FEM methods & multiphysics · 108
Shared FEM machinery: H1/H(div) elements, transfers, PML, solvers, and coupled multiphysics contracts. Method shown on hover; tolerance is the pass gate. back to QA
| Case | Tier | Reference | Gate |
|---|---|---|---|
fem.h1-curved-p3-transfer.v1 | physics-gate | same-mesh complete global-state identity | maximum absolute global-DOF transfer error <= 2e-10 |
fem.adaptive-lineage-transfer.v1 | physics-gate | linear H1 and H(curl) fields, zero RT0/RT1 states, stale-history rejection | H1 error <= 2e-12; H(curl) error <= 2e-10; two typed rejections |
fem.adaptive-lineage-chain.v1 | physics-gate | linear H1/H(curl) fields and finite zero RT0/RT1 states | H1 error <= 2e-12; H(curl) error <= 2e-10; two typed rejections |
fem.adaptive-mpc-transfer.v1 | physics-gate | constant H1 state, zero vector states, complex-phase and constraint-residual rejection | H1 error <= 2e-10; at least two typed MPC rejections |
fem.spatial-source-locator.v1 | physics-gate | analytic linear scalar transfer, design transfer finiteness, outside/non-finite rejection | scalar transfer error <= 2e-12; at least two typed rejection checks |
fem.complex-dirichlet-elimination.v1 | physics-gate | independent dense complex system with constant prescribed boundary state | boundary and interior errors <= 1e-12; at least two rejection checks |
fem.cross-space-assembly.v1 | physics-gate | integral of each affine P1 basis against an element-local constant state | local mass error <= 2e-12; at least one typed rejection |
fem.hdiv-curved-cross-space-assembly.v1 | physics-gate | finite/nonzero curved RT mass, swapped-order transpose parity, malformed-topology rejection | curved transpose error <= 2e-10; nonzero vector block; at least one typed rejection |
fem.hdiv-divergence-cross-space-assembly.v1 | physics-gate | finite/nonzero physical divergence block, swapped-order transpose parity, malformed-topology rejection | transpose error <= 1e-12; nonzero physical block; at least one typed rejection |
fem.hdiv-curved-divergence-cross-space-assembly.v1 | physics-gate | independent published Piola reference, finite/nonzero curved blocks, swapped-order transpose parity, malformed-topology rejection | transpose error <= 2e-10 in both dimensions; nonzero physical blocks; typed rejection in both dimensions |
fem.mixed-full-assembly.v1 | physics-gate | independent four weighted mass forms with analytic rectangular block dimensions | block/monolithic apply error <= 2e-12; lower-right block nonempty; at least one typed rejection |
fem.h1-mixed-order-transfer.v1 | physics-gate | analytic quadratic and cubic polynomial fields | both transfer errors <= 2e-12; at least two typed rejections |
fem.h1-curved-p2-transfer.v1 | physics-gate | same-mesh complete global-state identity with curved edge midpoint | curved transfer error <= 2e-11; at least two typed rejections |
fem.cross-source-vector-transfer.v1 | physics-gate | constant H(curl) parity, finite RT0/RT1 transfer, typed outside-domain rejection | H(curl) error <= 2e-10; at least one typed rejection |
fem.hdiv-rt1-refinement-transfer.v1 | physics-gate | finite multi-element refinement, 3-D identity parity, typed outside/order rejection | 3-D identity error <= 2e-8; two typed rejections |
fem.pml-open-boundary-reference.v1 | physics-gate | analytically continued outgoing plane wave with independent attenuation integral | physical error <= 2e-3; full error <= 3e-3; reflection ratio <= 3e-3; residual <= 1e-11 |
fem.pml-directional-contract.v1 | physics-gate | independent complex coordinate-stretch coefficient identities on all 2-D sides and XY corner | 28 finite coefficient/direction checks pass to 1e-12 |
fem.validation-matrix.v1 | smoke | explicit maturity/reference/gate records | exact |
fem.element-contracts.v1 | smoke | exact nodal partition/unisolvence, lowest-order DOF counts, edge traces, orientation parity, constant-curl property, and independent curl(grad) finite-difference check | exact counts; 2e-13 identity residual; finite-difference curl(grad) <= 5e-8 |
