{"id":"3794bcfa-cb6f-4139-ad3c-4316cd7694cb","arxiv_id":"1908.05771","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Unsteady solutions of the transient double porosity/permeability model are Lyapunov stable and grow at most linearly with time under homogeneous velocity boundary conditions.","lead":"This paper proves that unsteady solutions of a two-network porous media flow model remain stable and grow no faster than linearly in time. The result gives a simple curve that numerical simulations of fractured reservoirs, bone, or 3D printed porous materials should respect.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 3's Lyapunov-stability claim is not established for the general transient DPP model: the proposed functional requires a conservative-load potential and a true zero equilibrium, neither of which holds for the paper's own Case 1.","rationale":"The reader's weakest assumption correctly identifies the potential-energy construction in Section 3 as the load-bearing weak point. My reading agrees and sharpens it: the failure is not only that Πext is not constructed, but that the proposed equilibrium Υeq = 0 is not an equilibrium under non-conservative body forces, including the paper's own Case 1. This makes the Lyapunov claim as stated unsupported, while leaving the linear growth bound in Section 4 intact. The growth proof uses only dissipativity of the spatial operator and the Cauchy-Schwarz inequality, and the numerical example is a legitimate, if loose, illustration of that bound. I therefore maintain the reader's CONDITIONAL verdict: the paper should be accepted only with the stability claim repaired or restricted, since the growth bound and verification procedure remain useful. No ad hominem is intended; the issue is the mathematical argument itself.","tokens_in":9990,"tokens_out":8298,"duration_ms":88124,"concrete_test":"Check whether the zero field is an admissible solution in the stability setting: set v1 = v2 = 0 in (2.3)-(2.4) and take the curl of both sides. This yields the necessary condition curl(ρ1b1/φ1) = 0 and curl(ρ2b2/φ2) = 0. Evaluate this condition for the Case 1 body force in Table 1, b = (10 sin(πxt), 5 sin(2πxyt)); the curl is 10πyt cos(2πxyt) ≠ 0, so no pressure field can balance the load at zero velocity. If this calculation is confirmed, the Lyapunov-stability theorem as stated in Section 3 cannot apply to the general transient DPP model or to the paper's own numerical example, and the claim must be restricted or repaired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central stability assertion is that the zero state Υeq = {0,0} attracts unsteady solutions under the transient DPP model, proved via the functional V in (3.3). For this to work, V must be a positive-definite functional of the current state and non-increasing along solutions. The proof hinges on assertions (3.5)-(3.6): that Πext = Πeq_ext at every evaluation instant and that dΠeq_ext/dt = 0 while dΠext/dt ≠ 0. No such potential is constructed. If the body force is non-conservative, no potential Πext exists at all, so (3.4) is undefined. If the body force is time-dependent and conservative, the explicit time derivative of the potential should appear in dΠext/dt, so the cancellation of the ρb·v term used to obtain (3.11) is not justified. More fundamentally, Υeq = 0 is not an equilibrium of the DPP system unless the body force is a gradient field: at v1 = v2 = 0, equations (2.3)-(2.4) require φ1∇p1 = ρ1b1 and φ2∇p2 = ρ2b2, which is solvable for p only if curl(bi) = 0. The numerical Case 1 in Table 1 uses b = (10 sin(πxt), 5 sin(2πxyt)), whose curl is nonzero, so zero is not a solution and Lyapunov stability around it is not even well-posed for that case. The linear growth bound of Section 4 does not depend on this construction and appears sound; the stability portion of the paper needs either a genuine construction of the potential, a restriction to autonomous conservative loads with a fixed equilibrium, or a reformulation of the stability statement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the transient double porosity/permeability (DPP) model for incompressible flow in porous media with two pore networks. It claims two mathematical properties for unsteady solutions under homogeneous velocity boundary conditions: first, that the zero state Υeq = {0,0} is Lyapunov stable; second, that the weighted L2 