基于CatBoost算法的高炉铁水硫含量预测  

PREDICTION OF SULFUR CONTENT OF MOLTEN IRON IN BLAST FURNACE BASED ON CATBOOST ALGORITHM

作　　者：刘福龙 郝良元 靳亚涛 刘二浩 邱洪涛 张永升 Liu Fulong;Hao Liangyuan;Jin Yatao;Liu Erhao;Qiu Hongtao;Zhang Yongsheng(Strategic Research Institute of HBIS Group Co.,Ltd.,Shijiazhuang 050023,Hebei;Chengde Vanadium and Titanium New Material Co.,Ltd.,Chengde 067001,Hebei)

机构地区：[1]河钢集团有限公司战略研究院,河北石家庄050023 [2]承德钒钛新材料有限公司,河北承德067001

出　　处：《河北冶金》2023年第6期31-36,共6页Hebei Metallurgy

摘　　要：高炉铁水中的硫含量是描述铁水质量的一个重要指标。在出铁之前了解铁水中硫含量的高低,对于高炉实际生产具有重要作用,因此预测模型的建立非常必要。本文利用Catboost算法建立了高炉铁水硫含量的预测分析模型,采用国内某钢铁企业实际高炉生产数据进行学习和预测。运行结果表明,Catboost模型预测精度较高,计算时间较短,满足实际生产需求,同时模型的特征参量通过人工经验和相关性分析相结合的方法,相关关系结果与高炉冶炼理论基本吻合。测试结果表明,基于Catboost算法建立的高炉铁水硫含量预测模型在实际生产中能够起到很好的预测效果,对合理把控高炉铁水硫含量具有重要的参考意义。The sulfur content in blast furnace molten iron is an important indicator to describe the quality of molten iron.Understanding the sulfur content of molten iron before iron production plays an important role in the actual production of blast furnaces,so the establishment of predictive models is very necessary.In this paper,Catboost algorithm is used to establish a predictive analysis model for molten iron and sulfur content in blast furnace,and the actual blast furnace production data of a domestic steel enterprise is used to learn and predict.The operation results show that the Catboost model has high prediction accuracy and short calculation time,which meets the actual production requirements,and the characteristic parameters of the model are basically consistent with the blast furnace smelting theory through the combination of artificial experience and correlation analysis.The test results show that the prediction model of molten iron sulfur content of blast furnace based on Catboost algorithm can play a good prediction effect in actual production and has important reference significance for reasonable control of molten iron sulfur content of blast furnace.

关 键 词：高炉 硫含量 Catboost模型 相关性分析 特征工程 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]