{"id":"48d450b5-a89d-41ce-bdf8-d3c4d0784777","arxiv_id":"2411.14650","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper gives a convergence proof for gradient discretisation methods applied to the time-dependent Navier-Stokes equations coupled with the heat equation, with temperature-dependent viscosity.","lead":"This paper proves that a broad family of numerical grid methods converges to a true solution of the equations for hot, flowing fluids where the viscosity depends on temperature. It matters because it supplies theoretical backing for computer simulations of cooling, heating, and industrial flows.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4 passes through the viscosity term using V(S_m)→V(S) with only L∞ regularity on V; this requires a continuity/composition assumption that (2.3) does not provide, leaving the central convergence claim unsupported as stated.","rationale":"The reader's verdict is CONDITIONAL, and my analysis supports it. The most load-bearing gap is V regularity because it is the only place where the temperature coupling interacts with the velocity equation's principal part; if not repaired, the theorem cannot be accepted as stated. I do not select the pressure bound as primary because, although also missing, it is a standard inf-sup consequence and less tied to the paper's novelty. The numerical section cannot rescue the theorem: it uses V(S)=√(S^2+1)+2, which is continuous, and the manufactured solution does not satisfy the homogeneous boundary conditions of (1.1). Therefore the requested fix is to add a continuity/composition condition on V and a pressure estimate. Since these repairs are plausible and localized, the appropriate status remains CONDITIONAL, not REJECT.","tokens_in":10125,"tokens_out":4972,"duration_ms":52615,"concrete_test":"Construct V(ξ)=1+1_{ξ≥0} (bounded, a1=1, a2=2) and a sequence S_m→0 strongly in L2 but taking values ±1 on alternating sets with measure tending to 1/2 (e.g. S_m=1_{A_m}-1_{A_m^c}). Then V(S_m) converges weak-* to 3/2, not to V(0)=1. Re-run the proof of Theorem 3.4, specifically the viscosity term, with this V: verify that no available argument replaces V(S_m) by V(0); then add V∈C(R) and confirm the dominated-convergence step becomes valid. If continuity is required, Theorem 3.4 must be amended.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central step is the passage to the limit in Theorem 3.4 of the term ∫ V(eΠ_{D_m} S_m) ∇_{D_m} u_m : ∇_{D_m} w_m (arising in (3.10)). The proof states that \"the assumptions inforced on V enable us to apply the dominated convergence theorem\", but Assumptions (2.3) only say V ∈ L∞(R) and a1 ≤ V ≤ a2. Strong L2 convergence of eΠ_{D_m} S_m to S (from [11, Thm 4.14]) does not imply V(eΠ_{D_m} S_m) → V(S) in any norm unless V is continuous or some composition property holds. With V discontinuous, V(S_m) may converge weak-* to a different function, and the product with the only weakly convergent ∇_{D_m} u_m cannot be controlled. Thus the proof of Theorem 3.4 is incomplete: an L∞ bound is enough for weak-* compactness, not for the strong convergence needed at the product. The theorem requires an additional assumption such as V ∈ C(R) (bounded), or an analogous composition/monotonicity condition. This is a proof gap, not a disagreement with consensus; the likely fix is small.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the gradient discretisation method (GDM) to the evolutionary incompressible Navier-Stokes system coupled with a heat equation through a temperature-dependent viscosity V(S). After defining a space-time gradient discretisation and a fully implicit scheme (2.6), the author states coercivity, consistency, limit-conformity, trilinear limit-conformity and compactness assumptions, proves discrete energy estimates (Lemmas 3.1 and 3.3), and claims in Theorem 3.4 that any sequence of discrete solutions associated with a suitable sequence of discretisations converges, up to a subsequence, to a weak solution of the continuous problem (2.4), with strong L2 convergence of reconstructed velocity and temperature and weak convergence of gradients and pressure. Numerical experiments with hybrid finite volume discretisations on triangular and distorted meshes report first-order convergence rates.","tokens_in":10282,"tokens_out":8182,"duration_ms":80476,"significance":"If the proof gaps are repaired, the result is a useful generic framework: it covers families of polytopal schemes for a non-isothermal Navier-Stokes model and avoids artificial smoothness assumptions beyond boundedness of V. The numerical section provides concrete tests, although no code is shipped, and the paper does not contain machine-checked proofs. The extension over the stationary analysis in [4] and the abstract GDM machinery in [11] is incremental but within the normal scope of a numerical analysis journal. However, the claimed generality of Theorem 3.4 currently rests on an unstated continuity/composition assumption on V and on an unproved pressure estimate.","major_comments":[{"comment":"The passage to the limit in the viscosity term is not justified. Assumption (2.3) only gives V in L∞(R) with a1 <= V <= a2. Strong L2 convergence of ePi_{D_m} S_m