Issue 31

R.D.S.G. Campilho et alii, Frattura ed Integrità Strutturale, 31 (2015) 1-12; DOI: 10.3221/IGF-ESIS.31.01 10 Configuration 3 Following the method described in J-integral technique to measure G n c section, G n c was calculated identically to the previous cases, which considered  o instead of  p to obtain G n . The aforementioned method was applied to all tested specimens and the G n c results for each t A value and respective deviation are presented in Fig. 11. For the specimens with t A =0.1 mm, the obtained results gave G n c =1.83±0.24 N/mm. The increase of G n c from this point was of 14.5% ( t A =0.2 mm), 57.8% ( t A =0.5 mm), 105.6% ( t A =1.0 mm) and 195.9% ( t A =2.0 mm). 0 1 2 3 4 5 6 0 0.5 1 1.5 2 2.5 G n c [N/mm] t A [mm] Figure 11 : Average values and deviation of G n c as a function of t A by the J -integral. Regarding the available studies (for epoxy adhesives), Yan et al. [30] studied the influence of t A on the fracture properties ( G n c ) of DCB and Compact Tension (CT) joints with aluminium adherends and a rubber-modified epoxy adhesive. Using a large deformation Finite Element technique and the peak loads measured in the experiments, the critical value of the J - integral was calculated for different values of t A . A G n c increase was found up to t A =1 mm and a decrease afterwards. An identical conclusion was found by Khoo and Kim [31] for an epoxy adhesive between 0.2< t A <1.5 mm, with the maximum G n c being found for t A =1 mm. The increasing trend obtained in this work of G n c with t A is linear up to t A =2.0 mm, and this result is consistent with previous studies in this matter, except from a reduction of G n c for big values of t A that is common with less ductile epoxy adhesives. Another exception is the work of Marzi et al. [32], which attained a maximum G n c between t A =1 and 2 mm for the polyurethane SikaPower 498TM, a modern crash resistant epoxy adhesive, without a reduction tendency of G n c up to t A =2 mm, due to its large ductility. An identical trend to this work regarding the G n c - t A law was found by Banea et al. [33] with the high elongation polyurethane adhesive Sikaforce ® 7888, characterized with conventional fracture methods in the range of 0.2≤ t A ≤2 mm. In both this and the present work, the peak value of G n c is attained for a t A value bigger than 2 mm, but in this range of values the joints are more likely to have fabrication defects, and be more difficult to fabricate, which justifies its limited industrial applicability. Discussion of results The proposed technique, applied to the 3 joint configurations, showed that the proposed J -integral methodology can be a valuable tool to estimate G n c of adhesive joints. Moreover, with the measurement of  n , the cohesive law of the adhesive layer can be obtained as well. By analyzing the obtained results between the three tested configurations, a direct analogy cannot be formed between configuration 3 and configurations 1 and 2, because a different adhesive was considered (although both tested adhesives are ductile polyurethanes). In the comparison between configurations 1 and 2, it should be noted that, as depicted in Tab. 2, the value of h varied. This has a significant influence on the plastic zone size and, thus, also on the G n c measurements. The measured data for the bonded joint of configuration 1 gave G n c =1.182±0.215 N/mm ( h =5 mm), while for configuration 2 and h =4 mm the value of G n c =1.095±0.195 N/mm was obtained. These values agree quite well, although the difference in bending stiffness of the adherends has to be considered: the values of h between these two configuration differ and, additionally, the value of E of the adherends for configuration 2 (aluminium) is much higher than that of configuration 1 (jute-epoxy composite). Since the results of configuration 2 show that, for aluminium adherends, for h values above 3 mm the plastic zone effect ceases to affect the results, the G n c measurement theoretically should be identical between configurations 1 and 2. In view of this discussion, the 7.4% different between these two configurations is attributed to experimental scatter and related issues.