Skip to content

0.2.7 J2 Gap Closure

Baseline boundary

The 0.2.6 Owner decision qualifies small-strain J2 only within its recorded TET4/TET10/HEX8/HEX20 scope. Algorithmic tangent symmetry is not independently qualified, increment-refinement evidence is strongest on TET4, and finite-kinematic J2 remains research-only. This plan does not reopen or alter that decision.

Proposed work

WP11 may close only the evidence gaps that matter to the existing small-strain claim:

  1. repeatable stress/increment paths on all four solid families;
  2. finite-difference tangent checks over declared states, directions and perturbation sizes;
  3. increment-partition studies on families where the current evidence is thin;
  4. unload/reload and rollback checks within the existing material model;
  5. energy/residual diagnostics and explicit failure classifications;
  6. comparable external correlation where the constitutive convention matches.

Policy discipline

Existing policies are reused where applicable. A new tolerance, a change to the tangent, finite-kinematic formulation, default Newton path or rescue behavior requires an Owner decision and is outside this foundation plan. PROPOSED_OWNER_REVIEW is the only valid state for an unsupported threshold.

STOP/GO

GO requires all four families to have traceable cases, no unexplained divergence, deterministic replay and preserved qualified scope. STOP on a formulation change, a weakened criterion, or evidence that cannot distinguish material behavior from a harness/model error.

WP11 execution record

The targeted WP11 lot is recorded in qualification/0_2_7/wp11_j2_evidence.json with case definitions in qualification/0_2_7/wp11_j2_cases.json. It exercises the existing small-strain radial-return material and connected nonlinear route on TET4, TET10, HEX8 and HEX20. The lot covers elastic and plastic transitions, unloading/reloading, a simple cycle, finite-difference tangent checks, multi-element response, energy, rollback, increment characterization, Newton behavior and explicit failure modes.

The result is PASS_WITH_LIMITATIONS: the existing qualified bounded J2 scope is retained, with no promotion or demotion. Increment sensitivity is reported for all four families without adding a universal threshold; tangent symmetry is diagnostic only; modified Newton non-convergence is explicit and full Newton remains the accepted route. Finite-kinematic J2 and dynamics gaps remain outside WP11. The Code_Aster evidence is reused from the controlled 0.2.6 record and is not presented as a new WP11 external run.