DrivAer fastback: FVM and LBM sections at y = 0 (section 7.4, fig. drivaer)¶
| file | content |
|---|---|
y0_u.npz |
1-cm grid at y = 0, x from the rear: x, z, u_fvm, u_lbm (mean), u_inst (instantaneous, step inst_step) (m/s, NaN in the body), solid (LBM body cells) |
probes_y0.csv |
LBM u, w (m/s) at four points against time (s), every 10 steps (1.385 ms) for 2.216 s from step 8,000 |
probe_stats.py |
mean, standard deviation and spectral peak of u at the probes (numbers of sec. 7.4) |
domain.json |
FVM domain (blockMeshDict), LBM domain and refinement boxes (run log), inlet section |
make_y0.py |
how the two files were made from the raw output |
Raw output, Miyabi-G, 2026-10-14 (not in this repository):
- FVM: OpenFOAM training course DrivAer fastback case, simpleFoam, k-omega SST,
3,431,242 cells, rotating wheels, inflow 16 m/s, time 1000;
postProcess -func lbmCompare -> U_y0.raw. Cd 0.2571 = mean over iterations 501-1000 of
postProcessing/forceCoeffs/0/coefficient.dat (64 Grace ranks, 335 s). The job log's
0.2612 averaged the wrong iterations (awk compared iteration keys as strings).
- LBM: OpenRheo D3Q27 cumulant, wall model, moving floor, rotating rear wheels,
finest cell 1.1 cm (19,818,112 cells), inlet = FVM section at x = 1.6 m
(mean U and digital-filter fluctuations from k), 16,000 steps, mean over
steps 8,000-16,000; run run_nh32_ml2.3548845 -> slice_y0.csv.
- Instantaneous slice and probes: run run_nh32_ml2.3553357, same setup, 56,000
steps, DUMP_INST=1, PROBES=0.5:0.6;1.0:0.9;2.0:0.6;1.0:0.15, PROBE_EVERY=10,
PROBE_FROM=8000, slice at step 20,000 (openrheo-dev drivaer cce3dc0).
Probe statistics (probe_stats.py, 6.65 s, 7 half-overlapping segments of 1.6 s,
resolution 0.63 Hz): u mean / std 1: -1.81 / 1.75, 2: 11.67 / 1.25,
3: 12.24 / 1.86, 4: 2.57 / 3.44 m/s; spectral peaks 5.63, 5.00, 5.00, 0.63 Hz
(St = f H / U with H 1.418 m, U 16 m/s: 0.44 at points 2 and 3). The earlier
2.2 s record (run 3552365, 4 segments of 0.8 s) gave 6.25 Hz at point 2, which
the longer record does not confirm.
- Reverse flow behind the rear at y = 0: FVM 0.72 m, LBM 0.92 m
(0.93 m without wheel rotation). Underbody at 0.55 cm (finest box z < 0.355 m),
56,000 steps (run 3553360 continued from its checkpoint as 3556610), mean over
steps 12,000-56,000: underbody u at the rear 12.6 m/s (FVM 12.8), reverse flow
1.08 m; probe means 1: -3.20, 2: 3.45, 3: 6.17, 4: 9.47 m/s (FVM -1.99, 9.68,
11.00, 11.19) and no peak near 5 Hz (openrheo-dev drivaer
scripts/miyabi/job_drivaer_ml3long.pbs).