机器学习辅助2000MPa级弹簧钢成分和热处理工艺开发  

Development of Composition and Heat Treatment Process of 2000 MPa Grade Spring Steels Assisted by Machine Learning

作　　者：杨累 赵帆 姜磊 谢建新 YANG Lei;ZHAO Fan;JIANG Lei;XIE Jianxin(Beijing Laboratory of Metallic Materials and Processing for Modern Transportation,Institute for Advanced Materials and Technology,University of Science and Technology Beijing,Beijing 100083,China;Key Laboratory for Advanced Materials Processing(MOE),Institute for Advanced Materials and Technology,University of Science and Technology Beijing,Beijing 100083,China;Northeast Light Alloy Co.,Ltd.,Harbin 150060,China;Beijing Advanced Innovation Center for Materials Genome Engineering,Institute for Advanced Materials and Technology,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]北京科技大学新材料技术研究院现代交通金属材料与加工技术北京实验室,北京100083 [2]北京科技大学新材料技术研究院材料先进制备技术教育部重点实验室,北京100083 [3]东北轻合金有限责任公司,哈尔滨150060 [4]北京科技大学新材料技术研究院北京材料基因工程高精尖创新中心,北京100083

出　　处：《金属学报》2023年第11期1499-1512,共14页Acta Metallurgica Sinica

基　　金：国家自然科学基金项目No.52101118;中国科协青年人才托举工程项目No.2022QNRC001。

摘　　要：通过收集弹簧钢及其他典型淬火+回火型钢铁材料的文献数据,采用面向性能的机器学习设计系统(MLDS)结合实验优化,实现了具备超高强度和良好塑性的新型弹簧钢化学成分及热处理工艺参数的快速设计。所开发的2种新型弹簧钢的抗拉强度分别为2183.5和2193.0 MPa、屈服强度分别为1923.0和2024.5 MPa、断后伸长率分别为10.5%和9.7%、断面收缩率分别为42.4%和41.5%。新型弹簧钢的强化方式以晶界强化和位错强化为主,细小的晶粒尺寸和适量的奥氏体使得弹簧钢在具备超高强度的同时保持良好的塑性。与现有同等强度级别的超高强度钢相比,新型弹簧钢具有显著的成本优势和工艺优势。The rapid development of rail transit has led to the proposition of higher requirements for the mechanical properties of springs and spring steels.Thus,bogies have been identified as the key components for trains to achieve high speed since they are connected with train bodies and wheel sets through springs.Alternatively,since the properties of spring steel materials have an important effect on the safety and comfort of high-speed trains,the development of spring steels with ultra-high strength and good plasticity has attracted the attention of researchers and industrial circles.However,simultaneously improving strength and plasticity has remained an important challenge for the research and development of high-end steels.Notwithstanding,machine learning has recently made substantial progress in designing and predicting various materials,and is expected to become a powerful tool for clarifying the relationship between the composition,process,and properties of complex alloys like steels.Based on the above background,this study reports the realization of rapid chemical composition and heat treatment process-design parameters for new spring steels,using a performance-oriented machine learning design system with high strength and good plasticity(tensile strength(2050±50)MPa,elongation 10.5%±1.5%)after collecting literature data on spring steels and other typical quenched+tempered steels.Experimental studies were also carried out to obtain a further optimized heat treatment process(heating at 950oC for 30 min and oil quenching+tempering at 380oC for 90 min and water cooling).Investigations revealed that the tensile strengths of the two new spring steel materials developed were 2183.5 and 2193.0 MPa,their yield strengths were 1923.0 and 2024.5 MPa,their elongations after fracture were 10.5%and 9.7%,and the area reductions were 42.4%and 41.5%,respectively,with grain boundary strengthening and dislocation strengthening being the main strengthening mechanisms of the new spring steels.It was also observed that the fine gr

关 键 词：弹簧钢 超高强 热处理 机器学习 

分 类 号：TG142.33[一般工业技术—材料科学与工程]