{"id":"3bd05f80-523b-488f-8981-43a93e159d26","arxiv_id":"2411.17590","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Head-on quenching of thermodiffusively unstable lean hydrogen/air flames produces higher wall heat fluxes and smaller quenching distances than one-dimensional models predict, and these effects can be reproduced by an ensemble of one-dimensional flames at different local equivalence ratios.","lead":"This paper simulates, in two dimensions, a lean hydrogen flame that is naturally unstable and collides head-on with a cold wall. It finds that the unstable flame gets closer to the wall and deposits more heat than standard one-dimensional flame-wall models predict, and that local fuel-air mixture variations, not flame surface area, cause the difference.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Core 'ensemble of 1D HOQs' claim is inferred from correlations; the paper never constructs the ensemble from the 2D local-φ distribution and checks the predicted quenching statistics.","rationale":"Read in good faith, the paper's message is that thermodiffusive instabilities change FWI mainly through local mixture variation, and that 1D HOQs at different equivalence ratios can represent the unstable FWI. This is a strong, useful claim. The weakest step is the inference from correlation to representation. The paper never performs the ensemble reconstruction; it shows that the 2D data lie within the range of 1D results and that conditional means align, but this does not prove that the full distribution or the correlations between Φq and Peq are captured. A spurious correlation could arise because the local φ at the measurement point is itself affected by the same near-wall transport that sets the heat flux. The 1D comparison uses the unburnt equivalence ratio, so the x-axis variables match only loosely. The rare events in Appendix A are a reminder that the mapping is not one-to-one. These issues do not invalidate the physical mechanism proposed, but they do mean the central 'ensemble representation' conclusion is not yet demonstrated. The reader's CONDITIONAL verdict already calls for more support; our concern strengthens that by pointing to a specific missing validation step. If the reconstruction test passes, the paper's headline claim would be much stronger; if it fails, the claim would need to be downgraded to a qualitative correlation. Thus no verdict change is needed beyond the existing CONDITIONAL.","tokens_in":18065,"tokens_out":12111,"duration_ms":112730,"concrete_test":"Take the λ=100 2D runs and, for every wall position y, record the local equivalence ratio φ at x/δ_T=1 at the quenching time tq. Use the precomputed 1D HOQ series to define response functions Φq,1D(φ) and Peq,1D(φ). Construct the predicted PDF of normalized Φq and Peq, the joint distribution, and the predicted temporal profile of wall heat flux by mixture over the measured φ values, and compare with the 2D results using quantile-quantile plots or a Kolmogorov-Smirnov test. Also rerun the comparison using, for the 1D HOQs, the local φ at x/δ_T=1 at their own quenching time instead of φu,1D, so that the 1D and 2D conditioning variables are identical. If the reconstructed distributions agree with the 2D data within sampling uncertainty, the ensemble claim is verified; if not, local φ is insufficient and the claim must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central conclusion (Sec. 5) is that unstable flame-wall interaction 'can be represented by an ensemble of one-dimensional head-on quenchings at different equivalence ratios.' The supporting evidence is correlational: Fig. 8(right) shows 2D scatter of normalized Φq vs Peq colored by local φ at x/δ_T=1 aligned with 1D HOQ markers; Fig. 9 shows the 2D wall-heat-flux joint PDF bracketed by 1D HOQs at different φu,1D and a conditional mean matching φ=0.44; Fig. 11(right) shows mean Φq vs the reactivity factor I0 collapsing onto the 1D Φq vs sl/sl,ref curve. None of these actually constructs the ensemble. A direct test would map the measured 2D local φ distribution through the 1D HOQ response functions Φq,1D(φ) and Peq,1D(φ) and compare the predicted PDFs, joint distribution, and temporal evolution with the 2D data. The current comparison is also between two different variables: the 2D points are colored by the local φ at x/δ_T=1 at quenching time, whereas the 1D markers are plotted at the unburnt φu,1D; these are not the same observable, so the collapse may be partly coincidental. The Appendix A rare events (lean pockets, double quenching) show the representation is not exact, but their frequency and impact are not quantified. If local φ is not a sufficient statistic for quenching outcome, the modelling conclusion fails.