具有伸长率分配计算功能的轧制力预报智能模型研究  

Intelligent Rolling Force Prediction Model With Elongation Distribution Function

作　　者：何海涛[1] 刘宏民[2] 蒋岳峰[3] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004 [2]燕山大学机械工程学院,河北秦皇岛066004 [3]宝山钢铁股份有限公司冷轧薄板厂,上海200431

出　　处：《钢铁》2007年第1期55-58,共4页Iron and Steel

基　　金：河北省教育厅自然科学指导性计划(Z2005309)

摘　　要：针对双机架平整机的特性,以基态弯辊力下带材出口板形最好为准则,提出具有双机架平整机伸长率分配系数计算功能的轧制力模型;在此基础上,为了改善传统轧制力模型的预报精度,提出了先通过神经网络利用在线测得的实际数据预测变形抗力和摩擦因数,再与轧制力机理模型自学习过程相结合的轧制力预报新方法;并将其应用于宝钢1220双机架平整机的生产实践,结果表明此模型可以高精度地预报轧制压力。According to the characteristics of two-stand temper mill, a rolling force model with elongation distribution calculation function was established on the criterion that the best shape of strip is obtained under the base bending force. Moreover, in order to improve the prediction precision of rolling load, a new approach was proposed in which deformation resistance and friction coefficient are first predicted by neural network on measured data, then rolling force is predicted by combining above parameters with the self-adaption process of rolling force model. The new model was applied on 1220 two-stand temper mill at Baosteel, proving that the rolling force on temper mill can be predicted with high precision.

关 键 词：平整机 轧制压力 神经网络 自适应 

分 类 号：TG335[金属学及工艺—金属压力加工]