norm of the velocity pair grows at most linearly with time, with the explicit bound ‖Υ‖V ≤ t·fmax + c given in equation (4.24). The Lyapunov argument in Section 3 is based on a candidate functional V that combines kinetic energy with external-load potentials Πext and Πeq_ext. The growth argument in Section 4 shows that the linear operator L is dissipative under homogeneous boundary conditions, and then integrates the resulting differential inequality. A numerical example with two body-force cases is used to illustrate how the bound can serve as an a posteriori verification tool for numerical implementations.","tokens_in":10354,"tokens_out":3636,"duration_ms":36799,"significance":"If the claimed results hold, the linear growth bound provides a simple, parameter-free, and non-intrusive verification tool for numerical solvers of the transient DPP model, including anisotropic media. The derivation of the growth bound in Section 4 is explicit and appears mathematically sound: the dissipativity computation leading to (4.19) and the subsequent integration to (4.24) are correct under the stated homogeneous boundary conditions, and the numerical comparison in Section 5 is a genuine independent check of the bound because the slope fmax and intercept c are computed directly from problem data rather than fitted. The Lyapunov-stability portion, however, is not established by the manuscript as written. The proof relies on assumptions about the existence and time-behavior of external-load potentials that are neither derived nor satisfied by the paper's own numerical cases. The stability claim therefore needs either a genuine construction of the Lyapunov functional under precise hypotheses or a reformulation to a setting where a fixed equilibrium actually exists.","major_comments":[{"comment":"The Lyapunov functional V in (3.3) is not well-defined for the general transient DPP model. Equations (3.4)–(3.6) assert that a potential energy Πext for the external loadings exists, that Πext equals Πeq_ext at every evaluation instant, and that dΠeq_ext/dt = 0 while dΠext/dt ≠ 0. For a time-dependent or non-conservative body force such as Case 1 in Table 1, no such potential exists; for a time-dependent conservative force, the explicit time dependence contributes to dΠext/dt, so the cancellation of the ρb·v terms used to reach (3.11) is not justified. Moreover, if (3.5) holds at every t0, then Πext and Πeq_ext coincide as functions of time and their derivatives must coincide, contradicting (3.6). The proof needs either a precise construction of V under stated restrictions on the body force and boundary data, or a different stability argument. This issue is load-bearing because it is the only support for the paper's Lyapunov-stability claim.","section":"Section 3, equations (3.3)–(3.6)"},{"comment":"The equilibrium state Υeq = {0,0} is not a solution of the DPP system for a general body force. Setting v1 = v2 = 0 in (2.3) and (2.4) gives φ1 grad[p1] = ρ1b1 and φ2 grad[p2] = ρ2b2, which is solvable for scalar pressures only if curl(b1) = curl(b2) = 0; this condition fails for the paper's own Case 1, where b = (10 sin(πxt), 5 sin(2πxyt)) has nonzero curl and is time-dependent. For this case, the zero state is not an equilibrium, and Lyapunov stability of that state is not well-posed. The stability theorem must therefore be restricted to autonomous conservative body forces with a genuine fixed equilibrium, or reformulated in terms of a moving equilibrium trajectory.","section":"Section 3, equations (2.3), (2.4), (3.2)"},{"comment":"The numerical example does not validate the Lyapunov-stability claim, and in fact its Case 1 does not satisfy the hypotheses needed for that claim. The example only checks the linear growth bound (4.24), which is the Section 4 result. The abstract and Section 6 present both properties as established for the general transient DPP model, so the manuscript should clearly separate the two results and state the restrictive assumptions under which Lyapunov stability is claimed.","section":"Section 5, Table 1 and Figure 4"}],"minor_comments":[{"comment":"The second pressure term in the Green's-identity step is written as ∫ div[φ1v2]p2 dΩ, but it should be ∫ div[φ2v2]p2 dΩ to match the micro-network equation; the subsequent line (3.10) uses the correct form.","section":"Equation (3.9)"},{"comment":"The second component of the vector