热轧钢材组织演变和力学性能预测—半经验模型、机器学习到工业大模型  

Development of the technology to predict microstructurel evolution and changes of mechanical properties for hot rolled steels——the path from semi-empirical models through machine learning to industrial foundation models

作　　者：刘振宇[1] 吴思炜 丛津功 梁亮[3] 周晓光[1] 曹光明[1] LIU Zhenyu;WU Siwei;CONG Jingong;LIANG Liang;ZHOU Xiaoguang;CAO Guangming(State Key Laboratory of Rolling and Automation,Northeastern University,Shenyang 110819,China;Anshan Iron and Steel Group Co.,Ltd.,Anshan 114021,China;Technical Center,Hunan Hualing Lianyuan Iron and Steel Co.,Ltd.,Loudi 417009,China)

机构地区：[1]东北大学轧制技术及连轧自动化国家重点实验室,辽宁沈阳110819 [2]鞍山钢铁集团有限公司,辽宁鞍山114021 [3]湖南华菱涟源钢铁有限公司技术中心,湖南娄底417009

出　　处：《轧钢》2024年第5期24-37,50,共15页Steel Rolling

基　　金：国家重点研发计划项目(2021YFB3702404);国家自然科学基金青年科学基金项目(52104370);辽宁科技项目(2022JH25/10200001);兴辽英才资助计划项目(XLYC2203186);中信铌钢发展奖励基金资助项目(2022-M1824)。

摘　　要：钢材热轧过程中,轧件显微组织演变、表面氧化与轧制力能负荷相互影响,构成了“牵一发而动全身”的复杂黑箱系统。长期以来,国内外一直采用数学模型分离求解,导致只能近似解析而无法实现精确求解,制约了产品综合质量的进一步提升。通过深度挖掘实验数据和热轧工业数据,我国开发出集力能、组织及界面耦合的热轧工业大模型,实现了热轧主流程的全面、精准解析。介绍了我国热轧钢材组织演变和力学性能预测技术的发展历程及最新发展方向,为钢铁行业数字化转型提供借鉴和参考。During the hot rolling process of steels,the microstructure evolution,surface oxidation,and rolling force energy load of the rolled piece interact with each other,forming a complex black box system.For a long time,analytical models have been used for separation and solution both domestically and internationally,resulting in only approximate analysis and inability to achieve accurate solutions,which has constrained the further improvement of product comprehensive quality.Nowadays,a big industrial model of hot rolling that integrates force,microstructure evolution,and interface has been developed by deeply mining experimental data and industry data,achieving comprehensive and accurate analysis of the main hot rolling process.The development process and latest development direction of the evolution of steel microstructure and mechanical property prediction technology in China are introduced,the direction for the digital transformation of the steel industry are pointed out.

关 键 词：热轧钢材 组织演变 力学性能 半经验模型 机器学习 工业大模型 预测 数字化转型 

分 类 号：TG335.11[金属学及工艺—金属压力加工] TP18[自动化与计算机技术—控制理论与控制工程]