{"id":"7cae2a50-9dc0-4b28-845c-f4b4b7e86cc9","arxiv_id":"1908.06268","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A reorganized cracking-elements method, GCEM, models quasi-brittle fracture by making crack openings global unknowns through borrowed center-node degrees of freedom in 9-node quadrilateral elements.","lead":"This paper presents a new finite element scheme for simulating cracks in concrete-like materials, called the Global Cracking Elements Method. It turns the center point of a 9-node element into a crack-opening variable so fracture can be modeled without remeshing or tracking crack paths.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The symmetric tangent substitution in Eq. 16 is not derived from the residual in Eqs. 14–15 and is internally inconsistent with the claim that K_s,new is nonsymmetric; the claimed Newton stability/efficiency advantage of GCEM over CEM is therefore unverified.","rationale":"The reader's weakest assumption identifies the load-bearing concern: the symmetric tangent substitution in Eq. 16 is unproved, and the central claim of GCEM depends on it for the claimed Newton stability and efficiency advantage. I refine the concern by noting an internal inconsistency: under the paper's own center-point quadrature, the displayed K_new in Eq. 15 appears symmetric, contradicting the text's assertion that K_s,new is nonsymmetric. Thus either the printed Eq. 15 has a sign error and K_sym is not a consistent tangent, or the stated motivation for K_sym is invalid. Either way, the derivation as written is not self-consistent, and the Newton iteration counts in Fig. 8 cannot be taken as evidence of robustness until the tangent is verified. The benchmark results themselves (force-displacement curves, peak loads, crack patterns) are plausible, and the paper honestly states its limitations (quadrilateral-only elements, no precise crack-tip tracking, element-wise disconnected cracks). No code or quantitative mesh-convergence metrics are provided, which strengthens the need for a concrete tangent check. A conditional acceptance is therefore appropriate: the method's formulation and examples are promising, but the linearization and hence the core efficiency/stability claim must be verified before the claim can be accepted as established.","tokens_in":14473,"tokens_out":14192,"duration_ms":146873,"concrete_test":"Re-derive the tangent by automatic differentiation of the residual in Eq. 10 at a representative cracked element (e.g., L-shaped panel, mesh I, d=0.3 mm) and check whether K_s,new as printed is symmetric. Then run the L-shaped panel and Brazilian disk benchmarks with (i) K_sym of Eq. 16 and (ii) the exact tangent using an unsymmetric sparse solver, comparing force-displacement curves, iteration counts, and convergence under a perturbed initial guess. If both implementations give the same results, the symmetrization is harmless; if K_sym diverges or changes the response where the exact tangent converges, the claimed stability/efficiency advantage of GCEM is unsupported.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central efficiency/stability claim (fewer Newton iterations than CEM, Sec. 1, Figs. 7–8) rests entirely on replacing the consistent tangent K_s,new of Eq. 15 by the symmetric matrix K_sym of Eq. 16. No proof or convergence study is given for this replacement. In fact, the situation is more problematic than a missing proof. Under the center-point/one-point quadrature used for the energy in Eq. 12, the off-diagonal blocks of the displayed K_new in Eq. 15 are transposes: (∫ B^T C_e B_ζ dΩ)^T = V (B_ζ)^T C_e B, and the added diag(0, V/l_c D) is symmetric, so K_s,new as printed is symmetric. The text says it is not symmetric. Either Eq. 15 contains a sign error (e.g., the lower-left block should retain -l_c as in Eq. 14), in which case K_sym is not the linearization of the residual, or the nonsymmetry claim is wrong, in which case the motivation for introducing K_sym disappears. In both cases, the numerical behavior of GCEM under general loading paths is not established. Moreover, the reported advantage is measured only by iteration counts, not by wall-clock time, and no comparison against an unsymmetric-solver implementation is provided, so the efficiency claim is not quantified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes the Global Cracking Elements Method (GCEM), a reformulation of the Cracking Elements Method (CEM) for quasi-brittle fracture. The key idea is to treat element-wise crack openings as global unknowns by borrowing the degrees of freedom of the center node of a Q9 element, thereby producing a pseudo-Q9 formulation that fits the standard Galerkin/FEM assembly framework. The kinematics, mixed-mode traction-separation law, and crack-initiation/propagation criteria are adopted