Abstract This paper describes the development and application of a frictionless contact stress solver based on the cell-centred finite volume method. The contact methodology, implemented in the open-source software OpenFOAM, is derived from the penalty method commonly used in finite element contact algorithms. The solver is verified on two benchmark tests using the available Hertzian analytical solutions.

In this study the adhesive joint fracture behaviour of a nano-toughened epoxy adhesive was investigated. Two experimental test methods were used; (i) the standard tapered double cantilever beam (TDCB) test to measure the mode I adhesive joint fracture energy, GIC, as a function of bond gap thickness and (ii) a circumferentially deep notched tensile test to determine the cohesive strength of the adhesive for a range of constraint levels. It was found that the fracture energy of the adhesive followed the well-known bond gap thickness dependency [1]. SEM analysis of the TDCB fracture surfaces revealed significant plastic void growth. Finally, numerical modelling of the experimental tests suggested that most of the fracture energy was dissipated via highly localised plasticity in the fracture process zone ahead of the crack tip.

In the present study, the mixed-mode fracture toughness of an adhesively bonded composite joint system was examined using a variety of linear elastic fracture mechanics (LEFM) based tests. These tests include the mode I double cantilever beam (DCB), mixed-mode asymmetrical DCB (ADCB) and mode II end load split (ELS) test. The joint system was also evaluated using the wide area lap shear (WALS) test that is often employed by the aerospace industry. While lap shear type tests are relatively simple to perform and post-process compared to their LEFM counterparts, the results can often be misleading and are greatly dependent on the overlap length, thickness of substrate and type of fillet. The experimental tests were also simulated using OpenFOAM, a finite volume based software package. Through this combined experimental-numerical approach, a greater understanding of the influence of the peel ply surface treatment and scrim cloth on the behaviour of the WALS test was achieved.

Abstract Tapered-double cantilever-beam joints were manufactured from aluminium-alloy substrates bonded together using a single-part, rubber-toughened, epoxy adhesive. The mode I fracture behaviour of the joints was investigated as a function of loading rate by conducting a series of tests at crosshead speeds ranging from 3.33 × 10−6 m/s to 13.5 m/s. Unstable (i.e. stick–slip crack) growth behaviour was observed at test rates between 0.1 m/s and 6 m/s, whilst stable crack growth occurred at both lower and higher rates of loading. The adhesive fracture energy, GIc, was estimated analytically, and the experiments were simulated numerically employing an implicit finite-volume method together with a cohesive-zone model. Good agreement was achieved between the numerical predictions, analytical results and the experimental observations over the entire range of loading rates investigated. The numerical simulations were able very readily to predict the stable crack growth which was observed, at both the slowest and highest rates of loading. However, the unstable crack propagation that was observed could only be predicted accurately when a particular rate-dependent cohesive-zone model was used. This crack-velocity dependency of GIc was also supported by the predictions of an adiabatic thermal-heating model.