| This dissertation presents a mechanistic approach to characterize the complex nature of fatigue behavior and computational damage models for elastic-viscoelastic randomly oriented particulate composites such as asphalt mixtures. The dynamic shear rheometer was used to characterize fundamental linear viscoelastic properties of asphalt binders and mastics. Various dynamic mechanical tests using the dynamic mechanical analyzer were performed for cylindrical sand asphalt samples made with pure binders and/or mastics to estimate viscoelastic characteristics and fatigue behavior. Test results were analyzed using linear and nonlinear viscoelastic theories, analytical micromechanics models, a rheological composite model, and a fatigue prediction model based on continuum damage mechanics. The mechanical effects of additives, material aging, and material healing due to rest periods on fundamental material characteristics and fatigue resistance were investigated. In addition, a reasonable definition of fatigue failure was made. Material property measurements of each constituent in the asphalt mixture were then employed in a micromechanical finite element model that includes effects of viscoelasticity, material heterogeneity, and nonlinear damage growth. Simulation results were compared to testing data. The resulting finite element model can be successfully used for numerous types of composite solid media that exhibit large impact of matrix viscoelasticity and complex damage evolution characteristics within the matrix and along the matrix-particle boundaries. The proposed experimental, analytical, and computational approach is expected to be suitable for the material characterization, damage identification, and prediction of damage-induced performance. |
