PSI - Issue 47

J.P.M. Lopes et al. / Procedia Structural Integrity 47 (2023) 48–55 Lopes et al. / Structural Integrity Procedia 00 (2019) 000 – 000

54

7

is also possible to verify a disruption of  xy / σ avg stresses in the radius zone, due to the stiffness variation in radius zone, with emphasis to higher l . On the other hand, by increasing l , it is also possible to verify the increase of  xy stress in x / l =0. This fact is a result of higher transverse deformations due to the bond area increase, from 10 to 40 mm, which causes a bigger stress concentration in the overlap tips of the adhesive layer.  xy / σ avg peak stresses are highest for the AV138, equally to σ y / σ avg peak stresses, due to the stiffness of this adhesive. In this case, maximum  xy / σ avg values were 56 and 133 for l =10 and 40 mm, respectively.  xy / σ avg peak stresses were smaller for the 2015 and smallest for the 7752, in which case these reached 21 and 55 for l =10 and 40 mm, respectively. 3.3. Maximum load The joint strength results are presented as a function of l using the three described structural adhesives. All the data result from the load-displacement ( P –  ) curves obtained with Abaqus ® . As it can be seen (Fig. 8), increasing the l values results in higher P m loads in ductile adhesives – 7752 and 2015. However, with the AV138, which is a stiff and brittle adhesive, P m practically do not change across the tested l . Due to the high stiffness, plasticization was not allowed in this adhesive. In this case, P m loads were 1473 N and 1471 N for l =10 and 40 mm, respectively (nil relative improvement). On the other hand, for the adhesive with the most ductility behavior, the 7752, P m were 3452 N and 5878 N for l =10 and 40 mm, respectively (relative improvement of 70.3%). In conclusion, adhesives with ductility behavior had the higher P m improvement values due to the capacity to resist the peak stresses appearing in the adhesive layer before failing, and behave better than stronger although more brittle adhesives in this particular application, in which the adhesive is mainly loaded in peel due to the applied loading.

8

6

4

P m [kN]

2

0

0

10

20

30

40

l [mm] AV138 2015 7752

Fig. 8. P m as a function of l .

3.4. Failure energy

The energy dissipation at failure ( U ) was addressed for the T -joints, considering the different l values and the three adhesives. This parameter was calculated with the information provided by the numerical P - d curves obtained in Abaqus ® , using the area below the respective P -  curve, which in turn was estimated by discrete integration functions in Excel ® . The l parameter in adhesive bonding is traditionally one of the most significant parameters for generic strength improvement due to increasing the area of adhesive resisting separation between the adherends, i.e., l values directly impact the amount of adhesive available to transmit the loads. However, the evolution is not linear because of the variations of stress concentrations with l , as previously shown. Fig. 9 collects all U data as a function of the l value for each adhesive. It is found that, for the strong but brittle adhesive (AV138), the U values are kept practically constant with l variation, at approximately 3 J. For such adhesives, increasing l does not reflect on higher P m nor displacements up to failure, leading to the visibly negligible U variation. On the other hand, for ductile adhesives, the l increase results in a significant P m difference, and also displacements to failure, which reflects on higher U . While the joints bonded with the 2015 showed an intermediate performance behavior between the three adhesives, the joints bonded with the 7752 gave the best results in the 10≤ l ≤40 mm range, with U improvements between 12 J and 27 J, corresponding to a relative difference of 125%.