PSI - Issue 47

J.E.S.M. Silva et al. / Procedia Structural Integrity 47 (2023) 70–79 Silva et al. / Structural Integrity Procedia 00 (2019) 000 – 000

77

8

4.3. Failure path and maximum load The TSJ failure modes are presented for the different adhesives and as a function of  . Table 2 presents the obtained failure modes. Most TSJ exhibited a cohesive failure, except for the joints with  =3.43°and bonded with the adhesives AV138 and 2015. In these joint configurations, adherend plasticisation occurred for P m , not allowing to extract the full potential of the adhesive.

Table 2. TSJ failure modes as a function of the adhesive type and  .

Adhesive

45°

30°

20°

15°

10°

3.43°

AV138

Cohesive Cohesive Cohesive

Cohesive Cohesive Cohesive

Cohesive Cohesive Cohesive

Cohesive Cohesive Cohesive

Cohesive Cohesive Cohesive

Adherend Adherend Cohesive

2015 7752

Fig. 8 shows the graphical representation of the equivalent plastic deformation (PEEQ variable in Abaqus ® ) for  =3.43°, at the instant of P m . It should be noted that the maximum PEEQ values were 0.27% for the AV138 adhesive and 0.28% for the 2015 adhesive. In these joints, the PEEQ gradient is notorious. On the other hand, in the TSJ bonded with the 7752, the adherends do not suffer plastic deformation when P m is reached. The same happens for the remaining simulated adhesive joints. This behaviour arises from the strength of the adhesive being smaller than the adherends’ strength. a)

b)

c)

Fig. 8. Graphic representation of PEEQ in the TSJ with  =3.43° for the adhesives: (a) AV138, (b) 2015 and (c) 7752.

Fig. 9 graphically represents the extent of damage of the CZM elements of the adhesive layer at P m (measured by the SDEG variable). This variable assesses the damage caused to the adhesive layer, ranging between 0 (absence of damage) and 1 (failure). For a brittle adhesive such as the AV138, the adhesive joint only shows damage at the bond edges (Fig. 9 a), as the adhesive does not allow stress spreading along the bond length, originating stress concentrations that lead to premature failure of the joint. On the other hand, the 7752 (Fig. 9 c), which is highly ductile and flexible in the elastic regime, allows for a more uniform stress distribution along the bond length. Consequently, at P m , the adhesive is damaged virtually over the entire length of overlap. The 2015 (Fig. 9 b) presents an intermediate behaviour between the other adhesives and, thus, minor damage is detected at the edges of the adhesive, simultaneously with an intermediate bond region with damage. Fig. 10 shows the P m evolution as a function of  for the TSJ bonded with the three adhesives, thus enabling the visual comparative assessment between adhesives. Analysis of Fig. 10 shows that the AV138 provides the best performance for  =45°. The increasing trend of P m by reducing  is verified up to  =3.43°, with a value of P m =35.93 kN. However, the percentile increase between limit  is the smallest among the tested adhesives, with a value of 268.48% from  =45° to 10°, corresponding to P m =22.49 kN. This behaviour is justified by large stiffness and brittleness of the adhesive, resulting in the little tolerance to peak  y stresses at the bond ends, which consequently leads to premature failure. On the other hand, as already mentioned, the value of P m =35.93 kN, obtained for  =3.43° does not serve as a basis for comparison since this value is limited by the adherend failure. The TSJ bonded with the 2015 presents, for  =45°, approximately half the P m value (3.21 kN) of the AV138, essentially due to the lower strength of the bulk material. Similarly to the AV138, P m is limited by the adherends ’ strength, giving P m =35.92 kN for  =3.43º. The percentile increase in P m between  =45º and 10° (322.24%) is considerably higher compared to the AV138, corresponding to P m =13.56 kN for  =10º. The flexibility