Neural Sliding Mode Control of a DFIG Based Wind Turbine with Measurement Delay

Abstract

In this paper, a robust predictor neural sliding mode control algorithm for a Doubly Fed Induction Generator based Wind Turbine is proposed. To improve the robustness of the proposed controller in presence of the parameter variations and disturbances, a Recurrent High Order Neural Network identifier trained on-line using an Extended Kalman Filter is proposed. In addition, to compensate the measurement delay in stator and rotor current, a robust predictor-based controller is integrated with the control scheme. To show the importance of the proposed control scheme, different experiments are done such as ideal condition, measurements delay, and presence of parameter variations. Simulation results illustrate the effectiveness of proposed control scheme even in presence of reference changing, parameter variations, and measurement delay. In addition, the stability, decoupling, and convergence are achieved.