fem.dg-p3-contract.v1 | physics-gate | independent P3 nodal partition, contiguous DG numbering, expected face traces, and cell-bubble exclusion | element_checks >= 4; boundary_trace_checks >= 4; partition residual <= 2e-12 |
fem.h1-p3-curved-geometry-reference.v1 | physics-gate | independent cubic nodal coordinates, partition of unity, positive Jacobian, finite gradients, and symmetry | partition residual <= 2e-12; symmetry residual <= 2e-10; positive finite Jacobian |
fem.operator-invariants.v1 | physics-gate | constant-field diffusion conservation, transpose symmetry, unit-domain mass, positive quadratic energy | 5e-13 residuals; mass total = 1 within 5e-13; energy > 0 |
fem.high-order-orientation-patch.v1 | smoke | analytic quadratic polynomial reproduction and canonical Nedelec edge parity | P2 value/gradient residuals <= 2e-12; Nedelec parity residual <= 2e-13 |
fem.p3-tetra-trace-orientation.v1 | physics-gate | independent cubic polynomial and shared-face continuity reference | two reference checks pass below the focused trace threshold |
fem.p3-tetra-reproduction.v1 | smoke | analytic cubic polynomial on affine tetrahedral elements | cubic interpolation residual <= 5e-12 |
fem.p3-h1-triangle-quad-reproduction.v1 | physics-gate | independent cubic polynomial on affine P3 triangles and quadrilaterals | cubic interpolation residual <= 2e-11 |
fem.vector-laplacian-convergence.v1 | physics-gate | independent sine manufactured vector field with analytic vector-Laplacian source | L2 refinement rate >= 1.5 and H1-seminorm refinement rate >= 0.75 |
fem.waveholtz-direct-reference.v1 | physics-gate | separate assembled direct Helmholtz solve | temporal order >= 1.8 and all relative reference errors <= 1e-3 |
fem.nonlinear-tangent.v1 | physics-gate | quadratic residual Taylor remainder under deterministic state/direction | order >= 1.8 across four steps |
fem.mesh-transfer-adjoint.v1 | physics-gate | independent weighted primal/adjoint pairing identity | 1e-11 |
fem.mesh-transfer-design.v1 | physics-gate | source element design values and analytic area-weighted integral | 1e-11 |
fem.helmholtz-plane-wave-pollution.v1 | physics-gate | independent exact oblique complex plane wave | focused test convergence and finite residual/phase gates pass |
fem.helmholtz-mms-convergence.v1 | physics-gate | closed-form plane wave, MMS, and scattering fields | nonzero free unknowns on every leg; quadrature L2 <= 3e-3; far-field solved L2 <= 4e-2; decay mismatch <= 0.3; convergence mean rate >= 1.5 |
fem.biharmonic-3d.v1 | contract-gate | analytical mixed block structure | exact block structure |
fem.electrochemistry-3d.v1 | contract-gate | partition of unity and prescribed-coordinate bilinear energy identities | 2e-10 |
fem.semiconductor-3d.v1 | contract-gate | partition of unity and prescribed-coordinate bilinear energy identities | 2e-10 |
fem.thermoelectric-3d.v1 | contract-gate | partition of unity and prescribed-coordinate bilinear energy identities | 2e-10 |
fem.biharmonic-mms-3d.v1 | physics-gate | u=sin(pi*x)sin(pi*y)sin(pi*z), q=3*pi^2*u, f=9*pi^4*u | u <= 0.02; q <= 0.01; both u rates >= 1.7; both q rates >= 1.7; residual <= 1e-10 |
fem.aeroacoustics-drp-reference.v1 | reference-only | analytical characteristic phase/group dispersion and Fourier-mode propagation | convergence rate >= 3.5; Fourier-mode relative L2 <= 2e-5 |
coupling.conservative-power-contracts.v1 | smoke | constant preservation, weighted adjoint identity, analytic power accounting, typed invalid-input rejection | trace and power residuals <= 1e-12; two invalid inputs rejected |
coupling.body-fitted-monolithic.v1 | physics-gate | finite driven states, two-driver residual blocks, and boundary-spanning brace patch area | all residual blocks <= 1e-7; body-fitted mesh and brace patch non-empty |
coupling.body-fitted-ported-monolithic.v1 | physics-gate | production body-fitted FEM/BEM monolithic port continuity and driver blocks | all residual blocks and port pressure jump <= 1e-7 across 60/90/120 Hz; finite non-empty port flow |
