{"id":"6d486485-62df-482d-a6f0-17476b843f36","arxiv_id":"2507.04334","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An entropy-stable DGSEM of arbitrary order is formulated and validated on curvilinear hybrid meshes containing hexahedral, tetrahedral, prismatic, and pyramidal elements.","lead":"This paper constructs an entropy-stable discontinuous Galerkin spectral element method (DGSEM) of arbitrary order that works on curvilinear hybrid meshes combining hexahedral, tetrahedral, prismatic, and pyramidal elements. The approach uses collapsed coordinate transformations, Legendre-Gauss collocation, and modal time-stepping to obtain a stable and efficient scheme for compressible flow problems.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Triangular-facet projection in §3.2.1 lies outside the proof of Theorem 3.2, so entropy stability of the implemented hybrid scheme is not established for arbitrary face orientations.","rationale":"The reader's conditional verdict is appropriate, and my analysis independently identifies the same load-bearing gap: the implemented triangular-facet coupling is not part of the entropy-stability proof. The paper states this explicitly by deferring the non-unique-node proof to [74], so this is not a manufactured concern. I do not think the gap justifies rejection: the construction follows established weight-adjusted and mortar-based frameworks, and the numerical tests show plausibly entropy-conservative behavior on structured hybrid meshes. However, the central claim—entropy stability on arbitrary curvilinear hybrid meshes—is not fully proven until the projection operators in §3.2.1 are shown to satisfy the compatibility conditions needed for eq. (63), or a targeted numerical test demonstrates it for all possible face orientations. This is exactly the kind of missing support that warrants a conditional rather than an unconditional acceptance.","tokens_in":26824,"tokens_out":5949,"duration_ms":78800,"concrete_test":"Build a two-element Euler case with one tetrahedron and one prism sharing a triangular face, and enumerate all independent relabelings/reflections of the second element's local face vertex order so that the collapsed-coordinate face nodes are maximally non-coincident. For each orientation, evaluate the discrete entropy residual of eq. (63) at N=4 using the entropy-conservative Chandrashekar flux with no surface dissipation. If the residual is not at rounding-error level for at least one orientation, the §3.2.1 projection is not entropy-conservative and Theorem 3.2 does not cover the implemented scheme; if all orientations pass, repeat on a random unstructured hybrid mesh with dissipative Roe fluxes to verify the entropy inequality.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.2 proves entropy stability for the spatial operator (46), whose surface terms are assembled on each element's own (N+1)^2 collapsed-coordinate face nodes via V_f^T W_f f^*(w_f^+, w_f^-). On triangular faces between arbitrarily oriented tetrahedra, prisms, and pyramids, those nodes are not unique, so §3.2.1 replaces the native-node evaluation by: (i) interpolation of entropy-projected states to Xiao–Gimbutas quadrature nodes, (ii) evaluation of f^* there, and (iii) transfer of the flux back to the local element face nodes. The proof of Theorem 3.2 never introduces these three operators; it only states that the non-unique-node proof can be derived by following [74]. This is not cosmetic: unless the interpolation/transfer matrices satisfy a mortar-type compatibility condition, the telescoping of the entropy flux potential in eqs. (67)–(68) can fail, and eq. (63) need not hold for the code that is actually run. The TGV entropy tests use structured splits of hexahedra, so adjacent faces may have coincident or only mildly mismatched nodes; they do not certify the arbitrary-orientation case. The paper itself flags a second limitation—curved pyramids are excluded from convergence tests (§4.2)—but the face-coupling gap is the one that directly threatens the central theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an entropy-stable discontinuous Galerkin spectral element method (DGSEM) of arbitrary polynomial order on two- and three-dimensional hybrid curvilinear meshes containing triangles/quadrilaterals and tetrahedra, prisms, pyramids, and hexahedra. The construction uses a collapsed-coordinate (Duffy) transformation, Legendre-Gauss collocation with a generalized summation-by-parts operator, entropy-projected variables, and a modal time-stepping formulation based on a weight-adjusted inverse mass matrix. The hyperbolic operator is extended to the Navier-Stokes equations with