{"id":"18a9c098-8e66-4dbc-9d31-6df2f1690163","arxiv_id":"2502.00625","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Bubble growth in convergent cylindrical Rayleigh-Taylor instability is vorticity-driven at low Atwood and Mach numbers and compression-driven at high values; a three-term model reproduces DNS only when fed DNS-derived terms.","lead":"This paper simulates how bubbles of light gas grow into heavy gas at an unstable interface in cylindrical geometry, with acceleration acting either inward or outward, and maps when growth is driven by swirling vortices versus gas compression. The proposed three-term formula matches the simulations, but two of its terms are computed from those same simulations, so it describes rather than predicts.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The compressibility term v_d in Eq. (3.7) rests on a uniform-piston assumption that is unverified in cylindrical geometry; if the dilatational radial velocity at the bubble tip varies in angle, the model's late-time agreement and phase classification are not secure.","rationale":"The central claim is the three-term additive model (3.8) and its ability to reproduce DNS bubble velocity. All three terms must be sound: v_p comes from a known potential-flow ODE, v_vort has a calibrated efficiency factor, but v_d is derived from a geometric assumption that is unique to the cylindrical generalization. If the angular uniformity of u_r^d at r_b does not hold, Eq. (3.7) is not the correct compressibility velocity, and the model's agreement could be coincidental or artificially compensated by the tuned η. This is more fundamental than the circularity complaint: even a diagnostic decomposition needs each term to be the correct projection of the dynamics. The paper's own Sec. 4 admits that the model is not predictive, so the value of the paper lies in the physical classification and the decomposition; an incorrect v_d would undermine that classification. The proposed test directly measures the disputed assumption and would settle it. The DNS itself, the grid-convergence check, and the n=16 validation are credible and do not remedy the assumption; they only show that the numerics are reliable for the fields used as inputs to the model.","tokens_in":19309,"tokens_out":9247,"duration_ms":90781,"concrete_test":"Perform a Helmholtz decomposition of the DNS velocity snapshots (e.g., A_T=0.9, Ma=0.9, Γt≈6.5) into dilatational and solenoidal parts, and evaluate u_r^d(r_b, φ) along the circle r=r_b. Compute the exact boundary integral in Eq. (3.6) and compare with the piston approximation -2π r_b v_d; also report the angular spread (e.g., standard deviation divided by mean). If the approximation error exceeds 20% or the angular variation is comparable to the mean, the piston assumption fails and Eq. (3.7) should be replaced by the full boundary integral.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The derivation of Eq. (3.7) from Green's formula (3.6) is the load-bearing step for the compressibility contribution. It assumes that the dilatational radial velocity at the bubble-tip radius, u_r^d(r_b, φ), is independent of φ and equals v_d—the 'piston' picture inherited from planar RTI. In cylindrical geometry the interface is curved (mode B=8, r = r0 + η0 cos Bφ), so u_r^d at r_b is expected to vary with φ; the boundary integral ∮(-u_r^d r dφ) is then not -2π r_b v_d unless the angular mean happens to equal v_d. Since v_d is not independently measured but defined through this assumption, the model agreement in Figs. 6–7 and the compressibility-dominated phase classification in Fig. 5 are conditional on it. No angular uniformity check or sensitivity to the choice of r_∞ (boundary versus actual hydrostatic front) is reported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports direct numerical simulations of two-dimensional single-mode compressible Rayleigh–Taylor instability in cylindrical geometry with isothermal stratification, covering Atwood numbers A_T = 0.1–0.9 and Mach numbers Ma = 0.1–0.9. It documents the nonlinear and highly nonlinear bubble dynamics for convergent and divergent accelerations, classifies the late-time response into robust acceleration, transient acceleration, and non-acceleration phases, and proposes a composite model, Eq. (3.8), that adds a vorticity-accumulation term (3.5) and a compressibility term (3.7) to the incompressible potential-flow ODE of Zhao et al. (2020b), with a density-ratio correction at the bubble tip. The model is compared with DNS