Understanding the effects of stochastic morphology on compressive behavior in freeze‐casted alumina ceramic

Abstract This study developed a macroscale homogenized progressive failure analysis (PFA) model to connect the distribution of morphological parameters in freeze‐casted ceramics to variability in mechanical compressive behavior. The proposed homogenized anisotropic finite element failure model incorporated stochastic material features, such as lamellar wall misalignment and local material defects. The model was validated experimentally, allowing for the investigation of the impact of variable morphology that develops during fabrication on the failure behavior. Compressive testing of freeze‐cast alumina samples was used to determine the mechanical behavior in various material directions, while the proposed computational model accurately predicted the effective stiffness and strength. The incorporation of stochastic morphology was shown to influence the onset of damage and its propagation, allowing for a greater match with the experimentally observed strain distribution on the material surface. The average stress‒strain curve was compared using functional analysis of variance and showed good agreement between the experimental results and the modeling predictions. Confidence interval envelopes were constructed around the group mean stress–strain curves based on PFA and experimental data, allowing the calculation of the range for predicted strength with 75%–85% accuracy based on the proposed morphological features captured in the homogenization scheme. The proposed computational framework establishes a processing–internal structure–property relationship in freeze‐casted materials and demonstrates the importance of considering microstructural variability in predicting compressive mechanical behavior and associated variability.