基于机器学习与遗传算法的Cu-Ni-Co-Si合金性能预测及机理  

Prediction of properties and mechanism of Cu-Ni-Co-Si alloys based on machine learning and genetic algorithm

作　　者：张英凡 陈慧琴[1] 党淑娥[1] 陈娟 徐全 房晓天 石腾龙 代云云 Zhang Yingfan;Chen Huiqin;Dang Shue;Chen Juan;Xu Quan;Fang Xiaotian;Shi Tenglong;Dai Yunyun(School of Materials Science and Engineering,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China)

机构地区：[1]太原科技大学材料科学与工程学院,山西太原030024

出　　处：《金属热处理》2025年第1期255-265,共11页Heat Treatment of Metals

基　　金：中央引导地方科技发展资金自由探索类基础研究项目(YDZJSX2021A039);山西省研究生优秀创新项目(2022Y667)。

摘　　要：机器学习已经被广泛应用于材料研究领域,然而从成分、工艺多维度进行合金设计依旧是一个不小的挑战。提出了一种机器学习方案,结合材料的物理化学性质、成分、工艺进行合金设计,采用遗传算法对Cu-Ni-Co-Si合金性能预测进行优化;采取递归消除法探索特征与合金性能之间的潜在联系。研究发现,影响合金硬度和导电性的主要工艺是时效处理与冷轧变形。除此以外的物理化学特征主要通过影响自由电子密度和自由电子迁移的自由程,从而对合金的导电率产生影响;通过影响固溶强化和位错强化,从而对合金的硬度产生影响。Application of machine learning in materials research is extensive.However,the task of designing alloys based on many composition and process factors remains a significant difficulty.A machine learning approach to develop alloys by considering the physicochemical qualities,composition,and process of the material was proposed.The property prediction of the Cu-Ni-Co-Si alloy was then optimized by using a genetic algorithm.The recursive elimination method was employed to investigate the potential correlation between the characteristics and alloy properties.The results show that the primary factors influencing the hardness and conductivity characteristics of the alloys are the aging treatment and cold rolling deformation.Furthermore,the physicochemical properties primarily influence the conductivity of the alloy by impacting the density of free electrons and the free path of free electron migration.It affects the hardness of the alloy by exerting influence on solution strengthening and dislocation strengthening.

关 键 词：机器学习 遗传算法 时效处理 冷轧变形 性能预测 

分 类 号：TG146.11[一般工业技术—材料科学与工程] TP181[金属学及工艺—金属材料]