燕山大学学报2020，Vol.44Issue(4)：340-346,396,8.DOI:10.3969/j.issn.1007-791X.2020.04.002
基于时间序列和支持向量机耦合分析的 冷轧辊系热变形预报
Thermal deformation prediction of cold rolls based on time series and support vector machine coupling analysis
摘要
Abstract
It is a significance issue to predicate the thermal deformation of roll system in strip mill in order to the control of plate thickness and shape. The first, the trend thermal deformation and periodic thermal deformation of 300 experimental rolling mills are separated and extracted based on the analysis of the mechanism of roll thermal deformation and time series analysis method. The second, the total rolling force and rolling speed are taken as the influence factors of periodic thermal deformation, and the support vector regression method is used to predict the thermal deformation. Finally, the cumulative thermal deformation prediction value is obtained by superposing the predicted values of trend thermal deformation and periodic thermal deformation. The new prediction ability and accuracy of the periodic variation of thermal deformation are improved, which has a good application prospect.关键词
板带轧机/辊系热变形/时间序列/支持向量机回归Key words
strip mill/ thermal deformation of rolls /time Series/support vector regression分类
矿业与冶金引用本文复制引用
刘涛1,2*,赵明星1,2,李健通1,2..基于时间序列和支持向量机耦合分析的 冷轧辊系热变形预报[J].燕山大学学报,2020,44(4):340-346,396,8.基金项目
国家自然科学基金资助项目(51875498) （51875498）
河北省自然科学基金资助项目(E2018203339) （E2018203339）
河北省自然科学基金钢铁联合基金项目(E2017203079) （E2017203079）
