Finite Control Set Model Predictive Control for AC–DC Matrix Converter With Virtual Space Vectors

Abstract

AC–DC matrix converter is a kind of bidirectional converter features buck-type rectifying and boost-type inverting. Considering the coupling of ac current and dc current, and the requirement on safe commutation, the finite control set model predictive control (FCS-MPC), featuring multiple-objective capability and direct generation of gating signals without pulse width modulation (PWM) scheme, is very suitable for this topology. Also, fast transient can be obtained with the FCS-MPC. However, in conventional FCS-MPC, only one space vector in each sampling period is applied. The performance will be degraded due to the low switching frequency under limited sampling frequency in practice. In this article, an FCS-MPC with virtual space vectors is proposed for a bidirectional ac–dc matrix converter. This article uses several virtual space vectors, each formed by two real space vectors, to improve the control performance without increasing sampling frequency. To further reduce the computation burden, an efficient method to preselect space vectors is proposed. Furthermore, a thorough analysis of the effect of parameter mismatch in the proposed FCS-MPC is conducted to ensure robust performance. Both simulation and experimental results are presented to verify the effectiveness of the proposed control strategy.