用神经网络方法确定冷轧厚度预测模型灵敏度系数  

Application of neural networks in the determination of sensitivity factor for a thickness prediction model of cold rolling

作　　者：王春刚[1] 方一鸣[1] 范之远[1] 高文达[1] 

机构地区：[1]燕山大学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：《钢铁研究》2009年第2期45-47,共3页Research on Iron and Steel

摘　　要：利用神经网络方法确定灵敏度系数,建立了冷轧厚度预测模型。该模型能很好地克服外界干扰或参数扰动,是一种能确定厚度控制参数的算法,通过一个单机架上的应用实例表明该方案具有好的应用前途。A thickness prediction model of cold rolling has been developed using BP neural networks method to determine the sensitivity factor. The model can overcome the interference or parameter disturbance that exists in cold rolling process, therefore, it is a suitable algorithm to determine thickness-controlled parameters. Simulation results in a single stand rolling mill has proven that the proposed technique has good application prospects in controlling of thickness for cold rolling products.

关 键 词：BP神经网络 灵敏度系数 厚度预测模型 相关系数 

分 类 号：TG335.1[金属学及工艺—金属压力加工]