{"id":"378fcb77-8c31-4e1e-8882-929a04eebda9","arxiv_id":"2505.13320","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"New semi-analytical formulas give advection-dominated concentrations and Taylor-like enhanced diffusion coefficients for arbitrary velocity profiles, applied to Cross and Carreau fluids and matched to microscale simulations.","lead":"This paper derives semi-analytical solutions for velocity and passive scalar transport of generalized Newtonian fluids in capillary tubes and parallel-plate slits. The solutions avoid requiring a closed-form velocity profile and are validated against pore-scale OpenFOAM simulations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Velocity integrals Eqns (11)-(12) have reversed limits/sign and give negative Newtonian profiles; as printed the velocity and transport validation cannot be reproduced.","rationale":"The reader's weakest assumption, the quasi-steady Taylor ansatz in Eqns (47)-(48), is legitimate and I agree it is the principal theoretical approximation; it is standard, however, and is supported by the Pe=100 validation. My independent stress-test found a more concrete and actionable issue in the printed velocity solution: the integration limits in Eqns (11)-(12) are reversed relative to the positive shear-rate convention of Eqn (3), producing negative Newtonian velocities. The transport formulas Eqns (73) and (78) themselves pass the Newtonian limits (tube: R^2 U^2/(48 D); slit: 2 B^2 U^2/(105 D)), so the central mathematical claim is sound. The paper should be published only after the velocity equations are corrected; hence CONDITIONAL rather than REJECT or UNCHANGED.","tokens_in":23477,"tokens_out":25270,"duration_ms":252685,"concrete_test":"Take the Newtonian limit mu(gamma) = mu in Eqns (11) and (12) with a positive pressure drop Delta-p. If the printed expressions yield negative velocities, namely -(Delta-p/(4 mu L))(R^2 - r^2) for the tube and -(Delta-p/(2 mu L))(B^2 - z^2) for the slit, then the integration bounds must be reversed, or a sign inserted, to match the Poiseuille profiles shown in Figures 2-3. This single substitution settles whether the velocity solution is correctly presented as written.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Eqns (11)-(12) as printed are inconsistent with the reported validation. For a positive pressure drop, the shear-rate magnitude obeys gamma = -dv/dr in the tube and gamma = -dv/dz in the slit; with tau = mu(gamma)gamma, Eqns (8)-(9) imply v(r) = (2L/Delta-p) integral from gamma_r to gamma_R of gamma (d-tau/d-gamma) d-gamma, not the printed integral from gamma_R to gamma_r. The printed limits make the Newtonian limit of Eqn (11) equal to -(Delta-p/(4 mu L))(R^2 - r^2), and Eqn (12) equal to -(Delta-p/(2 mu L))(B^2 - z^2), i.e. minus the Poiseuille profiles shown in Figures 2 and 3. Because Eqns (27), (28), (73), and (78) are evaluated with these velocity profiles, a reader implementing the printed equations cannot reproduce the stated <1% agreement. The Taylor-like quasi-steady ansatz in Eqns (47)-(48) is a second and acknowledged limitation: no rigorous error bound is given, validation is only at Pe=100, and the authors note in Section VI.C that investigating the assumptions is future work.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives semi-analytical solutions for steady, pressure-driven laminar flow of Cross and Carreau generalized Newtonian fluids in capillary tubes and slit channels, and then uses these velocity fields to construct two classes of passive-scalar transport solutions: a Taylor-type advection-dominated solution for a step inlet condition, and an enhanced-diffusion (Taylor dispersion) coefficient expressed as integrals over the velocity profile. The central claim is that the transport formulas, especially Eqs. (73) and (78), require no closed-form velocity expression and no specific rheological model, so they can be applied to any GNF or even to experimentally measured velocity fields. The solutions are validated against 500 OpenFOAM flow simulations and 3,000 transport simulations, with reported RMSE below 1% in the target regimes (Pe=10^5 for advection-dominated transport and Pe=100 for enhanced molecular diffusion).","tokens_in":23683,"tokens_out":12561,"duration_ms":117154,"significance":"If the printed equations are correct, the paper provides a useful and genuinely general extension of Taylor's classic dispersion analysis: the enhanced diffusion coefficient is expressed directly in terms of quadratures of the velocity profile, eliminating the need for closed-form velocity expressions that are unavailable for Cross, Carreau, and many other non-Newtonian models. The derivation is self-contained, uses no fitted parameters, and is supported by an unusually large set of independent microscale simulations. The TCAT-based averaging route to the effective dispersion coefficient is a strength, as is the explicit statement that the machinery applies to other rheologies and to experimental velocity data. The main weakness is that the velocity formulas as printed contain a sign and bound-order error, and the quasi-steady Taylor ansatz in the moving frame is validated at only one Peclet number.","major_comments":[{"comment":"The printed velocity integrals have reversed limits and an inconsistent sign, so a reader implementing Eqns. (11) and (12) cannot reproduce the positive velocity profiles shown in Figures 2 and 3. For a Newtonian fluid, tau = mu*gamma and d(tau)/d(gamma) = mu, so Eqn. (11) as printed evaluates to (2L/Delta-p)(mu/2)(gamma_r^2 - gamma_R^2) = (Delta-p/(4 mu L))(r^2 - R^2), which is the negative of the Poiseuille profile. The source is the chain in Eqn. (10): for the tube, the shear-rate magnitude satisfies gamma = -dv/dr, not gamma = +dv/dr. The correct form is v(r) = (2L/Delta-p) * integral from gamma_r to gamma_R of gamma*(d tau/d gamma) d gamma, and similarly for Eqn. (12) with the slit geometry. Because Eqns. (27), (28), (73), and (78) are evaluated with these velocity profiles, the printed equations are inconsistent with the reported validation; this must be corrected and the figures and RMSE statements re-verified with the corrected expressions.","section":"II.A, Eqs. (10)-(12)"},{"comment":"The derivation of the enhanced molecular diffusion coefficient rests on the quasi-steady Taylor ansatz: dropping the time derivative in the moving frame and postulating C = g(x) + (dC/dx) f(z) with f independent of x. The manuscript validates this only at Pe=100 and provides no error estimate for other Peclet numbers; Section VI.C explicitly defers investigation of these assumptions to future work. Because the abstract claims a general computation of the enhanced diffusion coefficient, the paper should either add a domain-of-validity statement supported by a few additional Pe values (e.g., showing convergence or breakdown) or temper the claim so that it is explicitly restricted to the validated low-Pe regime.","section":"IV.C and VI.C, Eqs. (47)-(48)"}],"minor_comments":[{"comment":"There is a typo: \"flluid\" should be \"fluid\" in the sentence after Eqn. (2).","section":"II.A"},{"comment":"The text says \"25,0000 possible points of comparison\"; this should be 25,000 (500 simulations times 50 velocity values).","section":"V.E"},{"comment":"The sentence \"The RMSE between averaged microscale simulations and macroscale modeling exceeded 10% until Pe <= 100\" is self-contradictory given that the Pe=100 results in Figures 14-15 show RMSE below 1%. It should be reworded to state that the RMSE exceeded 10% for Pe > 100, with Pe=100 marking the upper bound of the enhanced-diffusion regime.","section":"VI.C"},{"comment":"The claim that the transport solutions apply as a \"straightforward extension\" to viscoelastic or viscoplastic fluids should be accompanied by a caveat that the advection-dominated solution assumes a monotonically decreasing velocity profile from the centerline; non-monotonic or plug-flow profiles may require modification of the a* construction in Eqns. (27) and (28).","section":"Abstract and VII"}],"recommendation":"major_revision","confidential_remarks":"The sign/order error in Eqns. (11)-(12) appears to be a presentation error rather than a fundamental flaw, since the rest of the derivation and the validation are internally consistent when the velocity profile is positive. However, the error is load-bearing: as printed, the core velocity equations cannot be reproduced and the transport validation is not reproducible. This is exactly the kind of issue that warrants a major revision rather than a reject. The quasi-steady limitation is acknowledged by the authors and can be addressed with a clarifying statement. I would be willing to accept after the authors correct the velocity integrals and verify the downstream formulas numerically."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a solid and useful extension of Taylor–Aris dispersion to Cross and Carreau fluids in tubes and slits, with a genuinely rheology-independent transport formula that does not require a closed-form velocity. The validation is serious: 3000 OpenFOAM simulations, RMSE below 1% in the target regimes. But the manuscript has a load-bearing typo in the velocity integrals. Eqns (11) and (12) as printed have the integration limits reversed, and the derivation in Eqn (10) carries a sign inconsistency. A reader implementing the printed equations gets negative Newtonian profiles, so the reported validation cannot be reproduced from the text. The figures show the correct physics, so this is almost certainly a transcription error, but it still must be fixed before the paper is usable as written.\n\nWhat is actually new: the transport solutions, Eqns (73) and (78), are derived for arbitrary velocity profiles and reduce to Taylor's result in the Newtonian limit. That is a clean generalization, and the advection-dominated solution using the a* parameter is a neat trick that avoids closed-form velocity. The derivation itself follows standard Taylor–Aris logic and the TCAT averaging is used appropriately; no fitted parameters, no circularity. The literature review is honest about prior open-form velocity solutions (Sochi, Kim, Wang).\n\nWhere the soft spots are: the quasi-steady Taylor assumption in Eqns (47)–(48) is standard and is validated at Pe=100, but there is no rigorous error estimate for other Peclet numbers; the authors themselves note this is future work. That is a known limitation of the method, not a fatal flaw. Code and data are not shipped, only available on request, which is a minor reproducibility issue. The self-citations are to prior work by the same group and are relevant, not padding.