北京航空航天大学学报 ›› 2008, Vol. 34 ›› Issue (8): 869-872.

• 论文 • 上一篇    下一篇

冗余直接驱动阀系统的余度控制策略

蔚永强,郭宏,谢占明   

  1. 北京航空航天大学 自动化科学与电气工程学院, 北京 100191
  • 收稿日期:2007-07-30 出版日期:2008-08-31 发布日期:2010-09-17
  • 作者简介:蔚永强(1980-),男,陕西西安人,博士生,yongqiangyu@asee.buaa.edu.cn.
  • 基金资助:

    教育部新世纪优秀人才支持计划资助项目(NCET-04-0163)

Redundancy control of direct-drive-valve servo system

Yu Yongqiang, Guo Hong, Xie Zhanming   

  1. School of Automation Science and Electrical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2007-07-30 Online:2008-08-31 Published:2010-09-17

摘要: 针对冗余直接驱动阀伺服系统中由于余度降级所造成的性能降低问题,提出一种神经网络自适应滑模余度控制策略.利用径向基函数神经网络RBFNN(Radial Basis Function Neural Network) 的在线学习功能,对系统发生的变化进行快速自适应补偿,使系统状态趋近于滑模面,提高跟踪精度和鲁棒性;并通过与比例微分PD(Proportional-Derivative)算法的并行控制,促进RBFNN的收敛,增强系统的稳定性.通过与PID(Proportional-Integral-Derivative)切换控制策略的对比研究,表明RBFNN自适应滑模余度控制方法不但设计简单,而且能够有效克服余度降级带来的系统性能下降的问题,极大地改善了系统的品质.

Abstract: A novel neural network adaptive sliding mode control strategy was proposed, which was applied to ensure tracking capability to direct-drive-valve (DDV) servo system in the presence of degrading redundancy. A radial basis function neural network (RBFNN) was adopted to realize sliding mode control. By means of compensating varieties of the system with adaptive learning algorithm, the control based on RBFNN decreased the tracking error and enhanced the robustness. Meanwhile, a proportional-derivative (PD) controller was designed as the other parallel control part, which improved the convergence of RBFNN, and enhanced the stability of system. Simulation results show that the proposed control scheme solves the problems brought by the degrading of redundancy effectively. It possesses better tracking performance than switching proportional- integral-derivative (PID) control, and can be designed easily.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发