Microstructure-based multiscale modeling of mechanical response for materials with complex microstructures

Complex microstructures are found in several material especially in biological tissues, geotechnical materials and many manufactured materials including composites. These materials are
difficult to handle by classical numerical analysis tools and the need to incorporate more details on the microstructure have been observed. This thesis focuses on the microstructure-based
multi-scale modeling of the mechanical response of materials with complex microstructures and
whose mechanical properties are inherently dependent on their internal structure. The conditions of interest are large displacements and high-rate deformation. This work contributes to
the understanding of the relevance of microstructure informations on the macroscopic response.
A primary application of this research is the investigation and modeling of snow behavior, it
has been extended to modeling the impact response in concrete and composite.
In the first part, a discrete approach for fine-scale modeling is applied to study the behavior of
snow under the conditions mentioned above. Also, application of the this modeling approach
to concrete and composite can be found in the appendices. The fine-scale approach presented
herein is based on the coupling of Discrete Element Method and aspects of beam theory. This
fine-scale approach has proven to be successful in modeling micro-scale processes found in
snow. The micro-scale processes are mainly intergranular friction, intergranular bond fracture,
creep, sintering, cohesion, and grain rearrangement. These processes not only influence the
overall response of the material but also induce permanent changes in its internal structure.
Therefore, the initial geometry considered during numerical analysis should be updated after
each time or loading increment before further loading. Moreover, when the material matrix is
partially granular and continuum, the influence of fluctuating grains micro-inertia caused by
debonding, cracking and contact have a significant effect on the macroscopic response especially
under dynamic loading. Consequently, the overall rate and history dependent behavior of the
material is more easily captured by discrete models. Discrete modeling has proven to be efficient
approach for acquiring profound scientific insight into deformation and failure processes of many
materials. While important details can be obtained using the discrete models, computational
cost and intensive calibration process is required for a good prediction material behavior in the
real case scenarios.
Therefore, in order to extend the abovementioned fine-scale model to real engineering cases a
coarse-scale continuum model based have been developed using an upscaling approach. This upscaled model is based on the macroscopic response of the material with a special regard to the
microstructure information of the material. Different strategies are presented for incorporating
the microstructure information in the model. Micro-scale related dissipation mechanisms have
been incorporated in the coarse-scale model through viscoplasticity and fracture in finite strain
formulation. The thesis is divided into nine chapters, where each is an independent paper
published or submitted as a refereed journal article.