Hull form study: how many facets should a chined hull have?#

Sunwake SW-15 · SW-FORM Rev 1 · August 2026

Companion to the lines plan and drawing sheet 3. This study tests the number of facets used in the early chined hull. At four strakes per side the stitch-and-glue hull gives up only a few tenths of a percent of bus power against round-bilge strip plank — less than tortured-plywood radius chine — and the earlier two-strake drawing was leaving well over a percent on the table. The result supports the plywood owner-build method with little hydrodynamic penalty. Everything below is produced by calc/sunwake/facets.py and can be re-run.

1. Method#

Michell's thin-ship wave-resistance integral, validated against the Wigley parabolic hull before use, plus ITTC-1957 friction with a form factor of 1.10, a roughness allowance, and a +15 % appendage margin on wetted surface common to every variant, plus the project's air-drag convention. Twin-hull interference at 5.00 m centres is included identically for all variants. Bus power is tow power divided by 0.58 × 0.90.

The implementation returns a Wigley wetted surface of 0.1488 L² against the published 0.1489, and reproduces the canonical hump-and-hollow structure of the wave-resistance curve — humps near Fn 0.23, 0.31 and 0.50 with hollows between. Both checks are asserted in the test suite.

Comparison basis. Every variant is derived from the solved hull, holding displacement, LCB, the sectional area curve, the design waterline and the draft identical. The round-bilge counterpart replaces each chined section with a superellipse of exactly the same area, waterline beam and draft. The faceted ladder then re-approximates that round section with N straight segments: vertices at equal girth fractions, pushed out along the local normal until the polygon area matches exactly. So every delta isolates section shape alone, and nothing else moves.

A note on counting. A strake is a plank of plywood running the length of the hull. This boat's fourth strake — the topside — begins at the boot seam above the waterline, so a four-strake hull has only three straight segments in the immersed section. The ladder is computed in segments and reported in strakes, because strakes are what a builder counts and what a kit nests.

2. The facet ladder#

Wetted surface is for one demihull, canoe body. The power tax is the range across 5–12 knots, against the round-bilge reference.

Chart: drawings/sheet3-sections-and-facets.svg.

Four findings.

1. Diminishing returns set in hard after four strakes. Going from three strakes to four recovers about three-quarters of the tax; four to five recovers a tenth of a percent more; four to fully round recovers the last few tenths. Four is the knee, and the knee is not subtle.

2. Chine placement matters as much as chine count. The as-drawn hull sits above the idealised four-strake ladder point — +0.90 % of wetted surface against +0.58 %. The closure solve that produced the offsets chose chine heights to carry the section area, not to minimise girth. A naval architect's refair can address the placement and may recover two or three tenths of a percent. It is carried as a refair item in the lines plan, finding 1.

3. Flat four-strake panels match or beat radius chine. Tortured constant- radius plywood at the bilge lands in the same few tenths of a percent but requires mould work and a different construction skill. The radius-chine option is therefore rejected.

4. Nothing here changes the boat's character. Friction is 70–80 % of total drag at the cruise speeds and the facet tax is mostly wetted surface, which is why the ladder is nearly flat across speed. No facet count makes this a different boat; it makes it a slightly slower or slightly faster version of the same one.

3. Build cost of additional strakes#

About 62 m of stitched, double-taped seam per boat per added chine, and 80–120 hours including fills and fairing at the seam.

The boot seam is nearly free. The 1.9 m-deep topside needs a longitudinal plywood joint somewhere regardless; putting it at z = +130 mm and taping it instead of butt-blocking it costs almost nothing and turns a joint into a design line — it is where the boot stripe goes.

Submerged strakes at four per side run 0.35–0.49 m wide amidships and total 1.20 m of girth, so three of them rip from one 1.22 m sheet. Kit nesting therefore improves slightly. The additional parts and stitching are partly offset by narrower panels that twist less and are easier to pull fair.

Compared with two strakes, four add roughly 150–250 hours to a 5,500-hour build, reduce bus power by about 1.5 %, and appear round-bilged at the waterline.

