基于DRNN自适应反馈补偿控制器设计与应用

在轧钢控制系统中，由于轧机两侧液压压下系统的不一致会导致带材横向厚度差分布
不均匀，直接影响带材的板形质量，针对此问题，本文提出将对角递归网络自适应PID 控制器用于反馈回路中的补偿控制方案，用以提高轧机两侧压下动态同步的品质。仿真实验结果表明，此种方案能满足被控对象对高精度同步控制的要求。
关键词：液压伺服；对角递归网络；板形控制；递推预报误差算法

Design and application of self-adaptive feedback compensation controller based on
Dynamic Recurrent Neural Network ZHAO Xu ， MENG Xiang-ping (1. Electric and Information Engineer School,Changchun Engineer Uniersity, Changchum，Jilin Province 130012, China ; 2.Electric and Information Engineer School ,Beihua University,Jilin Jilin province ,132021,China) Abstract: The vary character of both side of rolling mill hydraulic pressing system leads to an unequal horizontal thickness difference distribution of strip which affects strip shape directly .In order to resolve this problem, the neural network self-adaptive PID based on Dynamic Recurrent Neural Network was used in feedback circle of rolling mill hydraulic pressing system to improve the synchronization of both side of rolling mill hydraulic pressing system .The results of simulation tests indicate this method can satisfy the demands of the controlled object for the high precision synchronous control.
keywords: Hydraulics servo; Dynamic Recurrent Neural Network; shape control ;recursive prediction error algorithm