TARDIGRADE-EXAMPLES
Tardigrade-examples (LANL code O4925) is a repository of computational workflows that exercise the Tardigrade software package.
The Tardigrade software package is an implementation of Eringen’s micromorphic continuum theory with capabilities to support multiscale material modeling workflows. These capabailities include homogenization through the Micromorphic Filter, calibration of micromorphic materials models, and macroscale simulation in Tardigrade-MOOSE.
This repository investigates continuum upscaling of various direct numerical simulations (DNS) conducted in Abaqus finite element (FE), Ratel FE, and GEOS material point method (MPM) software. Verfication of the upscaling workflow is first investigated by considering DNS of trivial stress states for homogeneous materials, results of which indicate that classical continuum behavior is recovered as expected. DNS of heterogeneous materials are then considered.
Description
Tardigrade-examples includes upscaling analyses of a variety of DNS. The WAVES computational workflow tool is used to automate these studies in a reproducible manner primarily through a collection of Python scripts.
A full upscaling study considers four main analysis stages:
DNS
Filter
Calibrate
Tardigrade-MOOSE
The included DNS investigate simple compression of a homogenoeous cylinder. Alternatively, externally performed DNS results may be upscaled that consider more interesting materials and stress states. DNS results are post-processed into the XDMF file format required by the Micromorphic Filter. This repository includes Python scripts for XDMF conversion of the Abaqus output database (ODB), Ratel visualization toolkit (VTK), and GEOS MPM VTK formats.
The Micromorphic Filter homogenizes DNS results onto a macroscale mesh where each element is considered a “filtering domain.” Meshes are generated using Cubit. For users without access to Cubit, a collection of example meshes are included, but functionality is limited. Upscaling workflows consider macroscale meshes of varying element size using WAVES parametric study tools to analyze how homogenization changes as the number of filtering domains increase.
The homogenized, micromorphic stress and deformation measures output by the Micromorphic Filter are used to calibrate various material models. A unique calibration for each filtering domain (macroscale element) is produced. For verification studies of simple DNS, classical inear elasticity is calibrated, while several forms of micromorphic linear elasticity are calibrated for heterogeneous DNS. Calibration of micromorphic elasto-plasticity is performed for inelastic DNS conducted externally to the workflows contained in this repository. Externally performed DNS are accessed using the “–peta-data-copy” tool.
Finally, material calibration results are applied to macroscale simulations in Tardigrade-MOOSE with similar boundary and loading conditions as the original DNS. Python scripts are included that collect and summarize results across parametric studies for different analysis stages.
Documentation is included which details the upscaling methodology, steps for building and connecting required software, relevant micromorphic theory, how the Micromorphic Filter works, constitutive theories, and results of several upscaling studies.
Micromorphic Upscaling Workflow
Software Requirements
Quasi-static Verification
Dynamic Verification
Heterogeneous Upscaling
GEOS Homogenization
Micro-averaging Domain Studies
Model Repository
Help & Reference
Appendix - Detailed Theory