coupling.scheduler-preconditioner-contracts.v1 | smoke | scheduler quota/refund counters, CPU completion bridge, saddle-point identity application, typed rejection paths | exact counters and rejection counts; identity error <= 1e-12 |
fem.schwarz-pml-reflection.v1 | physics-gate | direct sparse Helmholtz solve on the identical mesh and RHS | relative error < 1e-3 for target reflection 1e-2, 1e-4, and 1e-6 |
fem.hdiv-rt0-reference.v1 | physics-gate | independent affine RT0 moments, Piola metric, and unit-flux divergence identity | mass <= 2e-11; divergence/orientation <= 2e-10; mixed conservation <= 2e-11 |
fem.hdiv-rt0-3d-reference.v1 | physics-gate | independent tetrahedron/simplex and tensor-product hexahedron RT0 moment integrals | mass <= 5e-11; divergence/permutation <= 5e-10; mixed conservation <= 5e-11 |
fem.hdiv-rt1-triangle-reference.v1 | physics-gate | independent edge/cell moment quadrature, finite-difference divergence, topology, orientation, and admission checks | five independent RT1 reference checks pass |
fem.hdiv-rt1-quadrilateral-reference.v1 | physics-gate | independent edge/cell moment quadrature, finite-difference divergence, topology, and shared-edge orientation | four independent RT1 quadrilateral reference checks pass |
fem.hdiv-rt1-tetra-reference.v1 | physics-gate | independent tetrahedral face/cell quadrature, finite-difference divergence, topology and Piola checks | three independent RT1 tetrahedron reference checks pass |
fem.hdiv-rt1-hexa-reference.v1 | physics-gate | independent tensor-product face/cell quadrature, finite-difference divergence, topology Piola checks | three independent RT1 hexahedron reference checks pass |
fem.hdiv-curved-p2-reference.v1 | physics-gate | independent P2 isoparametric physical-coordinate reconstruction and physical-divergence finite difference | two independent curved H(div) reference checks pass |
fem.piola-consumer-contract.v1 | physics-gate | production checked tabulation, topology-owned shared-edge orientation, and typed singular/non-finite Jacobian rejection | all accepted values finite; exact typed rejection variants; minimum checks 4/2/4 |
fem.natural-boundary-contract.v1 | physics-gate | independent unit-square/unit-cube boundary resultants and P1 Robin mass/RHS integrals | resultants/integrals <= 1e-12; two typed rejection paths leave systems unchanged |
fem.mpc-system-contract.v1 | physics-gate | independent P^H A P/P^H b fixtures, recovered periodic residual, and ordering/error checks | analytic values <= 1e-12; recovery residual <= 1e-14; three typed rejection checks |
fem.hanging-node-contract.v1 | physics-gate | independent affine edge parameters, conforming-mesh no-op, typed malformed/order/tolerance rejection | geometric admission tolerance 1e-10; interpolation errors <= 1e-14; four typed rejection checks |
fem.higher-order-hanging-node-contract.v1 | physics-gate | independent coarse-edge/triangular-face trace partition, polynomial recovery, typed rejection | geometric admission tolerance 1e-10; interpolation errors <= 1e-10; five typed rejection checks |
fem.higher-order-natural-boundary-contract.v1 | physics-gate | independent unit-square/unit-cube constant resultants, P2 Robin mass/RHS, P2/P3 curved trace finiteness | resultants/integrals <= 1e-10; finite geometry; transactional rollback |
fem.hdiv-rt1-mixed-conservation.v1 | physics-gate | independent quadrature of RT1 Piola divergence against constant P1/Q1 field | four domain checks constant-field conservation contract pass |
fem.solver-failure-contract.v1 | smoke | shared SolverStatus and AccuracyStatus mapping for FEM solver errors | all four status/accuracy mappings match the shared contract |
fem.prebaked-checked-contract.v1 | physics-gate | independent malformed-triplet admission and valid CSR merge contract | valid merge succeeds; invalid triplets reject |
fem.p2-poisson-convergence-3d.v1 | physics-gate | independent quartic manufactured Poisson solution and refinement rates | finite residuals and expected L2/H1 refinement rates |
