基于分层优化的异步电机直接转矩预测控制

Sensorless Predictive Torque Control Strategy of Induction Motor Based on Stratified Optimization Method

传统的异步电机直接转矩控制采用基于开关表的bang-bang控制器,每个控制周期施加单一电压矢量,并且存在控制时延,导致电机的转矩和磁链幅值波动较大,电流波形畸变严重。引入模型预测和电压矢量优化等方法可以有效改善DTC的控制性能,但一般的优化方法难以兼顾转矩和磁链幅值双重控制目标。提出一种基于分层优化的直接转矩预测控制策略。以转矩为优化控制目标,确定若干个可行的电压矢量及其作用时间;再按磁链幅值误差最小原则筛选出最优的电压矢量。该方法的实现算法简单,不但使转矩控制更为准确,而且磁链幅值的波动也大幅减小,电流波形明显改善。仿真和试验结果均验证了新提出控制策略的有效性。

The conventional DTC, using bang-bang controllers based on a switch table, has the disadvantage of high torque and current ripples due to a single voltage vector applied in a control period and control delay. The use of model predictive control and optimization of vohage vectors could improve the control performance significantly. However, the normal method was hard to solve the multi-objective optimization problem because both torque and flux should be considered. A modified predictive torque control strategy based on stratified optimization method was proposed to overcome these shortcomings. The accurate torque control was given priority when determining the workable voltage vectors and their acting time. Then, the flux control was carried out by choosing the optimal one from these workable vectors to minimize the flux error. The proposed strategy was convenient to realize, and could make the fluctuation of torque and flux amplitude obviously reduced, together with the harmonics in stator currents. Simulation and experimental resuhs validated its effectiveness.