{"id":"3eb0a8c7-c0c6-4c65-9f52-7fec2ce62585","arxiv_id":"2607.17864","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A Robin outflow condition informed by local linear stability analysis lets global eigenmodes of bluff-body wakes converge on substantially smaller computational domains, particularly in low-Reynolds-number stable wakes.","lead":"This paper tests how the outflow boundary condition affects global stability calculations of bluff-body wakes, and shows that a Robin condition built from local stability theory converges on much shorter domains than classical Neumann or Dirichlet outlets. The practical payoff is faster, more reliable stability analysis of cylinder and airfoil wakes, especially at low Reynolds numbers.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Robin BC is under-specified: the neutral point (ω0, α0) used in eq. (8) is never defined for damped wake profiles, so the claimed convergence could depend on an undisclosed tuning parameter.","rationale":"The paper's central comparison is credible: eigenvalue convergence with x_o is improved by the Robin condition, the Re=40/70 contrast is physically plausible, and the airfoil results match He et al. across β. I therefore do not object to the numerical trend. However, the load-bearing link between 'local linear stability theory' and the outlet condition is eq. (8), a Taylor expansion about (ω0, α0). For a damped wake, no neutral point is defined; the only new algorithmic content is eq. (14), which presupposes α0. The manuscript gives no procedure for obtaining α0, and Figs. 4–5 show the local growth rate being recovered from the global eigenvalue, not independently prescribed. If the chosen α0/ω0 is actually derived from the global mode, the Robin condition is not an independent physical input and the comparison is partly circular. If it is chosen ad hoc, the method has a hidden free parameter, so the claimed robustness across x_o and Re is not established. This is exactly the reader's weakest assumption, and it should be made explicit and tested. If the α0 sweep shows insensitivity, the concern is resolved and the paper could be accepted; as it stands, CONDITIONAL is appropriate.","tokens_in":18997,"tokens_out":6230,"duration_ms":64764,"concrete_test":"Fix Re=40, cylinder wake, x_o=52. Compute the Robin-C least-stable eigenvalue for a sweep of expansion centers: α0 ∈ [0, 2] (or over the local LST branch if one is defined), with ω0 = ω_r from a long-domain reference. If the least-stable eigenvalue shifts by more than the claimed 0.2% convergence band, the unspecified neutral point is load-bearing and the central claim is underdetermined.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim requires that the Robin operator in eq. (13) be a physically derived outlet condition, not an adjustable suppressor. That operator is built from c0 and α0 via the Taylor expansion eq. (8) about a 'neutral point' (ω0, α0). The paper never states how (ω0, α0) is selected for a wake outlet. In the boundary-layer papers cited, the neutral point is an unambiguous zero-growth mode of the local Orr–Sommerfeld problem; for a damped Re=40 cylinder wake there is no such neutral mode, so the expansion center is undefined. Eq. (14) defines c0 only after α0 and the local eigenfunction are known, and §3.2.1 shows α_i being backed out of the global eigenvalue via eqs. (9)/(11) — which suggests the local parameters are not chosen a priori. If α0/ω0 are inferred from the global spectrum, the Robin BC inherits the answer it is supposed to produce, and the convergence comparisons in Figs. 2 and 9 are partly circular. If instead an ad hoc neutral point is chosen, the claimed 0.2% eigenvalue variation with x_o may merely reflect one favorable choice of a free parameter.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a matrix-forming, finite-difference global linear stability solver for incompressible bluff-body wakes and uses it to compare outflow boundary conditions (Dirichlet, Neumann, extrapolation, stress-free, sponge, and a Robin condition derived from local linear stability theory). The central claim is that the Robin condition yields eigenvalues and eigenfunctions that converge with respect to outlet location on significantly shorter domains, and that this is essential in the low-Reynolds-number globally stable regime where the global eigenmode amplifies spatially without bound. Two Robin variants are considered: Robin-R, which neglects the imaginary part of the group velocity, and Robin-C, which retains it. The solver is validated against published growth rates and frequencies for cylinder and NACA0015 airfoil wakes, and the Robin condition is shown to drastically reduce the required domain size, especially for the airfoil case at Re=200/220.","tokens_in":19401,"tokens_out":6522,"duration_ms":69540,"significance":"If the central claim holds, the paper makes a practically important contribution: it identifies a regime (low-Re stable wakes) where classical outlet conditions cannot be rescued by extending the domain, and it provides an OBC that is both physically motivated and compatible with matrix-forming eigenvalue solvers. The quantitative comparisons with independent literature values (Barkley 2006, Mittal 2010, He et al. 2017) are a genuine strength, as is the systematic comparison of the real and complex Robin variants. The paper also usefully documents that pressure eigenfunctions are far more sensitive to the outlet condition than