{"id":"7fbb0509-1c98-4771-8e8e-0355582081ab","arxiv_id":"2501.02095","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"OTD mode analysis of vortex gust-airfoil interactions shows the most amplified perturbation regions moving from the leading edge to forming vortices and into the wake, with strong gusts concentrating growth on shed vortex cores.","lead":"This paper applies optimally time-dependent mode analysis to a NACA 0012 airfoil struck by vortex gusts of different strengths, showing where small disturbances grow as the flow evolves. The findings map how the most sensitive regions shift from the leading edge to shed vortices, which could guide flow control during gust encounters.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The OTD amplification measure uses an unweighted Euclidean norm on conservative variables, mixing units and ignoring cell volumes, so the reported 'most amplified region' may reflect mesh density rather than physics.","rationale":"Reader's weakest assumption points to the L2 norm on conservative variables. I agree partially: the units-mixing is real, but I sharpen it: the norm is also an unweighted sum over discrete grid points, which in an unstructured finite-volume mesh is not the discrete L2 norm—cell volumes are omitted. This makes the location claim potentially mesh-dependent. This is more load-bearing than the boundary-condition issue because the headline claim is about where (spatial region) perturbations amplify. The boundary conditions, while underspecified, likely affect quantitative g_i less directly; the norm determines the very geometry of the OTD subspace. The paper includes convergence checks on time step and number of modes but no check on inner product weighting. A volume-weighted/Chu-norm recomputation is feasible with modest code changes. My verdict remains CONDITIONAL: the qualitative physics might survive, but the current evidence does not rule out a norm artifact.","tokens_in":18874,"tokens_out":5688,"duration_ms":56761,"concrete_test":"Rerun the OTD evolution for G=0.5 with a weighted inner product: replace the Euclidean projector in Eq. 2.12 by I - U_r (U_r^T W U_r)^{-1} U_r^T W, where W is a diagonal matrix whose entries are the local cell volume times a physical energy weight (e.g., Chu norm weights for primitive variables, or at minimum the cell volume alone). Recompute the leading OTD mode's vorticity fields at tau = -0.5, 0, 0.5, 1 and g_1(t). If the leading mode remains concentrated in the LEV and trailing-edge wake with the same transition timing, the norm sensitivity is benign; if its spatial support or ranking changes, the paper's central claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2 defines perturbations in the conservative state vector [rho, rho u, rho v, rho w, e] and achieves orthonormality of OTD modes in the Euclidean L2 norm, with amplification g_i = ||q'_*i||_2 / ||q'_0i||_2 (Eqs. 2.19-2.21). Two properties of this norm are load-bearing for the central claim that OTD modes locate the most amplified region. First, the Euclidean norm over state variables mixes heterogeneous physical units (density, momentum, energy), so the dominant mode need not correspond to the largest physical perturbation energy. Second, and more concretely, the discrete norm treats every grid point equally; CharLES uses an unstructured mesh with cell volumes varying by orders of magnitude (fine near the airfoil/wake, coarse in the far field). The paper's 'most amplified region' could therefore be biased toward the finest-mesh regions, and the reported transition from the leading-edge vortex sheet to the wake for G=0.5 might be a grid-resolution artifact rather than an intrinsic instability property. The paper does not test sensitivity of the OTD modes to the inner product or to volume weighting. Since the central descriptive claim is about spatial location of maximal amplification, this norm choice is exactly what the claim rests on.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript applies optimally time-dependent (OTD) mode decomposition to two-dimensional compressible DNS of a NACA 0012 airfoil at Re = 400, angle of attack 12°, and M = 0.1 interacting with a Taylor vortex of gust ratio G = -1, -0.5, 0.5, or 1. The unsteady vortex-airfoil flow is taken as the time-dependent base state, and OTD modes are evolved under the linearized Navier-Stokes operator extracted from the DNS. The leading OTD modes, singular values, and the quantities g_i are used to identify spatiotemporal regions of maximum perturbation amplification, with qualitative transitions reported (e.g., for G = 0.5 the most amplified region moves from the leading-edge vortex sheet to the forming leading-edge vortex and later to the wake). Appendix A provides convergence checks in time-step size, number of OTD modes, initial evolution time, and random initial conditions. Appendix B discusses the most amplified initial perturbations.","tokens_in":19132,"tokens_out":4945,"duration_ms":52107,"significance":"If the norm and definitional issues are resolved, the paper is a useful demonstration of OTD analysis for strongly unsteady aerodynamic gust encounters. The study has clear strengths: it uses standard OTD evolution equations, contains no fitted parameters, validates the DNS against previous lift data, and ships explicit convergence checks in Appendix A (temporal convergence, mode-number convergence, initial-time insensitivity, and