Abstract: Based on multicomponent regression analysis, the model of prediction for end temperature of molten steel T_L in CAS-composition adjustment by sealed argon bubbling refining has been established: T_L= t  -(-920.64  +  0.0643625X₁+0. 187162X₂+  0.581293t  +0.003662X₃),where  X₁-  molten   steel  laying  aside  time  /min;X₂-  refining teated time/min; X₃- amount of added alloy /kg; t- temperature of molten steel at beginning of treatment /℃. The re-  sults of statistic analysis on prediction temperature showed that the hitting probability of prediction deviation of±5 ℃ and  ±10 ℃ was 87% and 95% respectively.

Key words:  , CAS Refining,Temperature of Molten Steel,Multicomponent Regression,Model of Prediction

摘要： 采用多元回归分析建立了 30 0tCAS精炼 (密封吹氩调整成分 )终点钢水温度 T_钢包 ℃的预测模型 :T_钢包 =t - (- 92 0 6 4 +0 .0 6 4 36 2 5X₁ +0 .18716 2X₂ +0 .5 812 93t +0 .0 0 36 6 2X₃) ,式中X₁钢水搁置时间 min ;X₂ 精炼处理时间 min ;X₃ 合金加入量 kg ,t 处理开始钢水温度 ℃。对预测温度统计分析结果表明 ,该模型对CAS精炼终点温度的预测误差在± 5℃和± 10℃时的命中率分别为 87%和 95 %。