首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于神经网络的不稳定时滞对象控制
引用本文:张晓迪,蔡云泽,何星,张卫东.基于神经网络的不稳定时滞对象控制[J].上海交通大学学报,2014,48(7):1033-1038.
作者姓名:张晓迪  蔡云泽  何星  张卫东
作者单位:(上海交通大学 自动化系,上海 200240)
基金项目:国家自然科学基金资助项目(11072144,61025016,61034008,61221003)
摘    要:针对过程控制工业中的一类不稳定时滞对象存在的难以稳定、鲁棒性差、对输入的变化和扰动十分敏感的问题,采用了基于神经网络的双自由度控制结构.在利用内环控制器镇定对象并且提高系统反应速度、减轻输出的振荡的同时,结合Guillermo等提出的比例-积分-微分(PID)参数镇定区域理论优化设计外环BP神经网络控制器的学习范围、学习方式和初始参数,改善系统设定值跟踪和扰动抑制的性能,提高系统鲁棒性.仿真结果表明,即使在建模有误差的情况下,该控制结构仍能比传统双自由度PID控制有更好的控制效果和鲁棒性.

关 键 词:不稳定时滞过程    神经网络控制    双自由度控制    鲁棒性  
收稿时间:2013-08-05

Neural Network Control for Unstable Processes with Time Delay
ZHANG Xiao-di;CAI Yun-Ze;HE Xing;ZHANG Wei-dong.Neural Network Control for Unstable Processes with Time Delay[J].Journal of Shanghai Jiaotong University,2014,48(7):1033-1038.
Authors:ZHANG Xiao-di;CAI Yun-Ze;HE Xing;ZHANG Wei-dong
Institution:(Department of Automation, Shanghai Jiaotong University, Shanghai 200240, China)
Abstract:To deal with the stabilizing and robustness problem of the unstable process with time delay, a novel structure with PID type controllers based on BP neural network (BPNN) was proposed. The controller in the inner loop was used to stabilize the unstable process while the outer loop controller which was designed with the BP neural network and PID stabilizing theory purposed by Guillermo was used to improve the performance of the overall system. Simulation result shows that the proposed BPNN PID has good performance in both set point tracking and disturbance rejection. Key words:
Keywords:unstable process with time-delay  back propagation (BP) neural network  2DOF proportion integration differentiation (PID) control  
     robustness  
本文献已被 CNKI 等数据库收录!
点击此处可从《上海交通大学学报》浏览原始摘要信息
点击此处可从《上海交通大学学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号