二自由度交流主动磁轴承最小二乘支持向量机解耦控制

Decoupling control of 2 degrees-of-freedom alternating current active magnetic bearing using least squares support vector machines

采用最小二乘支持向量机(LS–SVM)理论,研究了二自由度交流主动磁轴承这一多变量、非线性、强耦合的控制对象的动态解耦问题.根据主动磁轴承的基本结构,利用等效磁路法推导了悬浮力模型,建立了系统的状态方程,并对其进行可逆性分析;应用LS–SVM辨识原理推导出系统的逆模型;将逆模型与原系统串联,从而将原非线性耦合系统解耦成伪线性系统,并设计了附加控制器;采用MATLAB软件平台构建了磁轴承仿真系统,对系统进行了拟合、起浮、干扰及解耦仿真试验并进行了分析;最后构建了交流主动磁轴承实验平台,对转子起浮和解耦性能进行了实验.研究表明:采用LS–SVM逆解耦控制策略,能够实现交流主动磁轴承的动态解耦,系统具有良好的动静态性能.

Using the least squares support vector machines （LS-SVM）, we investigate the dynamic decoupling of the 2 degrees-of-freedom （DOF） alternating current （AC） active magnetic bearing （AMB） system which is a multivariable, nonlinear and strong coupling object. According to the basic structure of the AMB system, we use the equivalent magnetic circuit method to develop the mathematical model for the suspension forces and setup the state-space equation for the 2 DOF AC AMB system. After the invertability analysis, we derive the inverse model of the system based on the identification principle of LS-SVM. By connecting the inverse system in series with the original system, we decouple the original nonlinear system into quasi-linear systems, for which additional controllers are designed. The simulation of the AMB system is carried out on the MATLAB software platform. On this platform, we perform the tests of set up, fitting, floating, disturbing and decoupling; the results are analyzed. Finally, the experimental platform of the AMB was established, on which the suspension of the rotor and the decoupling performance test are performed. The research results show that the control strategy based on LS-SVM can realize dynamic decoupling control for the 2 DOF AC AMB system with desirable dynamic and static performances.