Improved Finite Control Set Model Predictive Current Control for Five-Phase VSIs

Model predictive current control (MPCC) is widely studied and applied in five-phase VSIs, and virtual voltage vectors (V3s) are constructed to improve steady state performance and simplify computation complexity. However, V3s in five-phase VSIs were initially constructed using one large and one medium voltage vectors, and all existing works employed this method, which goes against the common-mode voltage (CMV) reduction. To address this problem, two improved MPCC (IMPCC) methods are proposed in this article. Firstly, the V3s are constructed in a new way using four adjacent large vectors, which enables the proposed IMPCC methods to reduce CMV and low-order harmonic currents inherently. Thus, there are no harmonic and CMV terms in the cost function. Secondly, the duty ratio optimization is introduced and dwell time of the optimal V3 is estimated to reduce current ripples. Then, in order to achieve superior steady state performance, two switching patterns are designed, namely the asymmetrical and symmetrical ones. The asymmetrical pattern aims to reduce the switching frequency while the symmetrical one focuses on harmonics suppression. Finally, experimental comparisons between the proposed IMPCC methods and conventional methods are presented. Experimental results have verified that the proposed schemes can mitigate CMV, suppress current harmonics and reduce computation complexity, simultaneously.