to S does not imply strong, or even almost-everywhere, convergence of V(ePi_{D_m} S_m) to V(S), and the discrete gradients ∇_{D_m} u_m are only weakly convergent, so the product cannot be controlled. The proof needs an additional assumption such as V continuous and bounded, or a monotonicity/composition condition ensuring V(ePi_{D_m} S_m) -> V(S) in some L^q space, before the dominated convergence argument in the proof of Theorem 3.4 can be applied. Without such an assumption, the central convergence claim is unsupported.","section":"Theorem 3.4, Eq. (3.10)"},{"comment":"The weak convergence χ_{D_m} p_m -> p̄ in L2(Ω×(0,T)) is asserted but never proved. No L2 bound on the reconstructed pressure is derived from the scheme; Lemmas 3.1 and 3.3 only give estimates for velocity and temperature. One needs to use the discrete inf-sup constant B_D from (2.5) together with the momentum equation (2.6a) to bound χ_{D_m} p_m in L2(0,T;L2(Ω)) before one can pass to the limit in the pressure term of (3.10). This is a load-bearing gap in Theorem 3.4.","section":"Theorem 3.4, third bullet and proof"},{"comment":"The algebraic inequality (a−b)·a >= 1/2(|a|²+|b|²) is false; the correct inequality is (a−b)·a >= 1/2(|a|²−|b|²). The subsequent displayed estimate in the proof uses the correct form, so the energy estimate is repairable, but the written statement must be corrected for the proof to be valid.","section":"Lemma 3.1, proof"}],"minor_comments":[{"comment":"Lemma 3.1 states that (u,p,S) is a solution to the discrete problem (2.4); it should refer to the gradient scheme (2.6).","section":"Lemma 3.1"},{"comment":"In Definition 2.3(2), the variable v is used twice ('for all v ∈ L²(Ω)^d and for all v ∈ L²(Ω)'); the second occurrence should be a scalar test function, for example r or φ.","section":"Definition 2.3"},{"comment":"In the weak formulation (2.4a), the pressure term −∫_Ω (div v)(x) p̄(x,t) dx is missing the dt inside the time integral.","section":"Eq. (2.4a)"},{"comment":"The proof claims that [11, Theorem 4.14] gives strong convergence in L∞(0,T;L2(Ω)); the usual GDM compactness result gives strong convergence in L2 and weak-* convergence in L∞. The statement of Theorem 3.4 only requires strong L2 convergence, so this should be corrected.","section":"Theorem 3.4, proof"},{"comment":"The numerical test uses V(S)=√(S²+1)+2, which is not in L∞(R) as required by Assumption (2.3). The test should use a bounded truncation or explicitly state that the unbounded growth is never attained on the range of the discrete temperature.","section":"Section 4"},{"comment":"There are numerous typographical errors ('inforced', 'trail functions', 'm =∈ {0,...,N}', 'Defintion') that should be corrected in a revision.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is an incremental extension of the author's earlier stationary GDM analysis [4], relying heavily on the abstract framework of [11] and [12]. This is not a disqualifying feature for a numerical analysis journal. The main theorem can likely be repaired with a continuity assumption on V and a pressure bound, but the current text contains a false algebraic inequality and two missing justifications in the central convergence proof, so it should not be accepted without changes. I have no concerns about circularity or hidden fitted parameters."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the extension of the gradient discretisation method to the transient coupled Navier-Stokes/heat system with temperature-dependent viscosity. The discrete framework is carefully assembled from the author's stationary paper and the standard GDM machinery, and the energy estimates in Lemma 3.1 are the right route. If the main theorem were fully supported, it would be a useful, if not spectacular, addition for GDM practitioners.\n\nThe trouble is that Theorem 3.4 is not fully supported as written. The passage to the limit in the viscosity term uses dominated convergence with only V in L^infinity. Strong L2 convergence of the reconstructed temperature does not imply V(S_m) -> V(S) unless V is continuous or has some composition property. That is a load-bearing gap, not a cosmetic one. The fix is likely small — assume V bounded and continuous, or Lipschitz — but without it the central claim does not follow. The pressure weak convergence is also asserted without deriving a discrete pressure bound; that missing piece should come from the discrete inf-sup constant, but it is not in the paper.\n\nThe numerical section is honest about implementation and shows the method works on triangular and distorted meshes, with rates close to one for the temperature-dependent viscosity case. But the exact solution does not satisfy the homogeneous Dirichlet boundary conditions assumed in the theory, so the tests do not actually validate the theorem as stated. That is a real soft spot, though it is secondary to the proof gap. Minor sloppiness in Lemma 3.1's summation step and a few typos round out the list.