    C. L. Lawson, R. J. Hanson, Solving least squares problems, SI AM, 1995

    1995

  30. [38]

    10th, 2024 (2022)

    NASA, Turbulence Modeling Resource: 2D NACA 0012 Airfoil Valida tion Case, accessed on Sep. 10th, 2024 (2022). URL https://turbmodels.larc.nasa.gov/naca0012_val.html

    2022

  31. [39]

    Yuan, Application and study of multiscale kinetic method applica ble for all flow regimes (in chinese), Ph.D

    R. Yuan, Application and study of multiscale kinetic method applica ble for all flow regimes (in chinese), Ph.D. thesis, Northwestern Polytechnical University (2021)

    2021

  32. [40]

    Zhang, S

    R. Zhang, S. Liu, J. Chen, H. Jin, C. Zhuo, C. Zhong, Implicit unifi ed gas-kinetic scheme for steady state solution of hypersonic thermodynamic non-equilibrium flows, Commun ications in Nonlinear Science and Numerical Simulation 140 (2025) 108367

    2025

  33. [41]

    Zhang, J

    Y. Zhang, J. Zeng, R. Yuan, W. Liu, Q. Li, L. Wu, Efficient parallel solver for rarefied gas flow using gsis, Computers & Fluids 281 (2024) 106374. 28

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.