L[Υ,p1,p2] has a missing closing parenthesis; it should be (−µφ2²K2⁻¹v2 − φ2 grad[p2])/ρ2.","section":"Equation (4.3)"},{"comment":"The notation ⟨Υ;L⟩V is not precise because L is an operator; the argument should be written as ⟨Υ;L[Υ,p1,p2]⟩V throughout the dissipativity calculation.","section":"Equations (4.15)–(4.19)"},{"comment":"The division by ‖Υ‖V requires the nonzero case; the trivial case ‖Υ‖V = 0 at isolated times should be handled by a continuity argument or stated explicitly.","section":"Equation (4.23)"},{"comment":"Section 3 explicitly allows non-homogeneous pressure boundary conditions, while Section 4 assumes homogeneous boundary conditions on the entire boundary; the abstract and Section 6 should make clear which boundary setting applies to each claimed result.","section":"Sections 3 and 4, boundary conditions"}],"recommendation":"major_revision","confidential_remarks":"The Section 4 growth-bound derivation is sound and could be the basis of a useful verification tool. The Section 3 stability argument, however, is not supportable in its current form without substantially restricting the class of loads and clarifying the equilibrium concept. I would urge the editor to require that the revision state the stability theorem with explicit hypotheses (e.g., autonomous conservative body forces, existence of a fixed equilibrium) or remove the Lyapunov-stability claim from the main results until it is properly established; the numerical section should also be presented only as a check of the growth bound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe part of this paper worth keeping is Section 4. The dissipativity argument is clean: under homogeneous velocity boundary conditions, ⟨Υ,LΥ⟩_V ≤ 0, and the Cauchy-Schwarz step gives d/dt||Υ||_V ≤ ||f||_V, hence ||Υ||_V ≤ t f_max + c. That is a simple but correct a posteriori verification bound for transient double porosity/permeability simulations, and it holds under anisotropic permeabilities. The numerical example confirms the bound, and the fact that the theorem is parameter-free (no fitted constants) is a real point in its favor.\n\nThe trouble is Section 3. The Lyapunov functional V(Υ) = kinetic energy + Π_ext − Π_eq_ext is not actually constructed. The proof assumes Π_ext is a potential for the external loads, assumes Π_ext = Π_eq_ext at every evaluation instant, and assumes dΠ_eq_ext/dt = 0 even while Π_ext changes. For time-dependent conservative loads the time derivative of the potential has an explicit ∂/∂t term, so the cancellation that yields (3.11) is not justified. For non-conservative loads, no such potential exists at all. Moreover, Υ_eq = {0,0} is not an equilibrium of the DPP system unless b1 and b2 are gradient fields; at v1=v2=0, equations (2.3)-(2.4) require φ1∇p1 = ρ1 b1, which is solvable only when curl(b_i)=0. The paper's own Case 1 uses b = (10 sin(πxt), 5 sin(2πxyt)), which has nonzero curl, so zero is not even a solution there. This undermines the abstract's blanket claim of Lyapunov stability for the transient DPP model.\n\nThe good news: the growth bound does not depend on the flawed stability construction. So even though the paper overstates what is proved, the verification application survives. Soft spots beyond Section 3: no code or data are provided, and the f_max used in the numerical check is a loose upper bound, but those are minor for an illustrative result.\n\nMy recommendation: send it to peer review, but insist Section 3 be repaired. Either construct the potential explicitly, restrict the stability claim to autonomous conservative loads with a genuine equilibrium, or drop the Lyapunov claim entirely and keep the growth bound. A serious referee will want that fixed before publication.","headline":"Keep the growth bound, fix the stability claim: Section 3's Lyapunov functional is assumed, not constructed, and zero is not an equilibrium for the paper's own Case 1.","tokens_in":10873,"tokens_out":3572,"would_cite":true,"duration_ms":30485,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B35","76S05","65M60","35Q35"],"pacs":[],"model":"deepseek-v4-flash","headline":"Unsteady solutions of the transient double porosity/permeability model are Lyapunov stable and grow at most linearly in time, giving a simple norm check for numerical simulators.","keywords":["double porosity/permeability","Lyapunov stability","bounded