{"id":"480d0ab1-be81-4890-a8da-702e993f6893","arxiv_id":"2411.13077","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A point-source boundary condition is required for accurate drag at high incidence, and a simple Lagally-Filon correction approximates it for blocked sidewalls.","lead":"At high angles of attack, accurate airfoil drag simulations need a point source, not just a point vortex, in the far-field boundary conditions. A simple mass-conservation correction recovers most of the accuracy gain without the extra computational cost.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim rests on the point source adequately representing the deflected wake at A=30c; the paper shows qualitative velocity differences but never quantifies their effect on the CV force integrals.","rationale":"The reader's weakest assumption identifies the isotropic point-source representation as the key premise, and my analysis agrees that this is the most load-bearing element of the paper's argument. The paper demonstrates that PVSBC outperforms BC-3 and PVBC for a NACA 0012 airfoil at α=45° and A=30c, but it does so only by comparing against its own A=500c reference and against the other BCs; it never isolates the error caused by replacing the actual wake-induced far-field with a centered isotropic source. The manuscript contains internal evidence that the model is imperfect: Figs. 8–10 show discrepancies in the induced velocity field, and Table 10 reveals that the PVSBC creates vorticity layers along the top and bottom boundaries that substantially affect the moment balance when the full domain is used as the control volume. These observations do not disprove the central claim, but they leave open the possibility that the point source is merely a convenient first-order approximation whose success at A=30c depends on error cancellation. The proposed frozen-far-field test directly quantifies the error in the CV boundary integrals induced by the point-source approximation, thus settling whether the central claim is robust. I therefore maintain the reader's CONDITIONAL verdict: the paper is worthwhile and largely convincing, but this validation step is needed before the point-source BC can be trusted beyond the specific configuration tested. The unresolved moment inconsistency and the ad hoc Imai correction remain secondary concerns, but they are derivatives of the same underlying question about the completeness of the far-field model.","tokens_in":18410,"tokens_out":8830,"duration_ms":94781,"concrete_test":"Extract the converged velocity and pressure from the A=500c PVSBC simulation on the A=30c square boundary and impose them as Dirichlet conditions for velocity and pressure on the inlet, top, and bottom boundaries (using the same solver, grid, and turbulence model), with the outlet treated exactly as in the PVSBC run. Compute Cd, Cl, and Cm for this 'frozen far-field' case and compare with the PVSBC values at A=30c. If the differences in Cd exceed roughly 0.1% (i.e., a significant fraction of the 0.9% improvement PVSBC offers over BC-3), the point source model omits wake-induced components that matter at the domain boundary, and the central claim is not robust. If the differences are negligible, the point-source representation is confirmed as adequate for the force coefficients.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The PVSBC's superiority over BC-3 and PVBC depends on the premise that a point source of strength Λ = D/(ρU∞) placed at the airfoil center captures the irrotational far-field displacement effect of the stalled airfoil and its deflected wake. The paper's own Figs. 7–10 show that the actual u and v distributions at the domain boundaries differ from the PVSBC model, especially near and downstream of the wake, where the wake has grown and been deflected. These differences are acknowledged qualitatively (Section 4.1, 4.6) but never translated into an error estimate for the force and moment coefficients. If the distributed, deflected wake induces non-negligible velocity components at the top/bottom boundaries or a source centroid offset that matters at A=30c, then the boundary integrals in Eqs. (4)–(6) used to compute Cd, Cl, and Cm would be biased, and the observed closeness of PVSBC to A=500c could be coincidental or case-specific. Since the central claim is explicitly that the point source is more important than the point vortex for high-drag flows, the adequacy of this isotropic, centered-source representation is the load-bearing assumption; it is asserted via the Lagally-Filon relation but not independently validated against a more complete far-field model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates far-field boundary conditions for steady, incompressible, two-dimensional RANS simulations of a NACA 0012 airfoil at α=45° and Re=6×10^6. It compares a standard boundary condition (BC-3), a point-vortex boundary condition (PVBC), and a novel point-vortex-source boundary condition (PVSBC) in which the source strength is