1403 Сomputational Materials Design: Emerging Multiscale, Optimization and Reconstruction Techniques
Masoud Safdari, SIMSOFTS LLC
Kai James, University of Illinois
Ahmad R. Najafi, Drexel University
Majid Baniassadi, University of Tehran
With the advent of more powerful numerical methods, analysts have been able to simulate material systems that incorporate more detailed and complex geometrical models. On the other hand, increasing access to high-performance computing has turned techniques such as molecular dynamics, multiscale methods, topology and shape optimization methods into powerful tools. In conjunction with homogenization and advanced finite element (FE) schemes, these methods now can offer a step change in technology and turn design of materials, the holy grail of computational mechanics, into a reality. Furthermore, rapidly growing additive manufacturing methods can exploit complex designs obtained numerically into products with tailored microstructure that provide desired macroscopic properties. Computational materials design promises turnkey solutions to the most challenging problems facing engineers, such extremely light-weight composites for aerospace structures, materials for highly efficient clean energy conversion and storage, materials for thermal management in high-powered electronics, bio-degradable materials, and many more.
Parallel to the development of these computational methods, key advances have been made in the development of techniques to acquire high-resolution 2D/3D images of materials at different length scales, e.g., optical microscopy, X-ray micro-tomography, electron microscopy (SEM, FIB, TEM) etc. The image data and statistics obtained from these technologies now serves as the basis for the creation of realistic virtual models in the meso- and multiscale modeling of materials whether directly or to mimic key statistical features of the microstructures. For instance statistical correlation functions, a well-known class of statistical descriptors, can be used to perform computational reconstruction, homogenization and eventually perform materials design for material systems such as nanocomposites and ceramics. These methods offer a compelling alternative approach to materials design, one that takes the image data to extract key statistics and sheds light on the microstructure-property-performance relationship needed for design.
The objective of this symposium is to bring together researchers working on state-of-the-art computational techniques with direct application in materials design to exchange ideas, present novel developments and discuss recent advances. Topics of interest include, but are not limited to
• Novel numerical schemes for meso- and multiscale analysis of heterogeneous media including extended/generalized finite element (FE) methods, polyhedral FE methods, mesh free methods and coupled MD-FE multiscale methods
• Optimization methods with application in materials design
• Mathematical methods for realistic reconstruction of material microstructure
• Novel statistical techniques for modeling and design of materials
• Novel techniques for efficient yet accurate reconstruction of digital virtual models from image-based statistics
Parallel to the development of these computational methods, key advances have been made in the development of techniques to acquire high-resolution 2D/3D images of materials at different length scales, e.g., optical microscopy, X-ray micro-tomography, electron microscopy (SEM, FIB, TEM) etc. The image data and statistics obtained from these technologies now serves as the basis for the creation of realistic virtual models in the meso- and multiscale modeling of materials whether directly or to mimic key statistical features of the microstructures. For instance statistical correlation functions, a well-known class of statistical descriptors, can be used to perform computational reconstruction, homogenization and eventually perform materials design for material systems such as nanocomposites and ceramics. These methods offer a compelling alternative approach to materials design, one that takes the image data to extract key statistics and sheds light on the microstructure-property-performance relationship needed for design.
The objective of this symposium is to bring together researchers working on state-of-the-art computational techniques with direct application in materials design to exchange ideas, present novel developments and discuss recent advances. Topics of interest include, but are not limited to
• Novel numerical schemes for meso- and multiscale analysis of heterogeneous media including extended/generalized finite element (FE) methods, polyhedral FE methods, mesh free methods and coupled MD-FE multiscale methods
• Optimization methods with application in materials design
• Mathematical methods for realistic reconstruction of material microstructure
• Novel statistical techniques for modeling and design of materials
• Novel techniques for efficient yet accurate reconstruction of digital virtual models from image-based statistics
