Monday, February 25, 2019
Edited by Dr. Barbara Planas, and Franziska Arnold, Math2Market GmbH.
Specialized methods for direct numerical simulations in porous media. (Abstract)
3D imaging methods, such as micro Computed Tomography (µCT) or focused ion beam scanning electron microscopy (FIB-SEM), allow deep insights into the three-dimensional structure of porous materials. The resulting 3D data sets are very large, often exceeding 20003 or 8 billion volume elements called voxels. Researchers and engineers are interested in determining effective homogenized material properties based on these data sets to understand existing materials or to design new man-made materials. Recent advances in computer technology have made it possible to compute and visualize effective properties such as permeability, and thermal or electrical conductivity on these large images in very short times and using surprisingly little memory.
Classical finite-element-methods (FEM) or finite-volume-methods (FVM) are not suited to compute physical properties on these large images. The bottleneck of these methods is the mesh generation that must be done before the actual simulation can take place and can take longer than the solving of the discretized partial differential equations. Instead, complex microstructures are best dealt with by fast and memory efficient numerical methods that are explicitly designed for them. In this paper, we present state-of-the-art numerical finite-volume-based and fast Fourier transformation-based methods which do not require mesh generation and are designed to compute effective properties directly on very large 3D images. We also present relevant application areas where these methods are used successfully. We show how simulations on the microscale help to support computational material research and development.
Read the full report: https://doi.org/10.30423/REPORT.M2M-2018-01