Nonlinear adaptive control for wind energy conversion systems based on quasi-ARX neural networks model

A wind turbine, by itself, is already a fairly complex system with highly nonlinear dynamics. Wind speed and torque fluctuations can change the dynamic parameters of wind energy conversion systems (WECS), so that the parameter will be a function of time. The quasi-ARX neural networks are nonlinear models, while the multi-layer parceptron (MLP) network is an embedded system to give the unknown parameters of the regression vector. Unknown parameter is the coefficient of nonlinear autoregressive moving average (ARMA) models and consists of two parts, linear and nonlinear parts. With a quasi-ARX model as an identifier, we design an adaptive controller for WECS. Logic switch function is used to ensure the stability and control accuracy. In this paper, the objective of WECS controller is to track the maximum power point tracking (MPPT) is used to maximize the power output of the wind turbine. However, from user's point of view, there are two majors. First, quasi-ARX neural network model is used to identification and prediction of nonlinear system, and second, by using using minimum variance controller with switching law, the proposed model successfully is used to track MPPT of WECS.