Seastead Design Review — RIM Drive Coasting & Kite-Robot Assessment

1. Do RIM drives have a “spin freely” mode?

YES — INHERENTLY

A rim drive is a permanent-magnet synchronous machine (magnet ring rotor, stator in the duct). If the inverter simply opens all three phase legs (high-impedance coast), there is no current path and therefore no electromagnetic torque. The prop spins freely, resisted only by bearing friction, cogging ripple (which averages out in rotation), and the hydrodynamic drag of the blades themselves. There is no gearbox, no shaft seal, and no oil pump to complicate coasting — mechanically it is trivial. The catch: whether you get true coast depends on the vendor's firmware defaults, so make it a written requirement.

Three electrical states to distinguish

State	What the controller does	Drag result
Coast (what you want)	All gate switches off; phases floating (open-circuit)	Near-zero electromagnetic torque; only bearing friction + cogging + blade drag
Active zero-torque	FOC running with zero torque command	Nearly identical drag; costs a little control power. Acceptable substitute.
Brake / regen	Shorted phases or controlled generation	Significant drag. Bad as a default — but useful as a selectable feature (below).

Questions to put to any thruster vendor

Is there a documented coast / high-Z neutral, and is it the default “off” state?
On loss of control power, does the unit coast or brake? (It must coast.)
Rated maximum unpowered freewheel RPM for the bearings (surf/surge can back-drive the props).
Is selective regeneration available (per-drive enable, min RPM, current limits)?
Any published tow-test drag data for the freewheeling condition — or approval to tow-test?

Ballpark drag numbers (order-of-magnitude estimates — confirm by tow test)

Condition, per 18″ unit at 6–7 kt	Typical added drag
Freewheeling (coast mode)	~10–30 lbf
Locked / stopped prop	~2–4× the freewheeling figure
Duct + support structure	Present regardless of prop state

For your six drives, that suggests roughly 60–180 lbf total freewheeling drag versus several hundred if locked — meaningful when the kite is doing the pushing. Antifoul the props and ducts: fouling multiplies coasting drag. A single day of tow testing alongside a dock will replace all of these estimates with real numbers.

Bonus capability — selectable regen. Because these are PM machines, you can optionally switch individual drives to generate while kite-sailing. That gives you: (a) battery charging from wind propulsion, (b) an adjustable-drag speed limiter, and (c) asymmetric port/starboard regen as an extra steering/roll-trim channel that is failure-independent of the thrust commands. Make it switchable per drive so you can choose pure coast vs. regen at will.

Side note: you already operate a rim drive — the Yamaha HARMO on the dinghy is one.

2. What do I think of the kite idea?

KEEP IT

The concept is physically sound, every subsystem has maritime precedent, and it gives you a fourth, genuinely independent propulsion-and-steering channel — consistent with your triple-redundancy philosophy. The real engineering risk is not “does it work in principle” (it does) but operational complexity: managing a 20–50 kite stack reliably at sea. Prototype incrementally, starting with one kite.

2.1 What's right about it

Your steering-by-position physics checks out. A downwind-directed force applied aft of the rotational center swings the bow upwind; applied forward, it bears away. This is exactly CE-vs-CRL behavior on a sailboat (rake the mast aft → weather helm) and exactly why a windsurfer drops the rig toward the tail to luff up and stop. Your fore/aft robot position is a smooth, reversible steering input, and it composes cleanly with your differential thrusters.
Altitude advantage. Wind at 50–150 m is stronger and steadier than at deck level. No mast-mounted sail of reasonable size could match it on this platform.
Lateral resistance is credible. Symmetric foil sections work on both tacks, and batteries low in the legs deepen the CG — helping both leeway resistance and heel stiffness.
Safety geometry is correct. Lines always lead outboard of the railing on the downwind side; the rail empties completely for storms; kites are stowable fabric, unlike a mast.
Precedent exists for every piece: kite-towing of ships (SkySails), stacked sport kites on dual lines, and grooved-wheel trolleys on I-beam tracks (crane/festoon systems). The integration is novel; the components are not.

