{"id":"b677472a-4229-4832-b5c5-f6900c49721e","arxiv_id":"1908.07589","paper_version":6,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A peridynamic double-well fracture model is shown, under three hypotheses and a uniform convergence assumption, to converge to classical plane elastodynamics with traction-free crack lips and the classic kinetic relation Gc V = J.","lead":"This paper studies a nonlocal peridynamic model of brittle fracture and shows that, in the limit of vanishing nonlocality, it reproduces the classical equations of dynamic fracture, including the relation between crack speed and energy release. The key caveat is that the result holds under three explicit geometric hypotheses about how the failure zone behaves, which are supported by simulations but not proved.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Uniform convergence assumed in Proposition 7 controls the boundary term that becomes the kinetic relation; it is not derived, so the recovery of Gc V = J is conditional.","rationale":"The reader's weakest_assumption correctly identifies Hypotheses 1-3 and the uniform convergence assumption in Proposition 7 as load-bearing. My stress-test narrows the concern: the uniform convergence assumption is not merely one of several hypotheses; it is the precise step that converts the nonlocal boundary integral into the local J-flux and thereby produces the kinetic relation. Hypotheses 1-3 support Propositions 2, 4, and 5, but Proposition 7 needs stronger uniform control that is neither proved nor shown to follow. The paper explicitly states that (43) is imported from Freund and Clifton, so the genuine recovery stops at Gc V = J with J as a local energy flux; this is a real but bounded limitation. No internal contradiction was found, and the conditional framework is coherent, so outright rejection would be too strong. The verdict CONDITIONAL is appropriate: the advertised recovery of the kinetic relation is not complete until the uniform convergence is proved or replaced by a weaker justification sufficient for (55).","tokens_in":27766,"tokens_out":1439,"duration_ms":18685,"concrete_test":"Derive or disprove the uniform convergence hypothesis in Proposition 7: show from (15) and Hypotheses 1-3 that sup_{t in [0,T]} sup_{x in U_delta(t)} |S(y,x,u^eps) - E u0(x) e dot e| -> 0 as eps -> 0 on bonds not crossing the crack centerline. A numerical check: for the three horizons in Section 6, compute this sup-error on a fixed contour at sides 1-4 and verify a rate at least O(eps); if the error does not vanish, the limit (55) and the kinetic relation (58) are not justified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Proposition 7, which yields the kinetic relation (58) via the energy-rate identity (56), rests on an unproved uniform convergence assumption: u^eps, dot u^eps, and the strains S(y,x,u^eps) converge uniformly to u0, dot u0, and E u0 e dot e on the contour neighborhood U_delta(t) away from the crack tip. This uniformity is asserted in Proposition 7, not derived from Hypotheses 1-3 or from the nonlocal dynamics. The crucial limit (55), especially lim E^eps_n(Gamma_delta(t)) = -int_Gamma CEu0 n dot dot u0 ds, is evaluated by replacing the nonlocal flux integrand with the local flux (129); that replacement needs exactly this uniformity. Without it, the power-balance route to (57)-(58) is not established. Additionally, the explicit kinetic relation (43)-(44) is quoted from Freund and Clifton and Freund (1990); J is not computed from the nonlocal model, so the paper recovers the energy-balance identity but not a fully determined crack-tip velocity law. The momentum-equation convergence in Propositions 4-5 is plausibly within reach of the stated hypotheses, but Proposition 7 needs stronger quantitative control.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a bond-based peridynamic model with a double-well (cohesive) potential for a single-edge-notch specimen under mode-I loading. The authors claim that as the peridynamic horizon ε tends to zero, solutions of the nonlocal initial-value problem (15) converge to the classical dynamic fracture problem: linear elastodynamics away from the crack, zero traction on the crack faces, and the kinetic relation G_c V(t) = J. The kinetic relation is not postulated as a power balance but is derived by passing to the ε → 0 limit in the nonlocal energy-rate identity (Proposition 6), yielding Proposition 7 and Eq. (58). The analysis is conditional