文献类型：期刊文献

中文题名：基于机器学习的异构金属材料性能预测及结构设计

英文题名：Property Prediction and Structure Design of Heterostructured Metallic Materials Based on Machine Learning

作者：王晓坤[1];汪永纪[1];贾云飞[1];张勇[1];贺琛贇[1];董博[1];张显程[1]

机构：[1]华东理工大学,承压系统与安全教育部重点实验室,上海200237

年份：2023

卷号：47

期号：5

起止页码：72

中文期刊名：机械工程材料

外文期刊名：Materials For Mechanical Engineering

收录：CSTPCD;;北大核心:【北大核心2020】；CSCD:【CSCD_E2023_2024】；

基金：国家自然科学基金资助项目(52222505,51975211)。

语种：中文

中文关键词：异构金属材料;强韧性;机器学习;性能预测;结构设计

外文关键词：heterostructured metallic material;strength and ductility;machine learning;property prediction;structure design

摘要：异构金属材料因其特殊的微观结构,在具有较高强度的同时仍然能保持良好的韧性,但是复杂的结构参数使其性能预测和结构设计变得非常困难。机器学习(ML)在处理高维物理量之间的复杂非线性关系方面表现出强大的能力,从而成为异构金属材料性能预测和结构设计的有力工具。介绍了异构金属材料的特征,总结了ML算法及其相关的数据处理问题,综述了ML在异构金属材料性能预测和结构设计方面的应用研究现状,给出了ML辅助异构金属材料性能预测及结构设计中存在的关键问题,并对未来的研究方向进行了展望。
Heterostructured metallic materials can maintain good toughness,and have high strength because of their special microstructures,but their complex structural parameters make it very difficult to predict their properties and design their structures.Machine learning(ML)is a powerful tool for the property prediction and structure design of heterostructured metallic materials due to its ability to deal with complex nonlinear relationships between high dimensional physical quantities.The characteristics of heterostructured metallic materials are introduced.ML algorithm and the related data processing problems are summarized.The application status of ML in the property prediction and structure design of heterostructured metallic materials is reviewed.The key problems existing in the property prediction and structure design of heterostructured metallic materials assisted by ML are given out,and the future research direction is prospected.