{"id":"2d8bd9d5-74dd-4821-8faa-118061593418","arxiv_id":"2507.00720","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A simplified unified wave-particle method for diatomic gases now includes rotational and vibrational nonequilibrium, enabling efficient multiscale hypersonic flow simulation.","lead":"A new numerical method splits gas molecules into colliding and free-flying groups to simulate hypersonic flows that are far from thermal equilibrium. It adds vibrational and rotational energy effects to an existing wave-particle scheme, matching slower reference solvers while using far less memory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Type-F particle reconstruction samples from equilibrium distributions f_tr/f_rot/f_vib, potentially biasing the free-transport flux in the multiscale regime and weakening the accuracy claim for non-equilibrium flows.","rationale":"The reader's weakest assumption identifies the most load-bearing risk: the equilibrium-form reconstruction of Type-F particles directly affects fidelity of the free-transport flux in the multiscale regime the method targets. A simple conservation check shows that in the intermediate Knudsen regime the reconstructed fraction (w_free(1-w_free^2) of the cell total) is non-negligible, so the approximation can bias fluxes. However, the paper's evidence suggests the early rise in shock structures is shared by DUGKS, implying a kinetic-model limitation rather than a sampling artifact; the cylinder leeward discrepancy is more specifically tied to the SUWP-vib particle treatment. Thus the reader's identification of the sampling assumption is valid, but its attribution to the early rise is not fully established. A dedicated quantitative test (e.g., comparing the reconstructed distribution against DSMC data, or replacing equilibrium sampling with moment-based resampling) is needed to determine whether observed discrepancies are dominated by this assumption or by the model itself. The efficiency claim is presented with both wall-clock and core-hour figures, so it is not misleading in a correctness sense, though the speedup metric depends on hardware. The lack of code/data is a reproducibility limitation, not a mathematical flaw. Overall, the CONDITIONAL verdict stands: the central claim is plausible but not fully established until the reconstruction error is quantified.","tokens_in":27418,"tokens_out":21008,"duration_ms":232467,"concrete_test":"Quantify the reconstruction bias using the Ma=10 shock structure DSMC data from Section 4.2. In a cell near the shock where the early rise occurs, bin the DSMC particles to obtain the actual distribution f. Then synthesize the SUWP-vib reconstructed distribution by sampling from f_tr/f_rot/f_vib according to Section 3.2 using the cell's macroscopic variables (n, U, Ttr, Trot, Tvib, qtr, qrot, qvib). Compare the two distributions' moments, especially the translational heat-flux moment B = ∫ c_x (c_x^2/2) f dΞ and the normal stress P_xx. If the relative difference is comparable to the observed early-rise overshoot or the cylinder leeward discrepancy, the equilibrium-sampling assumption is confirmed as a load-bearing error source; if the difference is small (e.g., under 5%), the concern is refuted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central accuracy claim requires the particle solver to faithfully represent the free-transport (Type-F) component of the distribution. In Section 3.2, after removing Type-C particles, reconstructed Type-F particles are sampled from the equilibrium/Shakhov-like target distributions f_tr, f_rot, and f_vib (Eqs. 13-15) using w_free·Q_remain. This replaces the unknown initial distribution f(x0,u,eta) in Eq. (26)'s free-streaming term with an equilibrium-form distribution, discarding non-equilibrium deviations (stress anisotropy, higher-order moments) that Type-F particles are intended to carry. In the intermediate Knudsen regime where w_free and w_hydro are both significant, the reconstructed fraction is non-negligible, so this approximation can bias macroscopic fluxes. The paper acknowledges a leeward-side temperature discrepancy in cylinder flow (Section 4.3) specific to SUWP-vib (not present in DUGKS), and it is not quantified. The early rise in shock structures (Section 4.2) is shared with DUGKS, indicating a kinetic-model contribution, so the sampling-induced error is not isolated. Without a dedicated test separating these effects, the magnitude of the reconstruction bias remains unknown, leaving the claim that SUWP-vib 'maintains accuracy' in non-equilibrium flows insufficiently supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper extends the simplified unified wave-particle (SUWP) method to diatomic gases with rotational and vibrational non-equilibrium. Starting from the time-integral solution of a vibrational kinetic model, the authors derive a quantified model-competition (QMC) mechanism that splits molecules into colliding (Type-C) and free-transport (Type-F) parts, weights them by w_hydro = 1 - exp(-Δt/τ) and w_free = exp(-Δt/τ), and couples a three-temperature Navier-Stokes solver with a collisionless