a BR1 lifting procedure. The authors prove free-stream preservation (Theorem 3.1) and discrete entropy stability (Theorem 3.2) for the semi-discrete scheme, and validate the method through free-stream tests, manufactured-solution convergence studies, a weakly compressible Taylor-Green vortex, and an inviscid flow around the NASA Common Research Model.","tokens_in":27033,"tokens_out":4500,"duration_ms":51156,"significance":"If the central claim is correct, the paper delivers the first arbitrary-order entropy-stable DGSEM on curvilinear hybrid meshes with hexahedral, tetrahedral, prismatic, and pyramidal elements while retaining tensor-product computational efficiency. The combination of collapsed coordinates, generalized SBP operators, entropy projection, and weight-adjusted mass matrices is a natural and valuable extension of prior work by Chan and by Montoya and Zingg. The numerical evidence is substantial: free-stream preservation to machine precision, expected h- and p-convergence rates on curved non-simplex elements, entropy conservation/stability for the TGV test case, and a large-scale CRM simulation. The implementation in the open-source FLEXI framework is a practical strength. However, the proof of the central Theorem 3.2 leaves a load-bearing gap for the triangular-facet coupling implemented in Section 3.2.1, and the convergence study explicitly avoids curved pyramids, which narrows the advertised scope.","major_comments":[{"comment":"The entropy-stability theorem is stated for the spatial operator in eq. (46), whose surface terms are assembled on each element's native collapsed-coordinate face nodes through V_f^T W_f f^*. On triangular faces between arbitrarily oriented tetrahedra, prisms, and pyramids, Section 3.2.1 replaces this evaluation by a three-step procedure: interpolation to Xiao-Gimbutas triangle quadrature nodes, evaluation of the numerical flux there, and interpolation of the flux back to the local element face nodes. The proof of Theorem 3.2 never introduces these interpolation/transfer operators, and the statement that the non-unique-node proof 'can be derived by directly following [74]' is not a derivation; moreover, [74] concerns mortar methods on non-conforming hexahedral meshes, not the collapsed-coordinate projection used here. Unless a mortar-type compatibility condition is established for these projection operators, the telescoping of the entropy flux potential in eqs. (67)-(68) need not hold for the implemented scheme. The TGV entropy tests in Section 4.3.1 use structured splits of hexahedra, so the face-node mismatch is mild; they do not certify arbitrary face orientations. This is a load-bearing gap that the authors should close with a proof or with careful numerical experiments on genuinely unstructured hybrid interfaces.","section":"§3.2.1 and Theorem 3.2"},{"comment":"The convergence study does not support the full 'curvilinear hybrid mesh' claim for pyramids. The text states that 'convergence cannot be guaranteed for curved pyramids' and that pyramid nodes are left unperturbed in the PYRA (uncurv) and MIX cases. Consequently, the p- and h-convergence results in Fig. 3 demonstrate convergence on curved tetrahedra, prisms, and hexahedra, but only on straight pyramids. The authors should either prove or numerically demonstrate convergence and entropy stability on genuinely curved pyramids, or they should explicitly state in the abstract and conclusion that the method's robustness for pyramids is currently established only for rectilinear pyramidal elements.","section":"§4.2"},{"comment":"The proof of Theorem 3.1 contains a notational inconsistency that obscures the conservation argument. After invoking Q1=0 and the metric identities, the proof writes 'M(κ)˜q(κ)_t = −(1(nq))^⊤ Σ_ζ (V_f^ζ)^T W_f ...', which appears to equate a modal vector with a row-vector expression. The subsequent sentence 'with V 1(M(N)) = 1(nq)' does not resolve how the transpose of the modal-to-nodal Vandermonde enters. This intermediate step should be rewritten to show explicitly how the volume and surface contributions combine into the global conservation statement.","section":"§3.3, proof of Theorem 3.1"}],"minor_comments":[{"comment":"The basis function for hexahedral elements is written as 'P 0,0)' with a missing left parenthesis; it should read P^{(0,0)}_j(η_2).","section":"Eq. (34)"},{"comment":"The caption states 'Left: p-convergence ... Right: h-convergence', but the panels themselves are labeled '2D:h-convergence' and '2D:p-convergence' with the h-axis on the left and the N-axis on the right; the caption appears to reverse the panels.","section":"Figure 3 caption"},{"comment":"The theorem