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents two-dimensional direct numerical simulations of head-on quenching (HOQ) of a thermodiffusively unstable lean hydrogen/air flame (φu = 0.4, Tu = 298 K, p = 1 atm) in domains of varying lateral width λ = Ly/δ_T between 2 and 100. A freely propagating unstable flame is first generated in a periodic domain and then mapped into a wall-bounded domain with an isothermal inert wall. The authors report that, compared to one-dimensional HOQ at the same nominal conditions, the unstable flame has smaller quenching Peclet numbers and larger quenching wall heat fluxes, and that these changes correlate with local mixture-enrichment-induced reactivity rather than with flame surface area or consumption speed. They conclude that unstable flame-wall interactions can be represented by an ensemble of one-dimensional head-on quenchings at different equivalence ratios, and that local mixture variations associated with differential and preferential diffusion are the dominant governing factor.","tokens_in":18331,"tokens_out":7106,"duration_ms":63991,"significance":"The paper provides a novel, detailed dataset on flame-wall interaction of thermodiffusively unstable H2/air flames, a configuration that is increasingly relevant for lean hydrogen combustion systems. If the proposed mechanism is confirmed, the study has direct implications for flamelet-based modeling of near-wall hydrogen combustion and for the interpretation of experimental wall-temperature patterns such as those of Ojo et al. [35]. Strengths of the work include the use of detailed chemistry with Soret diffusion, a systematic domain-width sweep that varies the instability cell-size distribution, multiple HOQ realizations per configuration, and several complementary analyses (PDFs of quenching quantities, conditionally averaged wall heat flux, and the reactivity factor I0). The main limitation is that the central ensemble-representation claim is inferred from correlations rather than demonstrated by a direct predictive test, and one of the key comparisons uses different conditioning variables for the two-dimensional and one-dimensional data.","major_comments":[{"comment":"The comparison between the two-dimensional scatter points, colored by the local equivalence ratio at x/δ_T = 1 at the quenching time, and the one-dimensional HOQ markers, parameterized by the unburnt equivalence ratio φu,1D, is not an apples-to-apples comparison. As the paper itself notes in the footnote on page 11, differential diffusion alters the wall-normal equivalence-ratio profile even in one-dimensional HOQs, so the local φ at x/δ_T = 1 in a one-dimensional HOQ is generally different from its φu. Unless the two data sets are represented with the same conditioning observable (for instance, local φ at x/δ_T = 1 for both, or an equivalent unburnt-based quantity for the two-dimensional cases), the apparent collapse in Fig. 8 (right) could be partly coincidental. The authors should provide a comparison that uses the same observable for both data sets, or at least quantify the difference between local and unburnt equivalence ratios for the one-dimensional HOQs.","section":"§4.2, Fig. 8 (right)"},{"comment":"The statement that unstable flame-wall interactions 'can be represented by an ensemble of one-dimensional head-on quenchings at different equivalence ratios' goes beyond the evidence actually presented. The support consists of correlations (Figs. 8, 9, and 11) and conditional averages, but the paper never constructs such an ensemble from the measured two-dimensional local-φ distribution and then compares the predicted statistics (PDFs of Φq and Peq, their joint distribution, or their temporal evolution) against the two-dimensional DNS. Appendix A documents rare events (lean pockets, double quenching) that are exceptions to the representation, but their frequency and contribution to the overall statistics are not quantified, leaving the 'rare' qualifier unsubstantiated. Please either perform a direct ensemble test using the existing one-dimensional HOQ response functions, or soften the claim throughout the manuscript, including the Abstract and the novelty statement, to phrasing such as 'consistent with' rather than 'can be represented by'.","section":"§5 and Novelty and Significance Statement"},{"comment":"The grid resolution of 20 grid points per flame thickness is not supported by a grid-convergence study. Because the quantitative conclusions concern near-wall quantities (quenching wall heat flux and quenching distance) that are sensitive to the resolution of the thermal boundary layer and to hydrogen differential diffusion near the wall, a resolution study, or at least a demonstration that the two-dimensional-versus-one-dimensional comparison is independent of resolution, is needed to establish the robustness of the quantitative claims.","section":"§2.1.1"}],"minor_comments":[{"comment":"The text contains