机器学习在钢铁材料研究中的应用综述

Application of Machine Learning in Steel Materials:A Survey

钢铁是当今世界处于核心地位的金属材料,时代的快速发展对钢铁材料的性能有了新的要求。然而目前钢材的设计具有超过百万种元素和工艺参数的组合,通过传统实验试错法进行钢铁材料的设计与研发缓慢而昂贵。机器学习技术已广泛应用于指导材料设计中,成为材料研究的新兴方法和热门领域。对机器学习在钢铁材料研究中的应用进展进行综述,介绍了机器学习的工作流程和常用模型与算法,阐述了机器学习在钢铁材料特征选择、成分-工艺-性能预测、服役行为预测以及逆向设计方面的研究进展。最后,分析了机器学习技术在钢铁材料领域面临的问题并展望了其发展前景。

Steel is the core metal material in the world,and the rapid development of the times has new requirements for the properties of steel materials.However,the design of steel materials currently involves combinations of over millions of elements and process parameters,leading the design and development of steel materials by traditional trial-and-error method slower and more expensive.Machine learning technology has been widely used to guide the development and design of materials,which has emerged as a novel methodology and a trending domain in the field of materials research.In this paper,the application progress of machine learning in the research of steel materials were summarized,the working process and common algorithms of machine learning were introduced.Meanwhile,the research progress of machine learning in steel materials feature selection,composition-process-performance prediction,service behavior prediction and reverse design were reviewed.Finally,the problems of machine learning technology in the field of steel materials were analyzed and the development prospects were forecasted.