{"id":"d6497c2d-5954-4cbc-bb95-03628353ddee","arxiv_id":"2411.13168","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A multi-species entropy-conservative flux for DGSEM is presented that works with arbitrary tabulated internal-energy models, and is validated against reference solvers.","lead":"This paper extends a high-order discontinuous Galerkin method to high-enthalpy reacting gas mixtures, with a numerical flux that preserves entropy and kinetic energy even when the gas's internal energy is given by arbitrary tabulated data. The method is tested on blast waves, chemistry relaxation, and a hypersonic flow around a cylinder, matching an established DLR solver.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Entropy conservation is established only for the tabulated piecewise-linear thermodynamic model; no consistency condition between interpolated cv and ε tables is stated, and Eq. (15) as printed contains a wrong interval slope.","rationale":"The reader's weakest assumption concerns table resolution degrading conservation and accuracy. I largely agree, but the more load-bearing issue is thermodynamic consistency of the tabulated model, not merely resolution: independent linear interpolation of cv,c and ε_int,c can violate dε_int,c/dT = cv,c(T), which is required for the entropy variables used in the Tadmor condition to be those of the discretized model. This is exactly the 'arbitrary internal energies' scenario advertised in the abstract. The Eq. (15) slope typo reinforces that the manuscript's formulas are not self-contained and cannot be verified from the text alone. I do not see this as a fatal flaw: the public reproducibility repository and machine-precision entropy measurements suggest a working implementation. But the paper must state the consistency condition, correct Eq. (15), and provide a ΔT-resolution study before the central claim is fully established. This does not move the reader's CONDITIONAL verdict; it sharpens the conditions under which the method is trustworthy.","tokens_in":7784,"tokens_out":21414,"duration_ms":232848,"concrete_test":"Evaluate algebraically the Tadmor residual Δw·F* - ΔF^S_x for the two-point flux (9)-(12) using two states built from deliberately inconsistent tables: take ε_int,c(T) linear on one interval while cv,c(T) is a different linear ramp, so the secant slope of ε differs from the interpolated cv. If the residual is nonzero at machine precision, entropy conservation depends on an unstated consistency condition between the cv and ε tables; if it is zero, the missing condition reduces to a ΔT-resolvable approximation question, and a ΔT-convergence study on the weak-blast-wave test would be the appropriate follow-up.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim's advertised strength is that arbitrary internal-energy expressions are supported. In Section 3, however, the flux uses independently interpolated cv,c(T) and ε_int,c(T) tables on a ΔT grid (Eqs. 14-16), with the entropy integral η_c computed from the interpolated cv via Eq. (15). For Tadmor's entropy-conservation condition to hold for the discretized model, the entropy variables w = ∂S/∂u must be consistent with the way T is recovered from E = Σ ρ_c ε_int,c(T); this requires dε_int,c/dT = cv,c(T) for the reconstructed functions. When both tables are linearly interpolated independently, this equality fails except in the ΔT->0 limit. The paper does not state this compatibility requirement and gives no analysis of how coarse ΔT or independently filled tables degrade entropy conservation or solution accuracy. Additionally, Eq. (15) as printed uses cv,c(T) - cv,c,N where the interval slope cv,c,N+1 - cv,c,N is required by Eq. (16); as written, η_c(T) is incorrect for off-node temperatures. The machine-precision entropy production in Section 4.1 therefore only demonstrates consistency for the particular implementation, not for the general claim as stated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends a previously proposed entropy-conservative flux construction for high-enthalpy single-species gas flows to multi-species reacting mixtures within a Discontinuous Galerkin Spectral Element Method (DGSEM). The flux uses piecewise-linear tabulated reconstructions of species specific heats and internal energies, with closed-form entropy integrals computed from the tabulated data, so that the cost of evaluating the flux is independent of the complexity of the underlying internal-energy expressions. The method is implemented in Trixi.jl and tested on three problems: a non-reacting weak blast wave (for which machine-precision entropy conservation is reported), a spatially homogeneous chemical relaxation