Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -143,6 +143,14 @@ All notable changes to CableDyn are recorded in this file. The format follows
branch, reported as a sensitivity. The case's script is now `wang_scr_viv.py` for both parts
(it replaces `wang_scr_uniform_current.py`; the L3-W numbers are unchanged), and its data,
including Thorsen et al.'s simulated lines, are read by `extract_thorsen2017.py`.
- Validation case L3-O (`validation/experiments/fu_oscillatory`): cross-flow VIV of the SJTU
flexible cylinder of Fu et al. (2014) moved harmonically through still water (KC 178, U_R,max 4,
6.5 and 7.9), A/D at mid-span on the acceleration and deceleration paths against the measured
paths of their Figs. 15-17 (read from the figures by `extract_fu2014.py`), with three published
coefficient sets and a `VIV FRAME` row. Every run converges through the flow reversals. At
U_R,max 4 the wake oscillator gives the measured largest A/D (0.280 against 0.286); at 6.5 and
7.9 every set holds the lock-in while the flow slows, the gap between the paths 2-5 times the
measured one and the largest A/D 1.3-2.3 times the measured.
- Theory manual, Published coefficient sets: the VIV coefficient sets of Thorsen et al. (2016,
2017), Ulveseter et al. (2018), Andersen et al. (2024a, b) and Facchinetti et al. (2004), each
with its source, its test data and the line-type columns it maps to, and Thorsen et al.'s
Expand Down
4 changes: 2 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -171,9 +171,9 @@ layout is described in the architecture notes in the repository.
## Validation

CableDyn is verified against analytical solutions and compared with MoorDyn,
OrcaFlex, seven published physical experiments (Holcombe lazy-wave, Bergdahl dynamic
OrcaFlex, eight published physical experiments (Holcombe lazy-wave, Bergdahl dynamic
chain, Chaplin riser natural frequencies and stepped-current VIV, Wang catenary-riser VIV in
current and heave, NDP high-mode riser VIV, FLOTANT power-cable VIV), the IEA 15 MW VolturnUS-S reference systems, and coupled OpenFAST runs. Results,
current and heave, NDP high-mode riser VIV, FLOTANT power-cable VIV, oscillatory-flow cylinder VIV), the IEA 15 MW VolturnUS-S reference systems, and coupled OpenFAST runs. Results,
acceptance criteria, and reproduction commands are in
the [validation record](https://cabledyn.readthedocs.io/en/latest/validation.html); data and scripts are in
[`validation/`](https://cabledyn.readthedocs.io/en/latest/validation.html).
Expand Down
22 changes: 22 additions & 0 deletions VALIDATION.md
Original file line number Diff line number Diff line change
Expand Up @@ -88,6 +88,7 @@ article's methods, are compared in
| **L3-Wh** | SJTU steel catenary riser, heave-induced cross-flow VIV in still water (measured top motion, 5.96 s) | measured r.m.s. strain (25 stations, band-passed 0.67–50 Hz) and dominant frequency | no `VIV FRAME`: frequency 1.019 Hz vs close to 1 Hz; max r.m.s. strain +20.8%, station mean +33%, lower/upper half 1.71/0.93 (Thorsen et al. 2017: +5.6%, +8.7%, 1.44/0.71); `VIV FRAME` `Tf` 0.35–1 s: two branches, 1.01–1.02 Hz (max +7 to +13%) or 0.88–0.89 Hz (max −2.8 to −4.2%), `Tf` 0.5 s changes branch at dt/2 | published coefficient set, no fitted parameter; heave period about 6 VIV periods, below the `VIV FRAME` scale separation | Pass measured |
| **L3-N** | NDP 38 m high-mode riser in uniform flow, 0.3–2.4 m/s (22 tests), cross-flow VIV | measured maximum r.m.s. cross-flow strain, dominant frequency, spatial-mean σ/D | strain r.m.s. log error: Ulveseter et al. (2019) set 0.178, Andersen et al. (2024b) 0.186, generic set 0.277, Thorsen et al. (2016) 0.363 (their own simulation 0.187); f D/U 0.145 / 0.134 / 0.192 / 0.193 vs 0.143 measured | generic set blind; the other sets fitted by their authors to NDP tests (implementation check) | Pass measured |
| **L3-F** | FLOTANT OD 48 mm power-cable specimen with a 200 kg clump, tow tests at 0.025–1.0 m/s (40 tests), cross-flow and in-line VIV | measured A/D_rms (σ/OD) at 1/2L, 3/4L and L (de Wilde et al. 2021, App. 3) | CF r.m.s. log error (Ur ≥ 3): Ulveseter et al. (2019) set 0.731, generic set 0.594, wake oscillator 1.208; first peak at 1/2L 0.755 / 0.483 / 0.296 vs 0.776 measured; release Ur 7.06 / 5.89 / 5.30 vs 5.88 | published VIV sets, none fitted to these tests; structure identified from the specimen's in-air and in-water RAOs | Pass measured |
| **L3-O** | SJTU 4 m flexible cylinder moved harmonically through still water, KC 178 at U_R,max 4, 6.5 and 7.9, cross-flow VIV | measured A/D on the acceleration and deceleration paths at mid-span (Fu et al. 2014, Figs. 15–17); lock-in shares 17/34/50 % | A/D r.m.s. error: wake oscillator 0.088 / 0.177 / 0.132, generic set 0.138 / 0.339 / 0.345, Ulveseter et al. (2019) set 0.079 / 0.340 / 0.338; largest A/D at U_R,max 4: 0.280 / 0.472 / 0.124 vs 0.286; at 6.5 and 7.9 every set holds the lock-in while the flow slows: mean deceleration-acceleration gap 0.25–0.50 vs 0.09–0.12 measured | published VIV sets, none fitted to these tests; `VIV FRAME` Tf 0.2 s (sensitivity none, 0.1, 0.3 s) | Pass measured |
| **L-fig** | Along-arc curvature overlay (CableDyn / OrcaFlex / MoorDyn-F ±4×) | 3-code committed record | CableDyn peak 0.11% from OrcaFlex | <0.5% | Pass |
| **L4-1** | Coupled platform + mooring | synthetic reference | bit-for-bit | — | Pass |
| **L4-1b** | VolturnUS-S coupled platform + power cable | OpenFAST v5.0.0 (CompMooring 5 vs 3) | platform DOFs <0.6% (3-seed DLC-1.1); fairlead and anchor <0.001% static | <3% platform | Pass |
Expand Down Expand Up @@ -450,6 +451,27 @@ the tow tests it predicts). Inputs, reduction scripts, results, and source prove
Half the time step, two elements per table element and 120 s runs change no A/D_rms by more
than 0.006 at five speeds; the default run reproduces the committed record byte for byte.

