G05-B Family Coverage¶
Status¶
This document records the controlled internal prequalification campaign for
026-G05 modal, Newmark and harmonic routes. The official gate is closed as
PASS_WITH_LIMITATIONS; the bounded external limitation remains explicit:
SUPPORTED != TESTED != VERIFIED != QUALIFIED.
The generated numerical source of truth is
qualification/vnv/g05b_family_coverage/summary.json; its vnv_manifest.json
records the exact source SHA, runtime and artifact digests after a clean run.
Executed campaign¶
| Analysis | Executed | Contract target | Result |
|---|---|---|---|
| Modal | 14 | 14 | PASS_INTERNAL_PREQUAL |
| Newmark | 32 time-level cases | 16 | PASS_INTERNAL_PREQUAL |
| Harmonic | 12 | 12 | PASS_INTERNAL_PREQUAL |
The modal catalog contains one baseline solve for each family and six declared
two-mode variants. The Newmark count represents four actual integrations per
baseline family at 30, 60, 120 and 240 steps per period. The harmonic catalog
contains one baseline sweep per family and four refined-grid sweeps. Every
listed case is executed; no READY status is counted as a result.
Element x analysis matrix¶
| Element family | MOD | DYN | HAR | Qualification state |
|---|---|---|---|---|
| TET4 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| TET10 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| HEX8 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| HEX20 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| BEAM2 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| MITC3 / MITC3+ | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | MITC3+ is the route label, not a second registry element |
| MITC4 | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | INTERNAL_PREQUAL, not closed |
| discrete | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SUPPORTED / TESTED / VERIFIED | SPRING_MASS retained as legacy alias |
The compact family models exercise the common solve_model routes. They are
not a replacement for each element's static evidence, external correlation,
or the final Owner qualification decision.
Refinement evidence¶
Modal mesh¶
The campaign aggregates existing controlled mesh studies rather than claiming that one-element family models constitute a universal mesh proof:
qualification/vnv/external/code_aster_modal_refinement_048/reference/summary.jsonqualification/vnv/tet10_stable_refinement/reference/summary.json
The Owner-approved bounded policy is a final adjacent-level tracked-mode
relative frequency change of <= 1e-2 over at least three compatible levels.
It retains mode tracking and residual evidence; it is not a universal mesh
convergence claim.
Newmark time¶
All baseline families run at 30/60/120/240 steps per period. The final adjacent-level error is below the proposed one-percent band for the campaign; the maximum over all levels is retained as a diagnostic because coarse starts are intentionally less accurate. The proposed metric is
||u_(dt/2) - u_dt|| / max(||u_(dt/2)||, u_floor) <= 1e-2.
It applies only to the same physical interval, load history and compatible
output times. It is OWNER_APPROVED_BOUNDED; residual and energy criteria
remain separate requirements.
Harmonic frequency¶
Baseline sweeps include zero, off-resonance, resonance and post-resonance
points. TET4, HEX8, MITC4 and discrete receive a refined grid with local
sampling around the first resonance. The proposed final-grid amplitude band is
<= 1e-2, with phase and peak-bin movement reported separately. It is not
applied at singular exact resonance or across changed damping/material data.
This bounded policy is Owner-approved.
Observed internal bounds¶
| Family | First modal frequency (Hz) | Modal residual | Newmark final-study residual | Newmark energy drift | Harmonic residual |
|---|---|---|---|---|---|
| TET4 | 1.589281e3 | 1.78e-16 | 8.52e-08 | 5.76e-13 | 6.94e-11 |
| TET10 | 3.257056e3 | 1.28e-16 | 3.67e-08 | 4.73e-13 | 1.75e-16 |
| HEX8 | 1.591549 | 1.80e-16 | 1.54e-14 | 3.80e-13 | 7.02e-16 |
| HEX20 | 1.591549 | 1.80e-16 | 1.54e-14 | 3.80e-13 | 7.02e-16 |
| BEAM2 | 1.026597e1 | 1.18e-12 | 3.07e-12 | 2.36e-13 | 3.29e-11 |
| MITC3 | 8.454679 | 4.39e-09 | 6.56e-14 | 6.88e-13 | 2.09e-09 |
| MITC4 | 8.433214 | 1.32e-08 | 2.13e-12 | 3.55e-11 | 1.88e-09 |
| discrete | 1.591549 | 1.80e-16 | 1.54e-14 | 3.80e-13 | 7.02e-16 |
The identical HEX8/HEX20 values in this compact constrained block are an observed result of this particular model and are not evidence of formulation equivalence in general. Larger or element-specific studies remain necessary.
Bounded limitation and traceability¶
- The three refinement policies are
OWNER_APPROVED_BOUNDED. - Final aggregation is recorded in
g05_final_evidence.jsonandowner_decisions.json. - External correlation coverage remains bounded and is not element-family complete; the corresponding limitation is part of the Owner decision.
026-G05 is PASS_WITH_LIMITATIONS by Owner decision. This closeout does
not promote unsupported external coverage or broaden any element claim.