Modeling Equations in Wave-Particle Turbulence Simulation

Gaocheng Liu; Kun Xu; Xiaojian Yang

REVIEW 3 major objections 4 minor 16 references

Modeling Equations in Wave-Particle Turbulence Simulation

T0 review · 3 major / 4 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read Wave-particle turbulence simulation with a wall model predicts flat-plate transition on coarse grids by letting stochastic particles carry unresolved non-equilibrium transport.

desk verdict Clean first derivation of the WPTS equations plus a wall-modelled flat-plate demo that works only after the particle-generation coefficient is hand-tuned to the DNS transition location. read the letter →

arxiv 2607.10176 v1 pith:FNQFW6DS submitted 2026-07-11 physics.flu-dyn math-phmath.MP

Xiaojian Yang , Gaocheng Liu , Kun Xu This is my paper · ORCID

classification physics.flu-dynmath-phmath.MP

keywords wave-particleturbulencesimulationdecompositionflat-platetransitionwallmodelgas-kineticschemenon-equilibriumtransportsubgrid-scalemodelingcoarse-grid

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 derives the full set of model equations for wave-particle turbulence simulation (WPTS) by splitting the fluid distribution into a wave part that evolves the large, grid-resolved motions and a particle part that evolves the unresolved motions through free streaming, forcing, and annihilation. It then couples WPTS to a simple equilibrium wall model so that near-wall resolution can be relaxed. On a coarse mesh of roughly two hundred thousand cells the method recovers the skin-friction rise and the mean velocity log-layer of a forced flat-plate transition that previously required a multi-million-cell direct simulation, while a pure gas-kinetic scheme on the same mesh delays transition badly. The practical payoff is a multi-scale kinetic framework that automatically injects more particles only where the grid can no longer resolve the turbulence, offering a route to transitional and wall-bounded flows without extreme mesh refinement.

What carries the argument

Wave-particle decomposition of the distribution function: the wave component supplies the multi-scale interface flux for resolved motion while stochastic particles, generated and annihilated according to a local turbulent time scale, supply the non-equilibrium free-transport contribution that models sub-grid scales.

What would settle it

Rerun the identical coarse-mesh flat-plate case with the same wall model but a different free coefficient (or none) and check whether the skin-friction rise still locks onto the direct-simulation location and whether the mean-velocity profile still recovers a clean log layer.

Watch

Extended reading notes

Core claim

The complete kinetic model equations of WPTS follow directly from a wave-particle decomposition of the distribution function; once these equations are closed by a mixing-length turbulent time scale and coupled to an equilibrium wall model, the resulting scheme on a coarse mesh yields skin-friction and mean-velocity profiles for flat-plate transition that match direct-numerical-simulation data far better than the pure gas-kinetic scheme under identical grid and wall-model settings.

Load-bearing premise

The free coefficient that multiplies the mixing-length formula for the turbulent time scale must be hand-tuned so that particles appear at the right rate; without that tuning the method reverts to ordinary gas-kinetic behaviour and the transition is delayed.

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.

Desk Editor's Note

read the letter

The new material here is the explicit wave-particle kinetic equations (3–6) and the first wall-modelled WPTS run on a forced flat-plate transition. Both are useful. The equations are written carefully: they reduce exactly to GKS when the particle weight vanishes, the force terms cancel so that the combined distribution still satisfies the BGK conservation properties, and the free-transport flux is constructed so that surviving particles are not double-counted. That part is solid and was missing from the earlier WPTS papers.

On the coarse mesh (~0.23 M cells) the method with C_m^{2} = 0.032 recovers a skin-friction rise and a log-layer mean profile that sit close to the DNS reference and are clearly better than pure wall-modelled GKS under identical settings. The wall-model coupling itself is standard and does what it is supposed to do. So the numerical claim in the abstract is true for this one case.

