PSI - Issue 47

J.P.M. Lopes et al. / Procedia Structural Integrity 47 (2023) 48–55 Lopes et al. / Structural Integrity Procedia 00 (2019) 000 – 000

49

2

nor welding (Petrie 2000). As any other technique, some drawbacks can be mentioned, namely surface treatment prior to adhesive application, reduced resistance to temperature and humidity gradients, and the requirement to design the joint in the direction of the elimination of peel (  y ) stresses. The most studied is the single-lap joint (Adams 2005). Additional joining strategies, less used and studied, are the double lap joint, scarf joint, stepped-lap, T -joints and tubular joints. T -joints, the design selected for the present research, is used in different types of industry, such as aviation, to bond stiffeners to the aircraft skin, and in the automotive industry for the pillar-to-rocker connection. A typical design of T -joint consists of a substrate joined by fillet and over laminates. The main goal of a T -joint is to transfer loads between the web and the base panel (Li et al. 1997). A multitude of strength prediction methodologies have been made available along the years, namely analytical or numerical approaches. Starting from 1938 with Volkersen (1938), 1944 with Goland and Reissner (1944) and later 1973 with Hart-Smith (1973), analytical models were proposed to capture the joint strength. However, with the development of novel ductile adhesives, analytical methods were no longer a suitable methodology to evaluate the joints ’ performance. Finite element method (FEM) models proved to be extremely adequate to overcome those complex analyses. Few approaches evolved, as the case of continuum mechanics, which uses the maximum values of stress, strain or strain energy predicted by a FEM analysis, or by analytical methods, and compares these values with the corresponding material properties to assess failure. On the other hand, fracture mechanics evolved to a different stage since it does not assume that the structure and its materials are continuous, therefore considering structural discontinuities. The concept of cohesive zone was proposed to describe damage, under static loads, at the cohesive process zone ahead of the apparent crack tip. Since then, CZM were improved and tested to simulate crack initiation and propagation even in composite delamination. CZM prediction accuracy is dependent on a precise estimation of the cohesive strengths in tension and in shear ( t n 0 and t s 0 , respectively), and the fracture toughness in mode I ( G IC ) and mode II ( G IIC ). Moreover, the extended finite element method (XFEM), proposed by Belytschko and Black (1999), surpasses one of the CZM method drawbacks, since it allows to simulate crack initiation and propagation along an arbitrary path. In addition, applying XFEM it is no longer necessary to define the mesh to match the geometry of the discontinuities or even remeshing near the crack. Different research works addressing T -joints can be found in the literature. Li et al. (2016) evaluated the strength of aluminum and steel T -joints bonded with the brittle adhesive Loctite ® ESP110 under cyclic temperature. CZM modelling and an environmental degradation factor (Deg) was implemented in Abaqus ® to simulate damage and the environmental degradation process of T -joints subjected to cyclic-temperature environments. The factor Deg could accurately evaluate and predict the degradation progression and respective ultimate load reduction of T -joints under cyclic temperature conditions. Barzegar et al. (2021) conducted a CZM numerical study on the effect of the adhesive type and geometrical parameters on the T -joint strength under bending loads. Adhesives with different ductility were used to bond carbon/epoxy (IM7/8552) and glass/epoxy (E‐glass/Epon 828) adherends. By comparing all adhesives, the strength performance was as follows (from best to worst): Araldite ® AV138, Araldite ® 2015, SikaForce ® 7752, Cytec FM73 and Cytec FM300. It was also found that joints with sharper corners present higher stress (within those regions) and lower at the adhesive free ends, whereas a rounded corner offers the opposite behavior. In addition, by increasing the stringer length, the von Mises stress at both adhesive free ends diminished while the stress near the corners increased. Finally, increasing the of the stringer’s thickness and web origins an increase of the overall von Mises stress distribution along the overlap area. The present work consists of a numerical analysis on the adhesive type effect in aluminum T -joints under peel loads, by CZM. CZM validation is previousl y accomplished. The joints’ assessment initiates with stress analysis, and prediction of P m and U , while geometrical modifications are also tested, to propose the best solution for the joint. 2. Methods 2.1. Geometry and dimensions Fig. 1 depicts the T -joint architecture and its general dimensions in mm, namely the specimens’ length L T =200, T part free length L A =40, specimens’ width B =25, adhesive thickness t A =0.2, base plate thickness a =3, T -part thickness t =1.5, and T -part radius r =6. To comply with the proposed parametric study, i.e., the effect of varying the value of overlap length ( l ), the following values were evaluated: 10, 20, 30 and 40 mm.