PSI - Issue 51

P.A.R. Ferreira et al. / Procedia Structural Integrity 51 (2023) 115–121 P.A.R. Ferreira et al. / Structural Integrity Procedia 00 (2022) 000–000

118

4

located between the 1 st and 2 nd plies next to the adhesive layer; these interlaminar layers had a thickness of 0.02 mm, as shown by Correia et al. (2020). A triangular cohesive law was employed for both adhesive and interlaminar plies, as proven suitable in previous research (Campilho et al., 2013). Regarding the mesh density, the geometry was partitioned to allow for a single row of cohesive elements in the adhesive layer, the same was done for the interlaminar layers; therefore, the element sizes for these regions were 0.2 mm and 0.02 mm, respectively. The element size was biased to reduce the computational cost, in this case, the maximum element size was 1 mm in those locations far from the adhesive and interlaminar layers. In order to obtain stress distributions in the bond line, a second set of models was created and meshed only with solid elements (CPE4). In this case, the element size was smaller, 0.02 mm by 0.02 mm in the adhesive layer. The mesh in the other regions was adjusted accordingly, allowing to reduce computational costs. Experimental testing of this type of joint under internal pressure presents several challenges; therefore, it was decided to validate the numerical technique with the SLJ configuration, which is similar to the JLJ, as proposed by Liu et al. (2018). Therefore, the numerical technique was validated by numerically reproducing the experimental tests performed by Nunes et al. (2016). 3. Results Joint strength was obtained first for the joints with an internal radius R = 1000 mm as a reference. It was found that P max increases with L O ; however, such behavior is also influenced by t P , as shown in Fig. 2. Starting with the joints with the thinnest substrates, i.e., t P = 1.2 mm, P max increases approximately linearly until reaching an L O = 40 mm; then, the value of P max stabilizes to around 11 kN for the larger values of L O (Fig. 2). Considering the shortest L O as a reference, the maximum strength is observed in the L O =50 mm case, being 173.35% higher than the reference. Upon increasing t P to 2.4 mm, the effect of L O on P max is smaller than the one observed with the thinnest substrates, having an abrupt increase up to L O =20 mm and then increasing gradually until L O =80 mm. The maximum strength was achieved on the joint with L O =80 mm, 54.47% higher than the reference. A similar effect was observed on those joints with the thickest substrates. Again, the maximum strength was achieved on the joint with L O =80 mm, 62.47% higher than the reference. Although the increase in L O led to stronger joints, the joints with the thinnest substrates performed better than their thicker counterparts. Furthermore, a L O around 40 mm to 50 mm provide the strongest joints for this joint configuration and material combination.

Fig. 2. Joint strength ( P max ) as a function of overlap length ( L O ) and substrate thickness ( t P ). Data correspond to the cases with an internal radius ( R ) of 1000 mm.

Regarding R and its effect on P max , the differences between the three radii evaluated are relatively small, as shown in Fig. 3. Increasing R to 2000 mm, the joints with the thinnest substrates presented the larger effect, being the largest on the joint with t P =1.2 mm and L O =20 mm, being 4.89% and followed by the L O =30 mm with 4.78%. In the remaining L O , the differences remained between ±1% except for L O =80 mm where P max was reduced by 3.44% and L O =10 mm