velocity eigenfunctions. However, several load-bearing details of the Robin condition are underspecified, and one of the quantitative comparisons contains an inconsistency; these issues need to be resolved before the conclusions can be fully accepted.","major_comments":[{"comment":"The Robin condition is constructed from a Taylor expansion about a 'neutral point' (ω0, α0), but the manuscript never states how (ω0, α0) is selected for a damped wake profile. For Re<Re_cr there is no obvious neutral mode, so the expansion center is not self-evident. Eq. (14) requires α0 through the inner products, and Fig. 4 reconstructs α from the global eigenvalue using the same expansion. If α0 is inferred from the global spectrum, the BC inherits the answer it is meant to produce; if it is chosen ad hoc, the convergence comparisons in Figs. 2 and 9 may reflect a favorable tuning parameter. The authors must state the procedure for determining (ω0, α0) for each outlet and Reynolds number and provide a sensitivity study over that choice.","section":"§2.3.2, Eqs. (8)–(14)"},{"comment":"The validation of the local spatial growth rate is partly circular. In Fig. 4, α_r and α_i are recovered from the global eigenvalue using the same Taylor expansion, Eqs. (9) and (11), that defines the Robin BC. Figure 5 then compares the global mode's spatial slope with the α_i obtained in this way. To be an independent check, the authors should compare the global mode's spatial growth with roots of the local Orr–Sommerfeld dispersion relation computed at each outlet without invoking Eqs. (8)–(11) as both construction and diagnosis. Without this, the agreement in Fig. 5 does not establish that the local LST description is the correct one at the outlet.","section":"§3.2.1, Eqs. (9)–(11), Fig. 4"},{"comment":"The pressure-error metric ε_p is anchored to the Robin-C condition itself: p_ref is defined as the pressure mode obtained with Robin-C, so ε_p=0 for Robin-C by construction. This does not measure absolute accuracy; it measures agreement with the advocated method. The qualitative conclusion that Robin-C produces smoother pressure fields is supported by the contour plots, but the quantitative claim '0.06' and '0.17' etc. should be replaced or supplemented by an error relative to an independent reference, e.g., a solution on a much larger domain or with a different discretization.","section":"§3.2, Eq. (16)"},{"comment":"There is a Reynolds-number inconsistency: the text and Fig. 11 state the NACA0015 case is Re=200, while Table 2 is captioned Re=220 and compares with Table 7 of He et al. The convergence comparison in Table 2 is a central quantitative claim ('three orders of magnitude smaller'), but it is not clear whether the present computation was done at Re=200 or Re=220. The authors must correct the inconsistency and ensure that the comparison with the literature is made at the same Reynolds number.","section":"Table 2 / §3.4"},{"comment":"The claim that global eigenfunctions amplify indefinitely as x→∞ relies on the approximation ω=αU_c−i k^2/Re≈αU_c, i.e., neglect of the viscous term. For Re=40, the lowest Reynolds number considered, this approximation is not obviously valid unless |α|^2 ≪ Re|ω|. Since this approximation underpins the conclusion that 'further extension of the domain will not facilitate convergence' for low-Re stable wakes, the authors should quantify the neglected term for the parameters used, or provide a more careful asymptotic argument that the growing root of the full quadratic has positive spatial growth even when the viscous term is retained.","section":"§2.3.2, Eq. (12)"}],"minor_comments":[{"comment":"Typo: 'less affective' should be 'less effective' in the Conclusions.","section":"§4"},{"comment":"Please label the axes explicitly and state what α_r and α_i represent (units are presumably D^-1; the current caption is terse).","section":"Fig. 4"},{"comment":"The dashes for missing entries are not explained in the caption; state that the corresponding data are not available from the cited reference or were not computed.","section":"Table 2"},{"comment":"The notation ⟨q†,N q⟩ is confusing because N is both the operator and the inner-product bracket; please use a clearer notation, e.g., ⟨q†, N q⟩ with explicit definitions of the operators.","section":"Eq. (15)"},{"comment":"The discussion of the sponge layer shows strong sensitivity to σ0, which is good, but please also state whether the sponge is applied to both velocity and pressure, and whether the damping profile is applied in the same way as in Mani (2012).","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The main risk is the underspecified neutral point in the Robin construction. If the authors can provide a clear, reproducible procedure for selecting (ω0, α0) and demonstrate insensitivity to that choice, the paper's central claim will be substantially strengthened. The Re=200/220 inconsistency in Table 2 is the kind of error that must be fixed before publication. I do not see a need for rejection, as the independent literature comparisons in Figs. 8 and 11 suggest the core numerical results are sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth a serious referee. The central demonstration — a Robin outlet built from local linear stability theory makes global eigenvalues of cylinder and airfoil wakes essentially independent of outlet position, with clean pressure modes — is backed by a consistent set of convergence plots and independent literature cross-checks (Barkley, Mittal, He et al.). The claim that low-Re globally stable wakes are the hard case, because the eigenmode keeps growing spatially and longer domains do not fix classical Neumann/Dirichlet outlets, is supported both by the Re=40/70 contrast and by the uniform-stream argument around eq. (12). I found no evidence that the main results are fitted.