An Image-Based Co-Designed Experimental and Computational Framework for Heterogeneous Materials

Heterogeneous materials are indispensable to a variety of scientific and engineering applications. However, the design of these complex material systems has often been outpaced by the product to market cycle. In other words, these materials are not being understood quickly enough to exploit their full potential. This is often due to the lack of integration between necessary experiments and modeling capabilities. The modeling of these complex systems often relies on the microstructural information acquired from imaging techniques. To this end, this research focuses on establishing a co-designed framework which improves characterization techniques and image-based modeling protocols for investigating multi-scale behaviour of heterogeneous materials. Specifically, an image-based modeling concept along side complex three-dimensional imaging techniques is used to understand the influence of microstructure and local damage phenomena on the effective mechanical response of particulate reinforced composites.

The experimental framework encompasses utilizing the X-ray micro-computed tomography imaging technique to obtain a detailed characterization and statistical analysis of the evolution of microstructural features upon uniaxial compression loading. Furthermore, the nondestructive, three-dimensional image-based analysis protocol is leveraged in the development of representative unit cells (RUCs). RUCs enable the ability to perform complex computations on relatively smaller domains all the while conserving the statistical characteristics of the original microstructure. By properly representing the microstructure and calibrating relevant physical models, accurate numerical predictions are realized. Moreover, this co-designed framework is used in the development of a novel damage model to study particulate reinforced composites. Continuum damage mechanics are employed to develop the novel damage model that correctly accounts for the deviatoric and volumetric split of the strain energy density function and the damage variables in the finite strain setting. Numerical studies are included to demonstrate the full co-designed experimental and computational framework while using the novel damage model to investigate damage at the interface in a particulate reinforced composite.

Large three-dimensional damage simulations of particulate composites under high-strain rate loading are performed to address the effect of the microstructure on the overall transient mechanical behavior. Spheres as well as oblate, prolate, and plane strain ellipsoids are used and show that resolving the wide range of spatial and temporal scales in complex morphologies is essential to capture damage characteristics. In order to resolve the characteristic damage length-scale, a highly parallel finite element solver, PGFem3D, is used. This work shows that microstructural details as well as high-strain rates play an important role in addressing complex damage patterns and the overall material response. The goal of this dissertation is to further improve the fidelity of image-based modeling and enable high confidence in simulation-based engineering decision in design and analysis.