【目的】针对双三相永磁同步电机(DTP-PMSM)模型预测转矩控制(MPTC)的权重系数的设计整定和x-y平面的高次谐波电流的问题，提出一种基于虚拟矢量的模糊无权重系数改进MPTC方法。【方法】首先，用模糊动态的边界条件分别对转矩和磁链跟踪误差进行约束，在约束边界下进行电压矢量的选择，被选择的电压矢量需要分别使转矩和磁链跟踪误差保持在边界条件内。将符合两个边界条件的两个矢量集合按照设定的规则进行交集处理，以选择最优矢量。然后，引入由四矢量合成的虚拟电压矢量集作为备选矢量集合，对系统的谐波电流进行抑制。最后，通过仿真将本文所提改进MPTC与传统MPTC、基于虚拟矢量的MPTC以及基于相对误差率代价函数的MPTC进行对比分析。【结果】仿真结果表明，相较于传统MPTC、基于虚拟矢量的MPTC以及基于相对误差率代价函数的MPTC，本文所提改进MPTC的相电流总谐波失真分别下降了87.53%、26.57%和35.05%，磁链脉动均方根误差分别降低了69.23%、50%和20%，转矩脉动均方根误差分别降低了6.15%、4.95%和3.89%，负载起动时的调节时间分别降低了15.7%、22.9%和44.8%，稳态误差更小，响应更快。【结论】本文所提控制方法不仅实现了对DTP-PMSM转矩和磁链的有效控制，还消除了权重系数带来的不确定性，提升了系统的动态性能，同时有效抑制了x-y平面上的谐波电流，具有良好的可行性。

[Objective] Aiming at the problems of designing and tuning the weighting factors of the model predictive torque control (MPTC) of dual three-phase permanent magnet synchronous motor (DTP-PMSM) and the high harmonic currents in the x-y plane, an improved MPTC method with fuzzy unweighted factor based on virtual vectors was proposed. [Methods] Firstly, the tracking errors of torque and flux linkage were constrained by fuzzy dynamic boundary conditions, respectively, and the selection of voltage vectors was carried out under the constraint boundaries. The selected voltage vectors were required to keep the tracking errors of the torque and flux linkage within the boundary conditions, respectively. The two sets of vectors that meet the two boundary conditions were intersected according to set rules to select the optimal vector. Secondly, a set of virtual voltage vectors synthesized from four vectors was introduced as an alternative set of vectors to suppress the harmonic currents of the system. Finally, the improved MPTC proposed in this paper was compared and analyzed with the traditional MPTC, virtual vector-based MPTC and relative error rate cost function-based MPTC through simulation. [Results] The simulation results showed that compared with the traditional MPTC, virtual vector-based MPTC and relative error rate cost function-based MPTC, the total harmonic distortion of phase current of the improved MPTC proposed in this paper was reduced by 87.53%, 26.57% and 35.05%, respectively. The root mean square error of flux linkage ripple was reduced by 69.23%, 50% and 20%, and the root mean square error of torque ripple was reduced by 6.15%, 4.95% and 3.89%, respectively. The regulation time at load start-up was reduced by 15.7%, 22.9% and 44.8%, respectively, with smaller steady-state error and faster response. [Conclusion] The control method proposed in this paper not only achieves effective control of torque and flux linkage in DTP-PMSMs, but also eliminates the uncertainty introduced by the weighting factor, improves the dynamic performance of the system, and effectively suppresses the harmonic currents in the x-y plane, which has good feasibility.

国家自然科学基金项目(61603263)