人工智能驱动高性能钛基材料设计与制造的研究进展:机遇与挑战OA
Research progress on artificial intelligence-driven design and manufacturing of high-performance titanium-based materials:opportunities and challenges
高性能钛基材料因成分-工艺-组织-性能关系的敏感性与复杂性,其研发过程长期受限于高维非线性优化与高试错成本的双重挑战.人工智能(artificial intelligence,AI)作为一项渗透性极强的颠覆性技术,正在为高性能钛基材料这一战略领域引入从经验驱动向模型与数据双轮驱动的研发新范式.本文综述人工智能驱动高性能钛基材料技术(AI+Ti)的最新研究进展,聚焦 AI如何针对高性能钛基材料的成分复杂、相变多样、热工艺窗口窄、组织演化路径依赖性强等固有特点,提供新的解决方案.主要内容包括:AI在构建高精度相图-性能预测模型、实现性能目标→微观组织→成分/工艺逆向设计中的突破;在增材制造、热处理等关键过程中,实现从成形控制到组织与性能主动调控的智能升级;以及构建基于数字孪生的服役行为预测框架.在此基础上,进一步剖析 AI+Ti领域面临的数据、模型、验证、集成等核心挑战,并展望物理信息机器学习、自主实验平台等未来发展方向.最后,针对知识表示、人机协作模式、工程信任建立等非共识性问题进行讨论,并对该领域的未来发展趋势进行深入阐述:(1)复杂服役环境下的材料性能预测与多尺度耦合;(2)全流程工艺参数的智能协同;(3)钛合金专用物理信息感知模型的构建与演进.AI+Ti已超越工具应用范畴,升维为一场深刻理解并最终驾驭高性能钛基材料认知与范式的变革.
Due to the sensitivity and complexity of the composition-process-microstructure-performance relationship,the research and development of high-performance titanium-based materials have long been constrained by the dual challenges of high-dimensional nonlinear optimization and high trial-and-error costs.As a highly pervasive disruptive technology,artificial intelligence(AI)is introducing a new research and development paradigm for the strategic field of high-performance titanium-based materials,shifting from experience-driven modes to dual-driven approaches supported by models and data.This review summarizes the latest research advances in artificial intelligence-enabled high-performance titanium-based material technology(AI+Ti),focusing on how AI provides innovative solutions targeting the inherent characteristics of high-performance titanium-based materials,including complex compositions,diverse phase transitions,narrow thermal processing windows,and strong path dependence of microstructure evolution.The main contents include breakthroughs achieved by AI in constructing high-precision phase diagram and performance prediction models,as well as realizing the inverse design from performance objectives to microstructures and further to composition and processing parameters;the intelligent upgrading from forming control to active regulation of microstructures and properties in key processes such as additive manufacturing and heat treatment;and the establishment of an in-service behavior prediction framework based on digital twins.On this basis,this paper further analyzes the core challenges in the AI+Ti field regarding data,models,verification and integration,and prospects future development directions such as physics-informed machine learning and autonomous experimental platforms.Finally,it discusses controversial issues involving knowledge representation,human-machine collaboration modes and engineering trust establishment,and elaborates on the future development trends of this field:(1)material performance prediction and multi-scale coupling under complex service environments;(2)intelligent coordination of full-process processing parameters;(3)the construction and iteration of specialized physics-informed perception models for titanium alloys.Beyond simple tool application,AI+Ti has evolved into a transformative revolution that enables in-depth understanding and ultimate mastery of the cognition and research paradigm for high-performance titanium-based materials.
弭光宝;李培杰;王新宇;唐堰清;成浩;孙若晨;孙圆治;邱越海;谭勇;陈义斯;隋楠;肖文龙
中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095清华大学 新材料国际研发中心,北京 100084中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095||北京航空航天大学 材料科学与工程学院,北京 102206中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095||清华大学 新材料国际研发中心,北京 100084中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095||清华大学 新材料国际研发中心,北京 100084中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095||清华大学 新材料国际研发中心,北京 100084中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095||清华大学 新材料国际研发中心,北京 100084中国航发北京航空材料研究院 先进钛合金航空科技重点实验室,北京 100095北京航空航天大学 材料科学与工程学院,北京 102206
矿业与冶金
高性能钛基材料人工智能机器学习材料设计智能制造发展挑战
high-performance titanium-based materialartificial intelligencemachine learningmaterial designintelligent manufacturingdevelopment challenge
《航空材料学报》 2026 (5)
119-147,29
国家自然科学基金"叶企孙"科学基金(U2141222)
评论