PSI - Issue 83

R.J.B. Rocha et al. / Procedia Structural Integrity 83 (2026) 179–186

182

a)

b)

Fig. 1. Geometry and dimensions of the SLJ (a) and SJ (b) under 4PB loading.

The distance between the upper loading pins is fixed at l =30 mm, while the lower support pins are separated by L =170 mm. The upper pins have a radius of 2.5 mm, whereas the lower pins feature a larger radius of 15 mm to provide stable support conditions. The adhesive layer thickness is t A =0.2 mm, the adherend thickness t P =5 mm, and the joint width B =25 mm. All elements of the adhesive joints were manufactured and prepared following a controlled procedure. The adherends were produced via AM according to the parameters listed in Table 1, with a total fabrication time of approximately 1 h and 45 min per joint. The resulting surface irregularities were mitigated through mechanical abrasion using 80-grit sandpaper, followed by solvent cleaning to eliminate debris and surface contaminants. To impose a uniform t A =0.2 mm, two copper spacers with a diameter of 0.2 mm were positioned between the adherends prior to joint assembly. Manual application of the adhesive resulted in material overflow along the overlap edges. This excess was removed after curing using a rotary tool equipped with a grinding wheel to restore a smooth surface profile. Mechanical testing of the bonded joints was performed using the same universal testing machine employed for bulk material characterization, with a crosshead speed of 20 mm/min. To enable 4PB loading, custom spacers were manufactured by AM and bonded to the lower adherend, as illustrated in Fig. 2. For each joint configuration, four specimens were tested until failure or until a clear loss of load-carrying capacity was observed. Fig. 2 presents the experimental 4PB setup and a representative ABS SJ with L O =10 mm and 7752 during testing.

Fig. 2. 4PB test configuration of an ABS SJ with L O =10 mm and 7752.

2.3. Numerical modelling Two-dimensional (2D) analyses were conducted using Abaqus ® , taking advantage of its CZM crack growth module. Geometric nonlinearity was included in the simulations to account for the significant deformations under 4PB. The adherends were represented using homogenized material properties derived from tensile testing of ASTM D638 Type I specimens, as reported in Table 2. Plastic deformation was represented using a perfectly plastic model. The adhesive was discretized using triangular cohesive elements governed by a triangular traction-separation law (Rocha and Campilho 2018). Damage initiation was defined through a mixed-mode quadratic nominal stress (QUADS) criterion, expressed in terms of the normal and shear cohesive strengths ( t n 0 and t s 0 , respectively), taken equal to the corresponding material strength limits. Damage propagation was modelled using a linear energy-based evolution law, requiring specification of G IC and G IIC . All cohesive parameters employed in the simulations are listed