机器学习在高效连铸制造中的应用现状及展望  

Application Status and Prospects of Machine Learning in High-efficiency Continuous Casting Manufacturing

作　　者：宗男夫 刘宏亮 李爽 刘佳韵 杨军 石凤丽 Maryam Khaksar Ghalati 董洪标 ZONG Nanfu;LIU Hongliang;LI Shuang;LIU Jiayun;YANG Jun;SHI Fengli;Maryam Khaksar Ghalati;DONG Hongbiao(Technology Center of Ben Gang Group Corporation,Benxi 117000,China;National Local Joint Engineering Laboratory for Advanced Automotive Steel Development and Application Technology,Benxi 117000,China;School of Engineering,University of Leicester,Leicester LE17RH,United Kingdom)

机构地区：[1]本钢集团有限公司技术中心,辽宁本溪117000 [2]先进汽车用钢开发与应用技术国家地方联合工程实验室,辽宁本溪117000 [3]英国莱斯特大学数字化研发中心,英国莱斯特郡LE17RH

出　　处：《鞍钢技术》2024年第6期23-33,49,共12页Angang Technology

摘　　要：高效连铸以高拉速、铸坯无缺陷为基础,通过钢-轧界面间铸坯热装和轧制实现钢铁制造流程的绿色低碳发展。传统连铸制造需向智能化连铸转型,以连铸制造流程大数据量和机器学习算法为基础,挖掘有价值的信息或逻辑关系。构建具有决策能力和预测判断功能的智能化铸机,实现连铸高效生产和智能控制。通过介绍国内外机器学习在高效连铸制造中的应用现状,分析智能连铸在铸坯裂纹在线智能预报、结晶器漏钢智能预报和连铸过程的其他重要智能化的研究进展,提出机器学习在高效连铸制造中的研究展望。通过深度学习在连铸全流程中的应用、探索不均衡样本的智能连铸算法和结合工业机理模型与机器学习模型实现多目标任务的智能化连铸模型,实现连铸流程智能化和透明化控制,为国内相关钢铁企业提供智能化连铸的研发思路和参考。High-efficient continuous casting is based on high casting speed and defect free castings,achieving green and low-carbon development of the steel manufacturing process through hot charging and rolling of castings at the steel rolling interface.Traditional continuous casting manufacturing needs to transform towards intelligent continuous casting,based on the large amount of data and machine learning algorithms in the continuous casting manufacturing process,to explore valuable information or logical relationships.Build an intelligent casting machine with decision-making ability and predictive judgment function to achieve efficient production and intelligent control of continuous caster.By introducing the current application status of machine learning in efficient continuous casting manufacturing at home and abroad,analyzing the intelligent continuous caster in online intelligent prediction of cracks,intelligent prediction of mold breakout,and other important intelligent research in the continuous casting process,the research prospects of machine learning in high-efficient continuous casting manufacturing are proposed.By applying deep learning in the total continuous casting process,exploring intelligent continuous casting algorithms for imbalanced samples,and combining industrial mechanism models with machine learning models to achieve intelligent continuous casting models for multi-objective tasks,the continuous casting process can be intelligent and transparent controlled,providing research and development ideas and references for intelligent continuous casting for relevant domestic steel enterprises.

关 键 词：机器学习 智能连铸 自主决策 高效连铸 

分 类 号：TH133.33[机械工程—机械制造及自动化] TF777.2[冶金工程—钢铁冶金]