of an O2/O mixture (compared against a master-equation reference), and a Mach 10 flow over a cylinder (compared against the DLR TAU solver). The paper claims excellent agreement in all cases and argues that the approach enables entropy-stable high-order simulations of high-enthalpy reacting flows with arbitrary internal energy models.","tokens_in":8030,"tokens_out":6039,"duration_ms":57823,"significance":"If the claims hold, this is a practically useful contribution: it provides a route to entropy-stable DGSEM for multi-species high-enthalpy flows with complex real-gas thermodynamics while keeping the per-element flux evaluation cheap and independent of the analytical form of the internal energy functions. The paper reports machine-precision entropy production in the blast-wave test, verifies the chemistry implementation against an independent master-equation solution, and validates the full reacting-flow setup against the established DLR TAU solver. The availability of a public reproducibility repository is a concrete strength. However, the manuscript is not fully self-contained: the flux derivation is delegated to a companion preprint, and two of the central printed formulas (Eqs. (14) and (15)) are corrupted or incorrect as written. These issues must be resolved before the central claim can be verified from the paper alone.","major_comments":[{"comment":"The printed definitions of c_{v,c}(T*) and c_{v,c}(T**) are garbled and cannot be read as mathematical formulas. In particular, the expression given for c_{v,c}(T**) contains extraneous characters and appears to mix an arithmetic mean with a jump in an undefined way. Since Eq. (12) is the central entropy-conservative energy flux, these definitions must be stated cleanly; otherwise the flux cannot be implemented from the manuscript and the claimed property cannot be checked.","section":"Sec. 3, Eq. (14)"},{"comment":"The closed-form expression for the reconstructed entropy integral η_c(T) is incorrect as printed. For a piecewise-linear c_{v,c}(T) on the interval [T_N, T_{N+1}], the coefficient multiplying ln(T/T_N) must involve the fixed interval slope (c_{v,c,N+1} - c_{v,c,N})/ΔT, but Eq. (15) instead contains (c_{v,c}(T) - c_{v,c,N}), which is not constant on the interval. As written, η_c(T) is not the exact integral of the linearly interpolated specific heat. This matters because Eq. (15) feeds directly into the flux through Eq. (14) and is the basis for the machine-precision entropy conservation reported in Sec. 4.1. The authors should correct the formula or clarify the intended notation.","section":"Sec. 3, Eq. (15)"},{"comment":"The paper does not state a consistency condition between the tabulated c_{v,c}(T) and ε_int,c(T) reconstructions. For the discrete entropy identity to hold, the temperature recovered from the total energy E must be compatible with the entropy variables computed from η_c; in particular, the reconstructed functions should satisfy dε_int,c/dT = c_{v,c}(T). When c_v and ε_int are linearly interpolated independently on the same ΔT grid, this equality fails away from the limit ΔT→0. The numerical entropy-production result in Sec. 4.1 demonstrates the property for the specific implementation, but it does not establish the general claim in the abstract for arbitrary tabulated data. The authors should either impose a consistency constraint on table generation (e.g., deriving ε_int from the same piecewise-linear c_v) or provide an analysis of how ΔT degrades entropy conservation and solution accuracy.","section":"Sec. 3, Eqs. (14)-(16) and Sec. 4.1"},{"comment":"The multi-species flux in Eqs. (9)-(12) is presented without derivation; the text only states that it follows the procedure of the authors' prior single-species work [17]. Since the central contribution of this manuscript is precisely this multi-species entropy-conservative flux, the paper should include the key steps of the Tadmor entropy-conservation condition for the reacting multi-species system, or at minimum state explicitly which equations and assumptions are being carried over from [17]. As it stands, a reader cannot verify the entropy and kinetic-energy preservation properties without consulting a separate preprint.","section":"Sec. 3, flux derivation"}],"minor_comments":[{"comment":"There is a typo in the Introduction: \"next-generation CDF solvers\" should read \"next-generation CFD solvers.