Application of neural network controller for maximum power extraction of a grid-connected wind turbine system

Abstract

This paper presents a neural network (NN) pitch controller of a grid-connected wind turbine system for extracting maximum power from wind and proves that its performance using the NN controller would be better than that using a classical PI controller. It discusses the maximum power control algorithm for the wind turbine and presents, in a graphical form, the relationship of wind turbine output, rotor speed, power coefficient, and tip-speed ratio with wind speed when the wind turbine is operated under the maximum power control algorithm. The paper describes the modeling and simulation of the horizontal axis wind turbine system, which includes the drive train model, induction generator model, and grid-interface model for dynamics analysis. The control objective is to extract maximum power from wind and transfer the power to the grid. This is achieved by controlling the pitch angle of the wind turbine blades by the NN pitch controller and firing angles of the inverter switches. The simulation results performed on MATLAB show the variations of the generator torque, the generator rotor speed, the pitch angle, and real/reactive power injected into the grid, etc. Based on the simulation results, the effectiveness of the proposed controllers would be verified.