Skip to content

0.2.7 WP20 - Residual J2 and external V&V closure

WP20 closes the remaining Owner-review item attached to the WP11 small-strain J2 characterization. It is a qualification and governance closeout, not a new constitutive implementation or a new numerical campaign. The review started from 26a734d1656c1c824c27f4708a8783abfddde17c.

Owner decision

The Owner decision is OWNER_APPROVED_BOUNDED_KEEP_EXISTING_SCOPE with work-package status PASS_WITH_LIMITATIONS. TET4, TET10, HEX8 and HEX20 are each KEEP: the existing MAT-J2-SMALL maturity remains QUALIFIED_BOUNDED, with no promotion or demotion. MAT-FINITE-J2 remains EXPERIMENTAL/NOT_QUALIFIED.

The bounded claim is small-strain isotropic J2 with radial return on the four existing solid families through the accepted full-Newton nonlinear-static route. It does not qualify finite-kinematic J2, Total-Lagrangian J2, coupled nonlinear routes, modal/dynamic behavior or a universal physical response.

Evidence reviewed

The controlled WP11 evidence is preserved unchanged:

  • case catalog: qualification/0_2_7/wp11_j2_cases.json;
  • evidence: qualification/0_2_7/wp11_j2_evidence.json;
  • execution/evidence source SHA: 94461602dfd1782be57c20e1801a0d5d8e262ef1;
  • result digest: 5e4825625d40f2363ecefb8f96baa43acac42f23db7195d000ca0c09717ef536;
  • historical execution contract: qualification/0_2_7/wp11_state.json.

The evidence passes elastic prediction, yield detection, radial return, internal-variable and stress updates, unloading/reloading, a simple cycle, finite-difference tangent checks, multi-element paths, energy diagnostics, rollback/state digest restoration, full Newton and explicit failure modes on all four families. The maximum reported finite-difference tangent error is 2.120472111937634e-10 against the existing G06 limit 1e-6.

The 1/2/4 subdivision study characterizes increment sensitivity on all four families. It deliberately introduces no universal structural increment-independence threshold. Tangent symmetry remains a diagnostic only; modified Newton non-convergence remains an explicit diagnostic and is not silently promoted.

External evidence

The Code_Aster 18.1.0 correlation from qualification/0_2_6/g06_depth_evidence.json is reused as controlled, formulation-compatible constitutive evidence. Its source SHA is 8bd0f2d8fdce7bf27ffc4c28e6aa26e69288fa63 and its archived artifact digest is 86e238847ca45d4b7024c5bd9dcb15cb07ea5eb72444486204c4d9c3d3c657c2. This is a partial external-V&V result for WP20: no new structural external run is claimed, and it does not extend the increment-independence claim. External evidence is therefore a limitation, not a reason to invent a new tolerance or to relabel the internal evidence.

Failure and reproducibility policy

Invalid material inputs fail closed with deterministic diagnostics. The controlled rollback path restores the committed-state digest before retry. Non-convergence in the modified-Newton diagnostic is recorded as a controlled failure mode. No NaN/Inf or silent pass is accepted, and the WP11 result digest remains stable on replay.

No solver, material, element, formulation or tolerance was changed by WP20. Full regression is intentionally not run under the targeted WP20 policy. The next work package is WP21.