|
FhSim
3.1.0
Marine systems simulation
|
| ID | 0157 |
| Class | KNOWN-LIMITATION |
| Severity | 2 |
| Status | blocked |
Models: — (annotated on the fhsim_environment library group)
Found: 2026-11-18, validation case B8 (fhsim_validation, cases/hydrodynamics/net_panel_load/B8_gansel_cd)
Decision needed: the owner chooses whether the default panel law should match Gansel et al. 2015 at low twine Re and high solidity, or whether case B8 documents a model range limit.
Case B8 (tracked, per-row tolerance 5 to 13 %) compares the drag coefficient of clean nets with Gansel et al. 2015 (PLOS ONE, Fig 5): solidity Sn 0.18 to 0.43, U 0.1 m/s, theta 0, 11 rows. The twine diameters in fhsim_validation/sources/gansel2015/fig5.csv are 1.3 to 2.9 mm, so twine Re = U d / nu is about 130 to 290. Reference run (fhsim_validation/results/reference/validation_results_laws_tracked.json):
| law | rows passing | median error | min | max |
|---|---|---|---|---|
LocalThroughFlow | 2 of 11 | -0.209 | -0.251 | -0.028 |
ScreenKF2012 | 2 of 11 | +0.059 | -0.163 | +0.285 |
ScreenMF2022 | 0 of 11 | -0.386 |
The laws are in fhsim_environment/include/fhsim_environment/hydrodynamics/ (LocalThroughFlow.h, ScreenKF2012.h, ScreenMF2022.h).
For clean nets at 0.1 m/s the default law gives about 21 % too little drag. The cause is not verified in code. This Re is below the range of most fits, which may be a reason.
Case B8: all 11 rows pass for the default law.
Changing the default law moves the drag of every net, so regression baselines move and need an owner ruling.