BP神经网络辩识感应电机转子磁链和转速  

Rotor Flux and Speed Estimation of Induction Motor Using BP Neural Network

作　　者：吴建兵[1] 刘国海[1] 

机构地区：[1]江苏大学,镇江212013

出　　处：《电力电子技术》2002年第4期27-30,共4页Power Electronics

基　　金：江苏省教育厅自然科学基金资助项目(0 0KJB4 70 0 0 2 )

摘　　要：根据感应电机数学模型 ,提出了仅基于定子电流的人工神经网络转子磁链与速度的辩识方法 ,实现无速度传感器的交流调速系统的转子磁链和转速闭环控制。用BP算法对神经网络进行学习和训练 ,构建相应的多层前馈神经网络 (MFNN)。仿真和实验结果表明 。Sensorless vector control of the induction motor with closed loop of rotor flux and speed requires the know ledge of instantaneous magnitude and position of the rotor flux as well as rotor speed. This paper deals with the identification of the rotor flux and speed on the base of stator phase current and the delayed one. According to the fundamental equations of induction motor for vector control, the novel identification method of rotor flux and speed using neural network is presented. The structure of multi layer feed forward neural networks is trained with Back Propagation Levenberger Marquardt's method. The simulation and experiment results show the system with neural network identification model has better Performance.

关 键 词：BP神经网络 辩识 感应电机 转子磁链 转速 数学模型 矢量控制 

分 类 号：TM346[电气工程—电机]