ANFIS Wind Speed Estimator-Based Output Feedback Near-Optimal MPPT Control for PMSG Wind Turbine

This paper proposed an estimator-based adaptive near-optimal output feedback control algorithm for maximum power point tracking control of wind energy conversion system. Firstly, the adaptive neuro-fuzzy inference system (ANFIS) estimator is designed to estimate the wind speed instead of direct measurement. The parameters of ANFIS are updated using hybrid least square technique to reduce the computation. Next, the near-optimal controller is utilized for nonlinear system to track the optimal rotor speed. The unknown components in the dynamic of controlled system are approximated by two neural networks. The stability of the proposed controller as well as the convergence of updated parameters is mathematically proven. Moreover, the presented controller requires only the outputs feedback of the system to execute the algorithm, so this is model-free output feedback scheme. Finally, four simulation models with two scenarios are setup to compare the performances of the proposed scheme with other techniques using Q-learning estimator and classic PI controller. The simulation results show that the presented ANFIS estimator-based adaptive near-optimal controller guarantees the best performances with the lowest error of rotor speed and output power in comparison with others for both step wind speed and random wind speed.