RBF neural network sliding mode control of a PMSG based wind energy conversion system

Abstract

In this paper, a radial basis function (RBF) neural networks sliding mode controller is designed for a permanent magnet synchronous generator (PMSG) based wind energy conversion system (WECS). The aim is to ensure maximum power capture. Within this control scheme, the WECS nonlinear control affine model is transformed into the canonical form via a diffeomorphism transformation. Afterwards, a RBF neural networks based controller is built around the approximation of an ideal sliding mode controller to ensure reference tracking. In this controller the system parameters are approximated through an RBF neural network, and then these approximations are substituted into their counterparts from the ideal controller. The parameter update laws are derived based on Lyapunov synthesis. A compensation term is appended to the composite controller to ensure robustness against approximation errors. Stability and tracking properties are proved using Lyapunov analysis. A numerical simulation is carried out on a typical 3kW PMSG based wind turbine to access the effectiveness of the proposed controller. The results are then discussed to assess the effectiveness of the proposed neural networks controller.