||This symposium focuses on the development and use of computational and data intensive characterization capabilities used by experimentalists and modelers to investigate materials structure and mechanisms at varying length and time scales. Advancements in computational processing power; instrument and detector capabilities; and multi-scale experimental and modeling techniques are generating increasingly large datasets that have facilitated the discovery of quantitative descriptors that link structure to processing parameters and material properties. For example, experimental techniques such as 3D x-ray and synchrotron tomography; atom probe tomography; multi-modal imaging; and high frame rate imaging are generating enormous characterization datasets used to understand and quantify material structure and behavior. Similarly, atomistic and mesoscale simulations are generating large datasets that provide insights for the genesis and evolution of various microstructural features, and provide links to higher order models and experiments. Throughout the materials community, scientific discoveries using these large characterization datasets are being accelerated through the advancement and automation of analysis techniques such as machine learning and artificial intelligence. As these computational and data intensive characterization approaches advance, there is a call for deeper study to quantify their inherent uncertainty of structural descriptors.
This symposium intends to bring together experimental and theoretical experts in computational and data intensive microstructure characterization from both academia and industry, with a focus on the methods and techniques to effectively manipulate, reconstruct, analyze, and apply these data to develop improved predictive capabilities for multi-scale materials design. Suggested areas of focus for this symposium include:
• Theoretical and computational development of novel structural descriptors to characterize microstructural features (e.g. grain boundary atomic and crystallographic structure, crystallographic texture, distributions of triple junction types), and their application to quantitatively characterize experimental and simulation data, and develop new predictive microstructure-property models.
• Methods and algorithms for collecting, reconstructing, analyzing, and quantifying large experimental microstructural datasets collected from tools such as: atom probe tomography, x-ray computed tomography, or high-speed measurements.
• Methods and algorithms for the detection, analysis, and quantification of microstructural features predicted through atomic and mesoscale simulation data. Validation approaches for computational and theoretical models using structural descriptors and advanced experimental mechanics techniques. Methods to bridge modeling and experiment through computed characterization (e.g. virtual X-ray and electron diffraction and simulated microscopy).
• Application of advanced analysis techniques, such as machine learning and artificial intelligence, to develop multi-scale microstructure descriptors and provide greater insights into materials characterization data.
• Methods for quantifying the uncertainty inherent in manipulation, reconstruction and analysis of large sets of characterization data.