I have considered the basic just measuring voltage over a known resistor and monitoring with a scope. Seeing how much one turn on the wheel gets, i can calculate power output as a function revs/min i guess?
maybe it is better to see, how much a known capacitor will be charged over time?
It probably would be easiest to let it charge a large capacitor (large enough so there is no significant ripple under load), add a known resistive load, and measure the voltage. Then power = V²/Rload. You could try different loads to see which produces the most power.