For three weeks the marine-hull programme has gone deep on side questions: how a solar survey boat should correct its sunlight climatology, and what tack control is worth on a wing-sail drone. Its main question has been answered for a while, though, and until today the answer existed only as a Markdown draft. This step sets it as a paper in IEEE conference format, aimed at OCEANS (IEEE/MTS): paper/main.tex.
The result the paper carries
We put seven body types on one energy basis: a 200 kg craft, 25% of its mass battery, a 30 W payload. Every input we could not source was swept. Here is each rung against a 4 m Wigley-form monohull at its own best speed:
| Rung | Swept cases | Wins over the monohull |
|---|---|---|
| Catamaran (Michell waves, Havelock interference) | 6,561 | 216 (3.3%), none at equal length |
| Catamaran, form factor bounded by tank data | 6,561 | 0 |
| SWATH | 4,374 | 0 |
| Planing hull | 48 | 0 |
| Hydrofoil | 2,916 | 0 |
| Glider vs propeller AUV, equal hotel load | 2,430 | glider 0 |
| Glider vs REMUS-100-class AUV, vendor-implied hotel load | 810 per row | glider in 100% |
The ranking holds. The absolute range does not. The hull-form constants together spread best range by 20.6%, though none moves it by 10% on its own. With deck solar, the payload's power draw moves km/day more than any hull input does (-38.2% over 20 to 45 W, against -9.2% for the wetted-surface constant). The 5 m craft makes 133.4 km/day at the Equator and 69.4 in the Gulf of Alaska, and the site order held in all 500 joint draws of the power-side inputs. The practical reading for anyone buying a survey craft: measure the payload's power draw before worrying about the hull's drag.
What the conversion turned up
The step ran no experiment. Every number is copied, and the copying is now checked: paper/check_numbers.py pairs each of the 53 quoted figures with the results file it comes from, and a test fails if a rerun changes any of them. The one figure is drawn from the results file by a script, so nobody types its values.
Checking caught two places where the Markdown draft was looser than its files. The draft said the no-form-factor model lands within +1% to +5% of the one Wigley tank point. Over 10 to 25 °C the file says -1% to +5%. The draft's "0 of 729 equal-length" catamaran cases is a per-wave-scale count; the file's total is 0 of 2,187. Neither change touches a conclusion, and the paper uses the file's figures.
What it is not yet
It runs 3 pages, and OCEANS allows 6 to 10. It has no range-against-speed figure per rung. The form-factor bound rests on one Wigley point, one 1 m model and 22 keeled yacht hulls, not on a resistance curve of a keel-less survey hull, and the threats section says so first. Crossref was unreachable this session, so one citation written from memory (Michell 1898) carries no DOI and is flagged to look up. The solar-correction and wing-sail chains are left for companion papers. The draft has not been submitted anywhere.