反馈神经网络在动态过程数据校正中的应用  

Data Reconciliation for Dynamic Process Based on Recurrent Neural Network

作　　者：徐俊波[1] 潘吉铮[1] 许庆国[1] 赵文[1] 周传光[1] 

机构地区：[1]青岛科技大学化工学院,山东青岛266042

出　　处：《青岛科技大学学报（自然科学版）》2003年第3期251-253,276,共4页Journal of Qingdao University of Science and Technology:Natural Science Edition

摘　　要：以全混流反应器为例 ,对反馈神经网络在动态过程数据校正中的应用进行了研究 ,并对 Elman网络及其改进形式进行了比较。由校正结果得出 ,Elman网络引入自反馈后 ,能更好地反映系统的动态特性。In this paper, the use of recurrent neural network in dynamic process for data reconciliation is studied with a example of CSTR, and a comparison of the results produced between using Elman network and its modified form is presented. It can be seen from the reconciliation results that Elman network can give the dynamic characteristics of the system with improved quality when the self-recurrent introduced into the reconciliation process.

关 键 词：神经网络 动态过程 数据校正 全混流反应器 自反馈 

分 类 号：TQ02[化学工程] TP183[自动化与计算机技术—控制理论与控制工程]