A Study on the Strain-Rate- and Stress-State-Dependent Behavior of Advanced Ceramics: Experimental Mechanics and Multiscale Simulations

Abstract

Design and development of better-performing ceramic structures are vital in industry-driven applications where a wide range of strain rates and stress states are induced in the material (e.g., impact events). Accordingly, this thesis investigates the strain-rate and stress-state-dependent behavior of advanced alumina ceramics through combined experimental mechanics and multiscale numerical approaches. Experimentally, the effect of the strain-rate and stress state on the failure response of alumina ceramics was explored by designing and testing cuboidal, angled, and flattened Brazilian disc (FBD) specimens to induce compression, shear-compression, and tensile stress states in the material, respectively. Experimental testing was equipped with ultrahigh-speed imaging coupled with digital image correlation (DIC) to achieve the full-filed strains and capture the failure initiation and propagation sites. The material microstructure was characterized by state-of-the-art diagnostics to inform the multiscale models in terms of grain size distribution, porosity, and grain crystallographic orientations. Finite element (FE) models of the macroscale experiments were developed to provide a better insight into the failure progression in the material by quantifying the history of damage. A rate-dependent viscosity regularized version of the phenomenological Johnson-Holmquist-2 (JH2) material model (i.e., the JH2-V model) was implemented through a VUMAT subroutine in Abaqus FE software. Once validated, the model was exercised to quantitatively analyze the damage initiation and growth in the material to provide insights into the role of stress state on the failure response of alumina ceramics at the macroscale. Next, polycrystalline based representative volume elements (RVEs) of the additively manufactured (AM) alumina ceramics were generated using Neper software based on the data captured through the microstructural characterization. To account for the transgranular failure mechanism, the grains were constitutively modeled by the developed JH2-V model, and the grain boundaries were modeled by the bi-linear cohesive zone model (CZM) approach to account for the intergranular failure mechanism. Upon validation with the experimental data, the micromechanical model was leveraged to quantify the history of failure mechanisms, which is challenging to unravel by in-situ experimental approaches, particularly for brittle materials. The model was used to study the effect of microstructural features (e.g., porosity, grain orientations, and grain boundary properties) on the macroscale response of the AM alumina. The novelty and importance of this thesis stem from (a) Building on the limited previous studies, in light of the expansion in the application of AM ceramic structures as a potential replacement for conventionally made ones, this thesis experimentally investigates the mechanical performance of AM alumina ceramics across different stress states and strain rates, providing implications for model development and design of AM ceramic structures with tailored mechanical performance. (b) This work develops an experimentally validated microstructure-based FE modeling framework to explore the relationships between the microstructure and macroscale response of AM ceramics, which lays the foundation for developing machine learning (ML)-based surrogate models and ML-assisted cross-scale simulations to accelerate the design and optimization of AM ceramic structures that perform better.