a typographical error: 'realease' should be 'release'.","section":"Appendix A.2"},{"comment":"The statement that there is 'no discernible correlation' between the mean wall heat flux and the consumption speed or surface area is based on visual inspection; a quantitative correlation measure (for example, Pearson's or Spearman's coefficient) would make this claim more rigorous.","section":"§4.3, Fig. 11 (left)"},{"comment":"The paper does not include a data availability statement; given the potential value of the dataset for the community, making the simulation configurations and post-processing scripts available would improve reproducibility.","section":"Article front matter"},{"comment":"The Abstract states that the increased wall heat fluxes 'are caused by' enhanced reactivity, whereas the body of the paper uses 'suggest' and 'seems'; the level of certainty should be harmonized between the Abstract and the conclusions.","section":"Abstract and Conclusions"},{"comment":"The caption of Fig. 11 (right) contains the phrase 'for varying a equivalence ratio', which should read 'for varying equivalence ratio'.","section":"Fig. 11 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid and useful characterization study, and the authors are to be commended for the systematic domain-size sweep and the careful separation of kinematic versus differential-diffusion effects. However, the central ensemble-representation claim is stronger than the evidence currently presented. The variable mismatch in Fig. 8 and the absence of a direct ensemble test are the main reasons for requesting a major revision. The authors should be encouraged to perform the direct test with their existing data, since it would substantially strengthen the paper and bring the conclusions in line with the evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First things first: this is the first DNS of head-on quenching of a thermodiffusively unstable lean hydrogen/air flame, and the central qualitative result—instabilities raise wall heat flux and lower quenching distance relative to the 1D HOQ—looks solid. The authors also make a good case that the mechanism is local mixture variation from differential diffusion, not flame surface area: the reactivity factor I0 tracks the wall heat flux across domain widths, and the conditional mean wall heat flux matches a 1D HOQ at phi = 0.44.\n\nWhat I like: the analysis is unusually thorough. The PDFs of quenching heat flux and Peclet number, the temporal evolution of wall heat flux bracketed by 1D HOQs, the domain-size sweep, and the appendix on rare events (lean pockets and double quenching) all support the qualitative picture. The numerics are credible: detailed H2 mechanism, mixture-averaged transport with Soret, 20 points per flame thickness, and the code has prior validation. No fitted parameters in the 1D comparisons.\n\nThe soft spot is the strongest interpretive claim. The paper concludes that unstable flame-wall interaction 'can be represented by an ensemble of one-dimensional head-on quenchings at different equivalence ratios.' That is a modeling statement, and it is inferred from correlations, not tested. The 2D scatter is colored by local phi at x/delta_T = 1 at quenching time, while the 1D markers are plotted at unburnt phi_u,1D; those are not the same observable. The conditional mean matching phi = 0.44 is suggestive, but the paper never maps the 2D local phi distribution through the 1D PHI_q(phi) and Pe_q(phi) response functions to predict the PDFs and compare with the 2D data. Until that is done, I would treat the ensemble statement as a hypothesis rather than a demonstrated result. The rare events in Appendix A show the representation is not exact, and their frequency is not quantified.\n\nAlso, the statistics are thin at the main condition: only 4 HOQ realizations for lambda = 100. The PDFs are based on that, so the quantitative tails are not well-constrained. This is a minor issue for the main trend, but it matters if someone uses the numbers for model tuning.\n\nWho should read this: combustion modelers working on hydrogen flame-wall interaction, thermal loading, and flamelet manifolds. It deserves a serious referee and, I think, publication after the ensemble claim is either directly tested or softened. I would engage with it.","headline":"First 2D DNS of head-on quenching of thermodiffusively unstable lean H2/air flames; the qualitative findings are solid, but the central 'ensemble of 1D HOQs' claim is inferred from correlations rather than directly tested.","tokens_in":18867,"tokens_out":2872,"would_cite":true,"duration_ms":25741,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Thermodiffusively unstable lean hydrogen/air flames quench closer to the wall and drive higher wall heat fluxes than one-dimensional head-on quenching predicts, and the