Weighting Factorless Reduced-Complexity FS-MPC for Modified Packed U-Cell Inverter Topology

Inhalt:
This paper proposes a weighting factorless finite set model predictive control (FS-MPC) for grid-connected cascaded modified packed U-cell (MPUC) inverter. The considered inverter topology has 64 switching states with the ability to generate 49 levels in the output voltage. Hence, considering all switching states in the FS-MPC as in the conventional algorithm results in a heavy calculation burden. Furthermore, for multi-objectives control, tuning weighting factors is an issue to consider, which adds a complexity to the system and affects adversely the robustness. Therefore, the proposed FS-MPC not only eliminates the weighting factors, but also significantly reduces the computational load. Only three states are evaluated per sample instead of 64 as in the conventional method to achieve current control with switching frequency minimization. The proposed concept is validated and compared to the conventional and redundancy elimination (49 states) FS-MPC methods in MATLAB environment. The computational load of the three FS-MPC algorithms are investigated and compared using solver profiler tool in MATLAB/Simulink.