PSI - Issue 83

R.J.B. Rocha et al. / Procedia Structural Integrity 83 (2026) 187–195

190

The FDM printing time for each joint (two adherends) was approximately 1 h and 50 min. During SLJ printing, shims were incorporated at the adherend edges to ensure proper testing alignment due to adherends’ misalignment. Once printed, the bonding regions were mechanically prepared through sanding and cleaning. Surface abrasion was performed using 80-grit sandpaper, followed by solvent cleaning: isopropyl alcohol was applied to ABS specimens, whereas acetone was used for PLA and PETG samples. Prior to adhesive application, two copper wires with a diameter of 0.2 mm were positioned within the overlap region to control t A . During curing, clamps were placed on the unbonded portions of the specimens to assure the specimens’ alignment. All joints were cured under ambient conditions. Due to the manual application of the adhesive, material overflow at the overlap edges was frequently observed. This excess adhesive was removed after curing, as it represents an unintended geometric deviation from the idealized configuration shown in Fig. 1 and could otherwise compromise the consistency between experimental specimens and numerical models. Mechanical testing was carried out using the same testing machine employed for the bulk material characterization, with a crosshead speed of 1 mm/min. The displacement data for the load-displacement ( P -  ) curves were approximated using the crosshead displacement. For each testing condition, four specimens were evaluated. 2.3. Numerical conditions The static numerical simulations were performed using Abaqus ® . Geometric nonlinearity was considered, as the joints (mostly the SLJ) experience large rotations at the overlap region. All analyses were conducted using two dimensional (2D) models. The adherends were modelled with solid elements, with the plastic response defined through the relationship between stress and plastic strain. To reduce computational complexity, a perfectly plastic material model was adopted once the yield stress was reached. The adhesive was represented using triangular cohesive elements (Rocha and Campilho 2018). In addition to the longitudinal and shear stiffness components ( E and G , respectively), damage initiation was governed by a quadratic nominal stress criterion (QUADS), defined by the maximum normal and shear CZM law strengths ( t n 0 and t s 0 , respectively), corresponding to the tensile and shear strengths ( σ f and τ f , respectively, defined in Table 2). Damage evolution under mixed-mode loading was described using a linear energetic criterion based on the critical fracture energies in mode I and mode II ( G IC and G IIC , respectively). Boundary conditions were defined to reproduce the experimental testing setup. The left end of the specimen was fully constrained, while a prescribed horizontal  of 5 mm was imposed at the right end, with the vertical displacement simultaneously restricted. The mesh consisted of quadrilateral elements. As illustrated in Fig. 2 (a) for SLJ and (b) for SJ, mesh refinement was introduced near the overlap edges, where stress concentrations are known to occur, through the application of biasing strategies. The adhesive layer was discretized using a single element through the thickness, with its stiffness calibrated to represent the full adhesive layer response for the prescribed t A .

Fig. 2. Mesh of an SLJ (a) and SJ (b) with L O =10 mm.