机器学习辅助增材制造材料和部件力学性能评价研究进展  

Research Progress on Machine Learning Assisted Mechanics Performance Evaluation of Additively Manufactured Materials and Parts

作　　者：王皞 王百涛[1] 高帅龙 刘建荣 李述军[2] 吉海宾[2] WANG Hao;WANG Baitao;GAO Shuailong;LIU Jianrong;LI Shujun;JI Haibin(University of Shanghai for Science and Technology,Shanghai 200093,China;Institute of Metal Research,Chinese Academy of Sciences,Shenyang 110016,China)

机构地区：[1]上海理工大学,上海200093 [2]中国科学院金属研究所,沈阳110016

出　　处：《航空制造技术》2025年第7期40-55,共16页Aeronautical Manufacturing Technology

基　　金：国家自然科学基金(U2241245);沈阳市自然科学基金(23-503-6-05);航空科学基金(2022Z053092001)。

摘　　要：随着增材制造技术的不断发展,越来越多的增材制造材料和部件被应用于航空航天、汽车制造、医疗器械等领域。然而,传统的力学性能评价方法试验耗时长、成本高、数据量有限,难以有效评估增材制造材料和部件的复杂力学性能。机器学习技术通过高效的数据处理、多变量分析和特征提取,为增材制造材料和部件的力学性能评价提供了一种新颖且高效的解决方案。本文综述了机器学习在增材制造材料和部件力学性能评价中的最新研究进展,介绍了增材制造技术及其在力学性能评价中的挑战,探讨了机器学习在拉伸、压缩、疲劳、蠕变等性能以及断裂韧性评价中的具体应用。机器学习方法通过提高预测精度、降低试验成本、加快评价速度,有效克服了传统方法的局限性。最后,列举了机器学习在增材制造领域存在的若干挑战和待解决的问题,并对未来的研究方向进行了展望。With the continuous development of additive manufacturing technology,more and more additively manufactured materials and parts are used in aerospace,automotive manufacturing,medical devices and other fields.However,traditional mechanical property evaluation methods are difficult to effectively assess the complex mechanical properties of additively manufactured materials and parts due to the time-consuming experiments,high cost and limited data volume.Machine learning technology provides a novel and efficient solution for mechanical property evaluation of additively manufactured materials and parts through efficient data processing,multivariate analysis and feature extraction.This paper reviews recent research advances in machine learning for the evaluation of mechanical properties of additively manufactured materials and parts.First,the challenges of additive manufacturing technology in mechanical property evaluation are introduced.Then,specific applications of machine learning in the evaluation of tensile,compressive,fatigue and creep properties as well as fracture toughness are explored.Machine learning methods effectively overcome the limitations of traditional methods by improving prediction accuracy,reducing experimental costs,and accelerating evaluation.Finally,several challenges and pending issues in the application of machine learning in additive manufacturing are enumerated,and future research directions are envisioned.

关 键 词：增材制造(AM) 性能评价 力学性能 机器学习 预测模型 

分 类 号：TG142[一般工业技术—材料科学与工程]