Real‐time neural sliding mode field oriented control for a DFIG‐based wind turbine under balanced and unbalanced grid conditions

Abstract

This study proposes a real-time sliding mode field oriented control for a doubly-fed induction generator (DFIG)-based wind turbine prototype connected to the grid. The proposed controller is used to track the desired direct current (DC) voltage reference at the output of the DC link, to maintain constant the grid power factor at the step-up transformer terminals controlled by the grid side converter, and to force independently the stator active and reactive power to track desired values through the rotor currents controlled by the rotor side converter. This control scheme is based on a recurrent high-order neural network (RHONN) identifier trained on-line by an extended Kalman filter. The RHONN is used to approximate the DC link and the DFIG mathematical models. The adequate approximation helps to calculate the exact equivalent control part of the sliding mode controller and to eliminate the effects of disturbances and unknown dynamics appearing in the grid, which improves the robustness of the control scheme. This controller is experimentally validated on a 1/4 HP DFIG prototype and tested for variable wind speed to track a time-varying power reference and to extract the maximum power from the wind, under both balanced and unbalanced grid conditions.