基于机器学习的铜合金设计的研究现状  

Review on the Design of Copper Alloys Based on Machine Learning

作　　者：张燕宇 柴胜利 吴婷 刘超[1] 姜尚元 胡勇[1,2] ZHANG Yan-yu;CHAI Sheng-li;WU Ting;LIU Chao;JIANG Shang-yuan;HU Yong(Shanxi Key Laboratory of Metallic Materials Forming Theory and Technology,School of Materials Science and Engineering,Taiyuan University of Science and Technology,Taiyuan 030024,China;Shanxi Key Laboratory of Copper-Based New Materials,Shanxi North Copper New Material Technology Co.,Ltd,Yuncheng 044000,China)

机构地区：[1]太原科技大学材料科学与工程学院金属材料成形理论与技术山西省重点实验室,山西太原030024 [2]铜基新材料山西省重点实验室山西北铜新材料科技有限公司,山西运城044000

出　　处：《铸造设备与工艺》2024年第6期60-66,共7页Foundry Equipment & Technology

基　　金：山西省科技成果转化引导专项(202204021301025)。

摘　　要：科技的快速发展对铜合金的硬度、强度和导电性提出了更高的要求。然而,目前铜合金的设计主要依赖于传统的实验试错法,这种方法一直面临成本昂贵、效率低下的问题。作为一种新兴方法,机器学习技术已经成为研究的热点,并广泛应用于材料设计。本文介绍了机器学习的工作流程和常用算法的适用环境,阐述了机器学习在铜合金材料特征选择、成分-工艺-性能预测、逆向设计和成分-工艺-组织预测方面的研究进展。最后,分析了机器学习在铜合金设计领域面临的问题,并对其应用提出了建议。Due to the rapid development of technology,the higher requirements about hardness,strength,and electrical conductivity have been proposed.However,at present,the design of copper alloys mainly relies on the traditional trial-and-error method,which is always faced with the problems of high cost and low efficiency.As a new method,machine learning technology has become a popular in research,which has been widely used in materials design.In this review,the workflow of machine learning and the suitable environments for common algorithms were introduced.Meanwhile,the progress of machine learning in enhancing selection,predicting composition-process-property relationships,reverse design and predicting composition-process-microstructure relationships of copper alloys were summarized.Finally,the challenges of applying machine learning in copper alloys design were analyzed,and some suggestions for the application were put forward.

关 键 词：铜合金 机器学习 性能预测 成分设计 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]