Before deciding on a regulator I would determine the voltage and current at which the unit will develop maximum power. The hydrodynamics of the pipe and nozzle and the aero drag of the spinning fan may effect this. I would then build an efficient step up or step down switching regulator to transfer this power at a standard voltage (eg 14.2). As an example, say you get maximum power at 8V, above that the combination of nozzle velocity and air loss reduce the output. You may find that you can get 15V but have little capacity to deliver any current at that speed. In that case you build a step up switching regulator ...
http://www.linear.com/pc/productDetail.do?navId=H0,C1,C1003,C1042,C1031,C1115,P8850
and convert the 8V to 14.2 with minimum loss. I think that the fan, with all those blades hanging in air, will provide sufficent braking to prevent serious overspeed if load drops off. If it does not I would provide a latching solenoid valve (they are available, latching avoids a continuous use of power) to turn the water off on the increasing motor voltage that would signify excess RPM. Assuming that you are going to have a 12V battery and 120VAC inverter setup - to make use of the energy, you could also, at a point below where the water shuts off, turn on a 60W light bulb inside the house in winter, thereby helping with heat. In the summer DC current could be supplied to a thermo electric cooler and heat removed from the house. These loads could be applied until battery voltage dropped to a setpoint and then removed to allow recharge. Ultimately if you can develop a steady 50 watts I would consider building a small synchronous inverter to deliver it to the AC line ... slightly decreasing your electric bill.
I think I would cut the blades on the fan down to about 1/2 inch, that would make a better pelton wheel and get rid of a lot of unnecessary air drag and possible interference with water splashing off the blades.