statement contains the typo 'free-steam preserving'; this should be 'free-stream preserving'.","section":"Theorem 3.1"}],"recommendation":"major_revision","confidential_remarks":"The central novelty is incremental but well positioned for JCP: it extends existing generalized-SBP and weight-adjusted mass matrix frameworks to prisms and pyramids. The key issue is the proof gap in Section 3.2.1 for triangular-facet coupling; if the authors can either provide a compatibility proof for the Xiao-Gimbutas projection or demonstrate entropy stability on genuinely unstructured hybrid interfaces, the manuscript would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a real step forward, but the paper's central theorem does not cover the code that is actually run on arbitrary hybrid meshes. If that gap is closed in revision, this is a solid JCP paper.\n\nWhat is new: Keim et al. extend the tensor-product collapsed-coordinate DGSEM from triangles and tetrahedra to prisms and pyramids, and combine all four element types in one mesh. The machinery—generalized SBP on Legendre–Gauss nodes, entropy projection, weight-adjusted modal time stepping—is largely inherited from Chan and from Montoya–Zingg, but assembling it on mixed hex/prism/tet/pyramid meshes is genuinely new and practically relevant. The numerical campaign is thorough: free-stream preservation, h/p convergence, entropy behavior in the TGV, and a large CRM run as a proof of concept. The CRM case is correctly labeled as such. Citations to Chan, Montoya–Zingg, and the mortar paper are appropriate; self-citation is not an issue here.\n\nThe soft spot is real and matches the stress-test. Section 3.2.1 replaces the native face nodes on triangular facets by Xiao–Gimbutas nodes, evaluates the flux there, and maps it back. Theorem 3.2 proves entropy stability only for the operator in eq. (46), where surface fluxes are assembled on each element's own (N+1)^2 nodes. The projection and transfer operators never enter the proof; the paper says the non-unique-node case follows from [74]. That is a nontrivial leap. Without a mortar-type compatibility condition, the telescoping in eqs. (67)–(68) can break, and the discrete entropy inequality (63) is not established for the implemented scheme. The numerical tests use structured splits of hexahedra, so adjacent faces are either coincident or mildly mismatched; they do not certify arbitrary orientations. The curved-pyramid exclusion from the convergence tests is a second, smaller limitation, and the paper does mention it.\n\nNothing here looks like a load-bearing mathematical error. The framework is established, the derivation is not circular, and the evidence is consistent. The gap is a missing proof for a step that is central to the hybrid claim. That should be fixable, but it needs to be fixed before the claim about arbitrary curvilinear hybrid meshes is made.\n\nWho should read this: people doing entropy-stable DG on mixed-element CFD meshes, and anyone implementing tensor-product operators on collapsed coordinates. It deserves a serious referee. I would send it out, with instructions to focus on §3.2.1 and to ask for a genuinely unstructured hybrid test case, plus either a proof of the projection's entropy stability or a clear statement of the compatibility condition under which it holds.","headline":"Solid extension to prisms/pyramids, but the entropy-stability proof skips the triangular-face projection that the code actually uses.","tokens_in":27623,"tokens_out":4943,"would_cite":true,"duration_ms":52053,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65M60","65M70","65N30"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs an entropy-stable DGSEM of arbitrary order on curvilinear hybrid meshes of hexahedra, prisms, tetrahedra, and pyramids, and proves a discrete entropy balance in Theorem 3.2.","keywords":["entropy stability","discontinuous Galerkin spectral element method","collapsed coordinate transformation","summation-by-parts operators","curvilinear hybrid meshes","compressible flow equations","pyramidal elements","modal time stepping"],"falsifier":"Run the entropy-conservative variant, without interface dissipation, on a distorted hybrid mesh with tetrahedron-prism and tetrahedron-pyramid interfaces, and track the discrete integral entropy: if it drifts beyond round-off over the vortex-decay timespan, the triangular-face projection is violating the entropy-conservation condition. Separately, perform a manufactured-solution h-convergence study on a genuinely curved pyramid mesh: if the error stagnates rather than converging at the expected order, the metric terms from the collapsed pyramid mapping are