fem.curved-p2-h1-geometry.v1 | physics-gate | independent P2 nodal interpolation and reference-coordinate gradient finite difference | positive Jacobian and finite-difference Hessian error |
fem.mixed-hdiv-conservation.v1 | physics-gate | independently integrated RT0 divergence and P1 partition of unity | maximum conservation error <= 2e-12 at refinements 1, 2, 4 |
fem.curved-p2-assembly-reference.v1 | physics-gate | independent curved-P2 stiffness symmetry conservation energy checks | symmetry <= 2e-12; constant-field row sums <= 2e-11; positive energy |
fem.hdiv-curved-rt1-mixed-physical-reference.v1 | physics-gate | independent Q2 physical map and RT1/P1 mixed divergence quadrature | all 288 checks pass; max block error <= 2e-10; minimum Jacobian determinant >= 0.05 |
fem.vector-natural-boundary-contract.v1 | physics-gate | independent unit-square/unit-cube constant traction resultants, midpoint loading, and rollback | constant resultants <= 1e-10; finite callbacks; transactional rollback |
fem.higher-order-hexa-hanging-node-contract.v1 | physics-gate | independent four-patch coarse-face partition, constant reproduction, polynomial recovery, typed rejection | geometric admission tolerance 1e-10; interpolation and recovery errors <= 1e-9 |
fem.vector-h1-hanging-node-contract.v1 | physics-gate | independent scalar/vector constraint-count multiplication, polynomial recovery, zero-component rejection | geometric admission tolerance 1e-10; imaginary recovery error <= 1e-14 |
fem.hdiv-rt1-hanging-edge-contract.v1 | physics-gate | independent RT1 edge traces, signed orientation, MPC recovery, topology/order rejection | geometric admission tolerance 1e-10; coefficient/recovery error <= 1e-12 |
fem.hdiv-rt0-hanging-edge-contract.v1 | physics-gate | independent positive/negative half-edge checks, MPC recovery, and unsupported-input rejection | geometric admission tolerance 1e-10; signed-fraction errors <= 1e-12; imaginary recovery error <= 1e-14 |
fem.hdiv-rt1-hanging-tri-face-contract.v1 | physics-gate | independent owner-face quadrature, signed subface fractions, recovery, unsupported-input rejection | geometric admission tolerance 1e-10; coefficient/recovery errors <= 1e-11 |
fem.hdiv-rt0-hanging-tri-face-contract.v1 | physics-gate | independent one-third subface checks, conforming no-op, recovery, and unsupported-input rejection | geometric admission tolerance 1e-10; signed-fraction errors <= 1e-12; imaginary recovery error <= 1e-14 |
fem.hdiv-rt0-periodic-face-contract.v1 | physics-gate | cyclic/reversed face orientation, complex MPC recovery, typed rejection | phase/recovery residual <= 1e-12 |
fem.hdiv-rt1-periodic-face-contract.v1 | physics-gate | cyclic/reversed face mode change, complex MPC recovery, typed rejection | phase/recovery residual <= 1e-12 |
fem.hdiv-rt0-refinement-transfer.v1 | physics-gate | independent affine RT0 field edge moments, 3-D identity recovery, unsupported-input rejection | 2-D error <= 2e-10; 3-D identity error <= 2e-10; at least 4 rejection checks |
fem.hdiv-rt1-curved-p2-transfer.v1 | physics-gate | curved self-recovery, affine-to-curved finite/complete transfer, typed rejection | curved identity <= 2e-8; affine-to-curved recovery checks >= 2 |
fem.hdiv-rt1-periodic-quad-face-contract.v1 | physics-gate | identity, quarter-turn, reflected square maps, complex MPC recovery, typed rejection | phase/recovery residual <= 1e-12 |
fem.hdiv-rt1-periodic-mixed-face-contract.v1 | physics-gate | mixed pairing, seven-mode MPC recovery, 7x7 condensation, typed rejection | phase/recovery residual <= 1e-12 |
fem.hdiv-rt1-multielement-ownership.v1 | physics-gate | unit-cube tetrahedral generator, shared-face reuse, checked topology layout | exact integer topology and numbering checks |
fem.eigenphysics-wrappers.v1 | physics-gate | beam theory, PEC cutoff, and Rayleigh-quotient consistency | EM wavenumber error <= 5e-3; residuals <= 1e-6; all legs converged |
fem.biharmonic-validation.v1 | physics-gate | manufactured field and Navier 0.00406 simply-supported coefficient | fine u <= 10%; fine q <= 15%; rates >= 1.0/0.7 |