set by the Lagally-Filon relation D=ρU∞Λ. The main claim is that for domains as small as A=10c to 30c, the PVSBC yields lift, drag, and moment coefficients closer to an A=500c reference than the other two BCs, because the point source represents the blockage-correcting effect of the wake at high drag. The paper also derives a simple 'Lagally-Filon' correction for the computational sidewall blockage induced by BC-3/PVBC, and shows that this correction brings those results close to the PVSBC values. Finally, the PVSBC is shown to be inconsistent with the impulse moment equation, but the authors argue the residual correction is small.","tokens_in":18668,"tokens_out":16715,"duration_ms":144025,"significance":"The paper addresses an important practical problem: the choice of far-field boundary conditions in airfoil simulations at high incidence, where drag and moment are large. If the central claim holds, the paper makes a useful contribution by showing that a point source is a more important element of the far-field model than a point vortex in high-drag flows, and by providing a simple analytical correction for computational blockage that closely tracks a known wind-tunnel blockage correction (Section 4.1, Eq. 21 vs. Eq. 22). The derivation of the Lagally-Filon correction from mass conservation is clean and parameter-free, and the comparison with experimental blockage corrections is a strength. The paper also carefully documents grid convergence and uses a consistent control-volume framework. However, the main comparison (Table 5) is weakened by the fact that the PVSBC sets its vortex and source strengths from the running solution's Cl and Cd, and by the absence of a quantitative sensitivity check of the isotropic point-source representation of the wake.","major_comments":[{"comment":"The central claim that PVSBC is more accurate than BC-3 and PVBC rests on the assumption that the far-field effect of the deeply stalled airfoil and its deflected wake is adequately represented by an isotropic point source of strength Λ=D/(ρU∞) located at the airfoil center (Eq. 15). The paper's own Figs. 7–10 show that the actual velocity distributions at the domain boundaries differ from this model, particularly near and downstream of the wake, and these differences are acknowledged qualitatively in Section 4.3. However, the effect of these differences on the boundary integrals in Eqs. (1)–(6) is never quantified. Since the PVSBC sets its source and vortex strengths from the computed Cl and Cd, the agreement with the A=500c reference in Table 5 is partly a consistency check of the model rather than an independent validation. To establish that the point source is the dominant far-field effect at A=30c, the authors should provide a sensitivity analysis with respect to source location or distribution, or impose the actual A=500c velocity profiles as boundary conditions at A=30c and compare the resulting forces.","section":"§3.4, §4.3, §4.6"},{"comment":"The PVSBC's inconsistency with the moment equation is not as small as the abstract suggests. For A=30c, the impulse moment balance over the CV coincident with the computational domain has an error of 65.9% (Table 10), whereas moving the CV inward by 1c (y=±29c) reduces the error to 0.05%. The paper attributes this to vorticity layers at the top and bottom boundaries but does not analyze these layers in detail. The statement that 'the further correction for this inconsistency is shown to be very small' refers to the Imai correction for the logarithmic divergence, not to the 66% imbalance shown in Table 10. The authors should either quantify the effect of the vorticity layers on the reported Cm and on the force integrals, or explicitly restrict the consistency claim to lift and drag.","section":"§4.5, Table 10"}],"minor_comments":[{"comment":"The zero-grid-spacing values of Cd=0.9082 and Cl=0.9207 reported after the grid convergence study appear to correspond to the BC-3 A=30c case rather than the A=500c reference used in Table 5. Please clarify whether a separate grid convergence study was performed for the A=500c PVSBC reference, and report the corresponding uncertainties.","section":"§3.2"},{"comment":"The statement that the Lagally-Filon relation is 'largely unknown' is difficult to reconcile with the three prior uses cited (Kelmanson 1987, Dannenhoffer 1987, Allmaras et al. 2005). Consider softening the phrasing.","section":"§1"},{"comment":"The procedure of 'approximately doubling' the LF correction to match the experimental data of Sheldahl and Klimas and Critzos et al. is ad hoc. A quantitative basis for the extra factor of two would strengthen the comparison.","section":"§4.2"},{"comment":"Equation (26) is an empirical fit with two free parameters (a*, b*) per component. Since the paper states that Imai's streamfunction contains an error and is not used directly, the support for the smallness of the Imai correction would be enhanced by reporting the sensitivity of the fitted parameters to the fit