基于动态递归神经网络的超磁致伸缩驱动器精密位移控制.pdfVIP

26 卷 3 期 中 国 电 机 工 程 学 报 Vol.26 No.3 Feb. 2006 2006 年2 月 Proceedings of the CSEE 2006 Chin.Soc.for Elec.Eng. 文章编号：0258-8013 （2006 ）03-0106-06 中图分类号：TM 153 文献标识码：A 学科分类号：47040 基于动态递归神经网络的超磁致伸缩 驱动器精密位移控制 曹淑瑛，郑加驹，王博文，黄文美，颜威利 （河北工业大学电磁场与电器可靠性省部共建重点实验室，天津市红桥区 300130 ） Precision Position Control for Giant Magnetostrictive Actuator Based on Dynamic Recurrent Neural Network CAO Shu-ying, ZHENG Jia-ju, WANG Bo-wen, HUANG Wen-mei, YAN Wei-li （Province-Ministry Joint Key Laboratory of Electromagnetic Field and Electrical Apparatus Reliability, Hebei University of Technology, Hongqiao District, Tianjin 300130, China ） ABSTRACT: Due to the inherent hysteretic nonlinearity, the 是根据GMA 的滞回特性构造的，通过反馈误差学习方案在 giant magnetostrictive actuator (GMA) can cause position error 线学习GMA 的逆滞回模型。仿真结果表明该控制策略能适 in the open-loop systems, and cause instability in the 应GMA 滞回特性随机械负载、输入信号的变化，在线建立 closed-loop systems. To remedy this problem, a real-time GMA 的滞回逆模型，从而消除滞回非线性的影响，实现 hysteretic compensation control strategy combining a dynamic GMA 的精密控制。 recurrent neural network (DRNN) feedforward controller and a 关键词：超磁致伸缩驱动器；滞回非线性；反馈误差学习； proportional derivative (PD) feedback controller was proposed 动态递归神经网络；实时补偿控制 to implement the precision position tracking control of the 1 引言 GMA. The DRNN controller was constructed based on the hysteretic characteristics of the GMA, and on-line learned the 超磁致伸缩驱动器 GMA （Giant Magneto- inverse hysteresis model of the GMA by the feedback-er