Overset body dynamics VERIFIED by free decay — natural frequency within 0.4% of the reference and light, linear damping (ζ ≈ 0.126); the #1324 near-resonance reduction is a forced-wave measurement artifact, not body damping
The #1324 wave sweep found the overset CFD did not reproduce the potential-flow heave resonance (near-resonance RAO ≈ 0.7–1.0 vs an inviscid peak ≈ 2.2) and left the cause open — a forced-wave test cannot separate physical bluff-body damping from a gauge-to-body incident-wave decay confound or overset numerical damping. A free-decay test can: release the body from a still-water displacement and watch it ring down. There is no incident wave and no wave gauge, so the measured damped natural period and damping ratio are properties of the body's dynamics alone.
Both releases (30 mm and 15 mm above the Archimedes waterline) ring down through the equilibrium and settle in a few cycles. The dotted lines are the fitted exponential decay envelopes z_eq ± A₀ e−ζωnt.
| release (mm) | ζ (fit) | ζ (log-dec) | ω_n (rad/s) | T_d (s) | extrema | predicted peak RAO |
|---|---|---|---|---|---|---|
| 30 | 0.129 | 0.109 | 7.40 | 0.857 | 10 | 1.78 |
| 15 | 0.123 | 0.110 | 7.50 | 0.844 | 11 | 1.87 |
The damped natural frequency ω_n ≈ 7.45 rad/s matches the reference resonance frequency 7.48 rad/s to 0.4% — the overset body's mass, hydrostatic stiffness and added mass are correct, ruling out a setup error. The damping ratio is ζ ≈ 0.126, only ~21% above the potential-flow radiation value (0.104), and it is amplitude-independent (30 mm vs 15 mm) — i.e. linear damping (radiation + a small viscous / numerical increment), not the amplitude-dependent separation drag that would flatten a resonance peak. The overset method is not spuriously over-damped.
Substituting the measured damping into the reference transfer function (green) lowers the resonance peak only modestly, from 2.20 to ≈ 1.83. The #1324 forced-sweep near-resonance RAO (0.74, orange square) sits far below both curves. So the low forced response is not body damping — with its measured light, linear damping the body would amplify to ≈ 1.8 near resonance.
Conclusion. The #1324 near-resonance reduction is a forced-wave delivery / measurement artifact: the short incident wave is dissipated over the ~9.5–13 m fetch to the body and the RAO is normalised by the amplitude at a gauge ~3.5 m upstream, so the body is under-forced and the measured RAO is biased low. The body's own dynamics — natural frequency and damping — are verified and consistent with the potential-flow reference. Measuring the near-resonance RAO directly in forced waves would need a body-adjacent incident-wave reference or a much finer short-wave grid (further work); it is a limitation of the forced measurement, not of the overset body model.
The still-water case reuses the verified #1302 geometry (depth 0.4 m, body at x = 13 m, nz = 49) with the
wave off (wave_height = 0) and the body, its overset component and its centre of mass raised
together by initial_heave_offset — a clean release from displacement. Heave is parsed from the
sixDoF report on log; the ring-down is fit to a damped cosine (FFT-seeded, with an independent
log-decrement cross-check). Builders in
digitalmodel.solvers.openfoam.validation.wave_excited_body_decay.
rao_reference_capytaine.csv.Generated from real OpenFOAM v2312 still-water solves. All chart data is computed directly from the solved body-state histories — no values are hand-entered.