Prediction of bonded asymmetric metallic cross-tension and single lap shear joints using finite element model with material-level adhesive properties and cohesive zone method

Predictive finite element (FE) models of adhesive joints are needed to enable the design and evaluation of adhesively joined structures, particularly large-scale structures that may be costly to assess experimentally. Although models calibrated to coupon-level tests have provided an important first-step, models based on material-level characterization are needed to enable joint assessment in complex modes of loading. In the present study, a structural epoxy adhesive was characterized using rigid double cantilever (Mode I) and bonded shear (Mode II) specimens to provide material-level input properties. Three single-lap shear (SLJ) and seven cross -tension (CT) specimen configurations were fabricated with aluminum and steel sheet materials. The speci-mens included symmetrical (the same adherend material and thickness) and asymmetrical (dissimilar adherend material or unequal thickness) configurations, with three loading angles (0 degrees, 45 degrees, 90 degrees) for the CT specimens. Finite element models of the SLJ and CT specimens were developed using Cohesive Zone Modeling for the ad-hesive, with properties determined from the Mode I and Mode II characterization tests. The FE models of the SLJ and CT test specimens predicted the peak load within an average difference of 2%-19%. The joint strength varied between different test configurations, owing to adherend deformation, load eccentricity and mixed-mode loading. Importantly, the model parameters were not calibrated to the SLJ and CT tests. The FE models were able to predict joint response for varying test specimen geometry, adherend thickness, adherend material, and modes of loading.

Automated summary

This study characterizes a structural epoxy adhesive and develops finite element models to predict the performance of adhesive joints. The models are able to adapt to different test configurations and loading modes, accurately predicting the peak load.