paper traces this to local mixture variation rather than…","keywords":["flame-wall interaction","head-on quenching","thermodiffusive instability","hydrogen/air premixed flames","differential diffusion","wall heat flux","quenching distance","direct numerical simulation"],"falsifier":"Measure, in a controlled lean hydrogen/air flame-wall experiment near φ=0.4, 298 K, and 1 atm, the spatiotemporal wall heat flux and the quenching distance along the wall together with a local equivalence-ratio indicator near quenching. If the wall heat flux and quenching distance do not collapse onto one-dimensional head-on-quenching curves for the corresponding local equivalence ratio, or if they match the one-dimensional values at the global φ, the ensemble representation is wrong.","tokens_in":17853,"feed_emoji":"🔥","tokens_out":6759,"duration_ms":116078,"temperature":0.7,"pith_summary":"Lean hydrogen/air flames develop intrinsic thermodiffusive instabilities: hydrogen's fast diffusion wrinkles the flame and locally changes the fuel-air mixture. This paper uses two-dimensional direct numerical simulations of head-on quenching to ask how those unstable flames interact with a cold wall. It finds that, relative to a plain one-dimensional quenching flame at the same unburnt conditions, the unstable flame quenches closer to the wall and deposits more heat into it. The explanation is not the increased flame surface area; it is that instability-driven local enrichment raises the flame's reactivity before it reaches the wall. If correct, single one-dimensional quenching calculations cannot give reliable wall heat fluxes or quenching distances for lean hydrogen systems, and models must resolve the local mixture distribution.","feed_headline":"Lean hydrogen's unstable flames quench closer and heat walls more","feed_subtitle":"2D simulations show local mixture shifts, not flame folding, control quenching; 1D models miss this.","key_machinery":"The load-bearing machinery is the one-dimensional head-on-quenching curve as a function of equivalence ratio, used as a reference database, together with a streamline diagnostic that tags each flame segment by the maximum equivalence ratio along the local reaction-progress gradient. The authors separate mixture-composition effects from geometric effects with the reactivity factor $I_0 = (s_c/s_{l,\\mathrm{ref}})/(A/L_y)$, where $s_c$ is the consumption speed, $s_{l,\\mathrm{ref}}$ the reference laminar flame speed, and $A/L_y$ the normalized flame-surface area. The collapse of the two-dimensional quenching statistics onto the one-dimensional curves at varying φ identifies the local mixture variation, induced by differential and preferential diffusion, as the variable that carries the argument.","core_discovery":"The central claim is that the flame-wall interaction of a thermodiffusively unstable lean hydrogen/air flame can be described as an ensemble of one-dimensional head-on quenchings, each at the local equivalence ratio seen by that wall segment. In the two-dimensional simulations at φu=0.4, Tu=298 K, and 1 atm, the quenching Péclet number is lower and the quenching wall heat flux is higher than in the reference one-dimensional head-on quenching, and the scatter in these quantities collapses onto one-dimensional head-on-quenching curves for equivalence ratios roughly between 0.34 and 0.48. The authors show that the local equivalence ratio at one flame thickness from the wall correlates with both quenching distance and heat flux, that the conditioned average wall heat flux follows the one-dimensional case at φ=0.44, and that varying the lateral domain size changes quenching only through the reactivity factor, not through flame-surface area or consumption speed. They conclude that kinematic effects are minor and that the dominant mechanism is enhanced local reactivity from differential and preferential diffusion.","pith_inferences":["In three dimensions, with hydrodynamic instability and flame-generated turbulence adding strain, the local equivalence-ratio distribution may be broader or shifted, so whether the strict one-dimensional-ensemble collapse survives is an open question this paper does not settle.","Because hydrogen's high diffusivity lets mixture variations persist near the wall after quenching, a purely local steady mapping from φ to heat flux may miss memory effects in rapidly changing flame-front configurations; this could be tested by comparing predicted and simulated wall heat flux in strongly forced conditions.","The same reasoning suggests an experimental path: combining spatiotemporal wall-temperature measurements with a local equivalence-ratio marker would turn the qualitative wall-temperature-pattern observation into a direct quantitative test of the predicted heat-flux–φ correlation.","Part II's parametric sweep across pressure, equivalence