\n\nOverall: the project is sensible, the gaps are identifiable and probably repairable, and the paper deserves a serious referee. I would engage with it if I worked on GDM or coupled flow problems, but I would not cite it in its current form.","headline":"A legitimate GDM extension for the time-dependent Navier-Stokes/heat system, but the main theorem has a load-bearing regularity gap and the numerics test the wrong boundary conditions.","tokens_in":10867,"tokens_out":2744,"would_cite":false,"duration_ms":26960,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65M08","76D05","35Q30","80A19"],"pacs":[],"model":"deepseek-v4-flash","headline":"A generic numerical scheme converges to a weak solution of the coupled Navier-Stokes/heat equations with temperature-dependent viscosity.","keywords":["Navier-Stokes problem","heat equation","time-dependent problem","gradient discretisation method","gradient schemes","finite volume scheme","convergence analysis"],"falsifier":"Run the fully discrete scheme with a bounded discontinuous viscosity such as $V(S)=1$ for $S<0$ and $V(S)=2$ for $S\\ge 0$ on a manufactured temperature field that crosses the discontinuity, and test whether the viscosity term converges to the expected limit as the mesh and time step shrink; if that limit is wrong, or if the discrete pressures grow without bound, Theorem 3.4 needs an additional hypothesis.","tokens_in":9798,"feed_emoji":"🌡️","tokens_out":9666,"duration_ms":86528,"temperature":0.7,"pith_summary":"The paper designs a generic space-time discretisation, in the gradient discretisation framework, for the time-dependent Navier-Stokes equations coupled with the heat equation through a temperature-dependent viscosity, and proves convergence without imposing non-physical conditions. The main theorem states that any sequence of space-time gradient discretisations that is coercive, consistent, limit-conforming, trilinear limit-conforming and compact produces discrete solutions whose velocity and temperature reconstructions converge strongly in $L^2$, whose gradients converge weakly in $L^2$, and whose pressure converges weakly in $L^2$, to a weak solution of the continuous coupled problem. The proof works by deriving uniform energy estimates for the discrete solutions and then using compactness and the abstract consistency properties to identify the limit. Numerical experiments with a hybrid finite volume scheme on triangular and distorted meshes, for both constant viscosity and the temperature-dependent viscosity $V(S)=\\sqrt{S^2+1}+2$, show first-order convergence rates. If the theorem is right, a single checklist of five abstract properties certifies convergence for every concrete scheme that fits this format.","feed_headline":"Generic scheme converges for coupled Navier-Stokes/heat flow","feed_subtitle":"Under five abstract properties, discrete velocity, pressure and temperature converge to a true weak solution.","key_machinery":"The central object is the space-time gradient discretisation $D_T=(D,J_D,\\tilde J_D,(t^{(n)})_{n=0,\\dots,N})$, a tuple of finite-dimensional spaces and linear reconstruction operators that encodes how a concrete numerical method represents velocity, temperature, pressure, gradients and divergence. Its role is to separate the specific scheme (finite volume, virtual element, nonconforming finite element, and so on) from the mechanism of convergence: if the discretisation satisfies the abstract properties of coercivity, consistency, limit-conformity, trilinear limit-conformity and compactness, then the generic variational scheme (2.6) inherits uniform discrete energy estimates and can be passed to the limit. The positivity assumptions on the discrete convection forms $A_D$ and $B_D$ and the discrete inf-sup quantity $B_D$ in (2.5) are what make the energy estimates and the pressure limits possible.","core_discovery":"The paper claims that the fully implicit gradient scheme (2.6) for the coupled problem is convergent in this precise sense: for any sequence of space-time gradient discretisations satisfying the five abstract properties, any sequence of discrete solutions $(u_m,p_m,S_m)$ has a subsequence for which the reconstructed velocity and temperature converge strongly in $L^2(0,T;L^2(\\Omega))$, the discrete gradients converge weakly in $L^2(\\Omega\\times(0,T))^d$, and the reconstructed pressures converge weakly in $L^2(\\Omega\\times(0,T))$, and the limit triple is a weak solution of the continuous coupled Navier-Stokes/heat system (1.1)-(2.4).","pith_inferences":["Editorial inference: the proof can be made fully rigorous by adding an explicit continuity or strong-composition assumption on $V$; the numerical examples use smooth $V$, so they do not probe the borderline case.","Editorial inference: a missing uniform $L^2$ bound on the discrete pressure means the claimed weak convergence of the pressure rests on an estimate that the paper does not supply; adding a discrete pressure-stability inequality would close the gap.","Editorial inference: the abstract framework should extend to other nonconforming discretisations, such as discontinuous Galerkin or virtual element methods, provided they satisfy discrete inf-sup stability; a natural test is to implement one and verify the five properties.","Editorial inference: the convergence statement is for weak solutions and does not address uniqueness or time regularity; if the coupled system has multiple weak solutions, the theorem