Improved model predictive current control for single-phase NPC shunt active power filter

A model predictive current control (MPCC) of shunt active power filter (SAPF) requires tracking of distorted reference current based on the discrete mathematical model of the system. But the approximations considered in the development of mathematical model for the nonlinear system and also the uncertainty in the parameter values will deteriorate the performance of control algorithm. To enhance the current tracking performance, an improved MPCC for single-phase NPC converter based SAPF is presented in this paper. In the proposed control strategy, a mean squared predictive current error is used as a feedback in current tracking objective of the cost function. The optimal switching vector that minimizes the defined cost function is applied in the next sampling period. The robustness of the proposed control strategy for variation in grid feeder impedance and interfacing filter inductance is verified using matlab/simulink results. The results presented are compared with MPCC for absolute current tracking error and the percentage of total harmonic distortion (%THD).