A Reduced Order Observer Design for Sensorless DFIG

This study presents a Reduced-order Extended Kalman Filter (EKF) implementation for Doubly Fed Induction Generator (DFIG) that is controlled by Matrix Converter (MC). MC has many advantages such as no need for high value filters, capability of bi-directional energy conversion and no need for DC bus. However, due to the complexity of control algorithms of this topology, various control methods such as Model Predictive Control (MPC) have been developed. Although it is easy to implement, weakness of this control method is that it uses a mathematical model and affected by parameter changes due to the disturbances. In this study, Reduced-order EKF has been applied instead of full-order EKF to reduce the execution time of the algorithm and to ensure that the resistance values of DFIG are more robust to variations due to external disturbances. Moreover, the sensorless control approach has been applied with the Reduced-order EKF to robust speed estimation.