\n\nWhat is genuinely new: the systematic OBC comparison in a matrix-forming framework, the Robin-C variant that retains the imaginary part of the group velocity and matches the local spatial growth rate far better than Robin-R, and the finding that stable wakes are the difficult regime. The sponge-layer appendix is honest about parameter sensitivity, which strengthens my confidence in the paper's judgment.\n\nThe soft spots are real but not fatal. First, the Robin BC in eq. (13) is under-specified. The paper states that a Taylor expansion about the neutral point gives eq. (8), and gives a c0 formula via biorthogonality, but never says how (ω0, α0) is selected for a damped wake profile. For a boundary layer the neutral point is well defined; for a Re=40 cylinder wake it is not. As written, α0 could be an ad hoc or a posteriori choice, and the convergence comparisons in Figs. 2 and 9 would then be partly circular. The results still look robust across a wide range of x_o and against literature values, so this is a reproducibility gap rather than a load-bearing flaw, but it must be fixed in revision. Second, the pressure-quality metric in eq. (16) uses Robin-C itself as the reference, so Robin-C has zero error by construction; useful for ranking, but not an independent benchmark. Third, Table 2 says Re=220 while the rest of the paper uses Re=200; presumably a typo, but in a quantitative convergence table that is confusing.\n\nWho this is for: practitioners of global stability of external wakes, and anyone wanting to shrink domains for Floquet or compressible extensions. The matrix-forming details in the appendices are practical and the method looks implementable once the missing recipe is supplied.\n\nRecommendation: send to peer review. Ask for the neutral-point selection procedure, an independent pressure error measure, and the Re correction. If those come back clean, this is a cite-worthy contribution.","headline":"Solid methodology paper: the Robin outlet genuinely improves global-mode convergence in truncated domains, especially low-Re stable wakes; but the BC's expansion point is under-specified and one table has a Re typo, so it needs revision, not rejection.","tokens_in":19770,"tokens_out":3642,"would_cite":true,"duration_ms":36523,"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":"Classical zero-value or zero-derivative outlets cannot converge for globally stable but spatially amplifying bluff-body wakes; a Robin outlet from local linear stability theory closes the problem in small domains.","keywords":["global linear stability analysis","outflow boundary condition","Robin condition","bluff body wake","local linear stability theory","matrix-forming eigenvalue method","spatial amplification","eigenmode convergence"],"falsifier":"Compute the leading global mode for the Re = 40 cylinder wake in a box with outlet at x_o = 200 (or use a high-order spectral reference domain) and measure whether the Robin-C prediction matches the true spatial envelope all the way to the outlet; if the local linear-stability group velocity fails to reproduce the measured downstream growth rate, the Taylor assumption is the point that breaks. Equivalently, solve the full spatial Orr-Sommerfeld dispersion relation at x_o = 100 and compare its complex wavenumber with the Taylor estimate; a mismatch comparable to the reported eigenvalue variatio","tokens_in":18910,"feed_emoji":"🌊","tokens_out":7054,"duration_ms":71650,"temperature":0.7,"pith_summary":"Global stability analysis of bluff-body wakes asks how an infinitesimal disturbance grows in space and time, and the answer can depend on where you cut off the computational box. The paper argues that the standard way to close the box—Neumann or Dirichlet conditions at the outlet—is fundamentally mismatched for the hardest case: low Reynolds numbers below the instability threshold, where the global eigenmode keeps growing downstream forever even after the wake has died out. In that regime, making the box longer does not help, because the mismatch is in the boundary condition, not the domain. The paper's proposed fix is a Robin condition that feeds local linear-stability information about the wake profile at the outlet into the eigenvalue solver; it demonstrates that this condition yields converged eigenvalues and clean pressure eigenmodes in substantially smaller domains, for both cylinder and airfoil wakes. If true, this makes global stability calculations cheaper and more reliable in precisely the regime where older boundary conditions mislead.","feed_headline":"Robin outlet yields converged stability modes in tiny wake domains","feed_subtitle":"Neumann and Dirichlet outlets fail for stable wakes; a physics-based Robin condition closes them in small boxes.","key_machinery":"The central object is the Robin outflow boundary condition built from a first-order Taylor expansion of the local dispersion relation ω(α) ≈ ω0 + c0(α − α0) about a neutral point, where c0 is the complex group velocity of the leading local linear-stability mode at the outlet profile. Discretizing the Fourier relation iα q̂ = ∂q̂/∂x