Two versus three correction passes

Six physical steps compare two and three PIMPLE outer correction passes from identical Revision H steady-iteration-300 fields. Both retain two pressure correctors and one nonorthogonal correction. Native binary internal fields are compared directly.

The acceptance limits were written before each solve: RMS error at most 0.1% of reference RMS, maximum error at most 1% of reference maximum, plus 1e-6 absolute tolerance for U, p, k, omega and nut; maximum Courant 0.5; boundary flux imbalance ratio at most 1e-4.

Timestep	Field checks	Maximum Courant	Disposition	Evidence
25 µs	All passed	0.734	Above Courant target	Comparison · Preset criteria · Three-pass log · Two-pass log
12.5 µs	All passed	0.367	All criteria passed	Comparison · Preset criteria · Three-pass log · Two-pass log

At 12.5 µs, the largest absolute velocity-component difference was 0.001868 m/s; the largest pressure difference was 0.05241 m²/s², or 0.06289 Pa at the assumed density of 1.2 kg/m³. Neither comparison logged turbulence bounding.

The separate flowing-field transient uses two outer correction passes, starts at 12.5 µs, and adapts toward Courant 0.5 with a 25 µs maximum. Its source is the later iteration-400 field, explicitly unconverged, and it remains under monitoring.

Wall-time ratios are not isolated speed benchmarks. Other solves were running, and the steady preparation ended during the smaller-step comparison. The approximately 2.03 ratio in that report does not prove a twofold solver speedup.

These short coupling checks do not establish timestep independence, mesh independence, long-time shedding statistics or thermal performance. Source fields remain preserved locally.