[petlenz]

mean_field_strain splits its concentrations from its strains, so that with constant phases the localisation computes its concentrations once (part 3 of 3, split from #161; stacked on #164).

What

• Concentrations and strains are separate callbacks. The concentrations are declared constant over the phase reads, and the strains run after them:

  auto &concentrations{base::add_callback(&mean_field_strain::compute_concentrations, m_A).constant(phase_reads)};
  base::add_callback(&mean_field_strain::compute_strains, m_eps).runs_after(concentrations);

  With constant phases the concentrations are computed once and only the strains every update. A phase read over a local edge (nonlinear) recomputes them every pass. The strains are history and stay in the loop.
• docs/mean_field.md describes the split.

Measured

The same phases as in #164 through mean_field_strain, linear_stress per phase and mean_field_composite_local_linear, 2000 committed steps, -O2 -DNDEBUG.

kernel	before	after	stress checksum
Mori–Tanaka	4.51 µs / update	1.14 µs	identical
self-consistent	10.38 µs / update	1.64 µs	identical

Tests

• MeanFieldLocal.ConstantPhasesAreLocalisedWithConcentrationsComputedOnce: a counting Mori–Tanaka kernel is called once over five committed steps, with stresses and strains equal to plain Mori–Tanaka. It failed before the split.
• Ordering: strain_<inclusion> has no callback of its own. Without the writer edges from Constant properties: engine and API #163 it ran after stiff::stress in some hash orders. test_mean_field_local and test_short_fibre_composite were repeated 10× each, all passing.

Gates (GCC 14, fetched dependencies): 1182/1182 tests, each also under ASan, 0 warnings. The tree is identical to the last commit of #161, whose 14 docs/materials report checks passed.