PROJECT 003 · MH1 · REFERENCE STUDY · MARINE HYDRODYNAMICS
KCS Case 2.1, six speeds, against the tank.
The KRISO Container Ship at model scale, six Froude numbers, free to sink and trim, compared against the Tokyo 2015 towing-tank measurement. This is the study we run to show a marine buyer what our work looks like end to end, and it is the template every hull job follows.
against a field that averages 3.0 to 3.4 percent across all Tokyo 2015 submissions
Conditions
- Hull
- KRISO Container Ship, bare hull with rudder, no propeller
- Scale
- 1:31.6, Lpp 7.2786 m, even keel at T = 0.3418 m
- Condition
- Calm deep water, free to sink and trim
- Range
- Six Froude numbers 0.108 to 0.282, Re 5.23e6 to 1.36e7
- Solver
- OpenFOAM interFoam, k-omega SST, three-grid convergence study
- Extrapolation
- ITTC-78 to the 230 m ship, with its sensitivity band
The comparator
Validation means agreement with a measurement this studio did not produce. Ours is named, cited, and downloadable by the reader.
Tokyo 2015 Workshop on CFD in Ship Hydrodynamics, Case 2.1 towing-tank data: resistance, sinkage and trim at six Froude numbers
SOURCE: NMRI AND KRISO
https://t2015.nmri.go.jp/Instructions_KCS/Case_2.1/Case_2-1.htmlPublic, free, and downloadable by the reader. The table below is transcribed from the workshop's own spreadsheet so you can check this page against the source in about ten minutes.
| FR | C_T X 10^3 | SIGMA (CM) | TAU (DEG) |
|---|---|---|---|
| 0.108 | 3.796 | -0.090 | -0.017 |
| 0.152 | 3.641 | -0.275 | -0.053 |
| 0.195 | 3.475 | -0.599 | -0.097 |
| 0.227 | 3.467 | -0.944 | -0.127 |
| 0.260 | 3.711 | -1.394 | -0.169 |
| 0.282 | 4.501 | -1.702 | -0.159 |
Positive sinkage is upwards, positive trim is bow-up. Reference area S0 / Lpp squared = 0.1803.
How the study is run
- 01
Thresholds first
Every acceptance criterion is written and published before the first production run, each cited to an ITTC procedure number or to a published workshop statistic. Thresholds are not edited after a result is read.
- 02
A grid family built for the study, not for the picture
Three grids geometrically similar at a refinement ratio of root two, including the prism stack, because a refinement that changes the near-wall layer changes the wall treatment and the triplet stops measuring discretisation error.
- 03
Wall treatment audited over area
The y+ distribution is reported as a fraction of wetted area inside the valid band for the chosen wall function, rather than as a single maximum that hides the tail.
- 04
Attitude in absolute units
Sinkage in millimetres and trim in degrees. A percentage error on a trim of 0.017 degrees describes the size of the quantity, not the quality of the prediction, which is why the workshop's own mean error on trim came out between 48 and 60 percent.
- 05
Uncertainty stated twice
Grid uncertainty reported by both the grid convergence index and the ITTC factor-of-safety method, side by side, because they disagree and the reader is entitled to both.
- 06
The archive reproduces
One command from a clean checkout, a pinned container image digest, a SHA-256 for every file, and the wall clock and core hours for each run on named hardware.
The bar we hold to
16 acceptance criteria in three bands. Every threshold is cited to the procedure or the published statistic it came from, written into the SOW, and fixed before the first production run. A threshold with no basis is not worth stating, and a threshold that moves after a result is read is not a threshold.