\n\nBottom line: this paper deserves a serious referee. The typo is embarrassing but fixable, and the scientific content is worth publishing after correction. I would ask the authors to fix the bounds in Eqns (11)–(12), add a sanity-check line showing the Newtonian limit reduces to Poiseuille, and ideally release the code or at least the validation data. With that, it is a solid contribution. For you: worth reading if you work on non-Newtonian transport in simple geometries; I would probably cite it once the equations are corrected.","headline":"Genuinely useful Taylor–Aris extension for non-Newtonian fluids, but the printed velocity integrals have a sign/limit typo that must be fixed before the paper can be used as written.","tokens_in":24201,"tokens_out":4888,"would_cite":true,"duration_ms":43349,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.50.-d","47.56.+r"],"model":"deepseek-v4-flash","headline":"A single dispersion formula computes tracer spreading in tubes and slits from any velocity profile, with no closed-form rheology needed.","keywords":["generalized Newtonian fluid","enhanced molecular diffusion","passive scalar transport","capillary tube","parallel-plate slit","Cross model","Carreau model","semi-analytical solution"],"falsifier":"A decisive check is to run the same microscale transport simulation at a higher transverse Peclet number (say Pe around 1000) and a shorter observation time, and compare the breakthrough curve against the one-dimensional model using Eqs. (73) and (78). If the effective coefficient varies with the length of the system or the time window — or if the RMSE grows well beyond the mesh-convergence error — the quasi-steady ansatz has failed. Equivalently, a laboratory dispersion experiment in a capillary with a Carreau fluid at two different tube lengths would reveal any non-asymptotic dependence of the measured dispersion coefficient.","tokens_in":23269,"feed_emoji":"💧","tokens_out":9126,"duration_ms":89277,"temperature":0.7,"pith_summary":"This paper establishes that the enhanced molecular diffusion coefficient — the effective spread of a passive scalar along the flow direction — in a capillary tube or a slit between parallel plates can be computed from any known velocity profile through two integrals, Eqs. (73) and (78), without ever writing down a closed-form velocity law. The same idea gives the cross-sectionally averaged concentration in the advection-dominated regime through Eqs. (27) and (28), using the radial position at which the local velocity matches the front position. The authors apply the general formulas to Cross and Carreau model fluids, whose velocity profiles are available only semi-analytically, and verify the results against microscale numerical simulations of flow and transport. If correct, the result removes the main obstacle to analytical transport modeling for shear-thinning fluids in these reference geometries and opens the door to using measured velocity data in place of rheological models.","feed_headline":"Dispersion formula works for any fluid velocity profile","feed_subtitle":"Shear-thinning fluids in tubes and cracks now need only tabulated velocities, verified to within numerical error.","key_machinery":"The load-bearing object is the classical moving-frame ansatz in a frame moving with the mean flow: $\\tilde C(\\tilde x,z)=g(\\tilde x)+(\\partial \\tilde C/\\partial \\tilde x)f(z)$, combined with the quasi-steady neglect of the time derivative in that frame. Inserting this ansatz into the normalized advection-diffusion equation reduces the transverse problem to nested integrals of the velocity deviation $v-\\bar v$, denoted $I_{R1}$ and $I_B$; these integrals are then averaged against the deviation velocity to produce Eqs. (73) and (78). The same machinery yields the advection-dominated concentration through the inverse-velocity coordinate $a^*$, and the semi-analytical velocity profiles for Cross and Carreau fluids follow from a shear-stress integral $I_v$ that involves hypergeometric functions. The effective dispersion coefficient is therefore a function of the velocity profile alone, not of any particular rheological closure.","core_discovery":"On the paper's own terms, the central discovery is that the whole effect of the velocity field on longitudinal dispersion in these two geometries is captured by a doubly integrated velocity-deviation function. Defining $\\bar v$ as the cross-sectional average velocity and $v(z)$ as the local velocity, the enhanced diffusion coefficient is $\\hat D^{iw}= -\\frac{2R^2}{\\hat D_{m,i}}\\int_0^1 I_{R1}(z_R)(v-\\bar v)z_R\\,dz_R + \\hat D_{m,i}$ for a tube and the analogous one-dimensional integral for a slit. These expressions reduce to the classical tube-dispersion result when the Newtonian parabolic profile is inserted, and they require only numerical quadrature when the profile is known pointwise. For advection-dominated transport, the average concentration is obtained from the inverse-velocity position $a^*$, again computable from tabulated velocities. Applied to eight Cross and Carreau fluids, the predictions agree with microscale simulations to within the numerical error of those simulations, and the computed longitudinal dispersivity falls by up to a factor of six relative to the Newtonian case because the shear-thinning viscosity flattens the velocity profile.","pith_inferences":["Because