Neural Network-Observer-Based MPPT Control for Variable Speed Wind Turbine

The present paper introduces a novel method to optimizing energy production in variable speed wind turbines (VSWT) by integrating a neural network observer. The proposed method aims to improve dynamic performance and achieve efficient maximum power point tracking (MPPT) under variable wind conditions in the case of the states of VSWT are not directly measurable. In the first, a two-mass model is developed to accurately represent the dynamic interactions between the turbine rotor and generator. Then, a neural network (NN) observer is implemented to estimate the states of WT system. Using these state estimates, we then formulate a control law designed to adjust operational parameters, such as rotor speed and pitch angle, in order to operate the turbine at its optimal power point. Finally, the results of simulation are supplied to evaluate the efficiency of the proposed approach.