- **L3-O SJTU flexible cylinder in oscillatory flow, cross-flow VIV hysteresis**
([Fu et al., 2014](https://cabledyn.readthedocs.io/en/latest/references.html#ref-fu2014)):
the 4 m cylinder (D 24 mm, pretension 500 N) moved across its axis through still water, A_m
0.68 m (KC 178) at periods 16.5, 10.2 and 8.45 s (U_R,max 4, 6.5 and 7.9), so the relative flow
reverses twice per cycle. The cross-flow A/D at mid-span is compared against the instantaneous
U_R on the acceleration and deceleration paths with the measured paths of Figs. 15–17 (read from
the raster figures; the traced gaps match the figures' annotations 0.15D and 0.2D), and the share
of each half-cycle above √2/2 of its peak with the article's lock-in shares. Three published sets,
none fitted to these tests, each with a `VIV FRAME` row (Tf 0.2 s), without which the wake
oscillator does not shed for a cylinder moving through still water. Every run converges through
the flow reversals. At U_R,max 4 the wake oscillator gives the measured largest A/D (0.280
against 0.286), the generic set is 65% high and the Ulveseter et al. (2019) set barely responds.
At U_R,max 6.5 and 7.9 every set holds the lock-in while the flow slows: the deceleration path
stays near its peak, the mean gap between the paths is 2–5 times the measured one and the largest
A/D 1.3–2.3 times the measured; the wake oscillator comes closest (A/D r.m.s. error 0.18 and
0.13). Without the row the synchronisation sets respond less (the generic set 0.034 at U_R,max 4,
the best of any run) but keep the hold; with Tf 0.1 and 0.3 s their largest A/D moves by up to
21%, the wake oscillator's by up to 11%. Half the time step and 80 elements change the largest A/D
by at most 0.1% (generic set, wake oscillator) and 4.1% (Ulveseter et al. set); the default run
reproduces the committed record byte for byte.