range.","section":"§5"},{"comment":"Reference [25] lists the date for Critzos et al. as 1995; the NACA TN 3361 report is from 1955. Please correct this citation.","section":"References"},{"comment":"The notation in Eq. (13) mixes terms with and without U∞ factors (e.g., the term '−∫_O uvdy' has no U∞ factor while neighboring terms do). Please check the notational consistency and clarify that all terms are perturbation quantities.","section":"§2, Eq. (13)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a reasonable fit for a CFD methods journal. The most solid contribution is the LF blockage correction, which is derived cleanly and validated against a wind-tunnel correction. The PVSBC's advantage is plausible but would be more convincing with a stronger independence test, e.g., using actual far-field profiles as Dirichlet conditions or testing source-location sensitivity. The moment inconsistency revealed in Table 10 is a notable caveat that the authors should address in revision, either by quantifying its effect or by re-scoping the consistency claim. The manuscript is otherwise clearly written and the computational work appears careful."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a solid, focused CFD paper with a genuinely new practical result: at high incidence, where drag is comparable to lift, the far-field boundary condition needs a point source (the Lagally-Filon drag-source) more than it needs the usual point vortex. The authors demonstrate this cleanly and then show that a simple mass-conservation correction (Eq. 21) recovers most of the benefit for the cheaper standard BCs.\n\nWhat's new: the point source dominates boundary-condition accuracy at high drag, and the sidewall blockage in the computational domain mimics wind-tunnel blockage. The derivation of Eq. (21) from mass conservation and drag is not fitted; it is a real argument. The grid-convergence and the impulse-equation balances (Tables 7–9) give internal consistency: PVSBC is closer to the A=500c reference at all smaller domains. The paper is also honest about its limitations: it flags the moment inconsistency (66% error for the full CV at A=30c), the vorticity layers near sidewalls, and the unresolved Imai correction.\n\nThe main soft spot is the load-bearing assumption that a point source of strength D/(ρU∞) at the airfoil center captures the far-field displacement of the deflected wake. The paper shows velocity profiles that deviate near the wake but never quantifies the error directly against a more complete model. However, the impulse-equation balances do translate the integrated effect into force coefficients: errors of 0.02% for Cd and 0.006% for Cl at A=30c with PVSBC, which is strong indirect evidence. So this concern is real but not fatal.\n\nAlso: the PVSBC is mildly circular (it uses the computed Cl and Cd to set the BC strengths); convergence to the large-domain reference is evidence, not proof. The paper is a single case (one airfoil, one angle, one Re), so generalizability is unknown. No code or data are public, which hurts reproducibility. The experimental comparison is ad hoc: the 'further correction of similar magnitude' applied to Sheldahl and Klimas data is not derived, just scaled from Fig. 4.\n\nWho this is for: CFD practitioners doing steady RANS at high alpha, especially for vertical-axis turbines or wind turbines in deep stall. It deserves a serious referee; the core result is practically useful and the analysis is careful. I would send it to peer review, with requests to release the code/data and to make the experimental comparison more principled.","headline":"Careful, useful CFD study showing that high-incidence airfoil far-field BCs need a point source, not just a point vortex; the main weakness is the unvalidated point-source representation of the deflected wake.","tokens_in":19211,"tokens_out":3635,"would_cite":true,"duration_ms":34148,"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":"At high incidence, consistent far-field boundary conditions must include a point source of strength D/(ρU∞) alongside the usual point vortex; the paper shows this removes most of the drag error on small domains and derives a cheap…","keywords":["airfoil simulation","far-field boundary conditions","high angle of attack","point source","point vortex","Lagally-Filon relation","impulse equations","blockage correction"],"falsifier":"Compute the same case with PVSBC on a series of domains A = 10c, 20c, 30c, 50c, and 100c and compare with A = 500c; if the coefficients do not approach the reference within the quoted margins, or if the boundary velocity profile departs measurably from Eq. (15) outside the wake, the isotropic-source model is inadequate. A sharper check is to repeat at α ≈ 60°, where C_l C_d is maximal: the claimed smallness of the moment inconsistency predicts only mild growth in the residual, whereas a noticeable C_m error at small domains would falsify it.","tokens_in":18168,"feed_emoji":"✈️","tokens_out":9181,"duration_ms":93866,"temperature":0.7,"pith_summary":"Simulating