矢量控制神经网络线性二次型调节器的设计  

Neural Network Linear Quadratic Regulator for Field-oriented Controlled Induction Motor

作　　者：张兴华[1] 

机构地区：[1]南京工业大学,南京210009

出　　处：《电气自动化》2004年第1期19-21,共3页Electrical Automation

基　　金：江苏省教育厅自然基金(03KJB510041)

摘　　要：本文给出了一种矢量控制神经网络线性二次型调节器的设计方法。将感应电机模型近似成五阶线性时变系统,以此来设计线性二次型状态反馈控制器,并由神经网络来表示状态反馈增益与转速的映射关系,从而实现对感应电机的控制。系统无需自适应算法就可获得较强的鲁棒性。仿真结果表明该调节器可获得优良的动、静态调速性能。A neural network linear quadratic regutator(NNLQR) for field-oriented controlled induction motor is proposed. The induction motor is transformed into a 5th-order linear time-variable system model. Based on which a linear quadratic state feedback regulator is designed. Then, several neural networks are used to approximate the relation between the rotor speed and feedback gain. The system possesses good robustness without applying adaptive algorithm. Simulation results show that the regulator can achieve exceHent dynamic and static performance, , ? . '',_-?.

关 键 词：矢量控制 神经网络 线性二次型调节器 设计 感应电机 

分 类 号：TM346[电气工程—电机] TP214.1[自动化与计算机技术—检测技术与自动化装置]