solutions","transient response","flow through porous media","linear growth bound","a posteriori verification","anisotropic permeability"],"falsifier":"Run the DPP model with the paper's Case 1 body force, $b_1=b_2=(10\\sin(\\pi x t),5\\sin(2\\pi x y t))$, and evaluate $\\dot V$ from equation (3.7); if $\\dot V$ becomes non-negative while $\\Upsilon\\neq 0$, the general Lyapunov-stability claim fails for time-dependent forcing. Separately, if a converged numerical norm ever exceeds $t f_{\\max}+c$, the linear growth bound is false.","tokens_in":9759,"feed_emoji":"💧","tokens_out":15996,"duration_ms":129638,"temperature":0.7,"pith_summary":"The paper establishes two properties of the transient double porosity/permeability (DPP) model, which describes incompressible flow in a porous medium with two pore networks that exchange mass. First, unsteady solutions are stable in the Lyapunov sense, so small deviations from equilibrium do not amplify without control. Second, under homogeneous boundary conditions the solution norm grows at most linearly in time, with the explicit bound $\\|\\Upsilon\\|_V \\le t\\, f_{\\max} + c$, where $f_{\\max}$ is the largest norm of the forcing and $c$ is the norm of the initial data. Because analytical solutions for the DPP model are scarce, especially with anisotropic permeabilities, the linear bound gives a direct, non-intrusive after-the-fact check on numerical simulations: a computed norm that crosses the line $t f_{\\max}+c$ indicates a likely defect in the implementation.","feed_headline":"Unsteady double-porosity flows grow at most linearly in time","feed_subtitle":"A simple norm bound gives numerical simulators a built-in check for fractured-rock flow.","key_machinery":"The load-bearing mechanism is dissipativity of the evolution operator $\\mathcal{L}$ on the weighted product space $\\mathbb{V} = (L^2(\\Omega))^{nd}\\times(L^2(\\Omega))^{nd}$ equipped with the inner product $\\langle \\Upsilon;\\tilde{\\Upsilon}\\rangle_{\\mathbb{V}} = \\int_\\Omega(\\rho_1 v_1\\cdot\\tilde v_1+\\rho_2 v_2\\cdot\\tilde v_2)\\,d\\Omega$. The identity that carries the argument is $\\langle \\Upsilon;\\mathcal{L}\\rangle_{\\mathbb{V}} \\le -\\langle \\frac{\\beta}{\\mu}(p_1-p_2);(p_1-p_2)\\rangle \\le 0$, obtained by substituting the momentum equations, integrating by parts, and using the mass-balance equations (2.5)-(2.6). This single inequality does the work: it makes the Lyapunov functional non-increasing along trajectories and converts the growth calculation into an integrable differential inequality. The accompanying defined object is the Lyapunov functional of equation (3.3), whose potential-energy part $\\Pi_{\\mathrm{ext}}-\\Pi_{\\mathrm{ext}}^{\\mathrm{eq}}$ is assumed to satisfy the instantaneous-equality conditions (3.5)-(3.6).","core_discovery":"The central claim is that unsteady solutions of the transient DPP model, with time-independent volume fractions and fluid density, positive definite permeabilities, and homogeneous velocity boundary conditions, are Lyapunov stable and satisfy the linear growth estimate $\\|\\Upsilon(x,t)\\|_V \\le t\\, f_{\\max} + c$, equation (4.24), with $f_{\\max} = \\max_{t\\in[0,T]}\\|f\\|_V$ and $c = \\|\\Upsilon(x,0)\\|_V$. The proof works by showing that the evolution operator $\\mathcal{L}$ is dissipative on the weighted space $\\mathbb{V}$: the mass-balance coupling between the two pore networks contributes a non-positive term, $-\\langle \\frac{\\beta}{\\mu}(p_1-p_2),(p_1-p_2)\\rangle$, so that $\\langle \\Upsilon;\\mathcal{L}\\rangle_{\\mathbb{V}} \\le 0$. This dissipativity turns the evolution equation into the differential inequality $\\frac{d}{dt}\\|\\Upsilon\\|_V \\le \\|f\\|_V$, whose integration yields the bound. The paper also constructs a Lyapunov functional $V(\\Upsilon) = \\frac{1}{2}\\int_\\Omega(\\rho_1 v_1\\cdot v_1 + \\rho_2 v_2\\cdot v_2)\\,d\\Omega + \\Pi_{\\mathrm{ext}} - \\Pi_{\\mathrm{ext}}^{\\mathrm{eq}}$ and shows formally that $dV/dt<0$ for $\\Upsilon\\neq 0$, provided the external loads are conservative and the equilibrium-potential conditions (3.5)-(3.6) hold.","pith_inferences":["The linear-growth half of the paper is not affected by the potential-energy construction: the Section 4 dissipativity proof never invokes it, so a time-dependent body force that violates (3.5)-(3.6) would undermine the stability claim without touching equation (4.24).","A