a deeply stalled airfoil in a finite computational box needs a far-field boundary condition that represents drag as well as lift. This paper shows that at 45° incidence, where drag equals lift in size, the correct representation is a point source of strength Λ = D/(ρU∞) added to the usual point vortex; the source balances the mass deficit of the deflected wake. Boundary conditions that lack the source seal the top and bottom sides of the square domain, creating a blockage that inflates drag, much like wind-tunnel sidewall blockage. The paper derives a one-line correction, the Lagally-Filon correction, that brings ordinary boundary-condition results close to the source-based ones, and shows the residual moment inconsistency is small for this flow. If the result holds, airfoil simulations on small domains can be made more accurate or cheaper by matching boundary conditions to the physics of the wake.","feed_headline":"Drag source in far-field BCs fixes stalled-airfoil errors","feed_subtitle":"Adding a point source of strength D/(ρU∞) cuts drag error from about 1.2% to 0.3% on a 30-chord domain.","key_machinery":"The carrying object is the point-source contribution to the far-field velocity: (u, v) from a source of strength Λ = D/(ρU∞) at the airfoil center, combined with the point-vortex term (Eq. 15). The source represents the mass deficit of the deflected wake and restores mass conservation at the domain boundaries, so the top and bottom walls no longer block the flow. The paper's derived Lagally-Filon correction, u**/U_I = (1/2)(√(1 + C_d** c/(2A)) − 1) and C_d*/C_d** = (1 + u**/U_I)^{-2}, encodes that blockage in terms of the computed drag and domain size, turning a cheap blocked-boundary simulation into an approximation of the source-based one. For the moment, the key diagnostic is the vorticity integral ∫_O y²Ω dy, whose logarithmic growth, proportional to C_l C_d downstream, quantifies the inconsistency that the source condition cannot remove.","core_discovery":"The paper's central claim is that consistency with the impulse-form drag equation requires a point source in the far-field boundary conditions, with strength set by the Lagally-Filon relation Λ = D/(ρU∞), just as consistency with lift requires a point vortex of circulation Γ = L/(ρU∞). Using a NACA 0012 airfoil at Re = 6×$10^{6}$ and α = 45°, the authors compare three boundary-condition sets on square domains from A = 10c to A = 500c. The point-vortex-plus-source condition (PVSBC) gives drag, lift, and moment coefficients at A = 30c much closer to the A = 500c reference than either the point-vortex-only condition or a standard fixed/slip condition; for drag the error drops from about 1.2% to about 0.3%. Standard conditions produce an artificially reduced inlet velocity and sidewall blockage; the derived Lagally-Filon correction, applied to those results, brings them close to PVSBC. The PVSBC still fails exact moment consistency because the moment's vorticity integral diverges logarithmically with downstream distance as the wake deflects, but the residual Imai correction to the flow ahead of the airfoil is shown to be about 5×$10^{{-4}}$, negligible for the force and moment coefficients.","pith_inferences":["Editorial: The same sidewall-blockage mechanism should appear in any two-dimensional high-incidence simulation, steady or unsteady; a testable extension is whether time-averaged drag errors in unsteady stalled flows obey the same c/A scaling as the Lagally-Filon correction.","Editorial: Because the Lagally-Filon relation does not carry over to three dimensions, where induced drag lives in the Trefftz plane, a three-dimensional analogue would need a different construction, limiting the direct transfer of PVSBC to wings or rotors.","Editorial: Since C_l C_d peaks near α ≈ 60°, that angle offers the sharpest test of the paper's claim that the moment inconsistency remains negligible at small domains.","Editorial: The closeness of the Lagally-Filon correction to established wind-tunnel blockage corrections suggests a unified way to compare CFD and experiment in high-blockage facilities: correct both simulated and measured coefficients with the same formula, so remaining differences reflect turbulence-model error rather than domain effects."],"forward_implications":["At a fixed domain size, using PVSBC removes most of the drag error for high-incidence airfoils; at A = 30c the drag coefficient error falls from about 1.2% to about 0.3% relative to the A = 500c reference.","For users of standard boundary conditions, the Lagally-Filon correction offers a post-processing route to obtain accurate coefficients without the roughly 50% extra cost of PVSBC, provided the blockage is small.","The inlet velocity in a stalled-airfoil simulation should not be treated as U∞; the paper's u** analysis implies that optimization routines and database computations on small domains carry a systematic lift-to-drag bias unless corrected.","The remaining moment