基于神经网络逆系统的单自由度混合磁轴承控制器设计  

Design on Controller for Single Degree of Freedom Hybrid Magnetic Bearing Based on Neural Network α th-order Inverse System

作　　者：杨泽斌[1] 孙玉坤[1] 朱熀秋[1] 刘贤兴[1] 

机构地区：[1]江苏大学电气信息工程学院,江苏镇江212013

出　　处：《轴承》2008年第1期9-12,共4页Bearing

基　　金：国家自然科学基金项目(60174052);江苏省自然科学基金项目(BK2003049)

摘　　要：在介绍永磁偏置单自由度混合磁轴承的结构和运行机理基础上,导出了磁轴承吸力方程。建立磁轴承控制系统的数学模型并验证了其可逆性,构造出其神经网络α阶逆系统并与原系统复合成α阶伪线性复合系统,采用PID进行闭环综合。数据仿真和试验结果表明,该控制策略能有效改善永磁偏置单自由度磁悬浮轴承的控制效果,所设计的控制系统具有良好的动、静态性能。Based on the configuration and operation principle of permanent magnet biased single degree of freedom hybrid magnetic bearing, the equation of attraction is deduced. The control model of system is built. It is verified that this model is invertible, and Neural Network （NN） α th - order inverse system is constructed and is combined with the o- riginal nonlinear system to form a pseudo - linear composite system. The PID controller is designed for the pseudo - linear composite system. Both numerical simulation and experiment results show that this kind of control strategy can improve the control effect of the bearing and the whole control system has fine dynamic and static performance.

关 键 词：电磁轴承 永磁偏置 神经网络逆系统 

分 类 号：TH133.31[机械工程—机械制造及自动化]