PSI - Issue 41

T.J.S. Oliveira et al. / Procedia Structural Integrity 41 (2022) 72–81 Oliveira et al. / Structural Integrity Procedia 00 (2019) 000 – 000

73

2

1. Introduction Increasing technological innovation has driven the need for new technological developments in all areas. As such, the rigor in quantifying the loads and stresses involved in a structure is progressively imposed. Here the connections between materials are highlighted, where adhesive joints can be considered. Analyzing their characteristics, these prove to be a great solution when compared to other forms of connection, namely welding, bolting and riveting (da Silva et al. 2011), Adhesive bonds have benefits such as corrosion resistance, capability to join dissimilar materials and complex geometries, attainment of lighter and stronger structures, improved aesthetics (without bolt heads, rivets or welding beads), high fatigue strength, more uniform stress fields along the bonded area that empower a more efficient load transfer (Adams 2005). However, disassembly is not feasible for most situations, adhesive joints exhibit poor pullout resistance and sometimes high pressures and temperatures are required to cure (da Silva and Öchsner 2008). Despite these facts, the application of adhesive joints still has a somewhat limited use, since the most conventional application joints occur between non-straight components. Concerning the large number of joint architectures available, the most typical ones are single-lap joints (SLJ) and double-lap joints (DLJ). Only a small area of development is associated with tubular joints. Thus, it is important that its geometry is optimized to provide the highest possible mechanical strength, keeping in mind the weight reduction of the entire structure. It is necessary to provide data regarding the stresses and strains in this type of joint, according to the differentiated behavior over more common adhesive joints such as SLJ and DLJ. Therefore, it is necessary to implement tools that allow the realization of such analyses. FEM techniques allowed the ability of numerically predicting a bonded joints’ behavior , disregarding the load conditions, joint geometry or adhesives’ specifications. This method has revealed itself as a very flexible tool, with very high precision that can be used in both simple and more intricate models. The most used modelling technique is CZM, originally introduced by Barenblatt (1959) and Dugdale (1960). Lubkin and Reissner (1956) extended the study of tubular adhesive joints to the case of axisymmetric modelling, and these authors are considered the first to develop analytical stress studies on tubular joints under axial load. Lubkin and Reissner deduced three different sources responsible for the stress concentrations in the adhesive: differential straining, bending introduced by the non-collinearity of the overlapping tubes, and edge effects. In their work, the adhesive is an elastic medium, in practice a spring layer, transmitting shear (longitudinal) and peel (radial) efforts. Both stresses are constant over the thickness of the adhesive and are only a function of the axial coordinate. Circumferential shear stresses are not considered because they would imply joint torsion. The investigation conducted by Esmaeel and Taheri (2011) aimed to study the delamination in the adherend layer of hybrid adhesively-bonded tubular joints under torsion, and its influence on the distribution of stresses within the adhesive layer. Two systems were tested as inner tube materials, a glass/epoxy composite and a stiffer graphite/epoxy composite. A comprehensive parametric study was carried out, using a FEM simulation performed in ABAQUS ® regarding the effect of geometric and loading parameters on σ y (peel) and τ xy stresses. Among the parameters considered were the location and length of delamination, the mechanical properties of the composite material and fiber orientation of the composite adherend. It was concluded that the trend in the distribution of stresses remained comparable for both composite systems, although the magnitudes were significantly different. The numerical and experimental study conducted by Hosseinzadeh et al. (2007) aimed to characterize, in a simplified way, the torsional performance of metallic tubular joints connected by structural adhesives, for different L O . The failure threshold of the adhesive, for various joint lengths, was characterized by using the Ramberg – Osgood plasticity model. This plasticity model was fine-tuned through comparison with the results of the FEM simulation, by application of a limited number of known parameters. It was concluded that the developed plasticity model could simulate the joints behavior with various lengths with good accuracy. The results showed that the strength capacity of the joint was highly dependent on its absorbed strain energy, i.e., as L O increased, the joint absorbed more energy, even after the joint went into a completely plastic mode. The numerical study published by Oh (2007) predicted the transmission capacity of torsional moments in tubular adhesive joints with composite adherends, combining thermal and mechanical analyses. Three criteria were considered for joint failure: interfacial failure, the cohesive failure of the adhesive layer, and adherend failure. The results of the study were compared with available experimental results (Choi and Lee 1995). It was possible to conclude that both the failure of the adhesive and the adherend occurs at low composite stacking angles, where the influence of residual thermal stresses is insignificant. Joints with high stacking angles fail at the interface between the adhesive and the adhesive due to high residual thermal stresses.