Diagnostic study on head thickness deviation of hot rolled strip based on optimized RT-PLS
JING Fengwei,FENG Ying,ZHANG Yongjun,LI Wen
Author information+
National Engineering Research Center for Advanced Rolling Technology, University of Science and Technology Beijing, Beijing 100083, China
{{custom_zuoZheDiZhi}}
{{custom_authorNodes}}
{{custom_bio.content}}
{{custom_bio.content}}
{{custom_authorNodes}}
Collapse
文章历史+
出版日期
2021-07-20
摘要
摘要:头部厚度偏差是热轧带钢的重要产品质量指标,在板带轧制厚度控制中起着重要作用。实际生产中,基于多种原因带钢头部厚度常会出现偏差超限现象。为了分析头部厚差超限的主导原因,采用偏最小二乘法,结合马氏距离相对变换和潜变量优化选取方法,建立了基于优化相对变换偏最小二乘法(Relative TransformationPartial Least Squares,RT-PLS)的带钢头部厚差诊断模型。实例表明:优化RT-PLS诊断模型能够准确查找出导致带钢头部厚差超限的主要特征参数,指导生产现场的调节,成功降低了后续带钢的头部厚差,使厚度命中率由92.18%提升至97.13%,为带钢头部厚差的诊断研究提供了一种有效的诊断方法。
Abstract
Abstract: Head thickness deviation is an important product quality index of hot rolled strip, which plays an important role in the thickness control of strip rolling. During the production process, due to various reasons, the head thickness deviation of strip is often out of tolerance. In order to analyze the dominant reasons for the head thickness deviation, adopting the partial least squares method and combining the relative transformation of Mahalanobis distance and the optimal selection method of latent variables, a diagnostic model designed to control the head thickness deviation of strip was established based on the optimized Relative Transformation Partial Least Squares (RT-PLS). The results show that the optimized RTPLS diagnostic model can accurately find the main characteristic parameters that cause the strip head thickness deviation exceeding the tolerance and guide the adjustment of the production site, which successfully reduces the head thickness deviation of the subsequent strip, increasing the thickness hit rate from 92.18% to 97.13%, and provides an effective diagnostic method for the diagnostic study of the strip head thickness deviation.
JING Fengwei,FENG Ying,ZHANG Yongjun,LI Wen.
Diagnostic study on head thickness deviation of hot rolled strip based on optimized RT-PLS[J]. Journal of Iron and Steel Research, 2021, 33(7): 593-599 https://doi.org/10.13228/j.boyuan.issn1001-0963.20200184