A damage elasto-plastic discrete model for the nonlinear characterization of pullout tests

The examination of stress transfer phenomena at the interface of composite materials has long been a focal point in mechanical engineering. This investigation involves an in-depth study at these contact surfaces, aiming at exploring the mechanisms governing stress distribution and exchange across the distinct phases. In order to exploit the interaction between diverse materials within each phase, leveraging their complementary attributes, the bond behaviour analysis would help to optimize the overall performance and functionality of the composite material. Many documents are available in literature providing efficient computational simulations, reproducing the non-linear bond behaviour at the interface contact. Nevertheless, those models are sensitive to underlying assumptions and modelling choices, which can significantly influence the results. The aim of this research is to include all the modelling assumptions in the formulation of an elastic and a dissipation energy functionals within the context of a hemi-variational principle. For the sake of simplicity, the interface interaction is reduced to a single damage-elasto-plastic spring in series to an elastic one, which represents the deformation of one of the two phases. The irreversible formulation was developed considering the following steps: (i) definition of two irreversible kinematic descriptors, i.e. a damage index and a plastic displacement, (ii) assumption of an energy functional and (iii) postulation of a hemi-variational principle. Governing equations are therefore derived, including the Karush–Kuhn–Tucker conditions that predict the evolution of irreversible damage and plastic descriptors. The suitability and reliability of this model have been verified comparing numerical and experimental results for the pullout tests carried out for different systems of a single steel cord embedded in an inorganic mortar matrix. Likewise, the evolution of damage and plasticity is calculated. Finally, a parametric investigation is implemented to evaluate to what extent a single factor of the energy functional affects the overall performance.