ratio, and unburnt temperature will reveal whether the ensemble representation holds beyond φ=0.4, 298 K, and 1 atm or is restricted to the conditions studied here."],"forward_implications":["Quenching distances and wall heat fluxes for lean hydrogen/air flames cannot be reliably taken from a single one-dimensional head-on quenching at the global equivalence ratio; the instability shifts both systematically.","Combustion models for flame-wall interaction must carry the full local mixture distribution, not just flame surface area or progress variable, to reproduce thermal loads on walls.","Flame-surface-area growth and consumption-speed enhancement play only a minor role in determining quenching; the reactivity of the mixture arriving at the wall is what matters.","Restricting the lateral domain changes quenching statistics only through the reactivity factor, so instability cell-size distributions matter indirectly through their effect on local mixture variation.","The time history of wall heat flux during quenching is bracketed by one-dimensional head-on quenchings at varying equivalence ratio, so an ensemble of such solutions can serve as a cheap surrogate for the unstable case."],"supporting_citations":[{"why":"Prior DNS of lean planar hydrogen flames showing how domain height and Lewis number control nonlinear cell dynamics; motivates the domain-width sweep. ","marker":"[9]"},{"why":"Establishes that domain widths near 100 thermal flame thicknesses leave instability length scales unconstrained; sets the reference case. ","marker":"[11]"},{"why":"Supplies the streamline-of-YH2-gradient method and the finding that local consumption speed tracks local equivalence ratio; the paper's diagnostic core. ","marker":"[13]"},{"why":"Experimental phosphor-thermometry study of lean H2/air and CH4/air flame-wall interaction; the only experimental comparison, showing instability-induced alternating wall temperature zones. ","marker":"[35]"},{"why":"Provides the relationship between quenching distance, flame thickness, and wall heat flux used to interpret why local enrichment tightens quenching.","marker":"[47]"}],"fun_headline_variants":["Unstable H2 flames quench closer with higher wall heat flux","Lean H2 flame instability alters quenching via local mixture","Thermodiffusive instability cuts quench distance in H2 flames","H2 flame-wall: instabilities boost heat flux, not kinematics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that two-dimensional direct numerical simulation with periodic lateral boundaries, an isothermal inert wall with zero species flux, and 20 grid points per flame thickness faithfully represents the quenching physics, so the conclusions transfer to three-dimensional engines and gas turbines; the only experimental comparison shows qualitative wall-temperature patterns, not measured quenching distances or heat fluxes.","fun_headline_variants_meta":{"raw":{"variants":["Unstable H2 flames quench closer with higher wall heat flux","Lean H2 flame instability alters quenching via local mixture","Thermodiffusive instability cuts quench distance in H2 flames","H2 flame-wall: instabilities boost heat flux, not kinematics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000196,"raw_usage":{"total_tokens":1412,"prompt_tokens":1046,"completion_tokens":366,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":662,"completion_tokens_details":{"reasoning_tokens":294}},"tokens_in":662,"tokens_out":366,"duration_ms":4255,"temperature":1.0,"reasoning_tokens":294,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:56:33.541216+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure, in a controlled lean hydrogen/air flame-wall experiment near φ=0.4, 298 K, and 1 atm, the spatiotemporal wall heat flux and the quenching distance along the wall together with a local equivalence-ratio indicator near quenching. If the wall heat flux and quenching distance do not collapse onto one-dimensional head-on-quenching curves for the corresponding local equivalence ratio, or if they match the one-dimensional values at the global φ, the ensemble representation is wrong.","supporting_citations":[{"cited_title":"Altantzis, C","cited_arxiv_id":null,"evidence_quote":"Prior DNS of lean planar hydrogen flames showing how domain height and Lewis number control nonlinear cell dynamics; motivates the domain-width sweep."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental phosphor-thermometry study of lean H2/air and CH4/air flame-wall interaction; the only experimental comparison, showing instability-induced alternating wall temperature zones."},{"cited_title":"Poinsot, D","cited_arxiv_id":null,"evidence_quote":"Provides the relationship between quenching distance, flame thickness, and wall heat flux used to interpret why local enrichment tightens quenching."}],"review_version":1}