模型预测控制（MPC）动态性能优越、抗扰动能力强，但零电压矢量的使用可引起较大的共模电压，产生轴电流、电磁干扰等问题。基于等效零矢量的多矢量合成MPC虽可抑制共模电压，但需在同一周期使用多个有效矢量，导致开关频率高、系统效率低。因此，提出一种用于永磁同步电机（PMSM）的混合模型预测转矩控制（MPTC）策略，针对不同幅值的参考电压，分别使用虚拟矢量和多矢量合成2种电压矢量生成方式，可用较低的平均开关频率实现共模电压抑制，并保证良好的动、静态转矩控制性能。此外，通过改变控制模式的切换条件，可实现开关频率和转矩脉动的灵活切换。最后，仿真验证了所提方法的有效性。

Model predictive control (MPC) has excellent dynamic performance and strong disturbance rejection capabilitybut zero voltage vector can cause large common mode voltageresulting in axis currentelectromagnetic interference and other problems. Although the existed multi-vector MPC based on the equivalent of zero vector can suppress common-mode voltageit needs to use multiple effective vectors in a periodresulting in high switching frequency and low system efficiency. A hybrid model predictive torque control (MPTC) strategy for permanent magnet synchronous motor (PMSM) is proposedwhich uses two waysthe virtual vector or multi-vector synthesis wayto generate the voltage vector for different amplitudes of the reference voltage. This method can achieve common mode voltage suppression with low average switching frequency and ensure good performance of dynamic and static torque control. In additionflexible compromise between switching frequency and torque ripple can be achieved by changing the switching condition of control mode. Finallythe effectiveness of the proposed method is verified by simulation.

TM351

中央高校基本科研业务费项目(NS2021020)；江苏省"双创博士"项目(JSSCBS20210178)