DSMC-type particle solver. The N-S solver uses a kinetic inviscid flux (KIF) and explicit relaxation source terms for rotational and vibrational energies; the particle solver handles particle removal, reconstruction, transport, and boundary conditions. The method is applied to shock tube, shock structure, cylinder, blunt wedge, sphere, Apollo 6, and Space Station Mir cases, with a reported efficiency comparison against implicit DUGKS.","tokens_in":27589,"tokens_out":8792,"duration_ms":105544,"significance":"If the central reconstruction approximation is sound, the paper offers a practical low-memory multiscale solver for high-speed aerospace flows, with an original derivation of the QMC weights for the three-temperature diatomic model and an explicit Chapman-Enskog reduction to three-temperature N-S equations. The manuscript ships a substantial set of benchmark computations, including a 2.6-million-cell Mir reentry, and reports concrete CPU-hour and memory advantages. No new free parameters are introduced beyond standard rotational/vibrational collision numbers and VHS gas data. The main strengths are the algorithmic construction and the breadth of demonstration; the main weaknesses are that most validation is against another discretization of the same kinetic model and that the particle-reconstruction approximation is not quantitatively bounded.","major_comments":[{"comment":"In the reconstruction step, Type-F particles are sampled from the equilibrium target distributions f_tr, f_rot, and f_vib rather than from the actual non-equilibrium distribution f(x0,u,ηrot,ηvib,0) that appears in the free-streaming term of Eq. (26). Since Type-F particles are precisely the fraction that should carry non-equilibrium information, replacing f by these equilibrium forms discards at least the deviatoric stress and higher-order moments; the Shakhov-like heat-flux terms in Eqs. (13)-(15) restore only part of the heat flux. In the intermediate-Knudsen regime w_free is not small, so this approximation can bias the particle flux and therefore the claimed accuracy. The reported early rise of Ttr,n (Sec. 4.2) and the leeward-side temperature discrepancy (Sec. 4.3) are consistent with such a bias, but they are not quantified and are partly shared with DUGKS, so they do not isolate the reconstruction error. Please add a quantitative test, for example comparing the statistical moments of reconstructed Type-F particles with the pre-removal particle distribution, or a benchmark in which the reconstruction is replaced by direct particle statistics, to bound the error in the target flow regimes.","section":"Section 3.2, Eqs. (13)-(15), (26)"},{"comment":"Most validation is against the authors' own DUGKS solution [27] of the same vibrational kinetic model. This confirms consistency between two discretizations, but it cannot validate the physical model or the particle-reconstruction approximation independently. Independent DSMC comparisons appear in shock structures and cylinder/blunt-wedge cases, but the known discrepancies (early rise in Ttr,n, leeward-side temperature increase in cylinder flow, post-shock temperature decay) are described qualitatively and no error metrics are reported. For the paper's central claim of 'maintaining accuracy' in strong non-equilibrium, please provide quantitative errors against DSMC (for example L1 or L2 norms of profile deviations) for rotational and vibrational temperatures and for wall heat flux, and state the range of Knudsen and Mach numbers over which the stated accuracy holds.","section":"Sections 4.1-4.7"},{"comment":"The efficiency comparison is not controlled: the implicit DUGKS calculation used a different processor generation (Xeon Gold 6258, 2.70 GHz, 640 cores) than the SUWP-vib calculation (EPYC 7763, 2.45 GHz, 108 cores), and no memory usage is measured even though the abstract and conclusion claim one-to-two-orders-of-magnitude memory reduction. The reported speedup is based on CPU-hours (2092.8 vs 887.86 GHz·h), while the wall-clock time is actually larger for SUWP-vib (8.22 h vs 3.27 h); with different architectures and core counts this does not establish a robust efficiency advantage. Please report measured memory, provide a controlled or at least well-characterized comparison, and clarify whether the 60,000 KIF initialization steps are included in the SUWP total cost.","section":"Section 4.6, Table 10"}],"minor_comments":[{"comment":"There are numerous typographical errors, including 'desity' in figure captions, 'Fig. 3, Fig. 4 和Fig. 5' in Sec. 4.1, 'The the first term' in Sec. 3.1, and a stray 'equilibrium' in Eq. (A.7); a careful copyedit is needed.","section":"Throughout"},{"comment":"The fixed-point iteration as written uses T_eq^{n+1,i} on the left-hand side and K_vib(T_eq^{n+1,i+1}) on the right-hand side, which is index-inconsistent; please state the intended iteration and the initial guess.","section":"Eq. (43)"},{"comment":"The sentence 'The total computation takes 25000 time steps' after describing a 60,000-step KIF initialization is ambiguous; please clarify whether the KIF warm-up is included in the reported runtime and