A whole bunch of years ago I had a nifty Idea for a windmill design. I'll be describing it in a separate Idea from this one, later, and then add a link here. At the moment the relevant fact is that that old Idea employed sails, much like original Dutch-style windmills had fabric sails on wooden frameworks.
Well,
when the wind begins to blow too hard, it is important to "furl" the sails (basically, roll them up), to keep the windmill from rotating too fast (by reducing the total surface area of sail that the wind acts on). In the Age of Sailing Ships, furling the sails was basically done by hand, by folks who climbed up to where the sails were. Dangerous!
Now sails basically come in two types, square-ish and triangular. On a sailboat, the square-ish sails were usually furled from the bottom to the top, and then tied in place. Triangular sails are almost always right-angle triangles, and are either furled horizontally at the vertical mast, pulling the sail fabric toward the mast, or furled vertically at the horizontal boom, pulling the sail fabric down toward the boom.
Nowadays electric motors do the furling, but someone still has to press the button to turn those motors on or off, or to run them in reverse (to unfurl sails after the wind stops blowing too hard).
So, obviously we need an automatic system to furl/unfurl sails, that quickly responds to wind-conditions. This Idea is correct for the Windmill category, and not for sailboats, solely because only windmills incorporate rotary motion as part of their basic functionality, and this Idea depends on that rotary motion.
(Of course you might install a sail-type windmill ON a sailboat, replacing the ordinary sail system with something that powers a propeller, after which this Idea could be relevant to those "sail" boats, too.)
We start by noting that because wind tends to blow harder above tree-top level than near the ground, most windmills are very tall, so that their blades (sometimes called "sails" even today) are high above the ground. That's A Good Thing for This Idea, as you'll see shortly.
Naturally we assume that the type of windmill under discussion here employs actual sails that need to be furled occasionally (existing types "feather"/reorient their blades so they meet a hard wind edge-on instead of face-on).
Now I have to change the subject momentarily, to talk about an ordinary roll-up window shade. The shade/sheet starts wrapped around a roller/rod, much like a furled sail. The roller is attached to a "torsion spring", which is tensed-up as you pull the shade down, and which is able to roll the shade back up again, later. Note that the bottom part of the shade incorporates a stiff ("shaping") rod that keeps the bottom straight and parallel with the top (prevents foldings such as window drapes tend to have), and moves with the shade.
We want something of an opposite of that window-shade system here. We want the unfurled state of each windmill sail to be associated with an untensioned torsion spring. As we furl the sail, the spring tension increases -- if the spring was allowed to loosen, the sail would return to being unfurled.
The overall framework for holding the sail (as if we were discussing a ship's sail) will involve an upper horizontal part associated with the roller/rod, and a lower horizontal part that will be associated with two pulleys. Look at this ASCII sketch:
|O|
|--| (ignore the dashes)
|O|
If the upper "O" is an end-view of the roller/rod, and the lower "O" is a pulley, then we can have a simple looping line (thin rope or equivalent, the vertical bars in the sketch) around the roller/rod and the pulley.
With a pulley associated with each end of the upper roller/rod, we will attach the lower shaping rod to the two moving lines. This keeps the bottom of the sail from flapping in the breeze, and, as we get around to making the upper roller/rod rotate, the lower shaping rod and attached sail smoothly raises or lowers.
Note that verbiage such as "upper rod" and "lower rod" become less relevant when talking about rotating windmill sails! But the basic Idea should carry over satisfactorily.
Next, we need another line wrapped around the roller/rod. If we pulled on this line, it would unwrap, causing the roller to rotate, while simultaneously causing the sail to be furled around the roller, and tensioning the torsion spring.
The unwrapped end of this line will be attached to a free-weight. IT will do the pulling, trying to unwrap the wrapped end....
SO: As the windmill rotates, good old "centrifugal" force causes each sail's free-weight to pull outward from the center of rotation. The torsion spring is now the critical component. We want it to be tough enough to resist the force appled by the free-weight when the windmill rotates slowly.
As the wind picks up, though, the windmill starts to rotate faster. The free-weights pull harder. Now they start to unroll the lines wrapped around the roller/rods. The amount of sail exposed to the wind decreases, and the windmill's rotation becomes speed-limited thereby.
If the wind blows even harder, then the free-weights pull harder, and the sail area automatically diminishes even more. And of course as the wind slows, the free-weights pull with less force, allowing the sails to automatically unfurl.