\"","section":"Sec. 1"},{"comment":"The two cases for atoms and molecules are presented in a single run-on line; a proper case distinction would make the formula readable and avoid ambiguity about which expression applies to which species type.","section":"Sec. 2, Eq. (4)"},{"comment":"The physical entropy per unit mass and the mathematical entropy density are both denoted by the symbol s (with the mathematical entropy defined as −ρs). Using two different symbols (e.g., s and S̃) would prevent confusion in the subsequent derivation.","section":"Sec. 2, Eq. (8)"},{"comment":"The y-axis label \"|s|\" is undefined; the plotted quantity appears to be the absolute value of the entropy production rate, so the label should read something like \"|ds/dt|\" or \"entropy production rate.\"","section":"Sec. 4.1, Fig. 2"},{"comment":"The paper claims the flux is kinetic-energy-preserving, but none of the numerical tests directly reports a kinetic-energy conservation error. Adding such a measurement would strengthen the empirical support for that part of the claim.","section":"Sec. 4.1"}],"recommendation":"major_revision","confidential_remarks":"The garbled Eq. (14) and the erroneous interval-slope term in Eq. (15) may be typesetting artifacts, but they sit at the computational core of the paper and must be corrected. The missing consistency analysis for the tabulated thermodynamics is more substantive: without it, the paper overstates the generality of the entropy-conservation guarantee. I do not see a novelty disclosure problem; the extension from the single-species flux in [17] to the multi-species reacting case is a reasonable incremental contribution suitable for this venue, provided the derivation is made checkable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a useful paper. It extends the authors' single-species entropy-conservative flux to reacting multi-species flows with arbitrary internal energy models, using precomputed piecewise-linear tables. That is a real increment over prior work, which was restricted to NASA polynomials or simple oscillator models. The validation is convincing in broad strokes: the weak blast wave shows machine-precision entropy production for the EC flux, the homogeneous chemical relaxation matches a master-equation solution, and the Mach 10 cylinder agrees along the stagnation line with DLR TAU. The code and input files are public, which is a big plus.\n\nThe soft spots are real but addressable. The flux derivation itself is delegated to the authors' earlier paper [17]; that is defensible, but it means the present paper does not stand alone, and the few equations it does include are careless. Eq. (14) is garbled: T* and T** are defined via jumps but the notation is ambiguous and the second definition looks like a typo. The stress-test note about Eq. (15) checks out: as printed, the formula for eta_c(T) does not follow from the preceding linear interpolation—there's a missing factor or a mislabeled index. A reader cannot reproduce the implementation from the paper alone without guessing corrections.\n\nMore substantively, the paper claims the flux works for arbitrary internal-energy expressions, but entropy conservation is only established for the tabulated piecewise-linear approximation. Independently interpolating cv and epsilon tables without ensuring they satisfy d epsilon/dT = cv can break the entropy-consistency that the flux relies on. The paper does not state this compatibility requirement or analyze how table resolution degrades conservation. The machine-precision result in Sec. 4.1 shows the particular implementation behaves, but it does not prove the general claim for arbitrary tabulations. The TAU comparison is also code-vs-code, not a physics benchmark, since the thermodynamic models in TAU are not specified.\n\nThese are weaknesses in presentation and analysis, not demonstrated errors in the method. I did not find a load-bearing flaw. The paper deserves a serious referee; it is a useful contribution for people working on high-order DG for high-enthalpy reacting flows. I would recommend sending it to peer review with a request to fix Eqs. (14)-(15), add a short discussion on tabulation consistency, and at least acknowledge the TAU model-matching caveat. My own verdict would be conditional acceptance after those revisions.","headline":"A genuine multi-species extension of the authors' tabulated entropy-conservative flux with solid validation, but the presentation has gaps that need fixing before the method is fully reproducible.","tokens_in":8552,"tokens_out":2792,"would_cite":true,"duration_ms":28590,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.40.Ki"],"model":"deepseek-v4-flash","headline":"This paper extends entropy-conserving high-order DG fluxes to multi-species reacting high-enthalpy flows by tabulating internal energies.","keywords":["high-enthalpy flows","discontinuous Galerkin spectral element method","entropy-conservative flux","kinetic energy