random-initial-condition tests). It also makes falsifiable predictions about where perturbation growth is largest during vortex-airfoil interaction, which could guide time-varying flow-control experiments or simulations. The main caveat is that the quantitative amplification measure and, potentially, the spatial ranking of modes depend on the chosen inner product, which is not physically motivated or tested in the present manuscript.","major_comments":[{"comment":"The definition of g_i is algebraically inconsistent. The first equality in Eq. (2.19) defines g_1 as the ratio of L2 norms, ||q'_*1||_2 / ||q'_01||_2, which equals sigma_1 / ||q'_01||_2 because ||u_1||_2 = 1. The second equality gives sigma_1^2 / ||q'_01||_2^2. These cannot both be true. If g_i is intended as an energy amplification (a ratio of squared norms), the first expression and the surrounding wording should be adjusted; if it is an amplitude ratio, the right-hand side is wrong. Since Figures 5-8 report g_i as 'energy amplifications', every quantitative growth statement in Section 4 depends on correcting this definition.","section":"Section 2, Eqs. (2.19)-(2.21)"},{"comment":"The amplification measure is the unweighted Euclidean L2 norm of the conservative state vector [rho, rho u, rho v, rho w, e], and OTD orthonormality is imposed in this same inner product. This choice is load-bearing for the central spatial claim, for two reasons. First, the conservative variables mix different physical units, so without a nondimensionalization and a physical energy norm (e.g., a Chu or kinetic-energy norm) the dominant mode can be controlled by a particular component rather than by the physical perturbation energy. Second, the discrete Euclidean norm treats every grid point equally, while the CharLES mesh is unstructured with cell volumes varying by orders of magnitude near the airfoil and wake; a mode concentrated in a finely resolved region can have a large norm simply because it contains many degrees of freedom. The claimed transition of the most amplified region for G = 0.5, and the identification of vortex cores for G = ±1, could therefore be affected by mesh resolution rather than by intrinsic amplification physics. I request a sensitivity test: recompute the leading OTD modes and g_i with a volume-weighted L2 inner product, or with a nondimensionalized energy norm, and report whether the qualitative locations in Figures 5-8 persist.","section":"Section 3.1 and Section 2, Eq. (2.2)"}],"minor_comments":[{"comment":"The phrase 'the the case' should read 'the case'.","section":"Section 4, first paragraph"},{"comment":"The sentence 'This suggests that a strong vortex-airfoil interaction' is incomplete and should be finished or rewritten.","section":"Appendix B, final paragraph"},{"comment":"The Küssner reference is displayed as 'K ¥ussner' in the Introduction and 'K¨ussner' in the bibliography; the Rössler system is displayed as 'R¨ossler'. These encoding artifacts should be corrected.","section":"Introduction and References"},{"comment":"The statement that changing the OTD domain size 'does not affect the OTD modes and their coefficients' is asserted without a demonstration; either add a brief convergence check or soften the claim.","section":"Section 3.1"},{"comment":"The Frobenius norm is rendered with a nonstandard symbol in Eq. (2.8); the notation should be defined or replaced with a conventional \\|\\cdot\\|_F.","section":"Section 2, Eq. (2.8)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid, careful application of OTD mode analysis to a problem that needs it, and the main new observations are probably right. But the paper's headline—where perturbations amplify most—rests on an unweighted L2 norm of the conservative state vector, and that choice is not defended. The paper needs a sensitivity check before I'd trust the spatial ranking.\n\nWhat's new: OTD has been applied to pitching airfoils, Blasius boundary layer, and pulsating Poiseuille flow. This is the first OTD analysis of vortex gust-airfoil interactions at gust ratios above 0.5, where linear thin-airfoil models break down. For G=0.5 the leading mode moves from the leading-edge vortex sheet to the LEV, then to the wake; for G=±1 the modes track the shed vortex cores. Those are useful, concrete descriptions. The authors also did the verification work: temporal convergence of singular values, convergence with mode count, insensitivity to initial time, and a random-initial-condition test. That is real evidence the numerics are sound.\n\nSoft spots: The amplification measure g_i is the Euclidean L2 norm of [rho, rho u, rho v, rho w, e] with no cell-volume weighting. On an unstructured mesh whose cells vary by orders of magnitude, that norm is not a physical energy and can bias 'most amplified region' toward fine-mesh areas. Modes are orthonormal in this same inner product, so the ranking itself could change under a volume-weighted or Chu norm. The paper doesn't test this. I don't think it breaks the qualitative story, but it is the exact thing the central claim depends on, so it needs a response. The linear operator boundary conditions are described briefly but not in enough detail to reproduce; no code or data are offered. Appendix B has a truncated sentence ('This suggests that a strong vortex-airfoil interaction The normalized ...'), which looks like a copyediting slip.