If I am correct about how it works, the operating frequency of an induction generator, in the absense of a grid connection providing it with a hard voltage source at a controlled frequency, is solely controlled by the resonance of the inductance of the windings with the attached capacitors (if you dedicate one winding to excitation) or with the combination of the capacitors and the reactance of your load (if you wire all the windings to the load).

So if you multiply the capacitor value by N you'll divide the operating frequency and RPM by N.

So get three run capacitors (or three per phase if you want to hook up all three phases for max output), of a size intended for a motor of about that power.  Or get one (or one set of three) intended for a motor about three times as large.  Hang them across its output(s), crank it up, and see what it does.  You have some residual magnetization in the rotor already - you can tell because it puts out a LITTLE when you spin it.  When it spins up to the resonance the voltage should suddenly climb, the current in the coils climb drastically until the core saturates, and the speedup with increasing wind input shoud be drastically reduced as the squirrel-cage tries to spin faster than the pumped-up field's rotation and the slippage produces eddy currents.

Voltage will be reduced about in proportion to the frequency reduction, so you'll get something near 36 volts at load with a 20 Hz resonance - although the voltage will climb with RPM as the rotor slips more.  Current doesn't change since it's limited by the core saturation - which is good since it's ALSO limited by coil heating melting the motor's windings.  B-)  Core saturation is bad in a motor (where it produces a drastic current rise) but great in a (stand-alone) generator (where it puts a cap on current).

Once you get it generating, you can adjust the generation-start speed and voltage up or down by adjusting the amount of capacitance.  Higher capacitance, lower startup and lower voltage.  I think you'll want it to start generating at a voltage somewhere between that needed to produce the float voltage and that needed to perform an equalizing charge.

(Once you run the experiment please let us know whether I got this right.)

I recommend you don't run the genny for more than a few tens of seconds without a load, as the heating (both in the cores and the rotor squirrelcage) may exceed the motor's design limits if the wind gets too high.  Putting a charging load on the genny and getting your furling system working will hold things down to a level the motor can handle.

Also:  Unlike a magneto, shorting an induction generator can suck out the excitation and eliminate drag, letting the blades overspeed.  (I'm not sure whether that will happen when you short it, or if it will first drag the rotor speed down, possibly to a near-stop, and THEN let it spin up and run away.  I think it will do one or the other depending on whether you're using one coil just for excitation.  Either way it's not good.)  Induction generator drag is good for a speed governor.  But if you want an electric brake you have to feed it a bit of DC.