fem.electroacoustics-validation.v1 | physics-gate | independent decoupled block solve, physical scaled reciprocity under swapped forward drives, and power balance | reciprocity and power imbalance <= 1e-9 |
fem.electrostatics-point-charge-image.v1 | physics-gate | independent Coulomb source minus mirror charge, potential and electric field | potential <= 1e-2; field <= 1e-1; rates >= 1 and 0.8; residual/boundary <= 1e-9; source <= 1e-12 |
fem.electrostatics-two-cylinder.v1 | physics-gate | bipolar potential and field; full capacitance pi epsilon/acosh(d/(2r)); explicit finite-sector truncation | full capacitance <= 2e-2; finite-domain capacitance <= 3e-3; flux <= 4e-2; field <= 6e-2; rates >= 0.7; conservation/residual <= 1e-9 |
fem.piezoelectric-1d-axial-load.v1 | physics-gate | independent stress-charge constitutive relations and uniform loaded/driven rod fields; cD=cE+e^2/epsilon for open electrodes | all relative errors/residuals and symmetry <= 1e-9; 24 actually solved cases |
fem.piezoelectric-1d-thickness-mode.v1 | physics-gate | f=sqrt((cE+e^2/epsilon)/rho)/(2L), u=cos(pi x/L), phi=e/epsilon*(u-u(0)); coupling-off frequency ratio | frequency <= 5e-4; field <= 2e-3; second-order rates >= 1.8; residual/potential/stiffening <= 1e-9 |
fem.thermoelectric-3d-mms.v1 | physics-gate | independent strong equations with sigma=2, alpha=0.005, Pi=1.5, k=1.5; Lap(V)=-1.2, Lap(theta)=-80; heat source corrected by sigma|grad(V)|^2 | field rates >= 1.8; gradient rates >= 0.8; P2 field/gradient errors <= 1e-9; original residual <= 1e-9; missing-term controls require nonzero accuracy errors |
fem.acoustics-driven-sweep.v1 | physics-gate | independent exact one-dimensional Helmholtz pressure; fresh stiffness and mass assembly and direct solve at each sweep frequency | pressure 100/1Hz <=2%/1%; GMRES/direct <=1e-6; residual <=1e-8; sweep solution/matrix <=1e-9/1e-12; three swept solves |
fem.electroacoustics-cavity-bar.v1 | physics-gate | independent closed-form damped waves and interface impedance sum; weighted physical reciprocity and fluid/solid dissipation | pressure/displacement errors <=0.003 and rates >=1.8; residual/reciprocity/power <=1e-9; omitted coupling error >=0.9 |
fem.semiconductor-3d-coupled-mms.v1 | physics-gate | independent quadratic potential/electron/hole strong-form sources, Einstein diffusivities and signed drift; steady native E.grad(c) model | field/gradient rates >=1.8/0.9; fine errors <=0.005/0.1; P2/residual/charge balance <=1e-9; five omitted-term errors >=1e-3 |
fem.electrochemistry-3d-pnp-electrode.v1 | physics-gate | independent mixed-xyz quadratic fields and strong forms; inverse symmetric BV prescribed metal-voltage field, solved concentration activity, electric/ionic fluxes and electrode current | field/current spatial rates >=1.8; flux rates >=0.9; fine current <=0.01; P2/residual/weak balance <=1e-9; four omitted-term controls >=1e-3 |
fem.autofrettage-spatial.v1 | physics-gate | incompressible elastic-plastic interface solution and Lame elastic unloading | fine residual stress <= 1%; history stress <= 2%; convergence rate >= 0.7; equilibrium <= 1e-9 |
fem.axisymmetric-cylinder-harvest.v1 | physics-gate | generated independent OpenCourant QUAD N2D3D=1 pressure-hold history and Lame fields | engine displacement/energy <= 0.5%; Lame displacement <= 0.1%; refinement rate >= 1.5; equilibrium <= 1e-8 |
fem.necking-bar-spatial.v1 | physics-gate | independent Voce uniaxial reaction turning point and fine integration of the nonuniform equilibrium branch | peak extension <= 1%; peak force <= 0.1%; displacement/reaction <= 0.1%; refinement rate >= 1.5; free equilibrium <= 1e-9 |
fem.foam-laws-integrity.v1 | contract-gate | original LOIS engine golden series | failed_probes = 0; numerical_solve_executed = 0 |
fem.foam-law51-material-replay.v1 | physics-gate | original engine stress/density/energy histories and independent exact PG adiabatic energy | stress <= 2e-4; density <= 2e-5; native exact energy <= 2e-4; original golden energy band usage <= 1; frozen-energy control >= 0.05 |
