Maximum power/ Mechanical torque VS Wind velocity

[arturojortega]

Dear emerrf Thank for your contribution. It looks very interesting the application of reinforcement learning for wind turbines. Please, I have a couple of questions for you:

1. In the last figures for "real control" I can see that even the maximum power was reached, and the rotor speed and wind velocity continue increasing, the maximum power still keeps almost constant? I was expected different maximum power for different wind velocities and rotor speeds.
2. In the file wind_turbine_env.py, class WindTurbine, function init, the arrays: self.rc_wind = np.array([...]) self.rc_gen_torq = np.array([...]) Actual they represents the curve: Optimal_mechanical_torque VS Wind_velocity? Thanks in advance for your replay. Cordially, Arturo

[emerrf]

Hi Arturo, apologies for the very late reply and thank you very much for your questions and your interest in the project.

1. The NREL turbine is designed to operate at 5MW. The rated power is achieved at wind speed of 11.5m/s approx. If the wind speed increases, the controller should pitch the blades so that the exceeding wind is not captured while maintaining the generator torque at maximum. This strategy is called "full load control" and the objective is to keep the power constant and the loads under control. It is also expected that the rotor speed is constant. However, the "real control" figure shows that it takes a bit of time to stabilise the RPMs. If you don't pitch, the generator will burn or the blade will break or the tower may collapse.
2. The "real control" is just the references values obtained from the turbine specification document. The variables rc_wind and rc_gen_torq are just the values of the GenTq in figure 9-1 page 32. They represent the expected generator torque value to produce the maximum power at each wind speed. Pitch and generator torque are the only variables required to be controlled.