基于BP网络逆模型的轧辊偏心自适应逆控制  

Adaptive Inverse Roll Eccentricity Control Based on BPNN Inverse Model

作　　者：王亚静[1] 刘福才[1] 张艳欣[1] 窦春霞[1] 吴士昌[1] 

机构地区：[1]燕山大学西校区电院工业计算机控制工程河北省重点实验室,河北秦皇岛066004

出　　处：《化工自动化及仪表》2010年第6期9-12,共4页Control and Instruments in Chemical Industry

基　　金：国家自然科学基金资助项目(60874023);燕山大学博士基金资助项目(B111);河北省自然科学基金资助项目(F2010001320)

摘　　要：针对冷轧带钢生产中的轧辊偏心控制问题,将神经网络与自适应逆控制相结合,提出一种基于BP网络逆模型的自适应逆控制方法。将系统的动态特性控制和对象的噪声控制分成两个独立的过程,当动态性能达到最优时,对象扰动的影响也减到最小。将该方法应用到350 mm四辊可逆液压轧机上,仿真结果表明,系统具有良好的动态和稳态特性。For roll eccentricity in cold rolling mill,neural network was combined with adaptive inverse control theory and an adaptive inverse method based on BP neural network was proposed.The dynamic characteristic control and the noise elimination were distributed to two independent courses.When the dynamic characteristic achieved optimal,the object disturb played down to the least.Apply this method to 350 mm four-roller reversing hydraulic mill,simulation results show that the dynamic and steady characteristics of the system are all quite well.

关 键 词：轧辊偏心 神经网络逆模型 自适应逆控制 噪声消除 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]