not sufficiently accurate.","tokens_in":26585,"feed_emoji":"🧊","tokens_out":10041,"duration_ms":104606,"temperature":0.7,"pith_summary":"High-order discontinuous Galerkin methods are accurate and flexible, but on mixed unstructured meshes they can lose stability when flow features are underresolved. This paper constructs an entropy-stable discontinuous Galerkin spectral-element method (DGSEM) of arbitrary polynomial order on curvilinear hybrid meshes containing hexahedral, prismatic, tetrahedral, and pyramidal elements, and their two-dimensional counterparts. The scheme's central promise is that the tensor-product efficiency and the provable entropy balance of hexahedral DGSEM survive on the full set of element shapes that real meshes need. Theorem 3.2 states that the semi-discrete scheme is discretely entropy-stable: the discrete integral entropy changes only through numerical interface dissipation, which prevents unphysical states from developing in underresolved compressible flow. Numerical tests verify free-stream preservation, polynomial and grid convergence, and entropy conservation when an entropy-conservative flux is used.","feed_headline":"Entropy-stable DGSEM reaches prisms and pyramids","feed_subtitle":"A high-order scheme keeps provable entropy balance on mixed curved meshes of all four element types.","key_machinery":"The load-bearing device is the collapsed coordinate transformation: a map from the unit cube to each polytopal reference element (prism, pyramid, tetrahedron, or triangle) that collapses cube faces onto edges or vertices. This turns all element shapes into tensor-product quadrature structures on the cube, so a one-dimensional Legendre–Gauss quadrature and differentiation operator can be applied in each direction. The discrete summation-by-parts identity of the generalized SBP operator in eq. (45) is what converts local volume integrations into boundary terms and makes the entropy-balance proof go through. A second mechanism is the weight-adjusted modal formulation: coefficients of the orthogonal modal basis are advanced in time, with the inverse of the curved mass matrix approximated by eq. (60), which keeps tensor-product cost and avoids time-step collapse. To evaluate fluxes on shared triangular faces, the solution is projected onto a fixed triangle quadrature rule, producing unique nodes from either side of the interface.","core_discovery":"The paper claims to be the first entropy-stable DGSEM formulation that works on curvilinear hybrid meshes of hexahedra, prisms, tetrahedra, and pyramids. The construction maps every polytopal reference element to the unit cube via a collapsed coordinate transformation, then applies Legendre–Gauss collocation with a generalized summation-by-parts operator and entropy-projected variables. Time evolution uses modal coefficients and a weight-adjusted approximation of the inverse mass matrix to avoid the small time steps that collapsing usually causes; viscous fluxes are added with a lifting procedure. Theorem 3.2 proves that the semi-discrete scheme in eq. (61) with the spatial operator in eq. (46) is discretely entropy-stable and satisfies the entropy balance in eq. (63), with equality when the numerical flux is entropy conservative. The proof covers quadrature nodes on triangular faces after projection to a single triangle quadrature rule, while the analytic entropy-stability argument for such non-unique nodes is deferred to a cited earlier work and demonstrated numerically.","pith_inferences":["A natural next step, not pursued in the paper, is to extend the proof from semi-discrete to fully discrete stability by pairing the spatial operator with an entropy-stable time integrator; the current tests rely on small time steps to keep temporal error negligible.","The projection of triangular-face quadrature points to a single triangle rule is mortar-like in spirit, so the same construction could extend to nonconforming or hanging-node hybrid meshes, where face nodes from different levels of refinement must be matched.","Because convergence for curved pyramids is not guaranteed in the paper and pyramid nodes are left unperturbed in the convergence tests, a practical hybrid-mesh strategy may either avoid curved pyramids or split them; a pyramid-only curvilinear refinement study would settle how much of the claimed geometric flexibility survives.","The weight-adjusted modal formulation could be reused for p-adaptivity, since modal coefficients are the evolved unknowns and element-local polynomial degree changes map naturally through the same Vandermonde matrices; this is an extension the authors do not discuss."],"forward_implications":["Entropy-stable