for mode numbers B = 8 and 16 and reported to reproduce the bubble-velocity evolution from linear to highly nonlinear regimes.","tokens_in":19507,"tokens_out":7838,"duration_ms":68184,"significance":"The DNS dataset and the documented differences between convergent and divergent cylindrical RTI are a useful contribution, and the paper is careful in several respects: a grid-convergence check is performed at the most demanding parameters (A_T = 0.9, Ma = 0.9), the incompressible limit is validated against the nonlinear theory of Zhao et al. (2020b), and the model is exercised for two mode numbers. The proposed additive decomposition, if it can be supported by an independent validation, would provide a valuable physical interpretation of late-time bubble acceleration in convergent geometries. However, the model as presented is not predictive—the authors themselves state this in Section 4—and the validation is partly circular because the added terms are computed from the same DNS.","major_comments":[{"comment":"The central validation claim is weakened by circularity between the model terms and the DNS. The vorticity velocity v_vort in Eq. (3.5) is computed from the DNS-measured mean vorticity and the density ratio at the bubble tip, and the compressibility velocity v_d in Eq. (3.7) is computed from the DNS-measured mean dilatation. Therefore, comparing v_p + v_vort + v_d with the DNS bubble velocity in Figs. 6 and 7 mainly checks the internal consistency of the definitions, not the predictive power of the model. The statement in Section 4 that the model 'requires the simulation and thus is not predictive' is an honest acknowledgment, but it is in tension with the abstract's claim that the model is 'verified by numerical results.' The paper should either provide an independent test (e.g., calibrating eta on a subset of cases and validating on the remaining ones) or explicitly frame the model as a diagnostic decomposition.","section":"Sec. 3.2, Eq. (3.8), Figs. 6–7"},{"comment":"The derivation of v_d from Green's formula assumes that the dilatational radial velocity at the bubble-tip radius is angularly uniform and equal to v_d (the 'piston' picture). In cylindrical geometry with mode B = 8 the interface is curved, and u_r^d at r_b can vary with angle phi; the boundary integral in Eq. (3.6) equals -2 pi r_b v_d only if the angular mean of u_r^d happens to be v_d. No angular-uniformity check is reported, and no sensitivity study on the choice of r_infinity (the computational boundary rather than the actual hydrostatic front) is reported. Since v_d is central to the compressibility-dominated phase classification in Fig. 5, the model agreement shown in Figs. 6 and 7 is conditional on this unverified assumption. I ask the authors to measure the angular distribution of u_r^d at r_b and to test the sensitivity of v_d to r_infinity.","section":"Sec. 3.2, Eq. (3.7)"},{"comment":"The efficiency factor eta = 0.4 in Eq. (3.5) is introduced as an empirical constant 'to account for the attenuation of vortices in cylindrical geometry.' No derivation, prior calibration, or sensitivity analysis is provided. Since Figs. 6 and 7 use the same eta, the B = 16 cases provide a consistency check but not an independent test of this parameter. Please discuss the physical basis for eta = 0.4 and quantify how the model agreement degrades as eta is varied.","section":"Eq. (3.5), Sec. 4"},{"comment":"The phase diagram in Fig. 5 classifies the dominant mechanism in the highly nonlinear stage based on the time-averaged magnitudes of v_vort and v_d, both of which are defined from DNS fields. The finding that 'acceleration in divergent cases is purely induced by flow compressibility' is therefore a statement about the relative sizes of the model terms, not a direct measurement of the physical mechanism. The paper should clarify this distinction, and ideally corroborate the classification with an independent flow diagnostic (e.g., the correlation between local dilatation and bubble acceleration, or the work done by pressure dilatation).","section":"Fig. 5, Sec. 3.1"}],"minor_comments":[{"comment":"The expression for the initial density and pressure is difficult to parse after typesetting; please rewrite it with proper parentheses, for example by explicitly writing the exponent as sgn(a*) Ma^2 (1 +/- A_T) (r - eta(phi)) B/(2 pi).","section":"Sec. 2.2, Eq. (2.6)"},{"comment":"The definition of the averaged vorticity uses 'the volume V inside the bubble'; since the simulations are two-dimensional, please