基于柔性神经网络自适应PID的磁轴承径向力控制  

Flexible neural network-based self-regulation PID of magnetic bearing force control

作　　者：王喜莲[1] 葛宝明[1] 伍召莉[1] 

机构地区：[1]北京交通大学电气工程学院,北京100044

出　　处：《北京交通大学学报》2012年第2期85-90,共6页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家自然科学基金资助项目(50907004);高等学校博士学科点专项基金项目资助(20060004027)

摘　　要：针对电机磁轴承径向力控制的严重非线性,提出了利用神经网络自适应整定PID参数,从而直接调节磁轴承径向悬浮绕组电流实现转子径向稳定悬浮的控制方案.在利用BP神经网络结合PID控制实现转子径向稳定悬浮的基础上,为改善径向位移跟踪的动静态性能,提出了基于柔性神经网络的径向力控制,给出了详细的控制算法,并仿真比较了柔性神经网络控制与BP神经网络控制下转子在空载和突加负载时径向悬浮情况,仿真结果表明柔性神经网络控制具有更好的动静态性能,为智能控制的进一步应用研究提供了基础.Considering serious non-linearization of magnetic bearing radial force control,the control scheme based on Neural Network self-regulation PID parameters to directly adjust radial suspending current is presented,which achieves the rotor radial displacement control.According to dynamic and stabilization performance demands,a control scheme of combining flexible neural network and PID is proposed on the basis of realizing rotor radial stable suspending.The algorithms of two kinds of control schemes are presented in details.Simulations are carried out to compare rotor suspending performances under load and no load,and the results show that the rotor displacement can follow the given reference,which verifies the proposed method.And flexible neural network control shows better dynamic and stabilization performance than BP neural network control.Moreover,the paper provides the basis for intelligent optimization method in further application research.

关 键 词：磁轴承 电机 神经网络 PID 

分 类 号：TM352[电气工程—电机] TM133.3