preservation","reacting gas flows","tabulated internal energy","hypersonic flow","chemical relaxation"],"falsifier":"Run the weak blast wave with the same physical O2/O mixture tabulated at a coarse step, say $\\Delta T = 500$ K, and at a fine step, $\\Delta T = 1$ K, on identical grids and time steps; if the coarse-table pressure and temperature fields differ by more than discretization error, or if entropy production leaves machine precision when temperatures approach the table boundary, the method's guarantee is tied to the interpolation table rather than to the physical model.","tokens_in":7569,"feed_emoji":"🔥","tokens_out":10254,"duration_ms":100813,"temperature":0.7,"pith_summary":"The paper claims that high-order Discontinuous Galerkin solvers can be made entropy-conserving for multi-species high-enthalpy reacting flows without restricting the form of species internal energies. The key step is to pre-tabulate each species' specific heat and internal energy on a temperature grid and use the piecewise-linear reconstruction to write the entropy integral in closed form, so the numerical flux is entropy- and kinetic-energy-preserving and its cost does not depend on the thermodynamic model. The authors verify machine-precision entropy conservation on a weak blast wave, match master-equation chemistry on homogeneous relaxation, and reproduce a reference finite-volume solution for a Mach 10 reacting flow around a cylinder. If correct, the approach lets one high-order solver handle arbitrary gas models by swapping data tables rather than deriving new fluxes.","feed_headline":"One entropy-stable flux handles any internal-energy model","feed_subtitle":"Tabulated heat capacities give a closed-form entropy integral, so reacting high-enthalpy flows stay stable","key_machinery":"The load-bearing object is the closed-form entropy integral obtained from piecewise-linear tabulated data. For each species, $\\varepsilon^{\\mathrm{int}}_c(T)$ and $c_{v,c}(T)$ are tabulated on a temperature grid of step $\\Delta T$; linear reconstruction turns the entropy integral $\\eta_c(T)=\\int_0^T c_{v,c}(\\tau)/\\tau\\,d\\tau$ into Eq. (15), a sum of tabulated interval integrals $\\eta_{c,i}$ plus one logarithm. This $\\eta_c(T)$ feeds the entropy-conservative flux: logarithmic averages for species densities, geometric and arithmetic averages for temperature and internal-energy terms in the energy flux (Eq. (12)), so the discrete interface flux is consistent with the entropy pair and conserves kinetic energy. Precomputed prefix sums of $\\eta_{c,i}$ keep the evaluation cost at one lookup and one logarithm per species.","core_discovery":"The central claim is that a single numerical flux can conserve entropy and kinetic energy for a compressible gas mixture with arbitrary species internal-energy models. The construction replaces the true functions $c_{v,c}(T)$ and $\\varepsilon^{\\mathrm{int}}_c(T)$ by piecewise-linear tabulated reconstructions; the entropy integral $\\eta_c(T)=\\int_0^T c_{v,c}(\\tau)/\\tau\\,d\\tau$ then has the closed form of Eq. (15), so the entropy-conservative interface flux (Eqs. (9)-(12)) can be evaluated with table lookups and one logarithm per species. The paper verifies the flux on a weak blast wave, where discrete entropy production stays at machine precision, on homogeneous chemical relaxation, and on a Mach 10 reacting cylinder flow, matching an established finite-volume solver.","pith_inferences":["A natural next step is to derive entropy-dissipative viscous and heat-conduction fluxes in the same tabulated framework, since the paper's conservation guarantee covers the inviscid Euler fluxes only.","Table resolution $\\Delta T$ becomes a user-facing accuracy parameter; for a given physical gas, practitioners will need a resolution criterion tied to the temperature range of the simulation.","The same flux could ingest internal-energy tables from kinetic-theory or state-resolved chemistry data without a new derivation, coupling CFD directly to high-fidelity thermochemistry.","In reacting simulations, the chemical source term is not entropy-conserving, so the overall scheme's stability depends on source-term treatment as well as on the flux."],"forward_implications":["Any species internal-energy model can enter the same flux by tabulation, with no analytic entropy integrals required.","The cost per interface state is one table lookup and one logarithm per species, independent of thermodynamic model complexity.","The discrete entropy production stays at machine precision in the non-reacting verification for polynomial orders 2, 4, and 6.","Reacting high-enthalpy flows with strong shocks and adaptively refined non-conforming