CPU-hours.","section":"Section 4.6"},{"comment":"The heat-flux correction terms in f_tr, f_rot, and f_vib can make these distributions locally negative, but the paper does not describe how the sampling algorithm handles this; a brief reproducibility-oriented description is needed.","section":"Section 3.2, Eqs. (13)-(15)"},{"comment":"The blunt-wedge case specifies ω1=0.75, ω2=0.25, ω3=0.75, and ω4=0.4, but the model in Section 2.2 defines only ω0 through ω3 and no ω4; please clarify the mapping of these coefficients to Eqs. (13)-(15) and whether δ also changes.","section":"Section 4.4"},{"comment":"The definition of the KIF weighting parameter β refers to ΔP and M_back from Ref. [11] without a local summary; for a self-contained presentation, please state the smooth-function arguments and typical parameter values.","section":"Section 3.3, Eq. (52)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript presents a plausible and substantial algorithmic extension, and I would not reject it. However, the validation strategy is largely circular with respect to the kinetic model, and the Type-F reconstruction bias identified in Section 3.2 is a load-bearing concern for the accuracy claims in the non-equilibrium regimes the method targets. The authors should either quantify the reconstruction error or soften the 'maintaining accuracy' claim accordingly. I would also ask the handling editor to ensure that the novelty relative to the UGKWP-vib method in Ref. [63] is stated more explicitly in the revised manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper extends the SUWP wave-particle method to diatomic gases with both rotational and vibrational nonequilibrium. The QMC mechanism is re-derived for the Zhang kinetic model, the three-temperature KIF flux is constructed, and the Chapman-Enskog expansion is carried out carefully. The test suite is broad—shock tube, shock structure, cylinder, blunt wedge, sphere, Apollo 6, and SS Mir—and the method reproduces DUGKS results across most of them. That is genuine work, and I think the central derivation holds up.\n\nThe soft spots are real but not fatal. The biggest one is that most validation is against the authors' own DUGKS implementation of the same kinetic model. That mainly shows the new discretization is consistent with the old one, not that either is independently accurate. The DSMC comparisons in the shock structure cases show the known \"early rise\" in translational temperature, shared by DUGKS and therefore likely a model-level issue, so I would not blame SUWP-vib for that alone. But the leeward-side temperature discrepancy in the cylinder case (Section 4.3) is specific to SUWP-vib and is never quantified. The stress-test concern about Section 3.2 is legitimate: after removing colliding particles, the reconstructed free-transport particles are sampled from the equilibrium-like distributions f_tr, f_rot, and f_vib, which discards the non-equilibrium deviations those particles are supposed to carry. In the intermediate Knudsen regime this could bias fluxes; the paper acknowledges the symptom but does not isolate the cause. A dedicated sensitivity study would answer it.\n\nThe efficiency claim also needs care. Table 10 reports 2.36× speedup based on CPU-hours, but the wall-clock time is actually longer for SUWP (8.22 h vs 3.27 h) because it uses fewer cores. That is potentially misleading and should be reframed. No code or data are released, and the SS Mir case has no external benchmark, so it is a demonstration of capability rather than a validation.\n\nOverall, this is an honest, competent method paper. It deserves a serious referee, but the referee should push for a quantification of the reconstruction bias, a fairer efficiency comparison, and ideally public code or data so others can verify the consistency claim. I would not block it on the current evidence, but I would want those issues resolved before accepting it as-is.","headline":"A useful extension of the SUWP wave-particle framework to vibrational nonequilibrium, with a real but unquantified question about how free-transport particles are reconstructed.","tokens_in":28229,"tokens_out":2371,"would_cite":true,"duration_ms":28923,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper extends the simplified unified wave-particle (SUWP) method to diatomic gases with rotational and vibrational non-equilibrium, deriving a quantified model-competition (QMC) mechanism that splits molecules into free-transport…","keywords":["multi-scale flows","rarefied gas dynamics","wave-particle method","rotational non-equilibrium","vibrational non-equilibrium","quantified model-competition","hypersonic flow","DSMC"],"falsifier":"A refinement study of the Mach-15 nitrogen shock structure: increase cell resolution and particle count while holding the physical model fixed, and compare the normal translational-temperature profile with DSMC. If the early rise upstream of the shock maximum persists or grows under refinement, it is an artifact of the equilibrium-sampling reconstruction; if it shrinks, it was a discretization effect. A second check is the leeward-side cylinder flow at