from the authors' previous CEM publications. The new element is presented in matrix form, and numerical tests include an L-shaped panel and Brazilian disk tests with single and double slots, compared with experiments, XFEM, CEM, phase-field, and peridynamics results. The reported advantages are no remeshing, no enrichment, no crack tracking, and fewer Newton iterations than the original CEM.","tokens_in":14781,"tokens_out":6877,"duration_ms":63895,"significance":"If the formulation is fully justified, GCEM is a potentially attractive method for quasi-brittle fracture: the center-node borrowing idea is elegant, the implementation appears straightforward, the benchmark results match independent solutions and experiments, and the mesh-sensitivity evidence is favorable. The paper is also honest about current limitations, namely quadrilateral elements only and no direct crack-tip tracking. However, the central numerical claim, improved stability and efficiency over CEM, currently rests on an unproved replacement of the consistent tangent by a symmetric matrix, and the internal consistency of the tangent derivation is questionable. Thus the significance is conditional on resolving the tangent issue and on providing a more quantitative cost comparison.","major_comments":[{"comment":"The paper replaces K_s,new from Eq. (15) with the symmetric matrix K_sym in Eq. (16) with only the statement 'Fortunately, K_s,new can be replaced by the following symmetric matrix.' No derivation from the residual, no proof of solution invariance, and no convergence study are provided. This replacement is load-bearing because the Newton-Raphson iteration and the claimed efficiency and stability advantage in Section 3 (Figs. 7-8) depend on it. Please add a proof that K_sym corresponds to a consistent tangent or an equivalent potential, and a numerical study comparing convergence behavior and solutions obtained with K_sym against those obtained with the unsymmetric tangent.","section":"Section 2.3, Eq. (16)"},{"comment":"The statement that K_s,new is not symmetric is inconsistent with the quadrature implied by the formulation. With B^(e),1 evaluated at the center point and B_zeta constant, the off-diagonal blocks of K_new in Eq. (15) are transposes of each other, and the added diagonal block is symmetric, so K_s,new as printed is symmetric. If the intended matrix is instead the one in Eq. (14) with the lower-left block scaled by -l_c, then K_sym in Eq. (16) is not the linearization of the residual. Please clarify which matrix is actually assembled and resolve this sign and symmetry inconsistency; the current text leaves the reader unable to reproduce the tangent.","section":"Section 2.3, Eqs. (14)-(15)"},{"comment":"The efficiency claim is supported only by Newton-iteration counts for the L-shaped panel, not by wall-clock time and not by comparison with a solver that handles the unsymmetric tangent. Because GCEM moves the crack-opening unknowns into the global system, its per-iteration cost may be higher than CEM's element-level condensation. Please provide a cost comparison such as CPU time or total linear solves, or restrict the claim to iteration counts.","section":"Section 3, Figs. 7-8"}],"minor_comments":[{"comment":"The threshold fracture energy is fixed as Gf,0 = 0.001 Gf without comment or sensitivity study; since this parameter enters the initial stiffness and the tangent matrix D, please state the basis for this value and its influence on the reported results.","section":"Section 2.1, Eq. (2)"},{"comment":"There is a typo 'if if' before the convergence criterion; please also specify how the criterion is applied when the total elastic energy becomes small or changes sign.","section":"Section 3.1, Eq. (22)"},{"comment":"The crack-orientation criterion is adopted from [1]; it would be helpful to state explicitly whether n^(e) is updated during Newton iterations for an already-cracked element and to quantify the claim that the rotation of n^(e) is very small.","section":"Section 2.4, Eq. (18)"},{"comment":"The sentence 'Unlike methods based on equivalent-type theories. The GCEM does not introduce bonds/lattices/links...' is grammatically incomplete and should be rewritten.","section":"Section 4"},{"comment":"The paper would benefit from a short derivation showing how Eq. (12) and the residual in Eq. (14) follow from the weak form, in particular the appearance of the factor V/l_c in Eq. (15).","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's scope. The self-citation load is substantial because the core kinematics, traction law, and crack criterion come from references [1,15,30,31], but the global reformulation is a reasonable contribution if the tangent issue is resolved. I recommend major revision rather than rejection because the benchmark evidence is promising and the central issue is fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper shows a genuinely neat reformulation of the cracking elements method—borrowing the Q9 center-node DOFs to carry crack openings globally—and the benchmarks (L-shaped panel, slotted Brazilian disks) are credible, with small mesh dependency and reasonable agreement with XFEM, phase-field, and experiments. This is a legitimate engineering contribution, not a breakthrough, and it reads as honest incremental work.