摘要
针对异步电机矢量控制系统中模型预测控制(Model Predictive Control,MPC)在线计算量大且复杂等问题,提出了一种改进型MPC方法,该方法首先以基础电压矢量为角平分线将矢量平面分成6个大区,再利用无差拍控制理论和引入纵向压缩变换概念得到参考电压预测值及所在大区,并通过旋转因子转换到横轴所在的大区,最后通过逻辑判断得到最优电压矢量。该方法不需要8次滚动预测,其价值函数最优控制由原来的8次计算和7次判断简化为1次计算和4次判断,从而减小了计算量。仿真结果表明:采用改进性MPC和传统MPC的电机虽然都具有一致的调速性能,但前者电磁转矩的波动比后者小12.5 N·m,同时前者的定子电流两相静止坐标绝对值比后者小10A,因此跟踪效果更好。
Aimed at solving the problem that online calculation of model predictive control( MPC) for induction motor is large and complex,an improved MPC based on the asynchronous motor vector control system as the research object was put forward. First of all,the vector plane was divided into six regions based on basic voltage vector as angle bisector,at the same time,using the control theory of non- error and introducing the concept of longitudinal compression transformation get the predicted value of reference voltage and its region,and through the conversion of rotation factor to the horizontal axis region,then using the logical judgment to get the optimal voltage vector. This method does not need 8 times rolling forecasts,and the value function of optimal control reduced to 1 time calculation and 4 times judgment from the original 8 times calculation and 7 times judgment,thus reduced the amount of calculation. Simulation results show that motors using the improved MPC and traditional MPC have consistent performance of speed adjustment,but the former's electromagnetic torque ripple is 12. 5 N·m smaller than the latter,and the two stationary coordinates stator currents of the former is 10 A smaller than the latter,therefore the improved MPC has better tracking effect.
出处
《微特电机》
北大核心
2016年第7期66-71,共6页
Small & Special Electrical Machines
基金
西华大学研究生创新基金研究项目(ycjj2015209)
西华大学学生创新创业项目(苗子工程)(2015RZ0030)
攀枝花市太阳能光伏离/并网智能化控制逆变一体集成应用(2014CY-S-1-2)
攀枝花学院分布式光伏多逆变器并网控制研究(2014YB11)
攀枝花市科技计划项目(2015CY-C-5)
四川省高校重点实验室项目(2013TYNZ-02/TYN2015-09)
四川省电力电子节能技术与装备重点实验室项目(szjj2015-066)
关键词
异步电机矢量控制
改进型MPC
纵向压缩变换
参考电压预测
induction motor vector control
improved model predictive control(MPC)
longitudinal compression transformation
prediction of reference voltage