High-contrast homogenization: #
Provider.PolynomialHomogenization.Root.Dirichlet.GaugeWeakSolutionAndCellFamily
Imported from the Apache-2.0 HighContrastHomogenization development at commit
7a13dbcd8d6609264a713373f5c69ceeac870472.
Row 4′ (the gauge weak solution) and composed row 7 #
Rows 6 and 7 are built here. Row 7's family producer lives in
HCPoly.Provider.PolynomialHomogenization.Root.RowSupply.CellFamilyAndFluxRow.
The HCPoly tree carries only the non-_v2 producer, whose
conclusion is RuledCellPhysicalForcedFamily rather than the older predicate
that the module consumes — which is why the search failed.
The real obstruction was neither producer but a step neither row states: both
want the weak-solution premise in the gauge frame, at aHat on Uhat,
while the module (row 4) delivers it in the physical frame. That step —
row 4′ — is exists_frozenWitnessGaugeWeakSolution below, and it is exactly the
affine transfer isWeakSolutionOn_skewCentered_matSqrtPullback_iff
applied through huHat. With it, row 7 is one composition.
Row 4′ #
Row 4′ — the gauge weak solution. Module physical weak solution
transports to the gauge frame at aHat and uHat, with no new
analytic input: the skew-centred square-root pullback is an
equivalence, and the huHat premise is precisely its pullback gradient.
Composed row 7 #
Row 7 — aCell, wCell, hfamily. With the gauge weak solution of
row 4′ and the gauge-reduced source of
exists_gaugeReducedSource, the _v2 cell family exists over the frozen
witness's ruled system. This is the aCell premise/wCell/hfamily triple.