Modeling And Control Of A Wind Power System Based On Doubly Fed Induction Machine by Aerodynamic Power Coefficient Neural Network Approximation

Abstract

This article presents a method for estimating the Aerodynamic coefficient power by an artificial neural network. This network plays the role of a virtual system which principle is simple; it makes it possible to estimate the coefficient power from the wind speed and of the blade pitch angle in order to facilitate the integration of control algorithms in practical phase for numerical control systems such as FPGAs. Firstly, the modeling of a wind turbine at variable speed with the application of the flux orientation control will be addressed. Subsequently, the structure of a multilayer neural network for estimation of the Aerodynamic coefficient power will be presented. Finally, the results of wind power system simulation using a 3.3 kW doubly fed induction machine will be produced in the Matlab/Simulink environment.