Cohesive Zone Modeling of Composite T-Joints: Insights from Blind Analysis Challenge

Abstract

The goal of this paper is to predict the failure behavior of composite T-joint specimens, specifically C-channels bonded to a spar cap and skin, under T-Pull, Side-Bend, and Mixed-Shear loading. The work was performed as part of a blind prediction challenge organized by the Technical Cooperation Program, with testing led by the Air Force Research Laboratory. The objective is to simulate damage initiation and propagation using cohesive zone modeling in a commercial finite element software package (Abaqus) to provide a benchmark for prediction accuracy and computational cost. True blind predictions, derived strictly from adhesive characterization data, showed good agreement for adhesive-dominated failures in the T-Pull and Mixed-Shear configurations. However, the Side-Bend configuration proved challenging due to failure initiation within the composite “noodle” region. Since the interlaminar strength in this region is difficult to predict from double cantilever beam/end-notched flexure test data, the initial predictions deviated from the experiments, necessitating post-test calibration using data from the Tip Defect case. The T-Pull Tip Defect exhibited a complex “multiple-delamination” mode; incorporating a secondary delamination path reduced the initial peak load overpredictions from 27.4% to 3.5%. Mixed-Shear blind predictions achieved a 2.3% error, while the Side-Bend cases required calibration of the noodle interface properties (1.2–38.7% deviation). Ultimately, cohesive zone modeling accurately predicts adhesive-dominated failures ([Formula: see text] error), but interlaminar noodle failures necessitate multiple-delamination modeling.