共查询到20条相似文献,搜索用时 140 毫秒
1.
2.
3.
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
4.
5.
6.
7.
轧制力预测中RBF神经网络的组合应用 总被引:1,自引:0,他引:1
传统的数学模型无法达到冷连轧控制的尺寸精度要求。针对传统轧制力模型的固有缺陷,为提高冷连轧机组轧制力计算精度,合理选择、更新和预处理训练样本,采用RBF神经网络预测冷轧带钢屈服应力并把它用于传统轧制力计算模型,获得较高的轧制力预测精度。而后使用RBF长期数据修正网络和RBF短期数据修正网络得到长期数据修正网络和短期数据修正网络的修正系数,对轧制力计算值进一步修正,从而进一步提高轧制力预报精度。上述方法直接用于某冷连轧机组,轧制力预测误差在±6%之内。这充分证明RBF网络可以成功用于轧制过程控制并满足实际生产的需要。 相似文献
8.
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
9.
摘要:轧制力预报一直是热连轧过程控制模型的核心,浅层神经网络对复杂函数的表示能力有限,而深度学习模型通过学习一种深层非线性网络结构,实现复杂函数逼近。利用深度学习框架TensorFlow,构建了一种深度前馈神经网络轧制力模型,采用BP算法计算网络损失函数的梯度,运用融入Mini batch策略的Adam优化算法进行参数寻优,采用Early stopping、参数惩罚和Dropout正则化策略提高模型的泛化能力。基于上述建模策略,针对宝钢1880热连轧精轧机组的大量轧制历史数据进行了建模实验,对比分析了4种不同结构的前馈网络预测精度。结果表明,相比于传统SIMS轧制力模型,深度神经网络可实现轧制力的高精度预测,针对所有机架的预测精度平均提升21.11%。 相似文献
10.
11.
Work rolls in hot rolling mills are thermally and mechanically loaded; both of these loading aspects are difficult to measure. Laboratory tests can be used for the specification of the thermal load in the cooling area; however a thermal load in a roll gap is still difficult to measure. The paper describes an experimental technique developed for monitoring the work roll surface temperature by sensors embedded in the work roll. Continuous hot rolling pilot line trials were performed for different process conditions. One parameter, for example, roll cooling, rolling velocity, reduction, or skin cooling, can easily be changed during the trials, and the effect on the thermal cycle of the work roll can be directly measured. These thermal measurements give very detailed information about the temperature field. An inverse heat‐conduction model has been developed to compute the surface boundary condition from the measured temperatures. The heat flux and heat transfer coefficient distribution along the roll circumference can be obtained afterwards. The results for different rolling velocities and reductions (up to 50%) are shown. 相似文献
12.
13.
Because the structure of the classical mathematical model of rolling load is simple, even with the self adapting technology, it is difficult to accommodate the increasing dimensional accuracy. Motivated by this fact, an Innovations Feedback Neural Networks (IFNN) was presented based on the idea of Kalman prediction. The neural networks used the Back Propagation (BP) algorithm and applied it to the prediction of rolling load in hot strip mill. The theoretical results and the off line simulation show that the prediction capability of IFNN is better than that of normal BP networks, namely, for the prediction of the rolling load in hot strip mill, the prediction precision of IFNN is higher than that of normal BP networks. Finally, a relative complete rolling load prediction system was developed on Windows 2003/XP platform using the OOP programming method and the SQL server2000 database. With this system, the rolling load of a 1700 strip mill was calculated, and the prediction results obtained correspond well with the field data. It shows that IFNN is valid for rolling load prediction. 相似文献
14.
15.
16.
17.
18.
轧机特性的回归与参数估计 总被引:10,自引:1,他引:9
本文对AGC系统中的轧机特性问题进行了研究,介绍了本钢热连轧AGC系统中使用的轧机特性回归模型和参数估计方法,并给出实际控制效果。 相似文献
19.
20.
A combination of finite element method and neural network methods was used for rapid prediction of the roll force during skin pass rolling of 980DP and 1180CP high strength steels. The FE based commercial package DEFOEM-2D was used to develop a mathematical model of the skin pass rolling operation. Numerical experiments were designed with different process parameters to produce training data for a neural network algorithm. The friction coefficient was considered as an input parameter in the neural network but it was optimised using an iterative method employing an equation that relates the friction coefficient to the rolling force. The load prediction method described in this paper is sufficiently rapid that it can be used in real-time as an adjustment tool for skin pass rolling mills with error within 10% (based on plant data from POSCO). 相似文献