only guarantees convergence to one of them."],"forward_implications":["Any concrete scheme that can be written as a space-time gradient discretisation and that satisfies the five abstract properties converges to a weak solution of the coupled problem, so the theorem acts as a certification checklist for finite volume, virtual element and other polytopal methods.","The uniform energy bounds of Lemma 3.1 give a constructive numerical proof of existence of weak solutions, since the limit of the bounded discrete sequence is the desired continuous solution.","Because the scheme is implicit in time and no CFL-type condition appears in the abstract assumptions, convergence holds for arbitrary time-step/mesh-size ratios as long as the time step tends to zero.","In the numerical tests, relative $L^2$ errors for velocity, pressure and temperature decay at a rate close to one on both triangular and distorted meshes, for constant viscosity and for $V(S)=\\sqrt{S^2+1}+2$."],"supporting_citations":[{"why":"Supplies the abstract compactness, interpolation and discrete integration-by-parts results (Lemma 4.8, Theorem 4.14, equation (D.15)) used to pass to the limit in Theorem 3.4.","marker":"[11]"},{"why":"Defines the spatial gradient discretisation and the properties of coercivity, consistency, limit-conformity, trilinear limit-conformity and compactness, and gives the hybrid finite volume discretisation used in Section 4.","marker":"[4]"},{"why":"Establishes existence of a weak solution to the continuous time-dependent coupled problem (2.4), the target that the discrete solutions converge to.","marker":"[2]"},{"why":"Provides the adaptation of the space-time convergence properties for incompressible Navier-Stokes schemes that the paper adopts in Definition 2.3.","marker":"[12]"},{"why":"Supplies the fixed-point reasoning for existence of a solution to the nonlinear discrete elliptic problem at each time step, referenced in Section 2.","marker":"[1]"}],"fun_headline_variants":["Gradient method converges for heat-dependent Navier-Stokes","Coupled Navier-Stokes-heat scheme: convergence proven","Discrete scheme for heat-coupled flow proven convergent","Fully implicit gradient scheme converges for coupled system","Heat-coupled Navier-Stokes: generic scheme's convergence proved"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on an unstated regularity property of the viscosity $V$: the proof exchanges a limit in temperature with a limit in viscosity, which requires $V(S_m)$ to converge in a strong enough sense, while the stated assumptions only guarantee that $V$ is bounded between two positive constants.","fun_headline_variants_meta":{"raw":{"variants":["Gradient method converges for heat-dependent Navier-Stokes","Coupled Navier-Stokes-heat scheme: convergence proven","Discrete scheme for heat-coupled flow proven convergent","Fully implicit gradient scheme converges for coupled system","Heat-coupled Navier-Stokes: generic scheme's convergence proved"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000279,"raw_usage":{"total_tokens":1547,"prompt_tokens":727,"completion_tokens":820,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":343,"completion_tokens_details":{"reasoning_tokens":741}},"tokens_in":343,"tokens_out":820,"duration_ms":7969,"temperature":1.0,"reasoning_tokens":741,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:04:28.004722+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the fully discrete scheme with a bounded discontinuous viscosity such as $V(S)=1$ for $S<0$ and $V(S)=2$ for $S\\ge 0$ on a manufactured temperature field that crosses the discontinuity, and test whether the viscosity term converges to the expected limit as the mesh and time step shrink; if that limit is wrong, or if the discrete pressures grow without bound, Theorem 3.4 needs an additional hypothesis.","supporting_citations":[{"cited_title":"Droniou, R","cited_arxiv_id":null,"evidence_quote":"Supplies the abstract compactness, interpolation and discrete integration-by-parts results (Lemma 4.8, Theorem 4.14, equation (D.15)) used to pass to the limit in Theorem 3.4."},{"cited_title":"Alnashri, The gradient discretisation method for the navier–stokes problem coupled with the heat equation , Results in Applied Mathematics, 11 (2021), p","cited_arxiv_id":null,"evidence_quote":"Defines the spatial gradient discretisation and the properties of coercivity, consistency, limit-conformity, trilinear limit-conformity and compactness, and gives the hybrid finite volume discretisation used in Section 4."},{"cited_title":"Agroum, C","cited_arxiv_id":null,"evidence_quote":"Establishes existence of a weak solution to the continuous time-dependent coupled problem (2.4), the target that the discrete solutions converge to."},{"cited_title":"Eymard, P","cited_arxiv_id":null,"evidence_quote":"Provides the adaptation of the space-time convergence properties for incompressible Navier-Stokes schemes that the paper adopts in Definition 2.3."},{"cited_title":"Agroum, S","cited_arxiv_id":null,"evidence_quote":"Supplies the fixed-point reasoning for existence of a solution to the nonlinear discrete elliptic problem at each time step, referenced in Section 2."}],"review_version":1}