Measuring of the torque for a generator or a wind turbine is rather difficult but there are several options. Measuring of the torque for a generator is rather easy if the generator has a shaft which is strong enough to take the generators weight if the generator is only connected to the shaft. The generator is provided by a lever and a thin rope at the end which is connected to a balance to measure the force in the rope. This option is described in my report KD 595 which can be copied for free from my website:
www.kdwindturbines.nl. The force F hast to be measured in N and the length of the lever in m to get the torque Q in Nm. The mechanical power P (in W) is then given by: P = Q * pi * n / 30. (n is the rotational speed in rpm so you also need a device to measure n).
In you want to measure the rotor of a wind turbine, you basically can use a PM-generator as load and measure the torque in the same way. If the wind turbine has a vertical shaft, you need an auxiliary wheel to make that the rope is going from horizontal to vertical if you use a balance which can only be used if it is positioned normally. Measuring of only the electrical output of the generator and calculating the mechanical power and so torque is only possible if you know the generator efficiency for every load condition.
Measuring of the whole Q-n curve of the rotor for a certain wind speed V might give problems which depend on the shape of the Q-n curve. Generally the Q-n curve has a maximum for a certain tip speed ratio. If the Q-n curve of the load is a horizontal line, like we have with a brake of Prony, you only find stable working points for the part of the Q-n curve of the rotor which lies at the right side of the maximum. For the left side of the Q-n curve, the rotor curve will increase faster than the load curve at increasing rotational speeds and this means that the load is not able to maintain the rotational speed which belongs to the point of intersection of both curves.
If you use no brake of Prony but a generator as load, the Q-n curve of the generator will depend on the electrical load. It may be rising at increasing rotational speeds but if the Q-n curve of the rotor is increasing faster, you will still find no stable working point and this means that a part of the Q-n curve of the rotor can't be measured. I have measured several scale models of wind turbines in the wind tunnel of the University of Delft and a part of the Q-n curve of the rotor could not be measured for fast running rotors because of this problem.