作　　者： (张晓光）; (张亮）; (侯本帅）;

机构地区： 北京市变频技术工程技术研究中心（北方工业大学）,北京市石景山区100144

出　　处： 《中国电机工程学报》 2017年第16期4800-4809,共10页

摘　　要： 为了减少传统模型预测转矩控制(model predictive torque control,MPTC)的计算量并消除转矩与磁链之间的权重系数,优化MPTC方法,提出一种改进模型预测转矩控制方法,利用转矩与磁链无差拍预测控制原理在线计算下一周期预施加的参考电压矢量,根据参考电压所处扇区进一步确定待选电压矢量,无需对所有电压矢量作用下的系统行为进行预测,将传统方法的8次预测减小到2次,显著减小计算量;在其基础上,构建电压误差代价函数取代传统转矩与磁链误差代价函数,选取与参考电压误差最小的矢量作为最优电压矢量,消除了MPTC的权重系数,进一步简化系统结构。仿真与实验结果表明,提出的优化MPTC方法能够显著降低传统算法计算量,并且具有良好的控制性能。 An improved model predictive torque control （MPTC） was proposed in this paper, in order to reduce the computation complexity and eliminate the weighting factor in traditional model predictive torque control. The deadbeat control principle was used to predict reference voltage vector which should be applied at the next switching period. Then, the sector of reference voltage can be obtained to determine the range of optimal voltage vector, which reduce the number of prediction from seven to two, compared with traditional MPTC. And then, the cost function of torque and flux in traditional MPTC was displaced by novel cost function of voltage error. The voltage vector that can minimize the cost function was selected as the optimal vector. The improved MPTC eliminate weighting factor and simplify system structure effectively. Simulation and experimental results show that the improved control method can evidently reduce computation time and has satisfied control performance.

关 键 词： 永磁同步电机 模型预测转矩控制 权重系数 基于电压的价值函数