感应电机的新型神经网络广义逆系统解耦控制  

Decoupling control of induction motor based on a new type of ANN generalized inverse

作　　者：刘陆洲[1] 肖建[1] 王嵩[1] 

机构地区：[1]西南交通大学电气工程学院,四川成都610031

出　　处：《电机与控制学报》2009年第A01期72-77,共6页Electric Machines and Control

基　　金：国家自然科学基金资助项目(60674057);高等学校博士学科点专项科研基金资助项目(20060613003)

摘　　要：针对感应电机这一多变量、强耦合系统,提出了一种新型神经网络广义逆系统解耦控制方法,给出了新型广义逆系统的存在性证明。它与感应电机复合,将转子磁链与转速的解耦成两个独立的伪线性子系统,并且可以随时通过调整反馈参数改变两个伪线性子系统的配置极点,为控制器的设计提供了便利。仿真结果表明,通过合理的选择配置极点和控制器参数,整个控制系统对电机参数变化和负载扰动有较强的鲁棒性和动态性能。A decoupling control method based on a new type of artificial neural network （ANN） generalized inverse is proposed for induction motor which is a muhivariable and high coupling system. The existence of the new type of generalized inverse was proved. The composition of the generalized inverse and the inductor motor decoupled the rotor flux and rotor speed into two independent pseudo-linear subsystems, the poles of which can be assigned anytime by regulating the feedback parameters which were beneficial to the design of controller. Simulation results show that the whole control system has strong robustness to variation of motor parameters and load torque disturbance by assigning poles and choosing parameter of the controller appropriately.

关 键 词：解耦控制 广义逆系统 神经网络 鲁棒性 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]