on three hypotheses stated in Section 5 (geometric structure of the softening zone, immediate bond failure after softening, and ε^{-1} strain growth across the crack with a positive limit jump) and on an additional uniform-convergence assumption in Proposition 7. The paper also presents numerical simulations for three horizons that are offered as qualitative support for the hypotheses.","tokens_in":28214,"tokens_out":9254,"duration_ms":541787,"significance":"If the conditional results are accepted, the paper provides a substantive bridge between peridynamic fracture and classical dynamic fracture mechanics. The nonlocal divergence theorem (Proposition 9) and the exact nonlocal power-balance identity (Proposition 6) are genuine technical contributions, and deriving the energy-rate identity G_c V = J from an ε > 0 balance rather than postulating it is a meaningful step. The numerical experiments for three horizons are a useful sanity check. However, the strength of the central claim is substantially tempered by the fact that the main theorem and the kinetic relation depend on Hypotheses 1–3 and on an unproved uniform-convergence assumption in Proposition 7, and by the fact that the semi-explicit kinetic relation (44) relies on the classical formula (43) imported from Freund and Clifton rather than being computed from the nonlocal model.","major_comments":[{"comment":"The central convergence claims—Propositions 2, 4, 5, and 7—are all conditional on Hypotheses 1–3, and Proposition 7 adds a further unproved uniform-convergence assumption on u^{ε}, dot u^{ε}, and S(y,x,u^{ε}) to u^0, dot u^0, and E u^0 e·e away from the crack tip. The key limit (55), especially the replacement of the nonlocal flux integrand by the local flux (129), requires exactly this uniformity; it is not derived from Eq. (15) or from Hypotheses 1–3. Hypothesis 3 itself, Eq. (34), postulates an ε^{-1} strain growth across the crack line, which is a singular-behavior input rather than a consequence of the dynamics. Since the kinetic relation (58) is derived through Proposition 7, the paper's main claim is conditional. The authors do state the hypotheses explicitly, but the title, abstract, and conclusions do not carry this qualifier, and the numerical section provides only qualitative support (Figures 5–8), not a quantitative verification of the ε^{-1} scaling or of the uniform-convergence assumption.","section":"Section 5, Hypotheses 1–3; Proposition 7 (Eqs. (55)–(58))"},{"comment":"The semi-explicit kinetic relation (44) is not recovered from the nonlocal model: formula (43) for J is quoted from Freund and Clifton (1974) and Freund (1990), not computed from the peridynamic model. What Proposition 7 actually establishes is the energy-rate identity (58), G_c V = J with J = lim_{δ→0} ∫_{Γ_δ} C E u^0 n · dot u^0 ds. The further reduction to (44), which is the practically useful crack-tip velocity law, uses an external elastodynamic calculation of the energy flux in terms of K_I and V. Thus the statements in the Introduction and Conclusions that 'the kinetic relation for crack tip velocity is recovered directly from the nonlocal model' overstate the result: the nonlocal model supplies the energy balance, while the dependence of J on K_I and V is classical input. The paper should be reframed accordingly.","section":"Section 5 and Section 11, Eqs. (43)–(44)"},{"comment":"The paper states in Section 4 that 'it is assumed as in [Lipton(2016)] that the magnitude of the displacement u^{ε} is bounded uniformly in (x,t) for all horizons ε>0.' This is an additional regularity hypothesis on the nonlocal solutions that is not listed among Hypotheses 1–3 and is not derived from Eq. (15). It is used in the compactness arguments leading to Proposition 3 and the SBD limit. The paper should either prove this bound under stated assumptions on the data or explicitly list it as a standing hypothesis, so that the reader can see the full set of conditions on which the convergence rests.","section":"Section 4, uniform bound assertion"}],"minor_comments":[{"comment":"The norm in Eq. (46) is written with a square '(∫ |w|^2 dx + ∫ |∇w|^2 dx)^2'; it should be the square root, i.e. the exponent should be 1/2.","section":"Section 8, Eq. (46)"},{"comment":"The identity 'µ = λ = M 1/4 h′(0)' is notationally ambiguous; it should read 'µ = λ = (M/4) h′(0)' or an equivalent explicit product.","section":"Section 3, Eq. (25)"},{"comment":"The body-force definition 