Datasets:
Possible mismatch between boundary fields and force coefficients for runs 105, 159, and 492
Hi, we found a possible inconsistency while checking the AhmedML surface data against the per-run force coefficients.
We integrated the pressure and wall-shear-stress fields from boundary_{run}.vtp using the corresponding boundary_cell_area_{run}.npy and the constant reference area (0.112032 m²). The results were:
| Run | Integrated Cd | CSV Cd | Integrated Cl | CSV Cl |
|---|---|---|---|---|
| 105 | 0.33799119 | 0.33624586 | -0.04941088 | 0.20526946 |
| 159 | 0.33622369 | 0.33299628 | -0.19687363 | -0.21300720 |
| 492 | 0.31839300 | 0.31594544 | 0.56097920 | 0.55676681 |
We also confirmed that the per-run CSV values agree with force_mom_all.csv. For run 105, the surface-area totals from the supplied NPY array and the VTP geometry are nearly identical. In ParaView, run 105 shows some cell-scale irregularity in wall shear stress near a front corner, but we do not know whether it is related to the force mismatch.
Could you please clarify whether the boundary fields and force CSVs for these runs use the same surface, averaging interval, and force-coefficient convention? Is there a known issue or a recommended procedure for reproducing the CSV values from the surface fields?
Thank you!
Run 105 (up) vs. run 491 (down), WSS near the front corner:
Hi @neashton ,
Just following up on this issue. We would appreciate it if you could help clarify whether the discrepancies we observed are expected, or whether there might be an inconsistency between the boundary fields and the reported force coefficients for these runs.
We would also be happy to share our integration code or provide additional details if that would be helpful.
Thanks for your time and for making this dataset publicly available!