2.2 Order-of-magnitude numbers (estimates — verify by test)

Quantity	Rough estimate	Basis
Kite stack thrust	~400 lbf @ 20 kt apparent wind; ~600 lbf @ 25 kt (for ~320 ft² total area, C_L ≈ 0.9)	T ≈ ½ρV²C_LA — same order as your full six-thruster suite, so a credible sole backup in useful winds
Heel at full pull	Low single digits of degrees	~500 lbf × ~12–14 ft arm ≈ 6–7k ft·lb; corner-spread legs + deep batteries give righting moments an order larger at small angles. Verify with a dockside pull test.
Speed under kite alone	~5–8 kt in moderate wind	Foil-leg hull is low-drag; a few hundred lbf of surplus goes a long way
Weakest point	Upwind VMG	The 44×7 ft house is ~300 ft² of bluff windage dead ahead when beating. Expect excellent reaching/running, modest beating — use thruster assist for light-air upwind work.

2.3 Challenges and mitigations

Challenge	Mitigation
20–50 individual kites = many connections, slow setup, tangle risk	Cartridge architecture: pre-rigged groups of 5–10 kites on short line segments joined by soft shackles. Power is adjusted in cartridge units; each stows in its own bag on a rack.
Repositioning the robot while the kite is loaded	Unload-to-move procedure: fly the stack to zenith/window edge to dump tension, drive the robot, re-load. This lets the rail drivetrain stay light (friction drive + brake) instead of fighting full kite force.
Gust / snatch loads on rail and wall	Design rail anchors for ~2.5–3× steady pull; fit calibrated weak links that fail before the wall structure; elastomer end stops; strain/load telemetry from the robot fed to both autopilots (which also improves your convoy motion-damping).
Dual-line control at high loads; chafe	Spectra line with chafe gear at every contact point; scheduled inspection; spares. Consider a single tow line + servo-controlled bridle pod on the robot as a simpler alternative to dual-line stacks.
Launch/recovery at sea	Formal drill: point downwind (as you planned), minimum-tension attach, add cartridges incrementally; reverse to recover. Practice first in <10 kt.
Robot traction and the curved bow section	Crane-trolley-style wheel sets (grooved tread + side-thrust + up-stop wheels) handle curved I-beam tracks routinely; specify the minimum radius your bogie can negotiate. Spring-applied, fail-safe clamp brakes.
Robot power	Your instinct is right: skip regen-by-tugging-the-robot; a cord from any of the three buses is cheaper and simpler. Optionally add a small battery + trickle solar for autonomy, plus a “kill everything” state.
Lightning, corrosion, UV	Bond the rail; isolate or sacrifice fittings; don't fly in thunderstorms; rinse and rotate kite cartridges for UV life.
Convoy operations (two seasteads bridged)	Ops rules: restrict kite work to the aft vessel, or opposite rails, so lines never cross the walkway; share wind data between the two computers.
Tension-leg mooring	Interlock: land and depower the full stack before setting tension legs. A moored hull with a live kite is a load case you don't want.

2.4 Recommendations

Adopt the cartridge architecture before anything else — it converts “managing 50 kites” into “managing 5–10 modules.”
Codify your lean-limiting idea in software: a wind-speed → max-cartridge map, with automatic depower on gust or wind shift.
Put load cells / motor-current telemetry on the robot and route it into both autopilots — kite force becomes a measured input instead of a guess.
Tow-test the six coasting drives early (cheap, answers the drag question with real numbers, and validates the kite-mode drag budget).
Minor spec cleanup: your text says legs are 21.5 ft long but cites 0.5 × 14.5 ft for the half-submerged dimension. Worth reconciling — it sets freeboard, floor height, and therefore the kite heeling arm.

2.5 Suggested prototype roadmap

Dockside static pull test on a rail section: validate 2.5× design load and the weak link.
Single kite, manual control, thrusters off: prove steering-by-position.
One cartridge (3–5 kites), automated elevation/figure-eight control.
Full stack sea trials, then convoy operations.

3. Bottom line

RIM drives: Yes — coasting is inherent to a PM rim drive; get “open-phase coast as default neutral” in writing from the vendor, confirm fail-to-coast on power loss, and tow-test. Treat selectable regen as a free bonus: charging, drag trimming, and an independent steering channel.

Kite robot: Keep it. The steering-by-position principle is sound, the redundancy is real, and the altitude advantage is something no mast could give this platform. Spend your prototyping budget on stack handling and load management — cartridges, unload-to-move, weak links, telemetry — not on whether the physics works.