\n\nThe best part is the pseudo-Q9 construction: it turns element-wise crack openings into global unknowns without enrichment or crack tracking, and the resulting code fits a standard FEM framework. The stability improvement over CEM is plausible, and the L-shaped panel iteration counts (Fig. 8) do show fewer Newton steps.\n\nBut the formulation has a real soft spot at Eq. 15–16. The paper says K_s,new is not symmetric and therefore replaces it with K_sym. Under the paper's own one-point quadrature for the energy (Eq. 12), the off-diagonal blocks of K_new are transposes and the added diag(0, V/l_c D) is symmetric, so K_s,new as printed appears symmetric. Either the printed matrix has a sign error (the lower-left block should keep -l_c from Eq. 14), or the nonsymmetry claim is wrong. In the first case K_sym is not the linearization of the residual; in the second case the motivation for introducing it disappears. Either way, the numerical behavior under general loading is not established. That is a load-bearing concern because the efficiency/stability claim rests on this tangent.\n\nSecond, the efficiency claim is measured only by iteration counts, not wall-clock time, and there is no comparison against an unsymmetric solver. So the 'more efficient than CEM' claim is not quantified.\n\nThe paper also assumes Gf,0 = 0.001 Gf without sensitivity study, and the method is limited to quadrilaterals—both acknowledged.\n\nWho is this for? Anyone working with embedded strong discontinuities or smeared crack models in quasi-brittle solids. The pseudo-Q9 idea is worth knowing and citing. But the tangent inconsistency should be fixed or explained before I'd rely on it.\n\nWe didn't see code or data, and the paper is self-referential to prior CEM work, but that's not a problem per se. The benchmarks are external and the results compare well.\n\nRecommendation: worth a serious referee, but I'd ask for a corrected derivation of the tangent (or an explicit statement of the quadrature and why K_sym is consistent), plus at least one wall-clock comparison. If the authors can fix the tangent issue, this would be a solid journal paper.","headline":"Neat pseudo-Q9 DOF-borrowing reformulation of CEM with credible benchmarks, but the tangent symmetrization in Eq. 16 is unjustified—possibly inconsistent—and the efficiency claim is under-supported.","tokens_in":15286,"tokens_out":7146,"would_cite":true,"duration_ms":62009,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74S05","74R10","65N30"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims a Galerkin finite-element method that reuses the center-node degrees of freedom of a nine-node quadrilateral element as global crack openings, eliminating remeshing, enrichment, and crack tracking in quasi-brittle fracture.","keywords":["quasi-brittle fracture","cracking elements method","standard Galerkin formulation","strong discontinuity embedded approach","pseudo-Q9 element","crack propagation","traction-separation law","Brazilian disk test"],"falsifier":"A direct check would be to run the method on a few-element specimen under non-proportional loading (load, unload, then rotate the principal stress direction) and compare the Newton iteration counts and final crack openings with those obtained from the non-symmetric tangent version of the same element; if the symmetric version needs significantly more iterations or diverges, the central stability claim is not general.","tokens_in":14246,"feed_emoji":"🧱","tokens_out":13407,"duration_ms":107992,"temperature":0.7,"pith_summary":"This paper reorganizes the earlier Cracking Elements Method (CEM) so that it lives inside a standard Galerkin finite-element framework: each cracked element is treated as a nine-node quadrilateral whose center-node displacement degrees of freedom are reassigned to represent that element's two crack-opening components. Because the crack openings become global unknowns, the method no longer needs the CEM's inner element-level iteration. The authors argue that the resulting scheme captures crack initiation and propagation in quasi-brittle materials on irregular, relatively coarse meshes without remeshing, enrichment, or crack tracking, and with better numerical stability and fewer Newton-Raphson iterations than the CEM. They support this with an L-shaped panel test and Brazilian