基于灰关联分析的多元线性回归模型对铁液含硅量的预测

Prediction of Silicon Content in Molten Iron upon Ash Determination RelevancyAnalysis and Multiple Linear Regression Model

铁液含硅量是表征高炉炉温等不可准确检测状态的反应量,准确预测可帮助高炉冶炼人员及时了解炉内信息。本文运用灰关联分析方法,提取20个影响含硅量预测因素中的6个作为主要影响因素,然后建立多元线性回归滑动预测,预测命中率在90%左右,相对误差为0.0178。

By using ash determination relevancy analysis method, 6 main factors of 20 actors which have influenced the forecast of Si content have been selected. Then a multiple linear regression forecast model has been built for glide forecast, resulting in hit rate 90% with relative error 0.0178.