direct testable extension is to run the paper's Case 1 forcing, $b_1=b_2=(10\\sin(\\pi x t),5\\sin(2\\pi x y t))$, while monitoring the proposed functional $V$; seeing whether $dV/dt$ actually stays negative would separate the two results empirically.","The same dissipativity argument should carry over to pore networks with more than two networks, or to nonlinear couplings, as long as each permeability is positive definite and the mass-transfer terms enter as a non-positive quadratic form.","For verification practice, the bound is a necessary condition rather than a sufficient one: a compliant norm does not prove correctness, so the check is best combined with manufactured-solution or patch tests."],"forward_implications":["Every numerical solution of the transient DPP model, under the stated hypotheses, should keep $\\|\\Upsilon\\|_V$ at or below the line $t f_{\\max}+c$; crossing the line is a clear signal of an implementation error.","The bound remains valid for anisotropic permeability tensors, so it covers the regime where analytical DPP solutions are scarce and where verification by manufactured solutions is less natural.","Bounded forcing functions imply bounded growth in finite time: the solution norm cannot blow up faster than linearly, no matter how complex the pore-network mass transfer is.","The paper's finite-element example, using a stable mixed velocity-pressure pair, satisfies the bound, so the estimate is ready to be used as a routine after-the-fact verification procedure.","Under the conservative-loading assumptions of Section 3, the same analysis gives Lyapunov stability of the zero equilibrium in the dynamical-systems sense: nearby initial states do not undergo unbounded deviation."],"supporting_citations":[{"why":"Supplies the transient DPP model, its underlying assumptions, and the steady analytical solutions that this paper extends to unsteady bounds.","marker":"[Nakshatrala et al., 2018]"},{"why":"Provides the definition of Lyapunov stability that frames the Section 3 result.","marker":"[Dym, 2002]"},{"why":"Supplies the dynamical-systems stability theory used to justify the Lyapunov-functional argument.","marker":"[Hale and Koçak, 2012]"},{"why":"Licenses the conclusion that a non-increasing Lyapunov functional for an infinite-dimensional system implies Lyapunov stability.","marker":"[Luo et al., 2012]"},{"why":"Provides the stabilized mixed formulation and uniqueness discussion used in the representative numerical example.","marker":"[Joodat et al., 2018]"},{"why":"Justifies the LBB-stable P3-P1 interpolation used in the numerical test.","marker":"[Brezzi and Fortin, 2012]"}],"fun_headline_variants":["Double-porosity flows: stable and linearly growing","Lyapunov stability and linear growth for DPP unsteady solutions","Unsteady double-porosity solutions: stable with linear bound","New bound: unsteady DPP flows grow at most linearly"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The Lyapunov-stability claim rests on the assumption that at every instant the external-load potential energy can be equated to the equilibrium potential energy at that same instant, with only the nonequilibrium part changing over time; the paper states this assumption rather than proving it, and time-dependent body forces need not satisfy it.","fun_headline_variants_meta":{"raw":{"variants":["Double-porosity flows: stable and linearly growing","Lyapunov stability and linear growth for DPP unsteady solutions","Unsteady double-porosity solutions: stable with linear bound","New bound: unsteady DPP flows grow at most linearly"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1531,"prompt_tokens":1085,"completion_tokens":446,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":376}},"tokens_in":701,"tokens_out":446,"duration_ms":4497,"temperature":1.0,"reasoning_tokens":376,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:06:49.448487+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the DPP model with the paper's Case 1 body force, $b_1=b_2=(10\\sin(\\pi x t),5\\sin(2\\pi x y t))$, and evaluate $\\dot V$ from equation (3.7); if $\\dot V$ becomes non-negative while $\\Upsilon\\neq 0$, the general Lyapunov-stability claim fails for time-dependent forcing. Separately, if a converged numerical norm ever exceeds $t f_{\\max}+c$, the linear growth bound is false.","supporting_citations":[],"review_version":1}