Evaluation of the mechanical behaviour of SLJ joints using numerical methods and uncertainties

Adhesive joints provide an effective alternative for joining different structural components in contrast to mechanical fasteners such as rivets and bolts. For composite materials in particular, this technology is relevant, as it avoids drilling-induced damage and stress-concentration effects. However, predicting the strength of bonded joints depends on multiple factors, including geometry, material properties, manufacturing processes, and load conditions. Single Lap Joints (SLJs) are widely used as coupons to evaluate the influence of geometric, material, and toughness parameters on the joint strength. Within numerical modelling, Cohesive Zone Modelling (CZM) is a robust approach for implementation in Finite Element (FE) codes, enabling a comprehensive analysis of the joint behaviour. This work integrates geometrically nonlinear FE modelling, multiple mixed-mode CZM formulations and a Plackett–Burman fractional factorial DoE sensitivity framework to simultaneously evaluate fifteen geometric and constitutive parameter uncertainties in SLJs. Two and three-dimensional FE models were implemented in Abaqus/Standard and validated against available data on SLJ experiments. Parameter influence was quantified using normalised main effect sensitivity indices derived from the DoE matrix. Results showed that, despite mixed-mode loading, adhesive constitutive parameters associated with mode II fracture dominated the joint response. The reported percentages correspond to normalised main-effect sensitivity indices, quantifying how strongly each parameter influences the predicted strength. Averaged across all Traction–Separation Law (TSL) combinations, mode II parameters contributed roughly 40% to joint strength and 26.6% to the critical strain energy release rate. Geometric parameters had minimal impact, except for specimen width, which showed an average influence of 9.4%. For the SLJ geometry and tensile loading conditions considered, these trends indicate that only a limited subset of parameters requires detailed characterisation, a finding that can inform parameter-screening strategies in bonded-joint design more generally.