Model predictive‐based shunt active power filter with a new reference current estimation strategy

Abstract

This study presents a new reference current estimation method using proposed robust extended complex Kalman filter (RECKF) together with model predictive current (MPC) control strategy in the development of a three-phase shunt active power filter (SAPF). A new exponential function embedded into the RECKF algorithm helps in the estimation of in phase fundamental component of voltage ( v h ) at the point of common coupling considering grid perturbations such as distorted voltage, measurement noise and phase angle jump and also for the estimation of fundamental amplitude of the load current ( i h ). The estimation of these two variables ( v h , i h ) is used to generate reference signals for MPC. The proposed RECKF-MPC needs less number of voltage sensors and resolves the difficulty of gain tuning of proportional-integral (PI) controller. The proposed RECKF-MPC approach is implemented using MATLAB/SIMULINK and also Opal-RT was used to obtain the real-time results. The results obtained using the proposed RECKF together with different variants of Kalman filters (Kalman filter (KF), extended KF (EKF) and extended complex KF (ECKF)) and PI controller are analysed both in the steady state as well as transient state conditions. From the above experimentation, it was observed that the proposed RECKF-MPC control strategy outperforms over PI controller and other variants of Kalman filtering approaches in terms of reference tracking error, power factor distortion and percentage total harmonic distortion in the SAPF system.