运动平台磁悬浮直线同步电动机驱动系统RBF-PID自学习控制的研究  

Research on RBF-PID self-learning control of magnetic levitation linear synchronous motor drive system of motion platform

作　　者：彭敬淇 蓝益鹏[1] PENG Jingqi;LAN Yipeng(School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,CHN)

机构地区：[1]沈阳工业大学电气工程学院,辽宁沈阳110870

出　　处：《制造技术与机床》2022年第6期113-118,共6页Manufacturing Technology & Machine Tool

基　　金：国家自然科学基金资助项目(51575363)。

摘　　要：针对运动平台磁悬浮直线同步电动机驱动系统存在的非线性、耦合性以及外部扰动的不确定性问题,提出一种RBF神经网络PID自学习控制器。首先对磁悬浮直线同步电动机的结构及其运行机理进行详细介绍;并建立磁悬浮直线同步电动机的数学模型,推导出电压方程、磁链方程、水平推力方程以及水平运动方程的表达式;设计一种基于RBF神经网络的PID自学习控制器,RBF神经网络中一般采取梯度法对参数w、b、c进行整定,为能有效地缩短网络的学习时间,并减小系统的振荡,论文在RBF神经网络中引入动量因子项,通过MATLAB对控制系统进行仿真,验证RBF-PID自学习算法较PI控制器有更好的稳定性和抗扰动性。Aiming at the non-linearity,coupling and uncertainty of external disturbance in the magnetic levitation linear synchronous motor drive system of the moving platform,a RBF neural network PID self-learning controller is proposed.Firstly,the structure and operating mechanism of the magnetic levitation linear synchronous motor are introduced in detail;the mathematical model of the magnetic levitation linear synchronous motor is established,and the expressions of the voltage equation,flux linkage equation,horizontal thrust equation and horizontal motion equation are derived;a design based on The PID self-learning controller of RBF neural network.In the RBF neural network,the gradient method is generally used to adjust the parameters w、b、c.In order to effectively shorten the learning time of the network and reduce the oscillation of the system,the paper is in the RBF neural network.The momentum factor item is introduced in,and the control system is simulated through MATLAB to verify that the RBF-PID self-learning algorithm has better stability and anti-disturbance than the PI controller.

关 键 词：运动平台 直线电动机 磁悬浮系统 RBF网络 动量因子项 

分 类 号：TM383[电气工程—电机]