## L4 coupled OpenFAST

- **L4-1 synthetic** (`l4_platform_mooring`): 6-DOF platform moored by three taut
Expand Down
6 changes: 6 additions & 0 deletions doc/references.rst
Original file line number Diff line number Diff line change
Expand Up @@ -272,6 +272,12 @@ Verification and validation references
catenary riser from vortex-induced vibration caused by vessel motions. *Marine Structures*
**39**, 131–156. https://doi.org/10.1016/j.marstruc.2014.07.002

.. _ref-fu2014:

- **Fu, S., Wang, J., Baarholm, R., Wu, J. & Larsen, C. M. (2014).** Features of vortex-induced
vibration in oscillatory flow. *Journal of Offshore Mechanics and Arctic Engineering*
**136**\ (1), 011801. https://doi.org/10.1115/1.4025759

.. _ref-davies2026:

- **Davies, R., Lekkala, M. R., Michaloliakos, A., Wong, W.-Y., Hall, M., Vakakis, A. & Zuo, L.
Expand Down
2 changes: 1 addition & 1 deletion validation/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ redistributed.
| [`RELEASE_0_1_1.md`](RELEASE_0_1_1.md) | The same for v0.1.1 |
| [`RELEASE_0_1_0.md`](RELEASE_0_1_0.md) | The same for v0.1.0 |
| [`bodies/`](bodies/README.md) | Rigid-body, rod and shared-anchor cases with MoorDyn-C twins, stored references and pass limits |
| [`experiments/`](experiments/README.md) | Holcombe et al. (2025) static lazy-wave, Bergdahl et al. (2016) dynamic-chain, Chaplin et al. (2005) riser natural-frequency and stepped-current VIV, Wang et al. (2014) catenary-riser VIV in current and heave, Trim et al. (2005) NDP high-mode riser VIV and de Wilde et al. (2021) FLOTANT power-cable VIV comparisons |
| [`experiments/`](experiments/README.md) | Holcombe et al. (2025) static lazy-wave, Bergdahl et al. (2016) dynamic-chain, Chaplin et al. (2005) riser natural-frequency and stepped-current VIV, Wang et al. (2014) catenary-riser VIV in current and heave, Trim et al. (2005) NDP high-mode riser VIV, de Wilde et al. (2021) FLOTANT power-cable VIV and Fu et al. (2014) oscillatory-flow cylinder VIV comparisons |
| [`PERFORMANCE_0_1_0.md`](PERFORMANCE_0_1_0.md) | Wall time of CableDyn 0.1.0, MoorDyn-C and MoorDyn-F at equal accuracy on five mooring, cable, buoy and coupled cases |
| [`performance/`](performance/README.md) | Same-host wall-clock records of the L5 cases: the resolved lazy-wave spans, the one-hour VolturnUS-S mooring and the tension-only chain |
| [`PAPER_REPRODUCTION.md`](PAPER_REPRODUCTION.md) | Every CableDyn result in the journal article, re-run with the release defaults and in paper mode |
Expand Down
1 change: 1 addition & 0 deletions validation/experiments/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,5 +16,6 @@ committed results.
| [`wang_scr/`](wang_scr/README.md) | Wang et al. (2014), truncated steel catenary riser in uniform current and under heave motion of its upper end, cross-flow VIV | Finite-EI riser, synchronisation VIV |
| [`ndp_riser/`](ndp_riser/README.md) | Trim et al. (2005), NDP 38 m high-mode riser in uniform flow, cross-flow VIV with a generic and three published coefficient sets | Finite-EI riser, synchronisation VIV |
| [`flotant/`](flotant/README.md) | de Wilde et al. (2021), FLOTANT OD 48 mm power-cable specimen with a clump weight, tow tests and forced-oscillation RAOs, cross-flow and in-line VIV with three published coefficient sets | Finite-EI cable, synchronisation and wake-oscillator VIV |
| [`fu_oscillatory/`](fu_oscillatory/README.md) | Fu et al. (2014), SJTU 4 m flexible cylinder moved harmonically through still water, cross-flow VIV on the acceleration and deceleration paths with three published coefficient sets | Finite-EI cylinder, synchronisation and wake-oscillator VIV, `VIV FRAME` |