\n\nFor the right reader—unsteady aero, modal analysis, flow control—this is worth reading. I'd send it to review, but ask for a norm-sensitivity test and a clearer statement of the linear operator setup.","headline":"A credible OTD application to a tough gust-airfoil problem; the main caveat is an unweighted L2 norm that may bias the 'most amplified region' toward fine-mesh regions.","tokens_in":19657,"tokens_out":2831,"would_cite":false,"duration_ms":29335,"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":"For a vortex gust hitting an airfoil, the region where perturbations amplify most tracks the forming leading-edge vortex and then shifts to the wake; when the gust is strong, it rides on the shed vortex cores.","keywords":["optimally time-dependent modes","vortex gust","airfoil interaction","transient perturbation amplification","unsteady base flow","leading-edge vortex","NACA 0012","flow control"],"falsifier":"Recompute the leading OTD mode and $g_1(t)$ for the $G=0.5$ case using a kinetic-energy or Chu norm instead of the conservative-state $L^2$ norm, and compare the time at which the amplification hotspot shifts from the leading-edge vortex sheet to the forming leading-edge vortex. If the hotspot location or transition time moves substantially, the structural conclusions depend on the choice of norm. Alternatively, in a forced simulation, inject perturbations only in the predicted OTD hotspot at each time; if that localized forcing does not produce the largest response relative to equal-energy forcing elsewhere, the OTD ranking is not predictive.","tokens_in":18683,"feed_emoji":"🌪️","tokens_out":7780,"duration_ms":67602,"temperature":0.7,"pith_summary":"This paper asks where and when small perturbations can grow most during a strong vortex-gust encounter with an airfoil, when the base flow itself is changing rapidly. Using optimally time-dependent (OTD) modes, it tracks the leading perturbation-amplification structures in four vortex-airfoil interactions with gust ratios $G = -1, -0.5, 0.5$, and $1$ at a $12^\\circ$ angle of attack. The central finding is that the most amplified region is not stationary: it follows the forming leading-edge vortex, then shifts to the wake or to the cores of shed vortex pairs depending on gust strength and sign. This matters because it gives a time-resolved map of where the flow is most receptive to perturbations, which can indicate where and when to apply flow control on an unsteady wing.","feed_headline":"Tracking where a vortex gust amplifies airfoil perturbations most","feed_subtitle":"Time-resolved OTD modes show the most amplified region sliding from the leading edge to the wake as the gust hits.","key_machinery":"The engine is the OTD low-rank approximation $Q'(t) \\approx U_r(t)Y_r(t)^T$, where $U_r$ holds time-dependent orthonormal modes and $Y_r$ holds their coefficients. The modes and coefficients evolve by projecting the instantaneous linearized Navier-Stokes operator $L(t)$ onto the subspace, $dU_r/dt = LU_r - U_r(U_r^T L U_r)$ and $dY_r^T/dt = (U_r^T L U_r)Y_r^T$, with the skew-symmetric gauge freedom set to zero. A rotation based on the singular value decomposition of the correlation matrix ranks the modes by singular value so that the first mode is the most amplified direction at each instant. The key diagnostic is the energy amplification $g_i(t) = \\sigma_i^2(t)/\\|q'_{0i}(t)\\|^2$, giving the maximum factor by which a perturbation can grow by time $t$ relative to its initial amplitude.","core_discovery":"The study reports that for a moderate positive vortex gust ($G=0.5$) the leading OTD mode first marks the leading-edge vortex sheet, then the core of the developing leading-edge vortex, and finally the wake behind the trailing edge, with higher-order modes adding a secondary sensitive region in the trailing-edge wake. For strong gusts ($G=\\pm 1$) the dominant amplified structures coincide with the cores of the shed vortex pair, and the leading energy amplification $g_1$ grows roughly monotonically, indicating a persistent instability mechanism tied to high-vorticity shedding. For the moderate negative gust ($G=-0.5$) the leading mode evolves gradually with the deformed wake, while a secondary mode spikes in amplification when the pressure-side vortex sheet begins to roll up. The paper interprets these results as evidence that OTD analysis can expose the spatiotemporal receptivity of highly unsteady vortex-airfoil flows, including where and when perturbations are amplified relative to their initial amplitude.","pith_inferences":["The quantitative amplification rankings are computed in the conservative-state $L^2$ norm; translating them into a kinetic-energy or Chu norm would likely change the $g_i$ values and could reorder subdominant modes, so the reported magnitudes should be read as norm-dependent.","The appendix finding that concentrated upstream perturbations on the vortex core amplify more than sparsely distributed ones for $G=\\pm 1$ suggests a testable experiment: seed the flow with compact versus distributed perturbations of equal energy and compare their downstream amplification.","The same OTD machinery could be used to build a reduced-order observer for real-time gust-load prediction, since the subspace tracks the instantaneous danger directions rather than time-averaged ones."],"forward_implications":["For $G=0.5$, flow control should act at the leading edge while the gust approaches, then move to the trailing-edge wake once the leading-edge