This paper proposed an improved discrete space vector modulation-based (DSVM-based) model predictive current control (MPCC) for permanent magnet synchronous machines (PMSMs) to reduce current/torque ripple and computation cost. Firstly, the control set is extended from 7 to 37 voltage vectors based on DSVM technical. Although the current/torque ripple is reduced remarkably, the computation cost inevitably increases significantly and thus deteriorating the implementation on microcontrollers. Then, the candidate voltage vectors are decreased to 4 from 37 by the proposed simplified vector preselection, hence significantly reducing computation cost. Finally, the conventional MPCC, the all vector DSVM-based MPCC, and the proposed DSVM-based MPCC are experimentally compared in a 2kW PMSM drive experimental platform.

Model Predictive Current Control,permanent magnet synchronous machine,torque ripple