4. Effect on performance#

Changing from two strakes to a round-bilge hull changes solar-noon speed by about 0.07 knots. The weight ledger's exchange rate is 0.11 knots per 500 kg, so the full effect is comparable to removing about 300 kg. Rev 0a captures most of that benefit with two additional taped seams. Further work on facet count has little value compared with weight control or hull-spacing analysis.

5. Weight, structure and build method#

Plywood and epoxy taped-seam structure runs within about 150 kg of strip plank with carbon at this size, equivalent to roughly 0.03 knots. The build-method choice is therefore driven by cost and labour. The chines act as natural stringers, helping an ISO 12215-5 Category B scantling schedule work without exotic laminates. [ENG: the scantling book rules either way.]

Strip plank buys a dead-fair yard topside. That is cosmetic, not structural, and it is achievable in plywood with long boards and patience — which is exactly what the 950-hour fairing line in the build timeline is for.

The decisive line is the one that is not hydrodynamic at all: round-bilge strip plank re-creates a yard dependency — hundreds of thousands of dollars of shell scope, or thousands of owner hours plus mould building, against about 1,550 hours from a CNC kit. The stitch-and-glue premise is what makes this boat buildable by its owner.

6. Alternative deep round-bilge form#

Freed of developability and of the 0.95 m propeller-draft cap, a fresh-drawn slack round hull carrying its volume deeper shows a materially lower wave resistance at 8–10 knots in thin-ship theory — volume carried deep sheds divergent-wave energy. Treat that with care: Michell inflates deep-section benefits, its handling of an immersed transom is weakest exactly here, and slack-bilge form drag is unmodelled. Such a hull draws close to a metre at the propellers.

The Bahamas draft constraint rules it out. Under the current requirements, the achievable advantage of strip plank is limited to the values in §4. If the draft cap is removed for a future Maine-only version, the alternative should be tested with CFD before it is adopted.

7. Twin-hull interference#

The same model that arbitrates the chine question has something more useful to say. At s/L = 0.33, the two hulls' wave systems interfere strongly, and the effect dwarfs everything in §2:

A favourable dip of roughly 20–24 % sits exactly on the solar cruise, and an interference penalty of roughly 45–48 % sits exactly on the single-generator passage speed. Chart: drawings/chart-interference.svg.

Two conclusions follow. First, hull spacing is a first-order design variable on this boat and chine count is a third-order variable. Second, this is why the published power curve is a calibrated curve rather than a thin-ship prediction: the model runs 25–35 % hot against the design basis at 9–9.5 knots for exactly this reason, and it should be used for comparisons while CFD determines the absolute values.

[ENG: an interference sweep across s/L, at fixed beam overall, is the single highest-value CFD question in the package.]

8. Recommendation#

Use stitch-and-glue construction with four strakes per side. Two working chines put the hull within a few tenths of a percent of round-bilge strip plank — inside model noise and inside weight-ledger noise — for 150–250 added hours, while keeping the owner-build premise intact. The radius-chine option is rejected, and the remaining fraction of a percent does not justify a different build method.

Further effort should go to the weight ledger, chine placement during refairing, and the hull-spacing question.

9. Model caveats and [ENG] items#

Thin-ship theory on a hull of this beam is being used outside the assumptions it was derived under, and this document leans on it for comparisons rather than absolutes throughout. Re-integrating the published offsets closes displaced volume to within 0.2 %; the facet variants re-integrate to within 0.05 %.

[ENG] items, in priority order:

1. Variant CFD: four strakes against round bilge at matched hydrostatics.
2. The interference sweep at s/L (§7).
3. Panel development and unrolling of the four strakes — equal-girth chine lines are fair by construction, but developability has to be proven with a strain report per strake, and a designer who models freeform and unrolls with strain hidden in the middle of a panel produces a kit that fights the builder on the shop floor.
4. Transom and platform immersion in the 6–10 knot band.
5. Form factors per variant, rather than one 1.10 carried across all of them.