Doctoral Dissertations
Date of Award
12-2025
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Major
Civil Engineering
Major Professor
Timothy Truster
Committee Members
Dayakar Penumadu, Mark Denavit, Reza Abedi
Abstract
Adhesive bonding enables the creation of lightweight structures. However, predictive modeling of single-lap joints under process-to-service loading remains a non-trivial task. This dissertation develops and validates a process-aware interface framework that unifies cohesive representations of the adhesive layer and thermomechanical curing effects consistent with ASTM-standard testing, and anisotropic damage formulation for directionally dependent materials. Within ABAQUS, four adhesive-layer idealizations listed as a single-row cohesive element, a middle-row cohesive element, a solid continuum layer, and a property-graded layer are benchmarked for mesh objectivity and mode-mixity evolution at overlap ends and failure-mode transitions. Later findings show that the solid cohesive layer provides a reliable local peel/shear stress field and damage onset, while a properly tuned property-graded or “middle-CZ” representation reproduces global load displacement with markedly lower cost; single-row continuum interfaces under-resolve peak peel near adherend terminations. Furthermore, a coupled thermomechanical analysis quantifies cure-induced residual fields using embedded optical fiber cable as in-situ sensors, with cohesive parameters calibrated by inverse identification against ASTM single-lap shear tests. The calibrated models reconstruct residual-stress gradients that bias early mode mixity, and they assess sensor intrusiveness through a diameter-to-adhesive-thickness scaling. The optical fiber’s stiffness perturbs local fields near the mid-thickness yet leaves far-field joint strength essentially unchanged when placement and diameter are selected within identified bounds. The simulations also rationalize the optical fiber strain under-reading in short edge zones via a finite strain-transfer length before the fiber tracks the adhesive strain. Crack propagation spanning adhesive and adherend regions is resolved through a formulation that preserves path continuity across dissimilar elastic and damage properties. To extend failure prediction to directionally dependent materials, an anisotropic damage model is formulated by combining a positive/negative projection, cast as a linear complementarity problem, with an auxiliary isotropic mapping. For transversely isotropic cases, a single auxiliary Poisson parameter identified within a narrow interval restores most lateral-strain recovery upon unloading, enforces unilateral damage, preserves the symmetry of the fourth-order consistent tangent, and yields objective energy dissipation with stable crack-path evolution. Collectively, the framework closes the process–structure–performance linkage for bonded materials and paves the way for the modeling of composites and sandwich structures under realistic manufacturing histories.
Recommended Citation
Adediran, Ibrahim A., "Multiscale Interface Modeling for Structural Health Monitoring and Failure Prediction in Adhesive Joints and Directionally Dependent Materials. " PhD diss., University of Tennessee, 2025.
https://trace.tennessee.edu/utk_graddiss/13575