1580 mm热连轧带钢精轧轧制力模型研究与改进  

Research and improvement of finishing rolling froce model used in 1 580 mm hot strip rolling mill

作　　者：徐芳[1] 王文广[1] 艾矫健[1] 李东宁[1] 王元嵩 XU Fang;WANG Wenguang;AI Jiaojian;LI Dongning;WANG Yuansong(Shougang Jingtang United Iron and Steel Co.,Ltd.,Tangshan 063200,China)

机构地区：[1]首钢京唐钢铁联合有限责任公司,河北唐山063200

出　　处：《轧钢》2023年第4期86-90,112,共6页Steel Rolling

摘　　要：为了提高首钢京唐1 580 mm产线精轧轧制力预报精度,对精轧轧制力模型进行了研究,结合现场生产的典型问题,即同一钢种族内化学成分波动、薄规格带钢头部大张力引起精轧模型自学习趋势异常及变形抗力自学习层别跳变引起的轧制力设定偏差,对轧制力基础模型和自学习模型进行了改进。修正了钢种族的划分方法、回归整定了化学成分对变形抗力的影响因子、增加了实测张力修正精轧自学习的方法以及建立了基于双线性插值方法来获取变形抗力自学习系数的方法。改进措施实施后,各机架的轧制力预报精度均有不同程度的提高,且带钢通长的厚度标准差由12.22μm降低至10.5μm以内,指标精度得到显著提升。To improve the prediction accuracy of the finishing rolling force on 1 580 mm hot strip production line of Shougang Jingtang Company,the finishing rolling force model was studied.And the basic model of rolling force and the self-learning model were improved based on the typical problems,such as the fluctuation of chemical composition in the same steel family,the abnormal self-learning trend of rolling force model caused by the head large tension of the thin strip,and the reference deviation on rolling force cause by the different classification of the deformation resistance.The classification method to steel family was modified,the impact factors of chemical composition on deformation resistance were tuned by regression,the correction self-learning method on measured tension was added and the bilinear interpolation method of get the deformation resistance self-learning coefficient was established.The prediction accuracy of rolling force was improved on each stand,and the standard deviation of strip thickness was decreased from 12.22 μm to less than 10.5 μm,which was significant improvement.

关 键 词：热连轧带钢产线 轧制力 预报精度 变形抗力 模型 化学成分 自学习系数 张力 改进 

分 类 号：TG335.56[金属学及工艺—金属压力加工]