模糊小脑模型神经网络在多辊冷连轧机轧制力预报模型中的应用  

Rolling force prediction model of a multi-roll cold tandem mill by fuzzy cerebellum model articulation controller

作　　者：刘华强[1] 唐荻[1] 杨荃[1] 郭立伟[1] 

机构地区：[1]北京科技大学高效轧制国家工程研究中心,北京100083

出　　处：《北京科技大学学报》2006年第10期969-972,共4页Journal of University of Science and Technology Beijing

基　　金："十五"国家重大技术装备研制项目(科技攻关)计划资助项目(No.ZZ02-13B-03)

摘　　要：针对宽带钢多辊冷连轧机组特点,为提高轧制力的预报精度,在结合传统轧制压力模型的基础上把模糊算法和神经网络有机结合,设计出基于模糊小脑模型神经网络的多辊冷连轧机轧制力预报模型.通过对传统轧制力模型计算值、小脑模型预报计算值与实测值进行对比分析可知,基于模糊小脑模型神经网络的多辊冷连轧机轧制力预报模型具有较高的计算精度,更适合于多辊轧机在线计算机过程控制的应用,满足现场在线生产的要求,取得良好的板形板厚控制效果.According to the special characters of a multi-roll cold tandem mill, a roiling force prediction model based on fuzzy algorithm and cerebellum model articulation controller （CMAC） combined with the traditional model was designed to improve the precision of rolling force prediction. By comparing the values of the traditional rolling force model and the prediction model by CMAC with the measured ones, the precision of the prediction model by the fuzzy CMAC is much better than that of the traditional roiling force model and is more suitable for online use of a multi-roll cold tandem mill controlled by computer, which also meets manufacturing requirement of the field and receives better control effect of the shape and thickness.

关 键 词：冷连轧机 轧制压力 小脑模型神经网络 模糊算法 

分 类 号：TF301[冶金工程—冶金机械及自动化]