inconsistency is negligible for this case, but the paper identifies vertical-axis-turbine flows, where the moment is critical, as the case where the higher-order correction could matter."],"supporting_citations":[{"why":"First derivation of the drag-source relation D = ρU∞Λ used to set the source strength.","marker":"[1]"},{"why":"Independent derivation of the same drag-source relation, giving the Lagally-Filon name.","marker":"[2]"},{"why":"Code-verification benchmark for the same airfoil and Reynolds number, used to assert that the simulations are otherwise accurate.","marker":"[4]"},{"why":"Prior study establishing the point-vortex boundary condition for lower incidence and supplying the BC-3/PVBC comparison and grid-independence procedure used here.","marker":"[6]"},{"why":"Modern derivation of the Lagally-Filon equation (5.12.15) and the laminar-wake profile used in the wake analysis.","marker":"[9]"},{"why":"Explanation that the moment's logarithmic divergence comes from wake deflection opposite to the lift.","marker":"[15]"},{"why":"Higher-order far-field solution used to estimate the residual velocity correction and to show that it is small.","marker":"[16]"},{"why":"Source of the impulse-form lift, drag, and moment equations applied to the control volume.","marker":"[17]"},{"why":"Wind-tunnel blockage correction whose form is compared with the derived Lagally-Filon correction.","marker":"[22]"}],"fun_headline_variants":["Point source in far-field BCs reduces drag error","Lagally-Filon source key to accurate drag at high AoA","Far-field BCs need a point source for drag","Point source fixes drag error at high incidence","Missing point source in BCs causes drag error"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The assumption that carries the argument is that a single isotropic point source placed at the airfoil center, with strength D/(ρU∞), adequately represents the far-field effect of the stalled airfoil and its deflected wake; if directional or higher-order wake effects matter at the boundaries, the source boundary condition and its correction are incomplete.","fun_headline_variants_meta":{"raw":{"variants":["Point source in far-field BCs reduces drag error","Lagally-Filon source key to accurate drag at high AoA","Far-field BCs need a point source for drag","Point source fixes drag error at high incidence","Missing point source in BCs causes drag error"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001934,"raw_usage":{"total_tokens":7655,"prompt_tokens":1118,"completion_tokens":6537,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":734,"completion_tokens_details":{"reasoning_tokens":6461}},"tokens_in":734,"tokens_out":6537,"duration_ms":42029,"temperature":1.0,"reasoning_tokens":6461,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:52:26.356029+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the same case with PVSBC on a series of domains A = 10c, 20c, 30c, 50c, and 100c and compare with A = 500c; if the coefficients do not approach the reference within the quoted margins, or if the boundary velocity profile departs measurably from Eq. (15) outside the wake, the isotropic-source model is inadequate. A sharper check is to repeat at α ≈ 60°, where C_l C_d is maximal: the claimed smallness of the moment inconsistency predicts only mild growth in the residual, whereas a noticeable C_m error at small domains would falsify it.","supporting_citations":[{"cited_title":"Proceedings of the Royal Society of London","cited_arxiv_id":null,"evidence_quote":"Explanation that the moment's logarithmic divergence comes from wake deflection opposite to the lift."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First derivation of the drag-source relation D = ρU∞Λ used to set the source strength."},{"cited_title":"Proceedings of the Royal Society of London","cited_arxiv_id":null,"evidence_quote":"Independent derivation of the same drag-source relation, giving the Lagally-Filon name."},{"cited_title":"J Aircr.55(4), 1338–1351 (2018) https://doi.org/10.2514/1.C034856","cited_arxiv_id":null,"evidence_quote":"Code-verification benchmark for the same airfoil and Reynolds number, used to assert that the simulations are otherwise accurate."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior study establishing the point-vortex boundary condition for lower incidence and supplying the BC-3/PVBC comparison and grid-independence procedure used here."},{"cited_title":"Proceedings of the Royal Society of London","cited_arxiv_id":null,"evidence_quote":"Higher-order far-field solution used to estimate the residual velocity correction and to show that it is small."},{"cited_title":"PhD thesis, California Institute of Technology (1997)","cited_arxiv_id":null,"evidence_quote":"Source of the impulse-form lift, drag, and moment equations applied to the control volume."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Wind-tunnel blockage correction whose form is compared with the derived Lagally-Filon correction."}],"review_version":1}