Pitch-regulated Mechanism of the Neural Network Control based on Hebbina Supervised Learning Algorithm

Pitch-regulated mechanism servo of wind turbine system is a slow time-variable nonlinear system. According to the requirement of the 1 MW wind turbine with variable speed and constant frequency, the parameters of its PID controller need to be learned or adjusted at real time. Therefore, we design this neural net PID controller based on the Hebbina supervised learning algorithm to realize the self-study and self-regulating of the wind turbine pitch control with hydraulic system. According to the practical data collected from the operation of 1 MW wind turbine, the learning speed (eta_i) of Hebbina supervised learning algorithm can be fixed off-line so that the wind turbine can operate under good conditions. The paper gives the simulation results with comparison and the application of control strategy on the 1 MW wind turbine system.