FhSim  3.1.0
Marine systems simulation
Loading...
Searching...
No Matches
0158 — Panel laws low at normal flow against Tang 2018 normal frame (B2d): knotted nets and the level of LocalThroughFlow and ScreenMF2022
ID 0158
Class KNOWN-LIMITATION
Severity 2
Status blocked

Models: — (annotated on the fhsim_environment library group)

Found: 2026-11-18, validation case B2d (fhsim_validation, cases/hydrodynamics/net_panel_load/B2d_tang_cd_normal_frame). It replaces ENV-0155 (resolved 2026-11-17) as the issue of the B2d accepted misses (owner ruling 2026-11-18, option 3 of fhsim_validation/build/b2d-issue-report.html).

Decision needed: the owner chooses, after an analysis, whether the panel laws get a knot term for knotted netting, whether the normal-flow level of LocalThroughFlow and ScreenMF2022 changes, or whether B2d records these as a model limit. The B2d accepted misses expire on 2027-05-20.

Evidence

Case B2d (gate, 5 % relative, the HV-D3 minimum; the paper states no uncertainty) compares cd_proj (drag over 1/2 rho Sn A U^2, Tang Eq. 4) with Tang, Xu and Hu 2018 (PLoS ONE 13(2):e0192206), S1 File B.xlsx: normal (streamlined) frame, flow normal to the panel (theta 0), 8 nets, 68 rows, Re 445 to 3540. Reference run (fhsim_validation/results/reference/validation_results_laws_gate.json), error range per net over its speeds:

Net (Tang Table 1) Knot Sn LocalThroughFlow ScreenKF2012 ScreenMF2022
net-1 knotless 0.073 +9 to -9 % +9 to -9 % -16 to -19 %
net-2 knotless 0.081 +8 to -9 % +8 to -9 % -26 to -28 %
net-3 knotless 0.132 +2 to -7 % +7 to -2 % -25 to -26 %
net-4 knotless 0.147 -4 to -9 % +1 to -3 % -22 to -25 %
net-5 knotless 0.155 -1 to -8 % +6 to -1 % -20 to -22 %
net-6 knotless 0.187 -6 to -13 % -3 to +6 % -14 to -18 %
net-7 knotted 0.168 -22 to -26 % -14 to -20 % -33 to -36 %
net-8 knotted 0.177 -28 to -31 % -21 to -25 % -36 to -39 %

Rows passing: LocalThroughFlow 11 of 68, ScreenKF2012 35 of 68, ScreenMF2022 0 of 68.

  • The laws take solidity and twine diameter only (fhsim_validation/laws/LawTable.cpp:32-45 fills Netting with solidity and twineThickness). Nothing tells a law that a net is knotted. Tang computes the solidity of a knotted net with its knots (Tang Eq. 2). The knotted nets miss most in all three laws. This points to a knot effect, but the cause is not checked in the law code.
  • The ENV-0155 analysis (2026-11-16) found that LocalThroughFlow gives 0.94 times Tang Eq. 6 at normal flow.
  • ENV-0155 covered in-plane flow (theta 70 to 90 deg). B2d has theta 0 only, so ENV-0155 does not explain these misses.

Effect

Drag of knotted netting at normal flow is 14 to 39 % low in all three laws. Knotless nets are a few percent low in LocalThroughFlow and 14 to 28 % low in ScreenMF2022. Without a ruling before 2027-05-20, the B2d accepted misses expire and the fast-tier ctest entry fhsim_validation_laws_gate fails.

Possible fix

  • Analyse first: compare the knotted rows with the knotless rows of the same Sn and Re, and check how each law uses solidity at normal flow.
  • A knot term (knot type and knot size as netting inputs), if the analysis supports it. This needs an owner physics ruling.
  • Or record the knotted rows as a model limit in B2d (known_limit, as B2a to B2c).

Test that would prove it

The B2d rows of the chosen law pass the 5 % gate, run with ctest -R fhsim_validation_laws_gate in fhsim_validation.

Risk

A change of the normal-flow level or a knot term moves every net result that uses these laws (trawl and cage models in ME, fishery and trawlwriter). Report the moves in the moves.md of those repos.