The soft spot is exactly the one the stress-test flags. Particle generation and lifetime are controlled by a mixing-length eddy viscosity whose free coefficient C_m^{2} is dialled until the transition location matches DNS. Set it to zero and you recover the delayed GKS result. No a-priori estimate, no second transitional case, no comparison with any established wall-modelled LES, and no code or data release. The free-parameter list is longer than one would like (C_0, y_ref, density limiter, wall-model activation thresholds). That does not make the equations wrong; it does mean the present evidence is still a demonstration that a suitably tuned particle rate can restore transition, not yet a demonstration of predictive multi-scale modelling.

Who should read it: people already working with kinetic schemes or looking for hybrid wave-particle ideas for transitional wall flows. It is not yet a paper that reorganises the broader LES/RANS conversation. I would send it to referees; the derivation is clean enough and the numerical improvement real enough to deserve a proper look, provided the authors are pressed on calibration and additional cases. I would not cite it myself until the coefficient is constrained independently or more flows appear.

Referee Report

3 major / 4 minor

Summary. The paper derives the complete kinetic model equations of wave-particle turbulence simulation (WPTS) from an additive decomposition of the distribution function into wave (grid-resolved) and particle (stochastic, subgrid) components, interprets the generation/annihilation and forcing terms, and shows that the combined equation recovers a BGK-type relaxation that conserves mass, momentum and energy. It then couples WPTS to an equilibrium wall model and applies the resulting WM-WPTS method to a forced flat-plate transition at Ma = 0.7, Re = 5e4. On a coarse mesh (~0.23 M cells) the method, with a mixing-length closure for the turbulent time scale, produces a skin-friction curve and a mean-velocity profile that match DNS data more closely than wall-modeled GKS under identical settings, thereby claiming multi-scale predictive capability for transitional wall-bounded flows.

Significance. If the wave-particle decomposition and its non-equilibrium particle transport genuinely supply a grid-adaptive multi-scale model that automatically reduces to a kinetic NS solver in laminar regions, the framework would offer a useful alternative to conventional LES/RANS hybrids for transitional flows. The explicit derivation of the model equations (Section 2) and the first demonstration of wall-model coupling are concrete contributions that strengthen the theoretical foundation of earlier WPTS work. The numerical improvement over GKS on a deliberately under-resolved mesh is of practical interest, provided the free coefficients can be constrained independently of the target data.

major comments (3)

[§4.2, Eq. (21), Fig. 2] Section 4.2 and Eq. (21): the turbulent time scale that controls particle generation, annihilation and free-transport weight is closed by a mixing-length formula whose free coefficient C_m^{2} is hand-tuned to 0.032 so that the skin-friction rise coincides with the DNS target (Fig. 2). When C_m^{2} = 0 the method reverts exactly to the failing WM-GKS baseline. No a-priori estimate, grid-convergence study, independent calibration on a different flow, or sensitivity analysis is supplied. Consequently the claimed superiority of WM-WPTS rests on a single adjustable parameter fitted to the very quantity being predicted, which undercuts the assertion of predictive multi-scale modelling.
[§4, Abstract] The entire numerical validation consists of one forced-transition case at a single Mach and Reynolds number, compared on two meshes against one DNS reference. No additional transitional, fully turbulent, or separated wall-bounded flow is shown, nor is any demonstration that the same C_m^{2} (or a fixed procedure for choosing it) works elsewhere. This single-case, single-parameter success is insufficient to support the abstract’s claim of “considerable promise or transitional flow simulations.”
[§3, Eqs. (12)–(13)] Particle sampling (Eqs. 12–13) assumes the production of turbulent kinetic energy to be C_0 e^{-Δt/τ_n} ho e with C_0 = 1 and e = ½U^{2}. This choice is ad-hoc, dimensionally inconsistent with a true TKE production term, and is never varied or justified against measured production budgets. Because the particle fraction (and therefore the non-equilibrium transport) depends directly on this assumption, its influence on the reported skin-friction and mean-velocity results should be quantified.

