A Hybrid Multivector Model Predictive Control for an Inner-Interleaved Hybrid Multilevel Converter

Abstract

A hybrid multivector model predictive control strategy for an inner-interleaved hybrid multilevel converter is brought forward in this article, which can enable the separation of the low- and high-frequency stages. It initiates with the use of sign patterns of the reference vector in the stationary reference frame to determine switching states of the low-frequency stage. Then, the reference vector is converted to the inner virtual space vector diagram, which is further transformed into a 120° oblique coordinate system, where the adjacent vectors are selected. Further, the current tracking is realized through the duty cycle optimization of the chosen vectors. Finally, a symmetric switching sequence (SS), which serves as the best one for dc capacitors voltages balancing and circulating current suppression, is selected among all the SSs that belong to the three chosen vectors with optimal duty cycles. The proposed method reduces both current ripples and computational burden while achieving a constant equivalent switching frequency. Comprehensive experimental results performed on an all silicon-carbide prototype verify the effectiveness of the proposed control.