Advanced Switching Sequences Based Model-Predictive Control for Single-Phase NPC Converters

Abstract

This article presents a continuous control set model-predictive control (MPC) technique based on advanced switching sequences (SSs) for a grid-connected single-phase three-level neutral-point-clamped converter. The SSs are developed considering the tracking performance of the current feed to the utility grid, the dc-bus capacitor voltage balancing, and the switching loss of the semiconductor devices. In the proposed strategy, with the same sampling frequency, the number of pole state changes (SCs) per control cycle is only two, while in the optimal SS and the modulated MPC (pulse pattern similar to the conventional space vector modulation technique) schemes, the number of pole SCs is four. The reduction in the number switching SCs per control cycle considerably reduces the switching loss of semiconductor devices. Moreover, to avoid the challenges of utilizing the weighting factor, a single-objective cost function is adopted to track the reference current, and the dc-bus capacitor voltages are balanced using the redundancy nature of the voltage vectors. Furthermore, the proposed scheme can ensure a fixed-switching-frequency operation, which is advantageous for selecting filter parameters during real-life implementation. The effectiveness of the proposed strategy is validated by the results obtained from both the simulation in MATLAB/Simulink and the experiment carried out using a downscaled laboratory test rig.