minor comments (4)

[Figs. 1–2] Figure captions and text repeatedly use “steamwise” for “streamwise” (Figs. 1b, 2b and surrounding prose).
[§4.2] In §4.2 the label “WM-WPTSC 2 m=0” is a typesetting artifact; it should read “WM-WPTS with C_m^{2} = 0 (i.e., WM-GKS)”.
[§4.2, Eq. (21)] The high-density modification α(ρ) and the damping length y_ref = 0.15 appear without reference or sensitivity test; a brief justification or citation would help reproducibility.
[Table 1] Table 1 lists wall units for the DNS mesh R1 but does not state the friction velocity used to non-dimensionalize the coarser meshes M1/M2; consistency of the reported y+ values should be clarified.

Circularity Check

1 steps flagged · score 6.0 of 10

Transition-onset 'prediction' on coarse mesh is obtained by hand-tuning the free coefficient C_m^{2} that sets the particle-generation rate; zeroing it recovers the failing GKS baseline.

fitted input called prediction [§4.2, Eq. (21) and Fig. 2]
"Fig.2 presents the c_f distributions predicted by WM-WPTS with different C_m^{2} values. At C_m^{2} = 0.032, the transition delay is substantially alleviated, and the c_f curve agrees considerably with the DNS data in terms of onset location, and the transitional shape. ... Note that WM-WPTS will reduce to WM-GKS when C_m^{2} = 0."

τ_t = ρ ν_t / p with ν_t = α(ρ) C_m^{2} y^{2} |S| D_srs(y) directly controls the fraction and lifetime of stochastic particles that supply the non-equilibrium transport. C_m^{2} is varied until the predicted transition location matches the DNS target; the same coefficient set to zero recovers the inferior GKS baseline under identical grid and wall-model settings. The reported superiority is therefore forced by the fit of that single free parameter rather than emerging from the wave-particle decomposition itself.

full rationale

The model-equation construction in §2 is a consistent decomposition (sources and forces cancel by design to recover the BGK equation) and is not itself circular. The load-bearing numerical claim, however, is that WM-WPTS on mesh M2 recovers the DNS skin-friction curve and log-layer velocity while WM-GKS does not. That improvement appears only after the single free coefficient C_m^{2} inside the mixing-length closure for τ_t (Eq. 21) is set to 0.032; the paper itself shows that C_m^{2} = 0 reverts exactly to the delayed-transition GKS result. No a-priori estimate, independent calibration, or multi-case validation of C_m^{2} is supplied. Consequently the reported agreement with DNS is a fitted-input-called-prediction rather than an independent multi-scale prediction. Self-citations to the authors’ prior WPTS papers supply the algorithmic framework but are not the source of the circularity; the circularity is the post-hoc tuning of the particle weight that produces the claimed superiority.

Assumptions & free parameters 5 free parameters · 5 assumptions · 2 invented entities

The central numerical claim rests on a kinetic decomposition that is postulated rather than derived from the Boltzmann equation, on a mixing-length closure whose coefficient is fitted to the target DNS, and on a standard equilibrium wall model whose constants are taken from the literature. No independent experimental or theoretical constraint fixes C_m^{2}; the particle generation amplitude C_0 is likewise set by hand to 1.0.

free parameters (5)

C_m^{2} = 0.032
Coefficient in the mixing-length eddy viscosity that sets the particle generation rate and transport time; hand-tuned to 0.032 so that transition location matches DNS (Eq. 21, Fig. 2).
C_0 = 1.0
Amplitude of turbulent kinetic energy production used when sampling particle velocity deviations; set to 1.0 without further justification (§3).
y_ref = 0.15
Reference length that damps the mixing length away from the wall in scale-resolving mode; chosen as 0.15 (§4.2).
α(ρ) high-density modification = 1/max(1,1+100(ρ-ρ_ref))
Ad-hoc density-dependent factor that multiplies ν_t; functional form and coefficient 100 are chosen by the authors (§4.2).
wall-model activation y+_act and weighting ω_wm = y+_act≈19.5–23.4, ramp 2≤x≤3
Threshold and ramp that switch the wall model on; set to y+_act≈19.5–23.4 and a linear ramp between x=2 and x=3 (§4.1).

