PSI - Issue 80

5

D.C. Gonçalves et al. / Procedia Structural Integrity 80 (2026) 443–450 Author name / Structural Integrity Procedia 00 (2019) 000–000

447

3. Results Double Cantilever Beam (DCB) specimens are analyzed in this work. The geometry and boundary conditions considered are presented in Fig. 2. The adherends are aluminium AA6082 T51, with elastic modulus = 70.07GPa and Poisson’s ratio =0.33 , the unique material properties required in the present analyses. The adhesive material used was the strong and brittle adhesive Araldite AV138 with properties = 4.89GPa and =0.3 . To assess convergence of the NNRPIM in this application two nodal discretizations were considered with 4200, and 10675 field nodes, shown in Fig. 3.

Fig. 2. DCB Joint dimensions and boundary conditions

(a)

(b)

(c)

(d)

Fig. 3. Nodal discretizations: 4200 nodes model (a), 10675 nodes model (b), and close-ups on the crack tip region respectively (c,d)

Fig. 4 shows the load-displacement curves for nodal discretization 1 (Fig. 4a), and nodal discretization 2 and considering % (Fig. 4b), ( (Fig. 4c), and 9 (Fig. 4d) in the J integral calculation. The results are compared against experimental data obtained in previous works considering the same DCB specimens and materials. The numerical analysis predicts the experimental behaviour of the adhesive joints within the complete fracture propagation process. Experimental and numerical maximum test loads, with the respective percentual errors, are presented in Table 1. It is noticeable that nodal discretization 1 predicts ':; with smaller percentual error, demonstrating the efficiency of the NNRPIM in analyzing complex adhesive joint problems with relatively sparse nodal discretizations. Significantly increasing the nodal density does not significantly affect the percentual error and the method’s performance, proving the documented high rates of convergence of the NNRPIM (Belinha, 2014). Moreover, varying the radius of the contour domain results in a discrepancy of 1.3% in the ':; prediction. Such small variances are expected to the theoretical independence of the J integral from the radius of the contour selected.