Early on I started with something like this, just looking at the theoretical power available at the prop. The upper end of this chart corresponds to about a 27 mph wind speed, and the straight line “dump watts†are using our resistors, though we don't actually go near the upper limit because we furl long before these speeds.
By the time I actually built the thing we had the benefit of some output data. The goal is the same though, which is to simply match the power from the turbine, if otherwise unloaded, over the bandwidth where you might be dumping (and not burn anything up). If for example, if we want to dump about 3 kW at 141V (DC into the controller)… the math would be:
line volts = rms volts/ph x 1.732, and
DC volts = rms volts/ph x 1.732 x 1.4, therefore:
rms volts/ph = DC Volts / 1.73 / 1.4 = RMS Volts / ph
141 1.732 1.414 57.56
line volts
99.70
We have pairs of 6.8 ohm, 700 watt resistors, so 58Vph/3.4 ohms = 17 amps, then 100Vline * 17A * 1.732 = 2,941 watts, 2.94 kW. And, each resistor should be carrying just under 500 watts. This is just one point, but you can chart a line from this method.
Notice that in delta the same resistors would dump 8.8 kW, not good here, but other combinations are certainly possible.
If you are working with water heater elements you would need to see if the resistance is workable for your needs. A 120V 2,000 watt would be 16.8A, so 7.2 ohms. In pairs you could double that (in series), or cut it in half (in parallel). If my math is right in delta at 7.2 ohms you'd dump 4.2 kW (120% over the element rating); in wye you dump 1.4 kW.