Investigation of Different Controllers and Observers Combinations for Grid-Tied LCL Filter

The problem of controlling the LCL filter remains an active topic since single loop proportional integral (PI) controllers fail to guarantee the stability of the system under changes in the grid impedance. This limitation can be tackled by adding an active damping to the system through various control techniques. In this paper, the first controller is obtained using feedback linearization with disturbance-observer-based controller (DOBC), while the second is designed based on the linear quadratic regulator (LQR) theory. A reduced-order observer, full-order observer, and Kalman filter (KF) are designed for both controllers. The two controllers with the different observers are compared in terms of total harmonic distortion (THD), and their transient performance is tested against changes in the grid impedance. Simulations show that the LQR with KF and DOBC with KF are able to achieve low THD with lower number of sensors, and the latter can maintain its performance if the design is adapted according to the changes in the grid impedance. However, for non-adaptive designs, sensor-based controllers offer a higher robustness against changes in grid impedance.