Mach 5, where the paper reports faster temperature rise than DUGKS; quantifying that gap against DSMC would reveal the same mechanism.","tokens_in":27093,"feed_emoji":"🚀","tokens_out":8809,"duration_ms":85642,"temperature":0.7,"pith_summary":"This paper extends the simplified unified wave-particle (SUWP) method to diatomic gases in which rotational and vibrational internal energy modes are out of equilibrium, the regime encountered in hypersonic near-space flight. It splits molecules into free-transport particles and colliding molecules, with the split weighted by a newly derived quantified model-competition (QMC) mechanism built from the time integral of a vibrational kinetic model. The two groups are simulated by a collisionless particle solver and a three-temperature Navier-Stokes solver, respectively. Across shock tube, shock structure, cylinder, blunt wedge, sphere, Apollo 6 command module, and space station Mir cases, the authors report agreement with DUGKS and DSMC reference solutions while using substantially fewer CPU-hours and far less memory. If correct, the method gives an accurate and low-cost tool for multiscale diatomic gas flows with internal energy non-equilibrium.","feed_headline":"Wave-particle method cuts hypersonic re-entry cost by 57 percent","feed_subtitle":"It splits molecules into colliding and free-traveling groups, covering rarefied to continuum regimes in one solver.","key_machinery":"The quantified model-competition (QMC) mechanism is the central machinery. Starting from the integral solution $f(t) = e^{-\\Delta t/\\tau} f(\\mathbf{x}_0, \\mathbf{u}, t_0) + \\tau^{-1} \\int_0^{\\Delta t} g(\\mathbf{x}', \\mathbf{u}, t') e^{t'/\\tau} dt'$, the fraction $e^{-\\Delta t/\\tau}$ of molecules that never collide is treated as free transport and sampled as particles, while the complement is treated as hydrodynamic. A second-order Taylor expansion of the equilibrium state $g$ turns the hydrodynamic part into the N-S flux with the scale-dependent viscous coefficient $c_{\\text{vis}}$ multiplying the viscous flux. The reconstruction of Type-F particles draws from the three equilibrium distributions $f^{\\text{tr}}$, $f^{\\text{rot}}$, $f^{\\text{vib}}$ with probabilities set by the relaxation numbers $Z_{\\text{rot}}$ and $Z_{\\text{vib}}$, and the macroscopic flux is closed by a three-temperature kinetic inviscid flux (KIF) scheme.","core_discovery":"The central claim is that a second-order Chapman-Enskog expansion of the time-integral solution of the vibrational kinetic model collapses the colliding-molecule fraction onto the three-temperature Navier-Stokes equations, leaving a complementary free-transport fraction that can be handled by a collisionless particle solver. This produces an explicit weight $w_{\\text{free}} = e^{-\\Delta t/\\tau}$ for particles and $w_{\\text{hydro}} = 1 - e^{-\\Delta t/\\tau}$ for hydrodynamic molecules, together with a scale-dependent viscous coefficient $c_{\\text{vis}} = 1 - (\\Delta t/\\tau) e^{-\\Delta t/\\tau}/(1 - e^{-\\Delta t/\\tau})$. The paper argues that the resulting SUWP-vib method reproduces reference solutions across Knudsen numbers from $10$ to $10^{-3}$ and Mach numbers up to $25$, and that because it reuses an N-S solver and a DSMC component, existing results for both can be directly adopted.","pith_inferences":["(Editorial inference) Because Type-F particles are reconstructed from equilibrium distributions $f^{\\text{tr}}, f^{\\text{rot}}, f^{\\text{vib}}$, the non-equilibrium deviation they are meant to carry may be under-sampled in strong non-equilibrium zones, which would worsen exactly where the method is most needed.","(Editorial inference) The QMC mechanism depends only on the form of the integral solution, so the same weighting could extend to dissociating and ionizing air by adding chemical or electronic energy modes to the relaxation targets.","(Editorial inference) The reported speedup compares an implicit DUGKS run on one hardware setup with SUWP-vib on another; a same-machine comparison with matched particle counts would make the efficiency claim more direct.","(Editorial inference) Mapping $w_{\\text{free}}$ and $c_{\\text{vis}}$ onto local continuum-breakdown parameters could give a quantitative criterion for when the particle solver must take over, usable in other hybrid methods."],"forward_implications":["The SUWP-vib method adapts cell by cell between continuum and rarefied descriptions, so one solver covers flows that cross the entire Knudsen range.","Because it is built from standard N-S and DSMC components, improvements to either component (relaxation models, boundary conditions, collision sampling) carry over directly into the coupled method.","The Apollo 6 case shows drag and lift within 0.42% and 3.06% of the DSMC benchmark while using about 57% fewer CPU-hours than the implicit DUGKS solver on the same mesh.","The space station Mir case shows the method runs on complex three-dimensional geometries with millions of cells, keeping memory demand well below that of deterministic velocity-space solvers.","The paper identifies an 'early rise' of translational temperature in