Extended Kalman Filter for Parameters Estimation of Permanent Magnet Synchronous Generators in Wind Energy Systems

Abstract

In this paper, a model predictive control is applied to a two-level voltage source converter that drives a permanent magnet synchronous generator in a wind energy system. The MPC uses the system model and the available converter switching states to predict the current of the system for each switching state. Then, a predefined cost function containing the difference between the reference and the predicted currents is evaluated to select the best switching state to be applied to the converter. The stator resistance and inductance of the permanent magnet synchronous generator change during operation due to change of temperature and frequency with a consequent effect on the system model resulting in prediction errors. The extended Kalman filter is used for online estimating the resistance and inductance of the permanent magnet synchronous generator to compensate for prediction errors of the model predictive control. To obtain accurate estimation, the extended kalman filter is tuned using the whale optimization algorithm. The proposed system and control algorithm are implemented and validated using MATLAB/SIMULNK.