'b(x,t) = (0, f_0 h(t)/ε)' is missing a closing parenthesis, and the corresponding formula for the bottom layer should be written with matching parentheses for clarity.","section":"Section 6, first paragraph"},{"comment":"The prefactor (1+ν)/E in Eq. (43) should be checked against the plane-strain convention used elsewhere in the paper; for plane strain one generally expects an expression such as (1−ν^2)/E or an equivalent combination of elastic constants, and the present notation with α_t is not standard.","section":"Section 5, Eq. (43)"},{"comment":"Equation (39) mixes the limit ε→0 with an O(δ) term in a single display; it would be clearer to state the limit first and then bound the remainder uniformly for small δ, so that the order of the two limiting processes is unambiguous.","section":"Section 5, around Eq. (39)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's main theorem and kinetic relation are conditional on a substantial cluster of hypotheses and on a uniform-convergence assumption that are not proved from the peridynamic equations. The authors are transparent about most of these conditions, but the title and abstract claim 'recovery' without the qualifier 'conditional'. I would support publication after major revision if the authors (i) explicitly label the main results as conditional, (ii) prove or clearly isolate the uniform-convergence assumption in Proposition 7, (iii) present Eq. (44) as an imported classical formula rather than as a consequence of the nonlocal model alone, and (iv) either prove or state as a standing hypothesis the uniform L^∞ bound on u^{ε} used in Section 4. The numerical experiments are qualitative; a quantitative check of Hypothesis 3's ε^{-1} strain scaling would materially increase confidence. The paper is in scope for the journal and the core energy-balance derivation is a worthwhile contribution, but the current framing overstates the strength of the results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis is a real contribution, not a repackaging. Lipton and Jha show that a bond-based peridynamic double-well model, in the zero-horizon limit, yields not just the wave equation off the crack (already in their earlier work) but also the zero-traction condition on the crack lips and the energy-balance identity GcV = J. The nonlocal divergence theorem in Section 10 is a genuine piece of machinery, and the derivation of (56) from the nonlocal power balance is the right route: they avoid postulating power balance, which is what earlier convergence results did not do. The numerical section is honest supporting material for the hypotheses, not proof, and the paper says so.\n\nThe soft spot is exactly where the reader's report puts it. Propositions 2, 4, and 5 all lean on Hypotheses 1–3, and Proposition 7 adds an unproved uniform convergence assumption on u^eps, dot u^eps, and the strains near the contour. That uniformity is doing real work—replacing the nonlocal flux with the local flux in (124)–(129). Without it, the kinetic relation is not established. The authors acknowledge the hypotheses are corroborated only by simulation, so this is a conditional convergence theorem, not an unconditional recovery. The abstract's phrase 'recovered directly from the nonlocal model' overstates the current proof. Also, the explicit J(K_I,V) formula (43) is imported from Freund–Clifton, so what is derived is the energy-balance identity, not a fully determined crack-tip velocity law for general loading.\n\nIs this enough to reject? No. The conditional structure is clearly stated, the missing pieces are identifiable, and the limit statements are coherent. The paper deserves a serious referee who can check whether the uniformity assumption can be replaced by something derived from the peridynamic dynamics, and whether Hypotheses 2 and 3 can be weakened. For the peridynamics and computational-fracture community, this is a useful milestone even in conditional form. I would bring it to a reading group and would cite it for the nonlocal divergence theorem and the explicit zero-traction derivation.\n\nRecommendation: send to peer review. The referee should focus on Proposition 7 and the hypotheses, not desk-reject.