disk tests with single and double slots, comparing force-displacement curves, normalized peak loads, and crack patterns with experiments and with other numerical methods. The interest for a general reader is that if this holds, fracture simulation becomes as routine as ordinary nonlinear finite-element analysis, at least for problems that admit quadrilateral meshes.","feed_headline":"Crack openings borrow a node to skip remeshing","feed_subtitle":"Reusing a quadratic element's center node to track cracking cuts iterations and needs no remeshing","key_machinery":"The load-bearing object is the pseudo-Q9 cracking element, a strong-discontinuity embedded formulation in which each crack is represented by a displacement jump inside an element rather than by additional nodes. In it, the center-node degrees of freedom of a standard nine-node quadrilateral are borrowed, not deleted: they switch from describing the displacement of the center point to describing the normal and tangential crack openings ($\\zeta_n$, $\\zeta_t$) of that element. This promotion of crack openings to global unknowns is what removes the inner element-level iteration of the original CEM and yields a symmetric, sparse global stiffness matrix. The element's strain is written as $\\bar{\\varepsilon} = B U - B_\\zeta \\zeta$ (with $B_\\zeta$ built from the crack normal $\\mathbf{n}$ and the characteristic length $l_c$), and the practical stiffness is the symmetrized form $K_{\\mathrm{sym}}$ assembled from $[B, B_\\zeta]$, with the traction-separation tangent added in the crack-opening block. That symmetrized matrix carries the Newton-Raphson iteration in the numerical examples.","core_discovery":"The central claim is that a strong-discontinuity embedded approach to quasi-brittle fracture can be written so that the only extra unknowns are global, element-wise crack openings, and that this form is more robust and cheaper than its predecessor. The device is a 'pseudo-Q9' element: the perimeter nodes of a nine-node quadrilateral keep their displacement unknowns, and the center node's two displacement degrees of freedom are replaced by the normal and tangential crack openings $\\zeta_n$ and $\\zeta_t$ of that element. From then on the cracked element is an ordinary Galerkin finite element with a symmetric, sparse stiffness matrix, so no remeshing, enriched basis, or crack-tracking strategy is required. Crack orientation is chosen locally from the eigenvector of the strain, and propagation is found by expanding a propagation region rather than by following a crack tip. The numerical tests show the method reproducing the benchmark force-displacement curves and peak loads of the L-shaped panel and slotted Brazilian disks, with small mesh dependence and fewer Newton iterations than the CEM.","pith_inferences":["Beyond the paper, the center-node borrowing suggests a general recipe: any element with re-assignable degrees of freedom could carry kinematic enrichment inside a standard Galerkin code, so the idea may extend to shear bands or delamination without changing the global solver.","Beyond the paper, the loading paths shown are monotonically increasing, so an untested consequence is how the symmetric-tangent replacement behaves under unloading-reloading cycles or load-path rotations; those are exactly the situations where a non-symmetric tangent may matter.","Beyond the paper, a quantitative mesh-convergence study across several element sizes, rather than the two meshes shown, would pin down the claimed negligible mesh dependency."],"forward_implications":["Crack initiation and propagation in quasi-brittle materials can be simulated with no remeshing, no enrichment, and no crack-tracking strategy, using only standard Galerkin finite-element machinery and irregular meshes.","Because crack openings are global unknowns, the Newton-Raphson iteration needs fewer steps than in the original CEM; the paper reports this consistently across the benchmark meshes.","The pseudo-Q9 form keeps a symmetric sparse stiffness matrix, so existing sparse linear solvers and standard nonlinear finite-element codes can adopt the method without new global data structures.","The formulation can be changed to an enriched form of the crack-opening field when desired, since the pseudo-Q9 structure leaves a clear place for enrichment.","Presently only quadrilateral elements with nonlinear displacement interpolation are usable, so the method is not yet available for triangular or linear-element meshes."],"supporting_citations":[{"why":"Defines the original CEM whose element-wise crack openings, self-propagating criterion, and local crack orientation the GCEM reorganizes.","marker":"[1]"},{"why":"Supplies the strong-discontinuity kinematics, the energy-based characteristic length $l_c = V/A$, and the warning that the new