assumptions (5)

ad hoc to paper The single-particle distribution function may be additively decomposed into a wave component that lives on the grid and a particle component represented by discrete stochastic particles (Eq. 1).
Postulated at the opening of §2; not derived from the Boltzmann equation or from a filtering operation.
domain assumption The equilibrium state g is the Maxwellian formed with the total temperature that already includes all turbulent kinetic energy converted to heat (Eq. 4).
Standard kinetic-theory closure for the BGK collision operator, adopted without modification.
domain assumption Particle acceleration is given solely by the macroscopic pressure gradient −∇p/ρ (Eq. 10).
Stated as the dominant force; other forces (viscous, inter-particle) are neglected or treated by DSMC collisions.
ad hoc to paper The turbulent time scale τ_t is closed by a mixing-length hypothesis with an extra damping function D_srs appropriate for scale-resolving simulations (Eq. 21).
Introduced in §4.2 specifically for the plate-transition problem; the functional form is not derived from the kinetic equations.
domain assumption Near-wall momentum flux can be replaced by the solution of the equilibrium wall-model ODE (Eq. 19) with van-Driest damping.
Standard equilibrium wall-model assumption taken from the LES literature (Bose & Park, Wang & Moin).

invented entities (2)

wave component f_w and particle component f_p of the distribution function
purpose: To separate grid-resolved continuum transport from unresolved non-equilibrium transport that is carried by discrete particles.
The decomposition is the defining modeling step of WPTS; no independent experimental signature is given that would confirm the split outside the numerical method itself.
particle generation/annihilation source S_gen(f_w,f_p,τ_t,h)
purpose: To transfer mass, momentum and energy between the wave and particle representations according to the local turbulent time scale and mesh size.
Appears as an unclosed source term in Eqs. 3 and 5; its concrete realization is the sampling rule (Eqs. 12–13) that depends on the free parameters above.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Modeling Equations in Wave-Particle Turbulence Simulation." pith.science (2026). https://pith.science/paper/FNQFW6DS

@misc{pith2026260710176,
  author       = {Pith},
  title        = {Pith review of: Modeling Equations in Wave-Particle Turbulence Simulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FNQFW6DS}},
  note         = {Machine review of arXiv:2607.10176}
}

read the original abstract

Recently, the wave-particle turbulence simulation (WPTS) has been proposed as a novel framework for non-equilibrium turbulence modeling and simulation. In this work, for the first time the complete model equations of WPTS are explicitly derived from the perspective of wave-particle decomposition, and the physical mechanism of each term is clearly interpreted. To extend its applicability to wall-bounded flows, the WPTS coupled with wall model is developed, and the introduction of wall model substantially alleviates the near-wall grid-resolution constraint. In the bulk region, the wave component resolves the large-scale structures, whereas the particle component accounts for subgrid-scale modeling through the non-equilibrium transport mechanism. As a result, the coupled method enables accurate predictions of the flat-plate transition on coarse-grid. In particular, the computed skin-friction coefficient and mean velocity profiles in the fully turbulent region agree well with the reference data from direct numerical simulation, and the accuracy is markedly superior to that of the gas-kinetic scheme (GKS) under the identical grid. These findings underscore the considerable promise of the multi-scale WPTS method for transitional flow simulations.

Figures

Figures reproduced from arXiv: 2607.10176 by the authors.

