Layerwise Decohesion Model: A Computationally Efficient Delamination Model for T-Shaped Composite Structures Under Pull and Side-Bend Loading

Delamination is a critical failure mechanism in laminated composites, severely limiting the structural performance and durability. This paper introduces a computationally efficient, thermodynamically consistent layerwise framework for modeling delamination in laminated composite beams. The method provides considerable computational advantages over full 3D simulations due to the reduced order beam model utilized here. The proposed approach incorporates internal cohesive interfaces within the beam elements, eliminating the need for external interface elements. Interfacial degradation is modeled using a damage mechanics-based decohesion model, employing damage driving force derived from the Helmholtz free energy, which ensures irreversible damage evolution and non-negative energy dissipation. Both crack opening and closure are accurately captured using a smooth kinematic ramp function, eliminating the need for complex contact algorithms and penalty regularization. The displacement-based layerwise formulation for beams allows direct coupling between beam and two-dimensional finite elements, thereby facilitating modeling of intricate T-shaped structures. The T-joint region is modeled using two-dimensional quadrilateral elements within the existing Graph-based Finite Element Analysis (GraFEA) framework. Results reveal that combining these modeling frameworks significantly improves computational efficiency without sacrificing accuracy. Model parameters are calibrated using Mode I (DCB) and Mode II (ENF) experiments, and the predictive capability is assessed on T-shaped structures under pull and side bend loading with experimental data provided by the TTCP Blind Challenge 2025 hosted by the Air Force Research Laboratory, USA. The predictions show good agreement with the experimental data, accurately capturing both the global load-displacement response and the associated delamination characteristics.