term must remain in the stiffness to avoid stress locking.","marker":"[15]"},{"why":"Provides the mixed-mode traction-separation law used for the crack-opening behavior and an XFEM benchmark for the L-shaped panel.","marker":"[30]"},{"why":"Provides the experimental force-displacement data for the L-shaped panel that the GCEM results are compared against.","marker":"[35]"},{"why":"Provides the experimental Brazilian disk tests with single and double slots that define the benchmark peak loads and crack patterns.","marker":"[36]"},{"why":"Supplies phase-field simulation results for slotted Brazilian disks used as a comparison for normalized peak loads.","marker":"[37]"},{"why":"Supplies peridynamics simulation results for the same disk tests, giving a second comparison method for peak loads.","marker":"[38]"},{"why":"Describes the Newton-Raphson iteration process that the GCEM inherits and makes more efficient by globalizing the crack-opening unknowns.","marker":"[2]"}],"fun_headline_variants":["Global crack elements: no remeshing, no enrichment, no crack tracking","Quadrilateral trick: center node's DOFs become crack openings","New finite element method for quasi-brittle fracture without remeshing","Symmetric sparse stiffness matrix for fracture, no crack-tip chase"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method's stability and speed depend on an unproved simplification in which a non-symmetric matrix is replaced by a symmetric one in the element formulation; if that simplification is inaccurate for some loading paths, the claimed Newton-Raphson advantages may disappear.","fun_headline_variants_meta":{"raw":{"variants":["Global crack elements: no remeshing, no enrichment, no crack tracking","Quadrilateral trick: center node's DOFs become crack openings","New finite element method for quasi-brittle fracture without remeshing","Symmetric sparse stiffness matrix for fracture, no crack-tip chase"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001034,"raw_usage":{"total_tokens":4373,"prompt_tokens":982,"completion_tokens":3391,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":598,"completion_tokens_details":{"reasoning_tokens":3313}},"tokens_in":598,"tokens_out":3391,"duration_ms":23438,"temperature":1.0,"reasoning_tokens":3313,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:50:54.633223+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct check would be to run the method on a few-element specimen under non-proportional loading (load, unload, then rotate the principal stress direction) and compare the Newton iteration counts and final crack openings with those obtained from the non-symmetric tangent version of the same element; if the symmetric version needs significantly more iterations or diverges, the central stability claim is not general.","supporting_citations":[{"cited_title":"Cracking elements: a self-propa gating strong discontinuity embedded approach for quasi-brittle fractu re,","cited_arxiv_id":null,"evidence_quote":"Defines the original CEM whose element-wise crack openings, self-propagating criterion, and local crack orientation the GCEM reorganizes."},{"cited_title":"Strong disc ontinuity embedded approach with standard SOS formulation: Element formulation, energy-based crack- tracking strategy, and validations,","cited_arxiv_id":null,"evidence_quote":"Supplies the strong-discontinuity kinematics, the energy-based characteristic length $l_c = V/A$, and the warning that the new term must remain in the stiffness to avoid stress locking."},{"cited_title":"Energy-based modeling of cohesive and cohesionless cracks via X-FEM,","cited_arxiv_id":null,"evidence_quote":"Provides the mixed-mode traction-separation law used for the crack-opening behavior and an XFEM benchmark for the L-shaped panel."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the experimental force-displacement data for the L-shaped panel that the GCEM results are compared against."},{"cited_title":"Propagation and coalescence of quasi-static cracks in brazilian disks: an insight from a phase field model,","cited_arxiv_id":null,"evidence_quote":"Supplies phase-field simulation results for slotted Brazilian disks used as a comparison for normalized peak loads."},{"cited_title":"Numerical simulation of cra ck propagation and coalescence in pre-cracked rock-like brazilian disks usin g the non-ordinary state- based peridynamics,","cited_arxiv_id":null,"evidence_quote":"Supplies peridynamics simulation results for the same disk tests, giving a second comparison method for peak loads."},{"cited_title":"Cracking elements method for dyn amic brittle fracture,","cited_arxiv_id":null,"evidence_quote":"Describes the Newton-Raphson iteration process that the GCEM inherits and makes more efficient by globalizing the crack-opening unknowns."}],"review_version":1}