A Duty Cycle-Based Model Predictive Control of Matrix Converter-Fed PMSM Using Rotating Vectors

Abstract

To address technical challenges such as insufficient steady-state performance and common-mode voltage interference in the Matrix Converters (MC), this paper proposes a duty cycle-based Model Predictive Control (MPC) strategy that utilizes the rotating vectors of MC, which are characterized by generating zero Common-Mode Voltage (CMV). This strategy predicts the motor's operating state in real time, selects the optimal reference vector, and synthesizes a virtual vector in the same direction as the reference vector by combining two adjacent rotating vectors, whose duty cycles are then calculated based on the position of the virtual vector. Experimental results demonstrate that the proposed control strategy fully leverages the fast response advantage of MPC while effectively suppressing CMV amplitude and reducing current harmonic distortion, significantly improving the system's steady-state performance.