**Figure 2.** (a) The friction coefficient predicted by WM-WPTS under coarse grid M2 with different [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗

**Figure 3.** The predicted snapshots by WM-WPTS under coarse grid M2: (a) streamwise velocity [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references

[1]

Wall-modeled large-eddy simulation for complex turbulent flows.Annual Review of Fluid Mechanics, 50:535–561, 2018

Sanjeeb T Bose and George Ilhwan Park. Wall-modeled large-eddy simulation for complex turbulent flows.Annual Review of Fluid Mechanics, 50:535–561, 2018

2018
[2]

Direct numerical simulation of supersonic turbulent boundary layer flow.Chinese Physics Letters, 22(7):1709–1712, 2005

Hui Gao, De-Xun Fu, Yan-Wen Ma, and Xin-Liang Li. Direct numerical simulation of supersonic turbulent boundary layer flow.Chinese Physics Letters, 22(7):1709–1712, 2005

2005
[3]

World Scientific, 2015

Michael Leschziner.Statistical turbulence modelling for fluid dynamics-demystified: an introductory text for graduate engineering students. World Scientific, 2015

2015
[4]

Unified gas-kinetic wave-particle methods I: Continuum and rarefied gas flow.Journal of Computational Physics, 401:108977, 2020

Chang Liu, Yajun Zhu, and Kun Xu. Unified gas-kinetic wave-particle methods I: Continuum and rarefied gas flow.Journal of Computational Physics, 401:108977, 2020

2020
[5]

Constructing turbulence models using the kinetic fokker– planck equation.Journal of Fluid Mechanics, 1011:A44, 2025

Peng Luan, Haoyuan Zhang, and Jun Zhang. Constructing turbulence models using the kinetic fokker– planck equation.Journal of Fluid Mechanics, 1011:A44, 2025. 11

2025
[6]

Wall-modeled large eddy simulation of high speed flows.AIAA journal, 60(7):4302–4324, 2022

Balachandra R Mettu and Pramod K Subbareddy. Wall-modeled large eddy simulation of high speed flows.AIAA journal, 60(7):4302–4324, 2022

2022
[7]

Springer, 2016

Frans TM Nieuwstadt, Jerry Westerweel, and Bendiks J Boersma.Turbulence: introduction to theory and applications of turbulent flows. Springer, 2016

2016
[8]

Direct numerical simulation and analysis of a spatially evolving supersonic turbulent boundary layer at m= 2.25.Physics of Fluids, 16(3):530–545, 2004

Sergio Pirozzoli, Francesco Grasso, and TB Gatski. Direct numerical simulation and analysis of a spatially evolving supersonic turbulent boundary layer at m= 2.25.Physics of Fluids, 16(3):530–545, 2004

2004

Show all 16 references

[9]

Dynamic wall modeling for large-eddy simulation of complex turbulent flows.Physics of Fluids, 14(7):2043–2051, 2002

Meng Wang and Parviz Moin. Dynamic wall modeling for large-eddy simulation of complex turbulent flows.Physics of Fluids, 14(7):2043–2051, 2002
[10]

Model of incompressible turbulent flows via a kinetic theory.Journal of Fluid Mechanics, 1033:A12, 2026

Ziyang Xin, Zhaoli Guo, and Hudong Chen. Model of incompressible turbulent flows via a kinetic theory.Journal of Fluid Mechanics, 1033:A12, 2026

2026
[11]

A gas-kinetic BGK scheme for the Navier–Stokes equations and its connection with artificial dissipation and Godunov method.Journal of Computational Physics, 171(1):289–335, 2001

Kun Xu. A gas-kinetic BGK scheme for the Navier–Stokes equations and its connection with artificial dissipation and Godunov method.Journal of Computational Physics, 171(1):289–335, 2001

2001
[12]

A unified gas-kinetic scheme for continuum and rarefied flows.Journal of Computational Physics, 229(20):7747–7764, 2010

Kun Xu and Juan-Chen Huang. A unified gas-kinetic scheme for continuum and rarefied flows.Journal of Computational Physics, 229(20):7747–7764, 2010

2010
[13]

Wave–particle turbulence simulation of spatially developing round jets: Turbulent flow modeling and method validation.Physics of Fluids, 37(10), 2025