high-order simulations become possible on curved meshes that mix hexahedra with prisms, tetrahedra, and pyramids, without splitting elements into hexahedra.","The tensor-product cost scaling of DGSEM, roughly $O(N^{d+1})$ operations per element for polynomial degree $N$ in $d$ dimensions, carries over to the non-hexahedral element types.","Modal time evolution with the weight-adjusted inverse mass matrix removes the severe explicit time-step restriction that collapsed coordinates would otherwise impose.","With an entropy-conservative two-point flux and no surface dissipation, the scheme is discretely entropy conservative; adding a stable dissipative interface flux makes it entropy stable.","The extension to viscous compressible flows via a lifting procedure means the same hybrid-mesh entropy framework applies beyond the inviscid equations."],"supporting_citations":[{"why":"Establishes the collapsed-coordinate tensor-product SBP approach on triangles and tetrahedra that this paper extends to prisms and pyramids.","marker":"[41]"},{"why":"Supplies the generalized SBP operator and entropy-stable DG framework used to write the spatial operator in eq. (46).","marker":"[20]"},{"why":"Provides the weight-adjusted inverse mass matrix, eq. (60), and the weight-adjusted projection identities used in the modal time stepping and entropy proof.","marker":"[73]"},{"why":"The lifting procedure used to approximate viscous flux gradients for the hyperbolic-parabolic extension.","marker":"[58]"},{"why":"Triangle quadrature nodes are the fixed point set onto which triangular-face solutions are projected so the left and right flux evaluations share unique nodes.","marker":"[71]"},{"why":"Cited as the source of the analytic entropy-stability argument for non-unique quadrature nodes, which the paper defers to.","marker":"[74]"},{"why":"Entropy-conservative two-point flux used in the numerical entropy-conservation experiments.","marker":"[77]"},{"why":"The entropy-conservation condition for the two-point numerical flux that the scheme's construction relies on.","marker":"[21]"}],"fun_headline_variants":["First entropy-stable DGSEM for prisms and pyramids","All four element types now entropy-stable in DGSEM","Collapsed coordinates give entropy-stable DG on any mesh","Hybrid meshes, one entropy-stable high-order scheme"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the projection used to make triangular-face quadrature nodes unique preserves the discrete entropy balance and quadrature accuracy for every interface orientation, with curved-pyramid convergence left unproved and deliberately avoided in the tests.","fun_headline_variants_meta":{"raw":{"variants":["First entropy-stable DGSEM for prisms and pyramids","All four element types now entropy-stable in DGSEM","Collapsed coordinates give entropy-stable DG on any mesh","Hybrid meshes, one entropy-stable high-order scheme"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000869,"raw_usage":{"total_tokens":3828,"prompt_tokens":1069,"completion_tokens":2759,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":685,"completion_tokens_details":{"reasoning_tokens":2689}},"tokens_in":685,"tokens_out":2759,"duration_ms":20092,"temperature":1.0,"reasoning_tokens":2689,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:48:51.687914+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the entropy-conservative variant, without interface dissipation, on a distorted hybrid mesh with tetrahedron-prism and tetrahedron-pyramid interfaces, and track the discrete integral entropy: if it drifts beyond round-off over the vortex-decay timespan, the triangular-face projection is violating the entropy-conservation condition. Separately, perform a manufactured-solution h-convergence study on a genuinely curved pyramid mesh: if the error stagnates rather than converging at the expected order, the metric terms from the collapsed pyramid mapping are not sufficiently accurate.","supporting_citations":[{"cited_title":"Montoya, D","cited_arxiv_id":null,"evidence_quote":"Establishes the collapsed-coordinate tensor-product SBP approach on triangles and tetrahedra that this paper extends to prisms and pyramids."},{"cited_title":"Bassi, S","cited_arxiv_id":null,"evidence_quote":"The lifting procedure used to approximate viscous flux gradients for the hyperbolic-parabolic extension."},{"cited_title":"Chandrashekar, Kinetic energy preserving and entropy stable finite volume schemes for compressible Euler and Navier-Stokes equations, Commun","cited_arxiv_id":null,"evidence_quote":"Entropy-conservative two-point flux used in the numerical entropy-conservation experiments."}],"review_version":1}