数据驱动的热轧带钢边部线状缺陷智能预报模型

 边部线状缺陷是热轧带钢易发缺陷,不仅严重影响成材率,还可能对热轧下游工序生产过程造成影响。边部线状缺陷的影响因素复杂多变,建立精确的机理预报模型十分困难。为此,首先分析边部线状缺陷的主要影响因素;然后以智能方法为基础,分别建立了基于逻辑回归与神经网络的边部线状缺陷智能预报模型,并分析了2个模型的精度与泛化能力;最后,以神经网络智能预报模型为基础,对加热工艺参数进行优化,使缺陷发生率与封闭率均大幅降低。研究结果对提高热轧带钢表面质量具有实践意义,可推广应用于同类轧线。

Data-driven intelligent prediction model of edge seam defects for hot rolling strip

Edge seam defects are easy to occur in hot rolling strip, which not only seriously affect the yield, but also may affect the downstream process of hot rolling. It is very difficult to establish an accurate mechanism prediction model because of the complex and changeable factors affecting the edge seam defects. Therefore, first, the main influencing factors of edge seam defects are analyzed; then, based on intelligent methods, the intelligent prediction models of edge seam defects based on logical regression and neural network are established respectively, and the accuracy and generalization ability of the two models are analyzed; finally, based on the neural network intelligent prediction model, the heating process parameters are optimized, which makes the defect rate and defect closure rate are significantly reduced. The research results in this paper have practical significance for improving the surface quality of hot rolling strip, and can be applied to similar rolling lines.