共查询到20条相似文献,搜索用时 137 毫秒
1.
针对涟钢CSP生产线热连轧机组的轧制工艺条件,建立了各工位温度的数学计算模型以及温度一速度模型,并将得出的道次温度应用到轧制力预测系统中去预测道次轧制力。现场的应用结果表明,在不同出炉温度情况下轧制力的预测计算值与现场实测值相吻合。 相似文献
2.
摘要:轧制力是影响中厚板厚度精度和板型的关键因素。兴澄特钢中厚板轧机二级模型采用传统Sims公式计算轧制力,精度较低。为提高轧制力预报精度,首先基于大量历史生产数据,通过主成分分析法对影响轧制力的因素进行处理和分析,选出权重较大的影响因子;其次选取现场代表钢种进行热模拟压缩实验,在此基础上提出基于极限学习机(ELM)的综合神经网络轧制力预报模型,即先通过化学成分计算出基准变形抗力,再将其作为轧制力神经网络输入变量进行轧制力预报。建模采用10折10次交叉验证确定最佳网络隐层节点数,并用现场实际生产过程数据对网络进行训练与测试。综合神经网络模型投入现场生产,轧制力预报相对误差±10%以内占比提高15.61%,钢板头部厚度命中率提高1.9%。 相似文献
3.
4.
分析了一种中厚板的轧制力在线动态修正算法,该算法以实测轧制力为基础,通过道次实测轧制力和模型计算轧制力的值决定轧制力模型参数修正量的大小,真正做到以实测轧制力数据动态校正中厚板轧制力模型,大大提高了轧制力模型的预报精度并使其具有良好的自学习功能.该算法已经在现场获得应用,并具有良好的应用效果. 相似文献
5.
6.
7.
8.
9.
《钢铁研究学报》2020,(5)
轧制力是影响中厚板厚度精度和板型的关键因素。兴澄特钢中厚板轧机二级模型采用传统Sims公式计算轧制力,精度较低。为提高轧制力预报精度,首先基于大量历史生产数据,通过主成分分析法对影响轧制力的因素进行处理和分析,选出权重较大的影响因子;其次选取现场代表钢种进行热模拟压缩实验,在此基础上提出基于极限学习机(ELM)的综合神经网络轧制力预报模型,即先通过化学成分计算出基准变形抗力,再将其作为轧制力神经网络输入变量进行轧制力预报。建模采用10折10次交叉验证确定最佳网络隐层节点数,并用现场实际生产过程数据对网络进行训练与测试。综合神经网络模型投入现场生产,轧制力预报相对误差±10%以内占比提高15.61%,钢板头部厚度命中率提高1.9%。 相似文献
10.
11.
12.
针对平整轧制过程不同用途带钢对表面微观形貌的特殊要求,在批量跟踪电火花毛化轧辊、磨削轧辊和冷轧后带钢表面微观形貌的基础上,建立工作辊与带钢都可考虑真实表面粗糙峰的带钢表面微观形貌轧制转印生成模型,采用工业实验验证了仿真模型的准确性,并据此模型分析轧制前带钢已经具有表面粗糙度分别大于、等于、小于轧辊表面粗糙度时,带钢表面微观形貌的轧制转印行为与遗传演变规律。提出了负转印和转印饱和的概念,定义了两种极限轧制转印状态的描述指标— —负转印最大和转印饱和,研究发现当带钢表面粗糙度小于或等于轧辊表面粗糙度时,存在负转印最大点和转印饱和点;当带钢表面粗糙度大于轧辊表面粗糙度时,负转印最大点和转印饱和点重合。在此基础上,采用负转印最大点与转印饱和点对应的临界板宽轧制力,描述带钢表面微观形貌的遗传及演变规律,并系统仿真分析带钢屈服强度、带钢轧前表面粗糙度、轧辊表面粗糙度等工艺条件参数对于负转印最大点与转印饱和点对应的临界单位板宽轧制力的影响规律,发现随着带钢屈服强度增大和轧辊表面粗糙度增加,该临界单位板宽轧制力均增大;随着带钢表面粗糙度增大,负转印最大点对应的临界单位板宽轧制力增大,但转印饱和点对应的临界单位板宽轧制力却减小。 相似文献
13.
《Baosteel Technical Research》2020,(2)
Thickness,width,temperature,and profile are considered as control targets in process control of hot strip finishing rolling. The pre-calculated settings of the model information include rolling force,cooling water flow between stands,bending force,and roll shifting position. Without changing the load distribution,the interaction among thread speed,rolling force,rolling power,and cooling water flow between stands is comprehensively considered based on the quantifiable relationship among speed,force,and temperature. This paper proposes a full-length multi-point-setting model that uses the settings of strip head to combine the rolling speed diagram with the target finishing mill delivery temperature( FDT) to achieve the calculation of the control parameters along full length of strip. Traditional models cannot effectively predict rolling force of strip body or the maximum and minimum temperatures of FDT. It is also difficult for traditional models to suppress fluctuations in shape accuracy of full-length strip or improve the shape accuracy of the product. Calculation results show that the proposed full-length multi-point-setting model can provide the control parameters for temperature and rolling-force over full length of strip,predict the risk of rolling exceeding the equipment capability,and improve the shape accuracy and rolling stability of hot-rolled products. 相似文献
14.
15.
型钢生产中轧制压力模型的研究 总被引:1,自引:0,他引:1
周怡谋 《金属材料与冶金工程》2000,(4):10-13
针对型钢生产中轧制压力计算存在的难点,在板带轧制理论基础上,通过引入考虑三维的、不均匀变形的应力状态因子,并根据现场轧制压力产测值的回归分析,建立了型钢轧制压力统计模型。 相似文献
16.
17.
18.
19.
For ultra-thin strip rolling, the conventional rolling force models are no longer applicable. To obtain accurate rolling force in the shape and gauge control process, Fleck proposed a new roll flattening model. In this study, experimental analysis, finite element simulation, and theoretical analysis were conducted to evaluate the Fleck model. The experiments and simulations show a clear neutral zone in the deformation zone with decreasing strip thickness. The finite element simulation results show that the proportion of the elastic unloading zone is small, when an elastic unloading phenomenon appears in the neutral zone. Thus, to simplify the rolling force model, the effect of an elastic zone could be ignored. Based on this finding, we develop a rolling force model with quick calculation speed, high precision, and convenient online application. Finally, the accuracy of the simplified model is verified by the measured rolling force. 相似文献