clarify whether this is a surface area and how the bubble interior is delimited.","section":"Sec. 3.2, above Eq. (3.5)"},{"comment":"The text refers to 'figure 3(a,b)' and 'figure 3(c,d)' when describing vortex positions; according to the caption, the convergent cases at Ma = 0.1 are panels (a) and (c), and the divergent cases are panels (b) and (d). Please correct the panel references.","section":"Fig. 3"},{"comment":"The grid-convergence study is reported only for the most demanding parameters (A_T = 0.9, Ma = 0.9). A second convergence test at, for example, A_T = 0.1 and Ma = 0.9 would help establish that the resolution is adequate across the parameter range.","section":"Sec. 2.2, Fig. 1"},{"comment":"The word 'initial' is misspelled as 'intial' in Section 2.2 and Section 3.1; please proofread the manuscript for typographical errors.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of JFM and the DNS effort is solid, but the model's non-predictive and partially circular nature should be addressed before publication. The referee concerns are not about the quality of the simulations but about the strength of the claims made for the model. The methodology closely follows Fu et al. (2023) with the addition of a curvature factor; the novelty is incremental but acceptable if the model is properly qualified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: you should read this for the phase diagram, not for the model. The paper reports well-executed 2D DNS of single-mode compressible cylindrical RTI over A_T=0.1–0.9, Ma=0.1–0.9, and the headline result is a clear classification: in convergent geometry, late-time bubble re-acceleration is vorticity-dominated at low A_T and low Ma, compressibility-dominated at high A_T and high Ma; in divergent geometry there is no vorticity-dominated robust acceleration. That is new and, as far as I can tell, solid. The grid-convergence check at the hardest case and the recovery of Zhao et al.'s incompressible theory at Ma=0.1 give me confidence in the numerics.\n\nThe improved model (3.8) is best understood as a diagnostic decomposition. The authors implicitly know this—they say in Sec. 4 that it requires the simulation and is not predictive. The circularity the reader flagged is real: v_vort and v_d are computed from the same DNS that the model is judged against, and eta=0.4 is a tuned constant. The stress-test's piston-assumption concern also lands: Eq. (3.7) assumes the dilatational radial velocity at the bubble tip is angularly uniform, which is unlikely to be exactly true for a curved B=8 interface, and no check or sensitivity to r_infty is reported. None of this destroys the paper's qualitative value, but it should be presented more honestly as curve-fitting with physically motivated terms rather than verification.\n\nIf it were my call, I'd send it to review and ask for two things: a sensitivity study on eta and r_infty, and an angular-uniformity diagnostic for the dilatational velocity near the bubble tip. I'd also ask them to soften the 'well reproduces' language. The data are useful enough that the paper deserves a serious referee.","headline":"Careful DNS gives a credible phase diagram for bubble acceleration in cylindrical RTI; the additive model is diagnostic, not predictive, but the paper deserves refereeing.","tokens_in":20018,"tokens_out":2371,"would_cite":true,"duration_ms":24996,"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":"The paper proposes an additive three-part velocity model for bubble growth in compressible cylindrical Rayleigh-Taylor instability and verifies it against direct numerical simulation across Atwood and Mach numbers.","keywords":["Rayleigh-Taylor instability","cylindrical geometry","compressible flow","isothermal stratification","bubble acceleration","vorticity accumulation","flow compressibility","direct numerical simulation"],"falsifier":"Measure the angular dependence of the radial velocity along the heavy-fluid front immediately ahead of the bubble tip in DNS at $A_T=0.9$ and $Ma=0.9$; if $u_r$ varies substantially with angle, the uniform-piston assumption behind $v_d$ is violated and the model's compressibility term would need revision.","tokens_in":19035,"feed_emoji":"🌊","tokens_out":10976,"duration_ms":92889,"temperature":0.7,"pith_summary":"The paper claims that the nonlinear velocity of a single-mode bubble in compressible, isothermally stratified Rayleigh-Taylor instability in cylindrical geometry is the sum of three contributions: a potential-flow velocity, a