turns this dispersion relation into a linear relation between the eigenfunction and its streamwise derivative at the outlet, which is inserted directly into the matrix-forming global eigenvalue problem. The group velocity is evaluated by a biorthogonality formula involving the direct and adjoint local eigenmodes, so no externally prescribed convection speed is nee","core_discovery":"A wake eigenmode below the critical Reynolds number has negative temporal growth but positive spatial growth far downstream, because the two are linked by the downstream advection speed: near a uniform far-wake profile the dispersion relation gives ω_i ≈ U_c α_i, so a damped global mode must amplify in x. Classical outlet conditions assume the opposite—zero normal derivative, zero value, or zero stress at the outlet—so they either force the mode to stop growing and generate oscillations, or leave it unconstrained and become numerically sensitive. The paper's Robin outlet condition instead imposes a first-order relation between the eigenfunction and its streamwise derivative at the outlet, us","pith_inferences":["A testable corollary the paper leaves implicit: the spatial-amplification argument applies to any localized base flow with a nearly uniform far field, so the low-Reynolds-number failure of Neumann and Dirichlet outlets should be generic for wakes, jets, and mixing layers, not just cylinders and airfoils.","One could use Robin-C as a built-in consistency check: compare the envelope slope of the global eigenfunction with the local linear-stability prediction at every outlet; if they diverge for large x_o, the first-order Taylor description is failing and no boundary condition of this form can claim physical consistency.","A direct extension is to carry the Taylor expansion of the dispersion relation to second order in α, which would capture curvature of ω(α) and might reduce the residual outlet-position dependence seen at small x_o.","The appendix's sensitivity of sponge layers to tuning parameters suggests that the Robin outlet may also be preferable in nonlinear and time-stepping simulations of stable wakes, since it replaces an artificial damping coefficient with a locally computed group velocity."],"forward_implications":["If correct, classical Neumann and Dirichlet outlets should not be trusted at Reynolds numbers below the global-instability threshold; the paper's dispersion-relation argument says no domain length cures them.","The Robin outlet permits stable-wake global modes to be computed in small boxes (outlet at x_o = 10–20 rather than 50–100+), cutting the memory and time cost of matrix-forming eigenvalue problems.","Pressure eigenmode quality becomes a usable diagnostic: Robin produces pressure fields with no striped outlet pollution, a component that most previous global-stability studies did not examine.","The method transfers from cylinder wakes to airfoil wakes and from two-dimensional modes to spanwise-wavenumber (stationary and traveling) modes, so it likely applies across the family of incompressible bluff-body wakes.","Because the condition is cast in the matrix-forming finite-difference eigenvalue framework, the same Robin closure can be dropped into Floquet and compressible extensions, as the paper suggests."],"fun_headline_variants":["Robin outlet condition shrinks wake stability domains","Physics-based outlet boundary speeds up wake stability analysis","Wake modes grow spatially below critical Re—Robin outlet fixes it","Tiny wake domains with Robin outlet for global stability","Robin boundary outperforms sponge and stress-free outlets"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the first-order Taylor expansion of the local dispersion relation around a neutral point, with a constant group velocity, describes the far-wake global mode; the paper does not specify how (ω0, α0) is defined for a damped wake profile, so if the dispersion relation is not captured by this local form, the Robin outlet imposes an artificial structure instead of removing one.","fun_headline_variants_meta":{"raw":{"variants":["Robin outlet condition shrinks wake stability domains","Physics-based outlet boundary speeds up wake stability analysis","Wake modes grow spatially below critical Re—Robin outlet fixes it","Tiny wake domains with Robin outlet for global stability","Robin boundary outperforms sponge and stress-free outlets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000189,"raw_usage":{"total_tokens":1173,"prompt_tokens":743,"completion_tokens":430,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":487,"completion_tokens_details":{"reasoning_tokens":356}},"tokens_in":487,"tokens_out":430,"duration_ms":5819,"temperature":1.0,"reasoning_tokens":356,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T16:45:30.204194+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the leading global mode for the Re = 40 cylinder wake in a box with outlet at x_o = 200 (or use a high-order spectral reference domain) and measure whether the Robin-C prediction matches the true spatial envelope all the way to the outlet; if the local linear-stability group velocity fails to reproduce the measured downstream growth rate, the Taylor assumption is the point that breaks. Equivalently, solve the full spatial Orr-Sommerfeld dispersion relation at x_o = 100 and compare its complex wavenumber with the Taylor estimate; a mismatch comparable to the reported eigenvalue variatio","supporting_citations":[],"review_version":1}