Verification
6 CRITERIAThat the numerical answer stands up on its own terms: convergence, grid, wall treatment, reproduction.
| ID | CRITERION AND BASIS | THE BAR |
|---|---|---|
| V1 | Iterative convergence of forces at the design speedITTC 7.5-03-01-01 | Force variation under 0.5 % over the last 10 wave periods |
| V2 | Grid triplet geometrically similar including the prism stackRequired for a defensible grid convergence index | Refinement ratio root two, layer count preserved on all three grids |
| V3 | Observed order of accuracy in rangeITTC 7.5-03-01-01 | 0.5 <= p <= 2.5, monotonic convergence |
| V4 | Grid uncertainty on total resistanceTokyo 2015 participant U_G spread of 2.57 to 9.39 % | U_G at or below 3.0 % of the computed value |
| V5 | Free-surface resolution at the design speedWavelength scales with Fr squared, so this is set at the design speed. At Fr 0.108 the same rule would demand sub-millimetre cells across the whole wave field. | At least 80 cells per wavelength streamwise at Fr 0.260 |
| V6 | Wall treatment audit over wetted areaReported as a distribution, not a maximum | 95 % of wetted area inside the valid y+ band for the wall function |
Requirements
4 CRITERIAThe commitments we hold ourselves to on every job, including the ones about how we behave with a threshold.
| ID | CRITERION AND BASIS | THE BAR |
|---|---|---|
| R1 | Every run reproduces from a clean checkoutStudio release rule | One command, pinned container digest, SHA-256 per file |
| R2 | Thresholds unedited since publicationMachine-checkable, and the point of the exercise | Threshold file hash matches the pre-registration commit |
| R3 | Machine time per speed on named hardwareCost transparency, so a reader can price the same job | At or below 48 h per speed on the production grid |
| R4 | No coefficient tuned after seeing the comparisonStudio release rule | Form factor, correlation allowance and model constants fixed pre-run |
Validation
6 CRITERIAAgreement with a measurement we did not produce. This is the band most published CFD leaves empty.
| ID | CRITERION AND BASIS | THE BAR |
|---|---|---|
| Va1 | Total resistance coefficient at the design Froude numberTokyo 2015 all-submission mean absolute error of 3.0 to 3.4 %. This bar beats the field. | Agreement within 2.0 % of the measured value at Fr 0.260 |
| Va2 | Total resistance across all six speedsTokyo 2015 all-submission statistics | Mean within 2.5 %, maximum within 4.0 % |
| Va3 | Mean sinkage at and above Fr 0.227Absolute units, and declared where the measured value is large enough for the band to discriminate. | Within 2.0 mm |
| Va4 | Trim at and above Fr 0.227At Fr 0.108 the measured trim is 0.017 deg, so this band would admit zero trim and the wrong sign. Stating it there would not be a test. | Within 0.02 deg |
| Va5 | Wave cut against measurementCompared where the measurement exists, and stopping where it stops | RMS difference within 15 % of first-crest amplitude |
| Va6 | Nominal wake fraction on the propeller planeTokyo 2015 wake comparison spread | Within 8 % of the measured value |
What the study shows
- Agreement with a towing-tank measurement this studio did not produce, at six speeds, stated as a number against a threshold set first.
- Sinkage and trim in millimetres and degrees, with the sign convention reproduced verbatim from the source sheet.
- A three-grid convergence study with the observed order of accuracy and two independent uncertainty estimates.
- Wall clock and core hours for every run on named hardware, so a reader can price the same job.
- An archive that re-runs from a clean checkout.
What it does not show
- Full-scale validation. The ITTC-78 extrapolation to the 230 m ship carries its sensitivity band, and no full-scale measurement of this hull exists to compare against.
- Propeller or self-propulsion performance. This is a bare hull with a rudder, towed.
- Seakeeping, added resistance in waves, manoeuvring or shallow water.
- A claim about any other hull. A method shown to work on KCS at these Froude numbers has been shown on KCS at these Froude numbers.
Every study here states the limit of what it covers. A reviewer looks for that limit first.
The same method, on your geometry.
Marine hydrodynamics engagements run the ladder from a two-week threshold study through to design closure. The check matrix you see here is the artefact you get first, written for your case, before anyone runs anything.