高炉铁水硅含量序列的支持向量机预测模型  

Prediction Model of Silicon Content Series in Blast Furnace Hot Metal with Support Vector Machines

作　　者：袁冬芳[1] 赵丽[1] 石琳[1] 李江鹏[1] 

机构地区：[1]内蒙古科技大学数理与生物工程学院,内蒙古包头014010

出　　处：《太原理工大学学报》2014年第5期684-688,共5页Journal of Taiyuan University of Technology

基　　金：国家自然科学基金资助项目(61263015);内蒙古高等教育厅研究项目(NJ10104);内蒙古科技大学创新基金(2011NCL019;2011NCL031)

摘　　要：基于包钢6号高炉的在线生产数据,首先建立铁水硅含量序列的自回归AR(p)模型,分析其滞后阶数;然后对硅含量的自回归项及影响因素进行主成分分析,找到多元变量相互独立的有效信息作为输入变量;最后建立铁水硅含量的支持向量机回归预测模型。该模型对炉温预测的准确度达到88.2%,对在线监测高炉炉温具有一定的实用价值。In this paper,based on the data from No.6blast furnace of Baotou Steel,firstly,an auto-regression model,AR（p）,is proposed to analyze the lag of the hot metal silicon content.Secondly,Principal Component Analysis,based on the auto-regression factors and influcial factors of hot metal silicon content,is conducted to achive independent and effective information as input variables.Finally,a support vector machine regression model is established to predict the silicon content in hot metal.It can provide an accuracy of 88.2%in prediction of temperature in the blast furnace and has some practical use in online monitoring the silicon content in hot metal.

关 键 词：自回归AR(p)模型 主成分分析 支持向量机 炉温预测 

分 类 号：TF543[冶金工程—钢铁冶金] TP273.4[自动化与计算机技术—检测技术与自动化装置]