Wind-Assisted Propulsion I: The Technologies
Wind is back on merchant ships, not as a return to square rig, but as auxiliary power. The main engine stays, sized for calms and schedules; the wind device sits on deck and shaves load off the propeller whenever the weather cooperates. To evaluate one, start with the physics all four device types share, then the differences that decide which suits which ship.
Thrust from wind: lift, not drag
A device that simply catches the wind, a drag device, only helps when the wind blows from well behind the beam. Every serious modern system works the way a yacht beating to windward does: it generates lift, an aerodynamic force at right angles to the apparent wind, and lets geometry resolve part of that force along the ship's track. Lift devices produce useful thrust across a wide range of wind angles, including beam winds, and beam-ish wind is most of what a real route offers. Every newton of thrust the device delivers is a newton the propeller need not, and the engine throttles back accordingly. That is the entire saving mechanism; everything else is detail.
The four device types
Rotor sails are spinning vertical cylinders. Rotation drags air around the surface, speeding the flow on one side and slowing it on the other; the pressure difference is a lift force: the Magnus effect. Rotors deliver striking force from a small deck footprint, and their control input is one number: spin speed. The spin motors consume power, so the honest accounting is net of that draw.
Rigid wingsails are vertical aircraft wings, often with flaps or camber control. They are aerodynamically efficient and mechanically simple in operation, but they are tall fixed structures: air draught under bridges, shadowing between multiple wings, and interference with cranes and hatches are the binding constraints. Tilting and furling designs exist precisely to buy those constraints back.
Suction wings are short, thick wings with fans drawing air through a perforated skin. The suction keeps flow attached at angles where a plain wing would stall, so the device produces very high lift for its size: useful where deck length is scarce. Like the rotor, it spends some power to make more: the fans are its parasitic load.
Kites fly from the bow at several hundred metres, where wind is stronger and steadier than at deck level, and fly dynamic figure-eight patterns that multiply their apparent wind. Deck footprint is nearly nil, attractive where cargo gear rules the deck, but launch, recovery and autonomous flight control carry the operational complexity the deck-mounted types avoid.
A maturity signal, and matching device to ship
Bureau Veritas' survey of the field, rotors, wingsails, suction wings and kites, with costs and regulatory gaps, counted roughly 105 ships with wind-assist installations by mid-2024 (the report is free after registration). Small against the world fleet, but consider what the number means: a classed, insured, financed population across several device types and ship types. This is no longer a concept stage; class societies publish dedicated rules for these installations.
Which device suits which ship is mostly a deck-space negotiation. Bulkers and tankers, with long clear decks and no gear, take rotors and wingsails readily. Container ships fight for every deck slot, pushing interest toward kites and retractable designs. Ro-ro and ferry decks often suit rotors. Lloyd's Register's review of retrofit readiness walks through installation, safety and regulatory obstacles case by case (the report is open, with a downloadable full text).
One worry the technology has largely retired: none of these systems asks the crew to sail. Modern installations read wind and ship data continuously and trim themselves, spin speed, flap angle, suction, flight pattern, and depower automatically for squalls, port approaches and cargo operations. The bridge gets a status display and an override, not a sail plan. Heel and manoeuvring effects are assessed at the design stage under class rules written specifically for these installations.
What none of this yet tells you is the number the CFO wants: how much fuel a given installation saves on your route. That figure is a property of the route's wind statistics as much as of the hardware, which is why it gets the next lesson to itself.