Reduced-Sensor-Based Optimal Switching Sequence Model Predictive Current Control for Grid-Tied Inverter With LCL Filter

Abstract

This article presents an improved reduced-sensor-based optimal switching sequence model predictive current control (OSS-MPCC) algorithm for a grid-tied inverter with the LCL filter with only injected grid current measurements. The OSS-MPCC algorithm utilizes four types of signals, namely, the estimated grid voltage symmetrical components obtained from the virtual flux observer, the state estimates obtained from the full-state Luenberger observer, the reference of positive sequence injected grid current obtained from the SOGI-QSG-based symmetrical voltage estimations, and switching states of the inverter to perform predictive current tracking in grid-tied operation. The proposed algorithm offers several advantages simultaneously. It successfully limits the switching frequency spectrum of inverter voltage, reduces the number of sensors from nine to three, keeps the grid current total harmonic distortion (THD), even in the advent of voltage unbalancing, below 2.5%, demonstrates robustness against the filter parameter mismatches, and still keeps the computational time below $15~\mu s$ comparable with previous OSS-MPC algorithms. The results were numerically verified through simulations in MATLAB and validated using a 1.5-kW experimental setup with a two-level inverter, programmable ac supply as a grid, and Microlab box controller with EMC block set for pulsewidth modulator (PWM) generation.