APC propeller CFD — case index

Structured multiblock conformal hex meshes (chh_pipeline), OpenFOAM. 32 steady MRF cases + 1 Euler transient. Generated 2026-09-29.

Read this before quoting any number A converged solver run is not the same thing as a defensible physical result. The gamma-Re_theta cases below converged cleanly and still failed physically — laminar over the entire blade. They are listed separately and are valuable as a negative result, not as predictions.

Steady MRF — usable

Converged, force history flat over the last 500 iterations (sd/|T| < 1%). Fully-turbulent SST unless noted.

case	iters	thrust (N)	sd	torque (N·m)	note
10x5E_J0_7000	8000	6.1288	3.71e-04	0.0925
10x6E_J0_7000	8000	7.0397	5.85e-04	0.1167
10x7E_J0.2_7000	8000	6.8267	7.21e-03	0.1440
10x7E_J0.4_7000	8000	5.2925	4.33e-04	0.1353
10x7E_J0.6_7000	8000	3.1525	7.89e-04	0.1048
10x7E_J0_4000	8000	2.4734	1.52e-03	0.0475
10x7E_J0_4000_SSTlowTu	8000	2.4472	7.54e-03	0.0475	SST low-Tu, 4000 rpm.
10x7E_J0_5500	8000	4.6249	7.76e-03	0.0901
10x7E_J0_5500_SSTlowTu	8000	4.7349	7.55e-03	0.0893	SST low-Tu, 5500 rpm.
10x7E_J0_7000_SSTlowTu	8000	7.7268	2.43e-02	0.1443	Fully turbulent SST at low freestream turbulence. Reattaches too EARLY (Catalano) -- a known one-sided bias.
10x7E_J0_8500	8000	11.3475	3.14e-02	0.2109
10x8E_J0_7000	8000	7.8473	7.45e-03	0.1766
11x7E_J0_7000	8000	10.2106	4.09e-03	0.1928
11x8E_J0_7000	8000	10.7352	2.58e-02	0.2310
12x8E_J0_7000	8000	14.0074	4.80e-02	0.2990
7x5E_J0_7000	8000	2.0472	8.38e-04	0.0285
8x6E_J0_7000	8000	3.5894	4.66e-03	0.0597
9x6E_J0_7000	8000	5.3635	8.35e-04	0.0895

Steady MRF — part of a numbered study

Grid convergence is OSCILLATORY, not asymptotic Coarse / medium / fine do not fall monotonically, so the Celik GCI percentages are not reportable and must not be quoted. The defensible statement is the spread: 4.6 % in thrust, 3.1 % in torque — which is the same size as the effects being measured. A further limitation: the refinement ratio is only r ≈ 1.14 because two mesh directions do not scale.

case	iters	thrust (N)	sd	torque (N·m)	note
10x7E_J0_7000_gci_coarse	8000	7.5247	1.03e-02	0.1469	Grid-convergence level. See caveat: the study is OSCILLATORY, not asymptotic.
10x7E_J0_7000_gci_fine	8000	7.7561	1.18e-02	0.1424	Grid-convergence level. See caveat.
10x7E_J0_7000_gci_fine_g1129	8000	7.6783	2.59e-03	0.1423	Grid-convergence level, BL growth 1.129. Finished 2026-09-29.
10x7E_J0_7000_mesh_coarse	8000	7.5863	7.27e-03	0.1440	Mesh-resolution variant.
10x7E_J0_7000_mesh_fine	8000	7.6391	1.13e-02	0.1451	Mesh-resolution variant.

Steady MRF — negative results (physics failed)

These did not predict a separation bubble, and could not The Langtry–Menter model stayed laminar over the whole blade (gammaInt max 0.32, nut = 0 at the wall, 28/28 stations with unclosed separation, thrust ~6 % low). A frame-corrected fork (kOmegaSSTLMrel, correcting the model's use of ABSOLUTE velocity, which is 2.6–3.4× too small at J = 0) did not change the outcome. A 19× change in ambient turbulence moved gammaInt by 0.0002. High Tu and low ν_t/ν are not simultaneously reachable at a static rotor — decay lengths 0.15 mm against a 300 mm domain. That is a physics limit, not an implementation bug.

case	iters	thrust (N)	sd	torque (N·m)	note
10x7E_J0_5500_LMrel	8000	4.5620	1.49e-02	0.0920	As above, 5500 rpm.
10x7E_J0_7000_LM	8000	7.4453	6.77e-03	0.1469	gamma-Re_theta. Converged, but laminar over the whole blade; gammaInt max 0.32, 28/28 stations unclosed separation. A decisive NEGATIVE result, not a usable thrust prediction.
10x7E_J0_7000_LMrel	8000	7.4002	6.38e-03	0.1478	Frame-corrected gamma-Re_theta fork. Same outcome as LM: no transition, no closed bubble.
10x7E_J0_7000_LMrel_amb06	8000	7.4300	9.74e-03	0.1473	Ambient turbulence sweep. A 19x change in ambient moved gammaInt by 0.0002 -- the insensitivity IS the result.
10x7E_J0_7000_LMrel_amb18	8000	7.4306	5.59e-03	0.1473	Ambient sweep, see amb06.
10x7E_J0_7000_LMrel_amb37	8000	7.4211	7.33e-03	0.1476	Ambient sweep, see amb06.
10x7E_J0_8500_LMrel	8000	11.0604	1.12e-02	0.2149	As above, 8500 rpm.

Steady MRF — not converged

case	iters	thrust (N)	sd	torque (N·m)	note
10x10E_J0_7000	8000	7.6326	9.14e-02	0.2566	Force history not flat over the last 500 iterations (sd/\|T\| > 1%). Do not quote.
9x9E_J0_7000	8000	5.7866	7.10e-02	0.1724	Force history not flat over the last 500 iterations (sd/\|T\| > 1%). Do not quote.

Euler transient — 10x7E, J=0, 7000 rpm

Cannot produce a laminar separation bubble — by construction simulationType RAS, RASModel kOmegaSST: fully turbulent from the leading edge, linearUpwind convection, backward time, PIMPLE nOuter 3, fixed dt 2.976e-06 at Courant max 128.7, 48 ranks, 22.8 s/step. The y+ ≈ 1 resolution buys near-wall shear stress and SST's low-Re treatment — not bubbles. What it gives: rotor-scale unsteady loads, blade passing, wake structure.

EULER_10x7E_J0_7000_transient — 99 reconstructed writes (175 GB of field data, not served here; available on request). Served: sbatch scripts, SLURM logs, case dictionaries, rotor/blade force and moment histories.

What is in each case folder

system/ and constant/ dictionaries (the definition of the case), postProcessing/ force and moment histories, log.simpleFoam.headtail (first and last 200 kB; full logs on request), and RESULT.json with the extracted values. Mesh and field data are excluded for size.