基于神经网络速度辨识的直接转矩控制系统  

Direct Torque Control System based on Neural Network Speed Recognization

作　　者：刘文胜[1] 王一[1] 王新伟[1] 

机构地区：[1]河南科技大学,河南洛阳471003

出　　处：《电力电子技术》2007年第9期96-97,114,共3页Power Electronics

摘　　要：介绍了一种采用电压空间矢量脉宽调制(Space Vector PWM,简称SVPWM)的新型直接转矩控制(DirectTorque Control,简称DTC)系统。针对DTC系统存在的低速时转矩脉动问题,将电压空间矢量与DTC技术相结合,用PI调节的方法取代了传统DTC系统中采用的滞环控制,提出了一种基于SVPWM的DTC方法。在SVPWM-DTC系统中采用了神经网络速度辨识器,通过神经网络对电机的定、转子磁链和转速进行在线辨识。实验结果表明,这种新型DTC系统有着良好的动、静态性能和全速范围的调速精度。In view of the torque ripples of the Direct Torque Control （DTC） system,a new DTC system based on Space Vector Pulse Width Modulation （SVPWM） is presented.The SVPWM technology was taken into the DTC technology.The PI regulator was used instead of the hysteresis loop controllor,so the ripple of torque was relieved.Then, the neural network speed recoganization unit was put into this SVPWM-DTC system.So the flux of the stator and the rotator could be identified on-line.As a result,the DTC based on SVPWM without speed sensor come true.The experimental result shows that the new DTC system have good static and dynamic state performances,as well as the full-speed regulation accuracy.

关 键 词：转矩控制 矢量控制/神经网络 无速度传感器 

分 类 号：TM921.5[电气工程—电力电子与电力传动]