Neuro-Fuzzy Adaptive Direct Torque and Flux Control of a Grid Connected DFIG-WECS with Improved Dynamic Performance

This paper presents an adaptive neuro-fuzzy interface system (ANFIS) based direct torque and flux control (DTFC) scheme for grid connected doubly fed induction generator (DFIG) based wind energy conversion system (WECS). The proposed ANFIS based DTFC compares the actual developed torque and stator flux with their respective references and generate required PWM logic signals for the Rotor Side Converter (RSC) that enhance the dynamic performance of the DFIG based WECS. The ANFIS is utilized in this work due to its capability of handling nonlinear system accurately, fast convergence and incorporating the advantages of both the neural network as well as the fuzzy system. A hybrid training algorithm is developed to adapt the membership functions of the ANFIS structure to handle the WECS nonlinearities and wind speed uncertainties. The training data for the ANFIS is obtained from the conventional PI controller based DFIG system running at different operating conditions. The stability analysis of the proposed ANFIS based WECS is performed by approximating the system to a standard second order system which confirms the stability of the proposed WECS. The proposed scheme is simulated using MATLAB-Simulink software. The performance of the proposed ANFIS based adaptive DTFC scheme for DFIG-WECS is found superior to both the traditional fuzzy logic and PI controllers in terms of robust control over electromechanical torque and stator current at various wind speed conditions. The real-time implementation of the proposed control scheme for a laboratory prototype DFIG-WECS is currently underway.