\n\nBest,\n[You]","headline":"A serious paper that genuinely derives the crack-lip traction and kinetic-relation energy balance from a nonlocal model, but only under unproved hypotheses that the authors themselves flag.","tokens_in":28507,"tokens_out":595,"would_cite":true,"duration_ms":114609,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74R10","74H20","74B05","35L05"],"pacs":[],"model":"deepseek-v4-flash","headline":"As the peridynamic horizon shrinks to zero, a double-well cohesive fracture model converges to classical plane elastodynamics with a running crack and its crack-speed kinetic relation emerges from the nonlocal dynamics rather than being…","keywords":["peridynamics","dynamic fracture","nonlocal-to-local convergence","kinetic relation","double-well potential","mode I crack","fracture toughness","energy release rate"],"falsifier":"Run the single-edge-notch simulation with horizons $\\epsilon = 2.5, 1.25, 0.625$ mm and smaller, extract crack length versus time, and check whether the inferred crack-tip velocity satisfies (44) with the specified $G_c$; if the two sides of that relation diverge as $\\epsilon$ shrinks, the recovered kinetic relation (58) fails.","tokens_in":27517,"feed_emoji":"💥","tokens_out":5222,"duration_ms":53226,"temperature":0.7,"pith_summary":"This paper claims that a simple peridynamic model of brittle fracture—bond pairs interacting through a double-well potential over a horizon of length $\\epsilon$—has a well-defined limit as $\\epsilon\\to 0$, and that this limit is the classical dynamic fracture problem for a single edge notch in tension. In the limit, the displacement satisfies linear elastodynamics away from the crack, the crack lips are traction-free, and the crack tip moves according to the standard kinetic relation $G_c V(t)=\\mathcal{J}$. The authors' point is that this kinetic relation is not assumed and is not derived from an energy-balance postulate; it is recovered by taking the $\\epsilon=0$ limit of the nonlocal power balance once the softening zone is handled internally by the model. A sympathetic reader would care because it grounds the use of peridynamic simulations as a surrogate for classical fracture mechanics and explains how crack dynamics can emerge from a purely nonlocal field theory.","feed_headline":"Nonlocal fracture model recovers classic crack-speed law","feed_subtitle":"As the peridynamic horizon shrinks, the standard kinetic relation GcV = J emerges from the model instead of being assumed.","key_machinery":"The load-bearing object is the pairwise double-well cohesive potential $W_\\epsilon(S)=J_\\epsilon(|y-x|)\\epsilon^{-3}\\omega_2^{-2}|y-x|g(\\sqrt{|y-x|}S)$, with $g(r)=h(r^2)$ for concave $h$: linear elastic for small strain, softening past a critical strain, and zero force past $S_+$. Its horizon scaling makes the fracture toughness $G_c$ independent of $\\epsilon$, so the same material toughness is present at every horizon. The argument runs through a new nonlocal divergence theorem and the nonlocal power balance (53), together with the assumptions that the softening zone is a thin strip, that softened bonds immediately fail, and that the opening displacement stays positive; these allow the energy integral over the two sides of the crack face to deliver exactly $-G_c V(t)$ in the limit.","core_discovery":"The central claim is that solutions $u^\\epsilon$ of the nonlocal initial-boundary value problem (15), with forces from the double-well potential (3), converge in $L^2$ to a limit $u^0$ that is the weak solution of $\\rho \\ddot u^0 = \\operatorname{div}(C E u^0)+b$ on $D\\setminus J_{u^0}(t)$, with $C E u^0 n=0$ on the crack faces, and with crack-tip speed $V(t)$ tied to the energy flux $\\mathcal{J}$ by $G_c V(t)=\\mathcal{J}$. The paper proves this through Propositions 2\\textendash 5 (crack set equals jump set, momentum balance, zero traction) and Proposition 7, which passes the nonlocal power balance (53) to the limit and obtains $\\lim_{\\epsilon\\to0} \\frac{d}{dt}\\int_{P_\\delta^\\epsilon}(T^\\epsilon+W^\\epsilon)\\,dx = \\int_{\\Gamma_\\delta} C E u^0 n\\cdot \\dot u^0\\,ds - G_c V(t)+O(\\delta)$. Setting the internal-energy rate to zero gives the kinetic relation; substituting the classical flux formula (43) yields the semi-explicit relation (44). The paper emphasizes that this is a recovery of the kinetic relation from the model, not an input to it.","pith_inferences":["If this recovery extends to other loadings and geometries, peridynamic simulations with finite horizon could be used in reverse: measuring crack speed numerically and reading off the dynamic stress-intensity factor through the classical formula $\\mathcal{J}(K_I,V)$.","The paper notes that a Lennard-Jones-type potential