无轴承电动机神经网络逆系统解耦控制关键技术发展  

Development of Key Technologies for Decoupling Control of Neural Network Inverse System for Bearingless Motor

作　　者：刘奕辰[1] 朱熀秋[1] 杨泽斌[1] LIU Yichen;ZHU Huangqiu;YANG Zebin(School of Electrical and Information Engineering,Jiangsu University,Zhenjiang 212013,China)

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》2024年第1期28-40,共13页Bearing

基　　金：国家自然科学基金资助项目(62273168)。

摘　　要：无轴承电动机利用一套新的悬浮力绕组产生悬浮力,使无轴承电动机成为一种多变量、强耦合的非线性系统,如何实现转速与径向位移以及径向位移之间的动态解耦控制是无轴承电动机稳定运行的关键要素之一。神经网络逆系统具有以任意精度逼近非线性函数的能力,可以有效实现无轴承电动机的解耦控制。阐述了无轴承电动机悬浮力产生的原理以及神经网络逆系统解耦原理,从纯神经网络逆系统、神经网络逆系统结合智能控制器等方面分析了国内外无轴承电动机神经网络逆系统解耦控制策略的研究现状,归纳了无轴承电动机在使用神经网络逆系统进行解耦控制时的共同和不足之处,总结了无轴承电动机神经网络逆系统解耦控制的一般设计方法,并对神经网络优化、在线训练、抗干扰能力等关键技术发展趋势进行了展望。The bearingless motor uses a new set of suspension force windings to generate suspension force,which makes the motor become a multi-variable and strongly coupled nonlinear system.How to realize dynamic decoupling control of speed with radial displacement and x and y axis radial displacement is one of the key factors for stable operation of the motor.The neural network inverse system(NNIS)has the ability of approximating the nonlinear function with arbitrary precision,and the decoupling control of the motor can be realized effectively.The principles of suspension force generation of the motor and the decoupling principle of NNIS are described.The research statuses of decoupling control strategy for NNIS of the motor at home and abroad are analyzed from the aspects of pure NNIS,NNIS with intelligent controller.The commonalities and deficiencies of using NNIS for decoupling control of the motors are summarized,the general design methods for decoupling control of NNIS of the motor are concluded,and the development trend of key technologies is prospected,such as neural network optimization,online training and antiinterference ability.

关 键 词：滑动轴承 磁力轴承 神经网络 逆系统 解耦控制 智能控制 

分 类 号：TH133.31[机械工程—机械制造及自动化]