fem.foam-law33-material-replay.v1 | physics-gate | original independently generated and harvested OpenCourant single-brick history | stress/work <= 0.5%; density/geometry <= 1e-5; air-disabled stress gap >= 1%; plastic strain >= 1% |
fem.foam-law76-compression-replay.v1 | physics-gate | original independently generated and harvested OpenCourant brick explicitly selecting LAW76 | stress <= 0.2%; plastic strain/work <= 0.5%; density/geometry <= 1e-5; pressure-blind stress gap >= 1%; plastic strain >= 0.0005 |
fem.foam-law76-tension-replay.v1 | physics-gate | original independently generated and harvested OpenCourant brick explicitly selecting LAW76 | stress <= 0.2%; plastic strain/work <= 0.5%; density/geometry <= 1e-5; pressure-blind stress gap >= 1%; plastic strain >= 0.0005 |
fem.foam-law33-mixed-spatial.v1 | physics-gate | independent continuum stress/reaction, closed-form compressive inverse, integrated displacement and external work | fine displacement/work <=0.5%; spatial rate >=1.5; mixed finest-pair rate reported without bound; separate smooth case gates finest-pair >=1.5; J error <=1e-7; load refinement <=2e-4; free force <=1e- |
fem.foam-law33-spatial.v1 | physics-gate | independent continuum stress/reaction, closed-form compressive inverse, integrated displacement and external work | fine displacement/work <=0.5%; spatial rate >=1.5; finest-pair rate >=1.5; J error <=1e-7; load refinement <=2e-4; free force <=1e-9; reaction <=1e-8; stress <=1e-3; elastic gap >=5% |
fem.foam-law76-spatial.v1 | physics-gate | independent continuum stress/reaction, closed-form compressive inverse, integrated displacement and external work | fine displacement/work <=0.5%; spatial rate >=1.5; finest-pair rate >=1.5; J error <=1e-7; load refinement <=2e-4; free force <=1e-9; reaction <=1e-8; stress <=1e-3; elastic gap >=5% |
fem.crash-box-impact-harvest.v1 | physics-gate | independently harvested original CRASH box1_42 energy, impulse and displacement histories | nine history errors <= 1%; timestep changes <= 0.3%; energy residual <= 1e-6; elastic response change >= 50%; wall response change >= 10% |
fem.opencourant.hertz-sphere.native-parity.v1 | physics-gate | OpenCourant external engine health is recorded separately from native parity; OPENCOURANT_EXE or legacy RADISS_EXE selects the reference | all three full native/reference histories and final same-time probes <=3%; half-CFL changes <=0.3%; energy drift <=1%; final wall-force balance <=1%; original stiffness window 1e8..3e9 N/m; exact nati |
fem.opencourant.taylor-bar.native-parity.v1 | physics-gate | OpenCourant external engine health is recorded separately from native parity; OPENCOURANT_EXE or legacy RADISS_EXE selects the reference | full displacement/plastic fields and peak radius/length/plastic strain <=3%; half-CFL fields and peaks <=0.3%; exact native80us stop, complete external history, no erosion, numerical energy drift <=1% |
fem.opencourant.rail-crush.native-parity.v1 | physics-gate | OpenCourant external engine health is recorded separately from native parity; OPENCOURANT_EXE or legacy RADISS_EXE selects the reference | external analytical probe budgets remain case-specific; native_parity_executed must equal 1 before release success; unavailable optional engine yields structured SKIP, SONIUM_QA_REQUIRE_OPENCOURANT=1 |
fem.opencourant.bending-3pt.native-parity.v1 | physics-gate | OpenCourant external engine health is recorded separately from native parity; OPENCOURANT_EXE or legacy RADISS_EXE selects the reference | native/OpenCourant displacement <=3%; original Euler-Bernoulli continuum approximation budget 10%; spatial changes <=2% then <=0.5%, contraction <=0.6; timestep change <=0.3%; every-step independent d |
fem.opencourant.necking-bar.native-parity.v1 | physics-gate | OpenCourant external engine health is recorded separately from native parity; OPENCOURANT_EXE or legacy RADISS_EXE selects the reference | external analytical probe budgets remain case-specific; native_parity_executed must equal 1 before release success; unavailable optional engine yields structured SKIP, SONIUM_QA_REQUIRE_OPENCOURANT=1 |