机器学习在金属增材制造技术中的研究进展  

Progress in Application of Machine Learning in Additive Manufacturing Technology

作　　者：操宇恒 陈超越 郭铠 胡涛 徐松哲 王江 任忠鸣 CAO Yuheng;CHEN Chaoyue;GUO Kai;Hu Tao;XU Songzhe;WANG Jiang;REN Zhongming(State Key Laboratory of Advanced Special Steels,School of Materials Science and Engineering,Shanghai University,Shanghai 200444;Sino-European School of Technology,Shanghai University,Shanghai 200444)

机构地区：[1]上海大学材料科学与工程学院,省部共建高品质特殊钢冶金与制备国家重点实验室,上海200444 [2]上海大学中欧工程技术学院,上海200444

出　　处：《特种铸造及有色合金》2024年第11期1454-1465,共12页Special Casting & Nonferrous Alloys

基　　金：科技部重点研发计划资助项目(2021YFB3702502);上海市自然科学基金资助项目(23ZR1421500);国家自然科学基金资助项目(52271035)。

摘　　要：增材制造广泛应用于医疗和航空航天等领域,但仍面临材料种类受限、零件质量和性能不稳定,以及工艺参数与力学性能间的复杂关系等挑战。作为人工智能的关键技术,机器学习具备识别数据中模式和趋势的优势,成为推动增材制造研究的重要工具。综述了机器学习在增材制造中的应用,包括模型参数选择、冶金缺陷与性能预测、原位监控及合金成分、工艺流程和几何结构优化,旨在为增材制造的过程控制与性能优化提供指导。Additive manufacturing is widely used in medicine and aerospace fields.However,it still faces challenges such as limited material types,unstable part quality and mechanical properties,and the complex relationship between process parameters and mechanical properties.As a key technology in artificial intelligence,machine learning has the advantage of recognizing patterns and trends in data,promoting research in additive manufacturing.The application of machine learning in additive manufacturing was reviewed,including the selection of model parameters,prediction of metallurgical defects and performance,in-situ monitoring,and optimization of alloy composition,process,and geometric structure,aiming to provide guidance for process control and performance optimization in additive manufacturing.

关 键 词：机器学习 增材制造 实时监控 缺陷控制 力学性能 

分 类 号：G353.11[文化科学—情报学] TG665[金属学及工艺—金属切削加工及机床] TP181[自动化与计算机技术—控制理论与控制工程]