The results are summarised as cases L3-H, L3-B, L3-C, L3-D, L3-W, L3-Wh, L3-N and L3-F in [`VALIDATION.md`](../../VALIDATION.md).
116 changes: 116 additions & 0 deletions validation/experiments/fu_oscillatory/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,116 @@
<!-- SPDX-License-Identifier: Apache-2.0 -->

# SJTU flexible cylinder in oscillatory flow: cross-flow VIV hysteresis

This directory compares CableDyn's cross-flow VIV of a flexible cylinder moved harmonically through
still water with the tests of Fu et al. (2014), <https://doi.org/10.1115/1.4025759>. The relative
flow reverses twice per motion cycle and the shedding frequency sweeps from zero to its largest value
and back, so the case tests what a steady current cannot: the VIV model through flow reversal, and
the response on the accelerating and decelerating halves of each half-cycle. Fu et al. measure
different amplitudes on the two halves (hysteresis) and give the paths for three cases. No model
parameter is fitted to these tests.

## Model and data

- **Cylinder** (Table 1): length 4 m, diameter 24 mm, 0.69 kg/m, EI 10.5 N m², EA 6.67e5 N,
pretension 500 N; their calculated first natural frequency in water f1 2.68 Hz (with C_a 1)
defines U_R. The ends are universal joints (pinned here), 40 elements, dt 1 ms (the axial period in
8 steps), C_d 1.2 and C_a 1.
- **Motion:** the whole cylinder moved across its axis, y = A_m sin(2πt/T), A_m 0.68 m (KC 178), with
a one-period cosine ramp, 8 cycles after it. The three KC 178 cases are the periods of the captions
of Figs. 15-17, T 16.5, 10.2 and 8.45 s, U_R,max = 2πA_m/(T f1 D) 4.03, 6.51 and 7.86 (the
article's 4, 6.5 and 7.9).
- **Measured** (Figs. 15-17, station CF4): A/D against the instantaneous U_R = |V|/(f1 D) on the
acceleration and deceleration paths, averaged by Fu et al. over 10 middle cycles of 15 (their
Eq. 20). `extract_fu2014.py` reads the paths from the figures, embedded raster images, calibrated
on the plot frame and traced by colour, at U_R steps of 0.25 (`fu2014_hysteresis.csv`). The traced
gaps between the paths, 0.164 at U_R 3.5 (U_R,max 6.5) and 0.222 at U_R 4 (7.9), match the
figures' own annotations, 0.15D and 0.2D.
- **Model output:** the cross-flow (vertical) displacement at mid-span, band-passed 0.5-25 Hz, its
Hilbert envelope over D, averaged in U_R bins of ±0.125 over the half-cycles from two periods on,
on the two paths. CF4 is the fourth of seven cross-flow stations at 0.5 m spacing (Table 2), taken
as mid-span.
- **Lock-in share:** the share of each half-cycle the envelope spends above √2/2 of its largest value,
the article's definition (Sec. 4.6), against the shares it gives for KC > 100: about 17, 34 and 50 %.

## Coefficient sets

The sets of the FLOTANT and NDP cases, none fitted to these tests:

- **uls19:** Ulveseter et al. (2019), Table 1 (Cl 0.85, cF 0.144, dF 0.064; VIV ClX 0.75, alpha 0.15,
Chh 1.5), fitted by them to the NDP riser;
- **b1:** the generic set of the Delta Flume case (Cl 0.8, cF 0.18, dF 0.08; ClX 0.85, alpha 0.15,
Ulveseter et al. 2018, Table 3; VIV AMPLITUDE with the drag amplification of Chaplin et al. 2005,
Eq. 4; VIV DAMPING 0 0 1);
- **wake:** the wake oscillator of Facchinetti et al. (2004) (CL0 0.3, St 0.2, eps 0.3, A 12),
cross-flow only.

The cylinder moves through still water, so every set runs with a `VIV FRAME` row: without it the
wake oscillator never sheds (the ambient flow is zero) and the amplitude estimate and damping split
of b1 act across a flow that is not there. Its filter time cannot meet both of the frame's
conditions here: the motion period is 23-44 periods of f1, around the 30 the conditions need.
The primary runs take `Tf` 0.2 s (ωTf 3.4 at f1; ω_m Tf 0.08, 0.12 and 0.15 at the three periods), and
the record reports the runs without the row and with `Tf` 0.1 and 0.3 s.

