CAREER: Mesoscale Modeling of Defect Structure Evolution in Metallic Materials

该学院早期职业发展(CALEAR)计划项目专注于在极端环境中使用的结构金属材料的虚拟分析的高级计算力学方面的研究。它支持促进对控制缺陷及其结构演变的因素、其演变的微观机制以及它们对材料性能的集体影响的研究。了解运行过程中缺陷结构的演化与材料性能和生存能力之间的联系是材料力学中的一个关键问题。这项研究将明确为什么一种特定的材料会导致韧性提高或强度提高，或者两者兼而有之。这种洞察力将为下一代汽车、航空航天和国防应用的材料开发提供支持。虚拟分析通过补充物理实验来减少材料部署的成本和时间，从而有助于实现材料基因组倡议的目标。通过该奖项的教育倡议将侧重于本科生的积极参与和在康涅狄格大学建立计算材料科学专业的研究领域。材料科学、机械工程和计算机科学在这个框架内的整合将有助于激发参与其中的本科生接受理工科高等教育的兴趣。外联活动将通过专业协会地方分会的领导作用，将材料机械引入大学前教育，并鼓励代表性不足的群体积极参与，以促进科学和工程教育的多样性。这项研究的目的是深入了解微观结构和加载条件对缺陷结构(位错、孪晶、界面)的形核、积累和相互作用以及空洞的形核、长大和合并形成裂纹(损伤)的微观机制的影响。这项研究采用了一种新开发的准粗晶动力学(QCGD)方法，该方法能够保留变形和破坏过程中涉及的过程的原子尺度物理，但扩展了分子动力学模拟的时间和长度尺度能力。这种方法在原子模拟和连续介质模拟之间架起了一座桥梁，并且位于中尺度。机器学习算法将被用来将缺陷结构的演变和分布映射到宏观的应力-应变响应，并识别触发诸如损伤起始等关键事件的分布。这将允许变形过程中的微观组织演变与结构金属材料的强度和韧性响应之间的直接联系。这种能够提供对性能和生存能力的洞察的虚拟分析框架将导致当前材料建模技术的重大进步，并可以扩展到其他结构材料。

This Faculty Early Career Development (CAREER) Program project focusses on research in advanced computational mechanics for the virtual analysis of structural metallic materials for use in extreme environments. It supports research on advancing the understanding of the factors that control the evolution of defects and their structure, the micromechanisms for their evolution, and their collective influence on material performance. Understanding the links between the evolution of defect structures during operation and material performance and survivability is a key question in the mechanics of materials. The research will define a clear rationale for why a particular material results in improved toughness or improved strengths or both simultaneously. Such insight would support the development of materials for next generation automotive, aerospace, and defense applications. The virtual analysis contributes to the goals of the Materials Genome Initiative by supplementing physical experiments to reduce costs and time in materials deployment. Educational initiatives through this award will focus on active involvement of undergraduate students and the establishment of a field of study specialization in computational materials science at the University of Connecticut. The integration of materials science, mechanical engineering, and computer science in this framework will help to stimulate an interest in the undergraduate students involved to pursue higher education in science and engineering. Outreach activities will introduce mechanics of materials into pre-college education through leadership roles in local chapters of professional societies and encourage active participation of underrepresented groups to promote diversity in science and engineering education. The objective of this research is to establish insight into the effects of microstructure and loading conditions on the micromechanisms responsible for the nucleation, accumulation, and interaction of defect structures (dislocations, twins, interfaces) as well as nucleation, growth, and coalescence of voids to form cracks (damage). The research employs a newly developed quasi-coarse-grained dynamics (QCGD) method that is able to retain the atomic scale physics of processes involved during deformation and failure but extends the time and length scale capabilities of molecular dynamics simulations. This approach bridges the gap between the atomistic and continuum simulations, and is located at the mesoscale. Machine learning algorithms will be used to map the evolution and distribution of defect structures to the macroscale stress-strain response and identify the distributions that trigger critical events such as damage initiation. This will allow direct connections between the microstructural evolution during deformation and the strength and toughness response for structural metallic materials. This virtual analysis framework capable of providing insights into the performance and survivability will lead to significant advancements in the current state-of-art for materials modeling and can be extended to other structural materials.