vortex detaches.","For strong gusts ($G=\\pm 1$), actuators aimed at the cores of the shed vortex pair would target the region of maximum perturbation amplification.","Higher-order OTD modes reveal secondary receptive regions, such as the trailing-edge vortex sheet in the $G=0.5$ case and the pressure-side roll-up in the $G=-0.5$ case, so controlling only the leading mode may miss important dynamics.","The monotone growth of $g_1$ for strong gusts signals a persistent instability, whereas rises and falls for moderate gusts indicate transient instabilities that deposit energy into the wake.","The same OTD pipeline applies to other unsteady base flows, such as pitching airfoils or separated wakes, where time-invariant stability analysis is not valid."],"supporting_citations":[{"why":"Introduces the OTD mode decomposition and the minimization principle from which the evolution equations are derived.","marker":"Babaee & Sapsis 2016"},{"why":"Shows OTD modes converge to the dominant finite-time Lyapunov vectors, grounding the method's interpretation.","marker":"Babaee et al. 2017"},{"why":"Supplies the variational formulation and residual-minimization view of the OTD evolution equations used here.","marker":"Donello et al. 2022"},{"why":"Demonstrates OTD-based flow control strategies, which motivate the paper's control implications.","marker":"Blanchard & Sapsis 2019b"},{"why":"Defines the gust ratio and maps the parameter space for gust encounters used to choose the studied cases.","marker":"Jones 2020"},{"why":"Provides the compressible Taylor vortex model used to generate the gust.","marker":"Taylor 1918"},{"why":"Underlies the CharLES solver used for the direct numerical simulations of the base flows.","marker":"Khalighi et al. 2011a,b"},{"why":"Describes the unstructured CharLES solver employed for the compressible flow DNS.","marker":"Bres et al. 2017"},{"why":"Earlier vortex-impinged airfoil study that established the flow configuration and baseline wake data.","marker":"Zhong et al. 2023"}],"fun_headline_variants":["Vortex gust airfoil: amplified perturbation region shifts from leading edge to wake","Where vortex gusts most amplify airfoil perturbations: leading edge to wake","Time-resolved OTD modes reveal shifting amplification zones in vortex-gust airfoil flow","Amplified perturbation sites migrate from leading edge to wake in vortex gusts"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that amplification should be measured by the Euclidean $L^2$ norm of the conservative state vector $[\\rho, \\rho u, \\rho v, \\rho w, e]$, so perturbations in density, momentum components, and total energy are combined with equal weight despite having different physical units; if a different physically motivated norm were used, the ranking of which regions are 'most amplified' and the growth values could change.","fun_headline_variants_meta":{"raw":{"variants":["Vortex gust airfoil: amplified perturbation region shifts from leading edge to wake","Where vortex gusts most amplify airfoil perturbations: leading edge to wake","Time-resolved OTD modes reveal shifting amplification zones in vortex-gust airfoil flow","Amplified perturbation sites migrate from leading edge to wake in vortex gusts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000637,"raw_usage":{"total_tokens":2983,"prompt_tokens":1042,"completion_tokens":1941,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":1856}},"tokens_in":658,"tokens_out":1941,"duration_ms":11331,"temperature":1.0,"reasoning_tokens":1856,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:14:07.161142+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the leading OTD mode and $g_1(t)$ for the $G=0.5$ case using a kinetic-energy or Chu norm instead of the conservative-state $L^2$ norm, and compare the time at which the amplification hotspot shifts from the leading-edge vortex sheet to the forming leading-edge vortex. If the hotspot location or transition time moves substantially, the structural conclusions depend on the choice of norm. Alternatively, in a forced simulation, inject perturbations only in the predicted OTD hotspot at each time; if that localized forcing does not produce the largest response relative to equal-energy forcing elsewhere, the OTD ranking is not predictive.","supporting_citations":[{"cited_title":"& Sapsis, T","cited_arxiv_id":null,"evidence_quote":"Introduces the OTD mode decomposition and the minimization principle from which the evolution equations are derived."},{"cited_title":", Farazmand, M","cited_arxiv_id":null,"evidence_quote":"Shows OTD modes converge to the dominant finite-time Lyapunov vectors, grounding the method's interpretation."},{"cited_title":", Carpenter, M","cited_arxiv_id":null,"evidence_quote":"Supplies the variational formulation and residual-minimization view of the OTD evolution equations used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the gust ratio and maps the parameter space for gust encounters used to choose the studied cases."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the compressible Taylor vortex model used to generate the gust."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the unstructured CharLES solver employed for the compressible flow DNS."},{"cited_title":", Fukami, K","cited_arxiv_id":null,"evidence_quote":"Earlier vortex-impinged airfoil study that established the flow configuration and baseline wake data."}],"review_version":1}