shock structures and notes that modifying the relaxation time, as done in prior wave-particle work, is a possible future correction."],"supporting_citations":[{"why":"Supplies the vibrational kinetic model equation with translational, rotational, and vibrational relaxation, and provides the DUGKS reference solutions for shock tube, shock structure, cylinder, and sphere cases.","marker":"[27]"},{"why":"Introduces the original simplified unified wave-particle method and the quantified model-competition mechanism that this paper extends to vibrational non-equilibrium.","marker":"[71]"},{"why":"The prior SUWP method for diatomic gases based on the Rykov model, which the present work generalizes to include vibrational energy non-equilibrium.","marker":"[64]"},{"why":"The unified gas-kinetic wave-particle framework of which the QMC mechanism is a simplification.","marker":"[65]"},{"why":"Provides the kinetic inviscid flux (KIF) scheme used as the macroscopic inviscid part of the three-temperature N-S solver.","marker":"[11]"},{"why":"Supplies the rotational collision number formula used in the test cases.","marker":"[75]"},{"why":"Provides the implicit UGKS and DSMC reference data for the blunt wedge simulations.","marker":"[80]"},{"why":"Supplies the DSMC benchmark lift and drag coefficients for the Apollo 6 command module.","marker":"[81]"}],"fun_headline_variants":["Simplified wave-particle method handles non-equilibrium diatomic flows","Unified wave-particle solver speeds up hypersonic re-entry modeling","New method splits molecules to model rarefied gas accurately","Efficient wave-particle method for non-equilibrium diatomic gas","Wave-particle method covers multi-scale diatomic flow physics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The accuracy in strong non-equilibrium regions rests on the assumption that newly created free-transport particles can be sampled from the equilibrium distributions $f^{\\text{tr}}$, $f^{\\text{rot}}$, and $f^{\\text{vib}}$ instead of from the actual non-equilibrium distribution, because those particles are supposed to carry the non-equilibrium part of the flow.","fun_headline_variants_meta":{"raw":{"variants":["Simplified wave-particle method handles non-equilibrium diatomic flows","Unified wave-particle solver speeds up hypersonic re-entry modeling","New method splits molecules to model rarefied gas accurately","Efficient wave-particle method for non-equilibrium diatomic gas","Wave-particle method covers multi-scale diatomic flow physics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001027,"raw_usage":{"total_tokens":4371,"prompt_tokens":1028,"completion_tokens":3343,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":644,"completion_tokens_details":{"reasoning_tokens":3273}},"tokens_in":644,"tokens_out":3343,"duration_ms":23024,"temperature":1.0,"reasoning_tokens":3273,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:10:21.777723+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A refinement study of the Mach-15 nitrogen shock structure: increase cell resolution and particle count while holding the physical model fixed, and compare the normal translational-temperature profile with DSMC. If the early rise upstream of the shock maximum persists or grows under refinement, it is an artifact of the equilibrium-sampling reconstruction; if it shrinks, it was a discretization effect. A second check is the leeward-side cylinder flow at Mach 5, where the paper reports faster temperature rise than DUGKS; quantifying that gap against DSMC would reveal the same mechanism.","supporting_citations":[{"cited_title":"Zhang, S","cited_arxiv_id":null,"evidence_quote":"Supplies the vibrational kinetic model equation with translational, rotational, and vibrational relaxation, and provides the DUGKS reference solutions for shock tube, shock structure, cylinder, and sphere cases."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the original simplified unified wave-particle method and the quantified model-competition mechanism that this paper extends to vibrational non-equilibrium."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The prior SUWP method for diatomic gases based on the Rykov model, which the present work generalizes to include vibrational energy non-equilibrium."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The unified gas-kinetic wave-particle framework of which the QMC mechanism is a simplification."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the kinetic inviscid flux (KIF) scheme used as the macroscopic inviscid part of the three-temperature N-S solver."},{"cited_title":"Parker, Rotational and vibrational relaxation in diatomic gases, The Physics of Fluids 2 (4) (1959) 449–462","cited_arxiv_id":null,"evidence_quote":"Supplies the rotational collision number formula used in the test cases."},{"cited_title":"Zhang, S","cited_arxiv_id":null,"evidence_quote":"Provides the implicit UGKS and DSMC reference data for the blunt wedge simulations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the DSMC benchmark lift and drag coefficients for the Apollo 6 command module."}],"review_version":1}