## Results (primary runs, `Tf` 0.2 s)

| U_R,max | Set | A/D r.m.s. error | Mean gap dec. − acc. (measured) | Largest A/D (measured) | Lock-in share (measured) |
|---|---|---|---|---|---|
| 4 | uls19 | 0.079 | −0.005 (+0.009) | 0.124 (0.286) | 0.08 (0.17) |
| 4 | b1 | 0.138 | +0.123 (+0.009) | 0.472 (0.286) | 0.31 (0.17) |
| 4 | wake | 0.088 | +0.106 (+0.009) | 0.280 (0.286) | 0.27 (0.17) |
| 6.5 | uls19 | 0.340 | +0.426 (+0.089) | 1.007 (0.475) | 0.34 (0.34) |
| 6.5 | b1 | 0.339 | +0.364 (+0.089) | 0.874 (0.475) | 0.53 (0.34) |
| 6.5 | wake | 0.177 | +0.271 (+0.089) | 0.609 (0.475) | 0.46 (0.34) |
| 7.9 | uls19 | 0.338 | +0.498 (+0.123) | 1.043 (0.450) | 0.34 (0.50) |
| 7.9 | b1 | 0.345 | +0.380 (+0.123) | 0.937 (0.450) | 0.60 (0.50) |
| 7.9 | wake | 0.132 | +0.255 (+0.123) | 0.570 (0.450) | 0.50 (0.50) |

The largest measured A/D is the largest on the U_R grid; the measured deceleration line ends at 0.32 at
U_R 4.12 in Fig. 15.

- **Through the flow reversals** every run converges, the synchronisation phases and the wake
variable running continuously through U → 0.
- **U_R,max 4:** the wake oscillator gives the measured largest amplitude (0.280 against 0.286); uls19
barely responds (its band centre, 0.144, is reached only near the velocity peak); b1 is 65 % high.
- **U_R,max 6.5 and 7.9: every set holds the lock-in on the deceleration path.** The deceleration path
stays near its peak while the flow slows, so the mean gap between the paths is 2-5 times the
measured one and the largest amplitude 1.3-2.3 times the measured. The measured hysteresis is the
same in kind (the deceleration path above the acceleration path), smaller in size. The wake
oscillator comes closest.
- The same persistence shows in the FLOTANT case, whose lock-in releases late for the set that
reaches its peak.

## Sensitivity and convergence

- **Without `VIV FRAME`:** the wake oscillator does not shed (largest A/D 0.000-0.021). The
synchronisation sets respond less: b1 0.297, 0.769 and 0.760 (A/D r.m.s. error 0.034 at U_R,max 4,
the best of any run there; 0.229 and 0.212 at 6.5 and 7.9), uls19 0.080, 0.738 and 0.824. The
deceleration hold remains (mean gap +0.25 to +0.35 at 6.5 and 7.9).
- **`Tf` 0.1 and 0.3 s:** b1's largest A/D 0.404-0.511, 0.710-0.922 and 0.738-0.978; uls19
0.110-0.139, 0.945-0.969 and 1.096-1.141; the wake oscillator 0.273-0.284, 0.607-0.621 and
0.579-0.633. The synchronisation sets depend on the filter time because their amplitude estimate and
damping split take the filtered frame; the wake oscillator's wake frequency barely does.
- **Convergence (U_R,max 6.5):** half the time step changes the largest A/D by +0.03 % (b1), −0.1 %
(wake) and −4.1 % (uls19); 80 elements by +0.1 %, +0.07 % and −4.1 %.
- The default command, run into a new work directory, reproduces `response_comparison.csv` and
`summary.json` byte for byte.

## Limitations

- The measured A/D is reconstructed from strain by a sinusoidal modal fit with constant tension (Fu
et al., Sec. 3.2), while the tension varies with the motion; the model's A/D is the displacement.
- The submergence of the cylinder, the band-pass cut-offs and the structural damping are not given;
the model takes 2 m, 0.5-25 Hz and no structural damping.
- The paths are one station (CF4) and three cases; Fu et al. do not show the scatter between cycles.
- The lock-in shares are "about" values from the text, for KC > 100.

## Reproduction

```text
python extract_fu2014.py FU2014.pdf
python fu_oscillatory_viv.py path/to/cabledyn.exe WORKDIR --jobs 8 --summary .
```

The executable of the record is named by its SHA-256 in `summary.json`; `provenance.json` gives the
sources, the data used and the build.
Loading
Loading