An Improved Model Predictive Control Method Using Optimized Voltage Vectors for Vienna Rectifier With Fixed Switching Frequency

Abstract

The Vienna rectifier is an attractive topology for industrial applications. In order to obtain fixed switching frequency (FSF) while maintaining lower current distortion, an improved model predictive control (MPC) method using optimized voltage vectors and switching sequences is proposed in this article. First, the space vector diagram of the Vienna rectifier is divided into six sectors according to the input currents. Then using finite-set MPC (FS-MPC), the sector can be further divided into subsectors, which reduces the number of the candidate vectors from 19 to 7 simultaneously. Second, to achieve decoupling of the neutral-point (NP) voltage balancing and current distortion, the small regions are recombined into novel N- or P-type regions to compensate for the absent vectors. More importantly, the FSF is realized by the multiple vectors determined together by the fluctuation of the NP voltage and the cost function without weighting factors, which improves the current quality and it is beneficial for filter design. Finally, based on FSF-MPC, sequence optimization between adjacent sampling periods is further considered. Compared with FSF-MPC, the improved FSF optimized (FSFO) MPC can significantly reduce the switching loss. The effectiveness of the proposed FSF- and FSFO-MPC methods is verified by simulation and experiment.