vorticity-accumulation velocity, and a compressibility velocity. It shows by direct numerical simulation that in convergent geometry the late-time bubble acceleration is dominated by vorticity accumulation at low Atwood and Mach numbers and by flow compressibility at high Atwood and Mach numbers, while in divergent geometry only flow compressibility produces robust acceleration. The improved model, which reads the simulated vorticity and dilatation fields, reproduces bubble-velocity histories from the linear through the highly nonlinear regime for Atwood numbers 0.1-0.9 and Mach numbers 0.1-0.9, for perturbation modes 8 and 16. If correct, the result gives a mechanistic bookkeeping for late-time cylindrical Rayleigh-Taylor instability and shows that planar-based models can misjudge convergent geometry.","feed_headline":"Three additive terms capture cylindrical Rayleigh-Taylor bubble growth","feed_subtitle":"Model matches DNS for Atwood 0.1-0.9 and Mach 0.1-0.9, and separates convergent from divergent dynamics.","key_machinery":"The carrying objects are the three velocity contributions in Eq. (3.8) and the diagnostic fields behind them: the potential-flow amplitude ODE of the incompressible cylindrical theory with a local Atwood number, the bubble-averaged vorticity $\\bar{\\omega}$ with a density-ratio correction and efficiency factor, and the mean dilatation $\\bar{\\theta}$ in the annulus between the bubble tip and the undisturbed front. The compressibility velocity is derived from a Green's-formula identity relating the line integral of the dilatational velocity to the area integral of $\\nabla\\cdot\\mathbf{u}$, and the factor $(1+r_\\infty/r_b)/2$ encodes the interfacial curvature, recovering the planar formula when $r_b\\to r_\\infty$. The model is not predictive from initial conditions alone; it reads the vorticity, the dilatation, and the front position from the simulation.","core_discovery":"The central claim is that the bubble velocity in two-dimensional single-mode cylindrical Rayleigh-Taylor instability with isothermal stratification can be modelled as $v_b = v_p + v_{\\mathrm{vort}} + v_d$ (Eq. 3.8), where $v_p$ solves the incompressible cylindrical potential-flow ODE with the local Atwood number evaluated at the bubble tip, $v_{\\mathrm{vort}} = \\eta\\, r_\\rho\\, \\bar{\\omega} / (2n/r_b)$ accounts for accumulated vorticity inside the bubble, and $v_d = -\\bar{\\theta}\\, (r_\\infty - r_b)\\,(1 + r_\\infty/r_b)/2$ accounts for the compression of the heavy fluid ahead of the bubble through the mean dilatation. The paper verifies this additive decomposition against DNS for convergent and divergent accelerations across $A_T=0.1$-$0.9$ and $Ma=0.1$-$0.9$. In convergent cases the highly nonlinear acceleration is dominated by vorticity accumulation at low $A_T$ and low $Ma$ and by flow compressibility at high $A_T$ and high $Ma$; in divergent cases robust acceleration is purely due to flow compressibility. Density stratification acts to suppress acceleration at low $A_T$ and high $Ma$.","pith_inferences":["Beyond the paper: the curvature factor $(1+r_\\infty/r_b)/2$ implies that the compressibility contribution should strengthen as the convergent bubble approaches the centre and weaken as a divergent bubble moves outward; this predicted scaling could be tested directly from the DNS data by plotting $v_d$ against $1/r_b$.","Beyond the paper: the vorticity efficiency factor $\\eta=0.4$ differs from the planar value 0.45, suggesting that vortex attenuation at the bubble tip depends on geometry; a systematic sweep over mode number, Reynolds number, and initial amplitude could calibrate this factor and reveal whether it is universal.","Beyond the paper: if the same additive decomposition carries over to spherical geometry with an appropriately modified curvature factor, it would offer a route to modelling supernova and inertial-confinement-fusion mixing layers without fully three-dimensional simulation."],"forward_implications":["Convergent and divergent bubbles follow different acceleration mechanisms: vorticity accumulation drives robust acceleration in convergent cases at low Atwood and low Mach numbers, while it does not produce robust acceleration in divergent cases.","At high Atwood and high Mach numbers, flow compressibility becomes the dominant destabilizing effect and causes transient acceleration in divergent cases that is absent from the incompressible potential-flow description.","The phase diagram in the Atwood-Mach plane for cylindrical Rayleigh-Taylor