Xiaojian Yang and Kun Xu. Wave–particle turbulence simulation of spatially developing round jets: Turbulent flow modeling and method validation.Physics of Fluids, 37(10), 2025

2025
[14]

Wave-particle based multiscale modeling and simulation of non-equilibrium turbulent flows.Computers & Fluids, page 107162, 2026

Xiaojian Yang and Kun Xu. Wave-particle based multiscale modeling and simulation of non-equilibrium turbulent flows.Computers & Fluids, page 107162, 2026

2026
[15]

Subgrid-scale model for large-eddy simulation of transition and turbulence in compressible flows.Physics of Fluids, 31(12), 2019

Hao Zhou, Xinliang Li, Han Qi, and Changping Yu. Subgrid-scale model for large-eddy simulation of transition and turbulence in compressible flows.Physics of Fluids, 31(12), 2019

2019
[16]

Coherent structures in transition of a flat-plate boundary layer at Ma= 0.7.Chinese Physics Letters, 24(1):147–150, 2007

Ying Zhou, Xin-Liang Li, De-Xun Fu, and Yan-Wen Ma. Coherent structures in transition of a flat-plate boundary layer at Ma= 0.7.Chinese Physics Letters, 24(1):147–150, 2007. 12

2007

Pith tools

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

[1] [1]

Wall-modeled large-eddy simulation for complex turbulent flows.Annual Review of Fluid Mechanics, 50:535–561, 2018

Sanjeeb T Bose and George Ilhwan Park. Wall-modeled large-eddy simulation for complex turbulent flows.Annual Review of Fluid Mechanics, 50:535–561, 2018

2018

[2] [2]

Direct numerical simulation of supersonic turbulent boundary layer flow.Chinese Physics Letters, 22(7):1709–1712, 2005

Hui Gao, De-Xun Fu, Yan-Wen Ma, and Xin-Liang Li. Direct numerical simulation of supersonic turbulent boundary layer flow.Chinese Physics Letters, 22(7):1709–1712, 2005

2005

[3] [3]

World Scientific, 2015

Michael Leschziner.Statistical turbulence modelling for fluid dynamics-demystified: an introductory text for graduate engineering students. World Scientific, 2015

2015

[4] [4]

Unified gas-kinetic wave-particle methods I: Continuum and rarefied gas flow.Journal of Computational Physics, 401:108977, 2020

Chang Liu, Yajun Zhu, and Kun Xu. Unified gas-kinetic wave-particle methods I: Continuum and rarefied gas flow.Journal of Computational Physics, 401:108977, 2020

2020

[5] [5]

Constructing turbulence models using the kinetic fokker– planck equation.Journal of Fluid Mechanics, 1011:A44, 2025

Peng Luan, Haoyuan Zhang, and Jun Zhang. Constructing turbulence models using the kinetic fokker– planck equation.Journal of Fluid Mechanics, 1011:A44, 2025. 11

2025

[6] [6]

Wall-modeled large eddy simulation of high speed flows.AIAA journal, 60(7):4302–4324, 2022

Balachandra R Mettu and Pramod K Subbareddy. Wall-modeled large eddy simulation of high speed flows.AIAA journal, 60(7):4302–4324, 2022

2022

[7] [7]

Springer, 2016

Frans TM Nieuwstadt, Jerry Westerweel, and Bendiks J Boersma.Turbulence: introduction to theory and applications of turbulent flows. Springer, 2016

2016

[8] [8]

Direct numerical simulation and analysis of a spatially evolving supersonic turbulent boundary layer at m= 2.25.Physics of Fluids, 16(3):530–545, 2004

Sergio Pirozzoli, Francesco Grasso, and TB Gatski. Direct numerical simulation and analysis of a spatially evolving supersonic turbulent boundary layer at m= 2.25.Physics of Fluids, 16(3):530–545, 2004

2004