Microstructural Evolution via Stochastic Morphology Reconstruction from Limited Tomography Data: Modeling, Simulation, and Experimental Verification

技术文摘。该项目的目的是显著提高我们的能力，定量表征和预测实时微观组织演变的非均质材料，包括金属合金，陶瓷，复合材料和颗粒介质在各种条件下，基于有限的断层扫描数据中包含的形态信息。对复杂微观结构及其在各种条件下如何演变的内在理解和知识对于新材料的设计和实现最佳材料性能至关重要。传统的断层扫描重建需要大量的数据（单个静态微观结构需要几个100gb的数据），这极大地限制了它们在表征动态演变的微观结构方面的应用。因此，我们有动力寻找替代方法，通过少量测量获得的最小层析成像数据集来统计表征和预测原位微观结构演变。本研究的基本主题是系统地调查和量化现有实验程序中可获得的断层扫描数据的信息内容，以提高这些数据在微观结构表征和预测中的实用性，为数据分析和管理提供有效的方案，并为数据收集提出新的实验程序。我们建议使用时变空间相关函数来量化结构信息。我们提出了新的数学、计算和物理方法，通过随机形态重建从有限的断层扫描数据中建模、预测和实验验证材料微观结构的演变，这将导致开发免费的集成软件包，用于基于可用的断层扫描数据进行有效的定量4D （3D +时间）微观结构表征和预测。加强材料教育和公众对成像和可视化在材料研究中的重要性的认识是我们建议的一个重要组成部分。我们提出了一个与研究项目相结合的多样化的教育和推广计划，其中包括为K-12学生创建交互式微观结构可视化软件，招募和参与代表性不足的女性和少数民族，为亚利桑那州立大学的学生提供基于项目的活动和研究机会，以及创建一个动态微观结构可视化网站。这些活动将有助于传播、教育和各级学生以及广大公众的参与。非技术抽象。异质材料在自然界和人造结构中大量存在。例子包括复合材料、陶瓷、合金、石料和骨头。这类材料通常在大尺度和小尺度上都表现出复杂的微观结构，微观结构决定了材料的宏观性质和性能。新材料的设计和实现最佳材料性能依赖于我们在无数外部刺激（如热、机械和电）下表征和修改材料特性和行为的能力。因此，对复杂微观结构及其在各种条件下如何演变的内在理解和知识是极其重要的。传统的成像技术（如x射线断层扫描）通常需要大量的数据来呈现微观结构的单个快照。这极大地限制了它们的应用，以捕捉感兴趣的材料的整个进化过程。在拟议的项目中，我们将研究典型的断层扫描数据中包含多少有用的结构信息，以及是否可以设计出可以巧妙地利用最关键的结构信息来准确快速地呈现材料微观结构快照的方法。我们提出了新的数学、计算和物理方法来建模、预测和实验验证材料微观结构的演变，从有限的实验数据。加强材料教育和公众对成像和可视化在材料研究中的重要性的认识是我们建议的一个重要组成部分。我们提出了一个与研究项目相结合的多样化的教育和推广计划，其中包括为K-12学生创建交互式微观结构可视化软件，招募和参与代表性不足的女性和少数民族，为亚利桑那州立大学的学生提供基于项目的活动和研究机会，以及创建一个动态微观结构可视化网站。这些活动将有助于传播、教育和各级学生以及广大公众的参与。

Technical Abstract. The intent of this project is to dramatically improve our ability to quantitatively characterize and predict real-time microstructural evolution of heterogeneous materials including metallic alloys, ceramics, composites and granular media under various conditions, based on the morphological information contained in limited tomography data. An intrinsic understanding and knowledge of complex microstructures and how they evolve under various conditions is extremely important to the design of novel materials and achieving optimal material performance. The large volume of data (several 100 GB data for a single static microstructure) required in traditional tomography reconstructions significantly limits their application in characterizing dynamically evolving microstructures. We are thus motivated to find alternative methods to statistically characterize and predict in situ microstructure evolution with a minimal set of tomography data that can be obtained in a few measurements. The underlying theme of the proposed research is to systematically investigate and quantify the information content of the tomography data obtainable, via current experimental procedures, in order to improve the utility of such data for microstructural characterization and prediction, provide efficient protocols for data analysis and management, and suggest novel experimental procedures for data collection. We propose to quantify structural information using time-dependent spatial correlation functions. We propose novel mathematical, computational, and physical approaches to modeling, predicting, and experimentally verifying material microstructure evolution from limited tomography data via stochastic morphology reconstructions, which will lead to the development of freely available integrated software package for efficient quantitative 4D (3D + temporal) microstructure characterization and prediction based on available tomography data.Enhancing materials education and public awareness of the importance of imaging and visualization in material research is a crucial component of our proposal. We propose a diverse educational and outreach program that is integrated with the research program, which includes creation of interactive microstructure visualization software for K-12 students, recruitment and involvement of underrepresented female and minorities, project-based activities and research opportunities for students at ASU, and creation of a website on dynamical microstructure visualization. These activities will serve to disseminate, educate, and involve students at all levels as well as the public-at large.Non-technical Abstract. Heterogeneous materials abound in nature and man-made structures. Examples include composites, ceramics, alloys, stand stone, and bone. Such materials usually exhibit complex microstructures on both large and small scales, and the microstructures determine the macroscopic properties and performance of the materials. The design of novel materials and achieving optimal material performance rely on our ability to characterize and modify material properties and behaviors under a myriad of external stimuli, such as thermal, mechanical, and electrical. Thus, an intrinsic understanding and knowledge of complex microstructures and how they evolve under various conditions is extremely important. Traditional imaging techniques (such as x-ray tomography) usually require a large amount of data to render a single snapshot of the microstructure. This significantly limits their application to capture the entire evolution process of the material of interest. In the proposed project, we will investigate how much useful structural information is contained in typical tomography data and whether methods can be devised that can smartly utilize the most crucial structural information to accurately and rapidly render snapshots of material microstructures. We propose novel mathematical, computational, and physical approaches to modeling, predicting, and experimentally verifying material microstructure evolution from limited experimental data. Enhancing materials education and public awareness of the importance of imaging and visualization in material research is a crucial component of our proposal. We propose a diverse educational and outreach program that is integrated with the research program, which includes creation of interactive microstructure visualization software for K-12 students, recruitment and involvement of underrepresented female and minorities, project-based activities and research opportunities for students at ASU, and creation of a website on dynamical microstructure visualization. These activities will serve to disseminate, educate, and involve students at all levels as well as the public-at large.