PSI - Issue 41
J.E.S.M. Silva et al. / Procedia Structural Integrity 41 (2022) 36–47 Silva et al. / Structural Integrity Procedia 00 (2019) 000 – 000
43
8
where x’ , y’ , and x’y’ are the software stress outputs, whilst is the complementary angle to , x , y , and xy are the transformed stresses. Fig. 7 shows y (a) and xy (b) stress distributions in the bond-line for the scarf joints as a function of . It was found that y stresses are much smaller in magnitude than xy stresses for the smaller values of . The difference between y and xy reduces as increases, when =45° both stress components have a proximal magnitude. The peak y stresses at the overlap edges, compared to the inner overlap, reduce by increasing due to the corresponding reduction of localized adherends strains at the scarf tips. Nonetheless, the peak y / avg stresses increase from 0.89 ( =3.43º) to 1.23 ( =45º). On the other hand, xy stresses keep an identical behaviour along the bond length, without the marked peaks at the overlap edges, as found for y stresses, although the minor peak xy stresses observed at the overlap edges also tend to become more uniform with the increase of (Campilho et al., 2007). Actually, considering the scarf joint with =45°, xy stresses are practically constant. The peak xy / avg stresses increased from 1.04 ( =30º) to 1.13º ( =3.43º). This behaviour contrasts with overlap joints in which larger peak xy stresses are found at the overlap edges due to significant shear-lag effects (Nunes et al., 2016). Overall, these results indicate that smaller have an improved behaviour because of the reduction of normalized y stresses, although minor τ stress gradients appear for small . Furthermore, smaller also corresponds to exponentially increasing the bond-line length ( LS ) which leads to a larger P m (Campilho et al., 2011a).
a)
b)
Fig. 7 – Normalized y (a) and xy stresses (b) in the adhesive layer.
4.3. Damage analysis
Former to the strength prediction, damage assessment in the adhesive layer at P m is performed to serve as a link between the stress distributions and the P m - behaviour of the tubular joints. The stiffness degradation (SDEG) variable was used for this purpose, which varies between 0 (undamaged material; any point in the linear-elastic portion of the CZM law) and 1 (CZM element failure). To obtain the SDEG data, the model state was approximated to the analysis increment closest to P m . Fig. 8 plots the SDEG variable along x / L O in the tubular joints with different values of . Regardless of , the damage is mostly concentrated at the adhesive layer’s edges, with a higher incidence for smaller , which agrees with the stress concentrations regions visible in Fig. 8. Damage is mostly nil at the inner overlap. Although SDEG peaks at both bond edges, there is a slight preponderance in the edge at x / L O =1 due to being the tip with the smaller tubular diameter. Smaller tend to increase the peak SDEG at the bond edges. However, in the tubular joints with =3.43°, the adherends were plasticized. As a result, the highest magnitude of damage occurred in the joints with =10° (SDEG=0.53). Moreover, smaller concentrate the damage over a less significant area at the bond edges.
Made with FlippingBook - Online magazine maker