Uncertainty in debonding of layered structures with adhesive layers

The impact of uncertainty on the stochastic debonding process in adhesively bonded layered structure is examined in this paper. The emphasis and focus of this work are put on the effect of explicitly modeling the adhesive layer within the stochastic-structural framework. A pull-out test of a single composite strip adhesively bonded to a rigid substrate is considered as a model of the layered configuration. The stochastic analysis includes physical uncertainty attributed to 10 different input parameters that are represented by random fields. Three main methodologies are integrated. The first one combines a high-order modeling of the deformable adhesive layer and the modeling of its two interfaces by the cohesive interface approach. This enables the analysis to look into the stress fields in the adhesive layer and the tractions that develop across its two interfaces. The second methodology quantifies the impact of uncertainty on the structural response by the nonlinear perturbation based stochastic finite elements method. Following the highly nonlinear character of the problem and the snap back folds involved with the response, the third methodology is integrated by means of an extended pseudo-arc-length procedure that is formulated to solve the nonlinear problem. The stochastic analysis points at the different role played by each component of the structure and each uncertain parameter in the stochastic structural response. It also describes the way these uncertainties evolve along the structural process. The results of the high order model are compared to those of a simplified model in which the adhesive layer and its two interfaces are replaced by a single cohesive interface representing the entire bonding layer. The comparison points at significant differences between the two modeling approaches, evident both in the expected (or deterministic) behavior of the structure and in its stochastic scattering measures.