临钢6^#高炉含硅量预测的计算机在线实现  

On-line Realization of Predicting Silicon Content by Computer for 6# BF in Linfen Iron and Steel Company

作　　者：孙桂利[1] 李孔安 皇祯平[3] 

机构地区：[1]山西师范大学数学与计算机科学学院,山西临汾041004 [2]山西临汾钢铁公司,山西临汾041000 [3]西安航空职业技术学院,陕西西安710089

出　　处：《控制工程》2003年第1期78-81,共4页Control Engineering of China

摘　　要：临钢 6# 高炉铁水中硅质量分数的预测是“高炉专家系统”中的一部分 ,为了使预测模型能够在线进行 ,采用了限定记忆法进行模型参数的辨识 ,对辨识出来的参数用牛顿法进行优化 ,并且用鼓风动能和风量、风温 2种方法对预测模型的预测值进行了修正 ,使命中率提高了 4～ 5个百分点。给出了数据库系统的组织以及计算机实现的流程图。Predicting hot metal silicon for 6 # BF in Linfen steel works is a part of 'Expert System in Blast Furnace'. In order to apply these predicting models on line,'Recursive Fixed Memory method' was used to identify parameters which were optimized by Newton method and modified. Predictors were modified using two parameters:ventilation, hot air volume and temperature which can increase hitting ratio for 4? % ～5 % . Database and flow chart were given .The calculations got to the hitting ratio level off line.

关 键 词：高炉 含硅量 预测 计算机 铁水 山西临汾钢铁公司 递推最小二乘法 限定记忆法 

分 类 号：TF549.9[冶金工程—钢铁冶金]