meshes run stably and match the benchmark solver.","Chemical relaxation of an O2/O mixture reproduces the master-equation reference solution."],"supporting_citations":[{"why":"Supplies the single-species tabulated-interpolation flux construction that this paper extends to mixtures.","marker":"[17]"},{"why":"Establishes the use of tabulated gas data with piecewise-linear reconstruction inside a DG code, which the flux relies on.","marker":"[18]"},{"why":"Provides the weak blast wave test case and the shock indicator used for adaptive refinement.","marker":"[9]"},{"why":"Provides the subcell limiter used to keep the strong-shock cylinder simulation stable.","marker":"[10]"},{"why":"Supplies the cut-off harmonic oscillator internal-energy model and the master-equation reference solutions for chemistry verification.","marker":"[23]"},{"why":"Supplies the Arrhenius dissociation rate coefficients used in the reacting-flow test cases.","marker":"[24]"},{"why":"Supplies the reference finite-volume solver results for the Mach 10 cylinder comparison.","marker":"[26]"}],"fun_headline_variants":["One flux, any gas model: entropy-stable for all","Closed-form entropy integral from tabulated heat capacities","Single numerical flux conserves entropy for arbitrary species","Machine-precision entropy in reacting flow simulations","High-enthalpy flows stabilized by one universal flux"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entropy-conservation guarantee holds for the piecewise-linear tabulated thermodynamic model, not for the true physical internal-energy functions, so a coarse temperature table would conserve the wrong gas's entropy.","fun_headline_variants_meta":{"raw":{"variants":["One flux, any gas model: entropy-stable for all","Closed-form entropy integral from tabulated heat capacities","Single numerical flux conserves entropy for arbitrary species","Machine-precision entropy in reacting flow simulations","High-enthalpy flows stabilized by one universal flux"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000227,"raw_usage":{"total_tokens":1384,"prompt_tokens":773,"completion_tokens":611,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":389,"completion_tokens_details":{"reasoning_tokens":537}},"tokens_in":389,"tokens_out":611,"duration_ms":6668,"temperature":1.0,"reasoning_tokens":537,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:45:27.800171+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the weak blast wave with the same physical O2/O mixture tabulated at a coarse step, say $\\Delta T = 500$ K, and at a fine step, $\\Delta T = 1$ K, on identical grids and time steps; if the coarse-table pressure and temperature fields differ by more than discretization error, or if entropy production leaves machine precision when temperatures approach the table boundary, the method's guarantee is tied to the interpolation table rather than to the physical model.","supporting_citations":[{"cited_title":"Entropy-conservative high-order methods for high-enthalpy gas flows","cited_arxiv_id":"2403.16882","evidence_quote":"Supplies the single-species tabulated-interpolation flux construction that this paper extends to mixtures."},{"cited_title":"Simulation of real gas effects in supersonic methane jets using a tabulated equation of state with a discontinuous Galerkin spectral element method,","cited_arxiv_id":null,"evidence_quote":"Establishes the use of tabulated gas data with piecewise-linear reconstruction inside a DG code, which the flux relies on."},{"cited_title":"A provably entropy stable subcell shock capturing approach for high order split form DG for the compressible Euler equations,","cited_arxiv_id":null,"evidence_quote":"Provides the weak blast wave test case and the shock indicator used for adaptive refinement."},{"cited_title":"Subcell limiting strategies for discontinuous galerkin spectral element methods,","cited_arxiv_id":null,"evidence_quote":"Provides the subcell limiter used to keep the strong-shock cylinder simulation stable."},{"cited_title":"Nagnibeda and E","cited_arxiv_id":null,"evidence_quote":"Supplies the cut-off harmonic oscillator internal-energy model and the master-equation reference solutions for chemistry verification."},{"cited_title":"Review of chemical-kinetic problems of future NASA missions. I-Earth entries,","cited_arxiv_id":null,"evidence_quote":"Supplies the Arrhenius dissociation rate coefficients used in the reacting-flow test cases."},{"cited_title":"Validation of the unstructured DLR-TAU-Code for hypersonic flows,","cited_arxiv_id":null,"evidence_quote":"Supplies the reference finite-volume solver results for the Mach 10 cylinder comparison."}],"review_version":1}