摘要
以邯郸钢铁公司2000 m3高炉采集的数据为样本,采用主成分回归(PCR)方法研究了各运行参数对高炉铁水含硅量的贡献,实现了高维复杂数据的降维.考虑到偏最小二乘法(PLS)在处理多重共线性数据中的优势,利用PLS对高炉铁水含硅量进行预测.结果表明,主成分回归和偏最小二乘法在对高炉冶炼过程中产生的大量数据的处理具有其独到的优势,取得了显著的效果.
With datasets from 2 000 m3 blast furnace of Han Steel as a sample space, principal component regression (PCR) was used to investigate the contribution of operating parameters to the silicon content of hot metal, in blast furnace iron-making and so the complexity of data was reduced. Considering the advantages of partial least square (PLS) in dealing with collinear data, it is used to predict silicon content. The result showed that PCR and PLS had their own advantages for application in iron-making process.
出处
《浙江大学学报(理学版)》
CAS
CSCD
北大核心
2009年第1期33-36,共4页
Journal of Zhejiang University(Science Edition)
基金
国家科技部重点推广项目资助(No.2005EC000166)
关键词
高炉冶炼
高维复杂数据
主成分回归
偏最小二乘法
blast furnace iron-making
principal component regression
partial least square