is a natural next step; the same limit procedure could then deliver a sharp fracture model whose crack lips do not interpenetrate, a property the current bond-based well does not enforce.","A testable consequence is that crack speeds computed for successively smaller horizons should converge to the solution of (44); a systematic offset would identify which of Hypotheses 1\\textendash 3 or the uniform-convergence assumption needs modification."],"forward_implications":["For every horizon $\\epsilon>0$, the nonlocal evolution has a unique solution, so the convergence program supplies a well-posed path from nonlocal dynamics to the classical sharp-crack equations.","In the small-horizon limit, the crack set coincides with the jump set of the limiting displacement, and the normal traction on the crack lips vanishes.","The crack-tip velocity satisfies $G_c V(t)=\\mathcal{J}$ and the semi-explicit form (44), so the classic dynamic fracture criterion becomes a corollary of the nonlocal model rather than an extra postulate.","Because $G_c$ is independent of the horizon, shrinking $\\epsilon$ yields numerical crack evolutions that can be compared directly with classical predictions while avoiding explicit crack-tip tracking."],"supporting_citations":[{"why":"Supplies the earlier convergence of cohesive peridynamic evolutions to sharp-crack Griffith evolutions and the SBD compactness estimates used throughout.","marker":"[Lipton(2016)]"},{"why":"Establishes the small-horizon convergence of peridynamic dynamics to brittle fracture with bounded Griffith surface energy.","marker":"[Lipton(2014)]"},{"why":"Provides the semi-explicit energy-flux formula $J(K_I,V)$ that is substituted into the recovered kinetic relation to obtain (44).","marker":"[Freund and Clifton(1974)]"},{"why":"Defines the classical dynamic fracture model and kinetic relation that the limit evolution is shown to match.","marker":"[Freund(1990)]"},{"why":"Introduces the peridynamic reformulation of elasticity on which the nonlocal model is based.","marker":"[Silling(2000)]"},{"why":"Supplies the SBD function-space structure used to identify the limit jump set and to relate it to the crack set.","marker":"[Ambrosio, Coscia and Dal Maso(1997)]"},{"why":"Provides the dynamic stress-field representation underlying the energy-flux formula used for the kinetic relation.","marker":"[Atkinson and Eshelby(1968)]"}],"fun_headline_variants":["Peridynamic limit recovers classic crack speed law","Nonlocal model yields kinetic relation as horizon shrinks","Classic fracture dynamics as vanishing nonlocality limit","Crack kinetic relation recovered from peridynamic theory","Nonlocal to classic fracture convergence proves crack law"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the unproved Hypotheses 1 through 3 together with the uniform-convergence assumption in Proposition 7: the softening zone must have the assumed thin-strip shape, every softened bond must immediately fail, the crack opening displacement must stay bounded away from zero, and the nonlocal fields must converge uniformly away from the crack tip.","fun_headline_variants_meta":{"raw":{"variants":["Peridynamic limit recovers classic crack speed law","Nonlocal model yields kinetic relation as horizon shrinks","Classic fracture dynamics as vanishing nonlocality limit","Crack kinetic relation recovered from peridynamic theory","Nonlocal to classic fracture convergence proves crack law"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000583,"raw_usage":{"total_tokens":2725,"prompt_tokens":911,"completion_tokens":1814,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":527,"completion_tokens_details":{"reasoning_tokens":1739}},"tokens_in":527,"tokens_out":1814,"duration_ms":14452,"temperature":1.0,"reasoning_tokens":1739,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:02:46.454843+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the single-edge-notch simulation with horizons $\\epsilon = 2.5, 1.25, 0.625$ mm and smaller, extract crack length versus time, and check whether the inferred crack-tip velocity satisfies (44) with the specified $G_c$; if the two sides of that relation diverge as $\\epsilon$ shrinks, the recovered kinetic relation (58) fails.","supporting_citations":[{"cited_title":"Atkinson, C","cited_arxiv_id":null,"evidence_quote":"Provides the dynamic stress-field representation underlying the energy-flux formula used for the kinetic relation."}],"review_version":1}