热连轧过程大梁钢变形抗力预报模型  

Prediction model of deformation resistance of strip steel during hot continous rolling process

作　　者：尹宝良 Ahmed elmi Sahal 崔熙颖 令狐克志 么洪勇 邝霜 李学通 YIN Bao-liang;Ahmed elmi Sahal;CUI Xi-ying;LINGHU Ke-zhi;YAO Hong-yong;KUANG Shuang;LI Xue-tong(National Engineering Research Center for Equipment and Technology of Cold Strip Rolling,Yanshan University,Qinhuangdao 066004,China;Tangshan Iron and Steel Group,Tangshan 063000,China;Tangshan Iron and Steel Technology Center of Hebei Iron and Steel Group,Tangshan 063000,China;Materials Technology Research Institute,HBIS Group,Shijiazhuang 050023,China;State Key Laboratory of Metastable Materials Science and Technology,Yanshan University,Qinhuangdao 066004,China)

机构地区：[1]燕山大学国家冷轧板带装备及工艺工程技术研究中心,河北秦皇岛066004 [2]唐山钢铁集团有限责任公司,河北唐山063000 [3]河钢集团唐钢技术中心,河北唐山063000 [4]河钢集团材料技术研究院,河北石家庄050023 [5]燕山大学亚稳材料制备技术与科学国家重点实验室,河北秦皇岛066004

出　　处：《塑性工程学报》2024年第10期159-166,共8页Journal of Plasticity Engineering

基　　金：河北省高等学校科学技术研究资助项目(CXY2023012);中央引导地方科技发展资金资助项目(236Z1024G);石家庄市驻冀高校产学研合作项目(241010191A)。

摘　　要：为解决某钢厂热连轧机组轧制700L大梁钢时因变形抗力预报精度低导致成品带钢板形质量较差的问题,首先,根据实际轧制压力等历史生产数据确定带钢的变形抗力,然后对热连轧机组变形抗力数学模型中的系数进行回归,得到适用于该机组的变形抗力数学模型,同时将遗传算法和BP神经网络结合起来,建立GA-BP神经网络变形抗力预报模型,进一步将优化后的传统数学模型和GA-BP神经网络模型计算的变形抗力结果代入系统轧制力模型中,其模型计算误差分别为6%和3%。将GA-BP神经网络变形抗力模型应用到现场后,700L大梁钢的板形合格率由原来的91.7%提高到了99.3%,板形改善效果显著。In order to solve the problem of the poor plate shape quality of the finished strip steel caused by low prediction accuracy of deformation resistance during rolling of 700L beam steel by a hot continuous rolling mill unit in a steel mill.Firstly,the deformation resistance of the strip steel was determined according to the historical production data such as actual rolling force.Then,the coefficients in the mathematical model of deformation resistance of hot continuous rolling mill were regressed to obtain a mathematical model of deformation resistance suitable for the mill.At the same time,the GA-BP neural network deformation resistance prediction model was established by combining genetic algorithm and BP neural network.Furthermore,the deformation resistance results calculated by the optimized tradition-al mathematical model and GA-BP neural network model were incorporated into the system rolling force model,and the calculation errors of the models are 6%and 3%,respectively.After applying the GA-BP neural network deformation resistance model to the field,the plate shape qualification rate of 700L beam steel is increased from 91.7%to 99.3%,and the improvement effect of plate shape is significant.

关 键 词：热连轧 神经网络 变形抗力 轧制力 板形 

分 类 号：TG335.56[金属学及工艺—金属压力加工]