Eqs. (73) and (78) are linear functionals of the velocity-deviation profile, closed-form dispersion coefficients for any rheology with an analytical velocity profile (for example power-law or Bingham-like approximations) follow by quadrature; one test would be to compare those closed forms against the numerical integrals for limiting parameter values.","The quasi-steady moving-frame assumption should break down at short times or high transverse Peclet number; a testable extension is to compute the moving-frame time derivative and check whether the effective coefficient acquires an explicit time or averaging-length dependence before the asymptotic regime sets in.","The opposite signs of dispersivity change between straight channels and random porous media suggest a competition between velocity-profile flattening and tortuosity; a direct experiment varying only the wall geometry (straight versus wavy slit) could isolate these two mechanisms."],"forward_implications":["The longitudinal dispersion coefficient for a tube or slit can be computed from a tabulated velocity profile at a few hundred points, so closed-form rheological solutions are no longer a prerequisite for transport modeling.","The same formulas apply, as the paper states, to any non-Newtonian fluid — viscoelastic, viscoplastic, or uncharacterized — whenever a velocity profile is available from simulation, imaging, or velocimetry.","For Cross and Carreau fluids in these geometries, shear-thinning flattens the velocity profile and lowers dispersivity by up to a factor of six, reversing the direction of the effect reported for disordered porous media.","A one-dimensional advection-dispersion model using the derived coefficient reproduces microscale simulated breakthrough curves with less than 1% RMSE at Pe = 100.","The approach supplies a fast 'first-pass' transport simulator for fluids whose direct numerical transport simulation is expensive or unavailable."],"supporting_citations":[{"why":"Supplies the classical moving-frame dispersion analysis whose result the new integrals generalize to arbitrary velocity profiles.","marker":"[1]"},{"why":"Provides the method-of-moments framework for dispersion in arbitrary cross-sections that motivates the effective-coefficient derivation.","marker":"[8]"},{"why":"Supplies the analytical shear-stress-based velocity framework for Cross and Carreau fluids in pipes and thin slits that the semi-analytical velocity solution builds on.","marker":"[16]"},{"why":"Provides an open-form velocity solution for generalized Newtonian tube flow that supports the semi-analytical approach used here.","marker":"[52]"},{"why":"Defines the macroscale dispersion closure and averaging theorems used to turn the microscale concentration solution into the effective coefficient.","marker":"[26]"},{"why":"Supplies the microscale simulation and validation methodology for generalized Newtonian fluid flow and transport, plus the random-sphere-packing dispersivity trend used for comparison.","marker":"[24]"}],"fun_headline_variants":["Dispersion from any velocity profile via simple quadrature","Shear-thinning flow: dispersion solved by one integral","Velocity profile fully captures dispersion in tubes and slits","Generalized Newtonian dispersion: just integrate velocity once"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes that, in the frame moving with the average velocity, the cross-sectional concentration pattern reaches an instantaneous quasi-steady balance between transverse diffusion and longitudinal advection, so the time-derivative term in Eqs. (47) and (48) can be dropped; this is asymptotically valid at long times and low transverse Peclet number but is not backed by a rigorous error bound in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Dispersion from any velocity profile via simple quadrature","Shear-thinning flow: dispersion solved by one integral","Velocity profile fully captures dispersion in tubes and slits","Generalized Newtonian dispersion: just integrate velocity once"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000297,"raw_usage":{"total_tokens":1726,"prompt_tokens":951,"completion_tokens":775,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":567,"completion_tokens_details":{"reasoning_tokens":710}},"tokens_in":567,"tokens_out":775,"duration_ms":8074,"temperature":1.0,"reasoning_tokens":710,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:15:38.518168+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check is to run the same microscale transport simulation at a higher transverse Peclet number (say Pe around 1000) and a shorter observation time, and compare the breakthrough curve against the one-dimensional model using Eqs. (73) and (78). If the effective coefficient varies with the length of the system or the time window — or if the RMSE grows well beyond the mesh-convergence error — the quasi-steady ansatz has failed. Equivalently, a laboratory dispersion experiment in a capillary with a Carreau fluid at two different tube lengths would reveal any non-asymptotic dependence of the measured dispersion coefficient.","supporting_citations":[{"cited_title":"Agrawal , author J","cited_arxiv_id":null,"evidence_quote":"Provides the method-of-moments framework for dispersion in arbitrary cross-sections that motivates the effective-coefficient derivation."}],"review_version":1}