Skip to content

0.2.6 G10 Owner Closeout

Status: PASS_WITH_LIMITATIONS

Execution source SHA: 9b4a44d61132a28fe9161f4aa8f04e838afc5f32 Worktree at review: dirty=false

This closeout audits the existing nonlinear and research routes. It does not add physics, elements or solver behavior, and it does not close or promote G07. The selected external evidence is bounded evidence only.

Owner route classifications

Route Owner classification Boundary
Nonlinear static small-strain OWNER_APPROVED_BOUNDED Existing G06 bounded scope only.
Transaction/checkpoint/retry OWNER_APPROVED_BOUNDED Existing G09/G11 evidence and stated coverage.
Structured failure diagnostics OWNER_APPROVED_BOUNDED Existing G11 bounded failure contract.
Arc-Length continuation OWNER_DEFERRED Bounded external evidence is useful to G07; G07 remains unchanged.
Total-Lagrangian elasticity OWNER_DEFERRED TET4/HEX8 evidence is bounded and routed to G07.
Finite-kinematic J2 OWNER_NOT_QUALIFIED No formulation-compatible external constitutive qualification.
J2 plus geometry OWNER_DEFERRED No external evidence isolating constitutive and kinematic effects.
Geometry plus frictionless contact OWNER_EXPERIMENTAL_ONLY Existing contact law/surface limits remain.
J2 plus geometry plus contact OWNER_NOT_QUALIFIED Coupled external qualification is absent.
Modified Newton finite-kinematic OWNER_NOT_QUALIFIED Existing finite-route restriction remains.

No route is classified OWNER_FAILED. The deferred and unqualified routes do not block this audit closeout because they are excluded from any qualified claim and remain visible as limitations.

External evidence decision

The selected campaigns are formulation-compatible within their declared domains and are recorded as PASS_WITH_LIMITATIONS:

  • Arc-Length: 75 common interpolated points, one branch turn on each solver, relative turning-point differences of 1.33e-4 in load factor and 6.62e-3 in control displacement.
  • Total-Lagrangian TET4: stress relative error 8.54e-5 and column maximum relative difference 1.69e-9 over four points to 80% of the same-mesh critical load.
  • Total-Lagrangian HEX8: matched-point displacement difference 2.34e-9, sign-aligned reaction difference 1.32e-11, and QF residual 2.29e-13.

The HEX8 full QF instrumented path was not completed within the bounded campaign budget. External stress, energy and det(F) outputs were not mixed with incompatible QF measures. These limitations prevent a general claim; they do not create a numerical defect.

Requirements and regression policy

All nine G10 requirements are addressed at full or bounded level. The external-correlation requirement is satisfied as a recorded, bounded audit: available comparable evidence is attached and missing formulation-compatible routes remain explicit. No blocking requirement remains for the G10 audit scope.

No functional source changed and no numerical regression was detected. Full regression is therefore SKIPPED_BY_POLICY; the 73 targeted tests, registry, anti-forgetting, Ruff, compileall and diff checks are the applicable closeout controls.

Final boundary

026-G10 = PASS_WITH_LIMITATIONS.

G07 remains unchanged. Arc-Length and Total-Lagrangian promotion decisions remain G07-owned and deferred. Finite-kinematic J2 and coupled nonlinear workflows remain experimental, not qualified, or deferred as listed above. No general physical validation, industrial-solver equivalence or exhaustive route-combination claim is made.