instability differs from the planar stratified case, so planar-based late-time models should not be applied directly to cylindrical or spherical configurations.","The model reproduces DNS bubble-velocity histories for both $n=8$ and $n=16$ perturbation modes, supporting the additive decomposition as a general characterization for single-mode cylindrical compressible Rayleigh-Taylor instability.","Because the model requires simulated vorticity and dilatation fields, it is a diagnostic decomposition rather than a self-contained predictive model."],"supporting_citations":[{"why":"Supplies the incompressible cylindrical potential-flow nonlinear ODE whose solution gives the baseline bubble velocity $v_p$ that the improved model extends.","marker":"Zhao et al. (2020b)"},{"why":"Introduces the vorticity-accumulation mechanism and the bubble re-acceleration model on which $v_{\\mathrm{vort}}$ is built.","marker":"Betti & Sanz 2006"},{"why":"Adds the efficiency factor and the vorticity-averaging procedure for late-time single-mode Rayleigh-Taylor instability used in the cylindrical vorticity term.","marker":"Bian et al. 2020"},{"why":"Establishes the planar compressible re-acceleration model and the Green's-formula relation between dilatation and velocity from which $v_d$ is adapted.","marker":"Fu et al. 2023"},{"why":"Provides the potential-flow saturation velocity used as the nonlinear-regime reference and the basis for defining the local Atwood number.","marker":"Goncharov 2002"},{"why":"Establishes the linear stabilizing effect of density stratification that the present stratified cylindrical setup builds on.","marker":"Livescu 2004"}],"fun_headline_variants":["Three physical terms predict cylindrical Rayleigh-Taylor bubbles","Additive model matches RTI bubble growth across Atwood and Mach ranges","Cylindrical RTI bubble speed: potential plus vorticity plus dilatation","Convergent and divergent RTI bubbles share one three-term model","DNS-verified three-term law for cylindrical Rayleigh-Taylor bubbles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model's compressibility term assumes that a rising bubble of light fluid acts like a piston that uniformly compresses the heavy fluid in front of it, so the radial velocity at the bubble tip is uniform along the angular direction and equals $v_d$; if this piston picture fails, the compressibility contribution and the conclusion that flow compressibility dominates at high Atwood and Mach numbers lose their basis.","fun_headline_variants_meta":{"raw":{"variants":["Three physical terms predict cylindrical Rayleigh-Taylor bubbles","Additive model matches RTI bubble growth across Atwood and Mach ranges","Cylindrical RTI bubble speed: potential plus vorticity plus dilatation","Convergent and divergent RTI bubbles share one three-term model","DNS-verified three-term law for cylindrical Rayleigh-Taylor bubbles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000347,"raw_usage":{"total_tokens":1969,"prompt_tokens":1085,"completion_tokens":884,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":794}},"tokens_in":701,"tokens_out":884,"duration_ms":9028,"temperature":1.0,"reasoning_tokens":794,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T18:16:37.308285+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the angular dependence of the radial velocity along the heavy-fluid front immediately ahead of the bubble tip in DNS at $A_T=0.9$ and $Ma=0.9$; if $u_r$ varies substantially with angle, the uniform-piston assumption behind $v_d$ is violated and the model's compressibility term would need revision.","supporting_citations":[{"cited_title":"& Sanz, J","cited_arxiv_id":null,"evidence_quote":"Introduces the vorticity-accumulation mechanism and the bubble re-acceleration model on which $v_{\\mathrm{vort}}$ is built."},{"cited_title":", Aluie, H","cited_arxiv_id":null,"evidence_quote":"Adds the efficiency factor and the vorticity-averaging procedure for late-time single-mode Rayleigh-Taylor instability used in the cylindrical vorticity term."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the potential-flow saturation velocity used as the nonlinear-regime reference and the basis for defining the local Atwood number."},{"cited_title":"2004 Compressibility effects on the Rayleigh--Taylor instability growth between immiscible fluids","cited_arxiv_id":null,"evidence_quote":"Establishes the linear stabilizing effect of density stratification that the present stratified cylindrical setup builds on."}],"review_version":1}