PSI - Issue 61

G.J.C. Pinheiro et al. / Procedia Structural Integrity 61 (2024) 71–78 Pinheiro et al. / Structural Integrity Procedia 00 (2019) 000 – 000

76 6

3.2. Effect of the damage initiation criterion In all models, crack initiation took place at both adhesive edges and propagated towards the central region up to failure. The numerical P m , determined with six damage initiation criteria are compared with the experimental ones. Fig. 4 shows the evolution of P m with  , for the different damage initiation criteria and the different adhesives.

30

30

25

25

20

20

15

15

P m [kN]

P m [kN]

10

10

5

5

0

0

0

10

20

30

40

50

0

10

20

30

40

50

 [ ]

 [ ]

QUADS MAXE

MAXS MAXPE

MAXPS

QUADE

QUADS MAXE

MAXS MAXPE

MAXPS

QUADE

a)

b)

Exp

Exp

30

25

20

15

P m [kN]

10

5

0

0

10

20

30

40

50

 [ ]

QUADS MAXE

MAXS MAXPE

MAXPS

QUADE

c)

Exp

Fig. 4. P m comparison between different damage initiation criteria and experimental data: (a) AV138; (b) 2015; (c) 7752.

With the AV138, the numerical values obtained using QUADS initiation criterion overlapped the experimental ones, except for  =3.43°. The second criterion that comes closest to the experimental values is MAXS. Using the MAXPS criterion, P m are only relatively close to the experimental values between  =20° and 45°. For the QUADE and MAXE initiation criteria, both considerably exceed the experimental P máx values for all  except 3.43°. The inadequacy of these criteria is related with damage and failure being governed by the stresses of the adhesives rather than the deformations. For the MAXPE criterion, although closer than the last two criteria mentioned, the values still differ significantly. With the 2015, P m predicted by the QUADS initiation criterion practically overlap the experimental data. The MAXS criterion also performs relatively well for the majority of  . The MAXPS criterion is furthest away. Regarding the 7752, the values obtained using the QUADS criterion were very close to the experimental ones. For bigger  , the MAXS criterion is also very precise. Nonetheless, the values obtained using the MAXPS criterion show a discrepancy with the experimental values, which is more pronounced for smaller  . The QUADE, MAXE and MAXPE do not accurately predict the experimental values. Overall, the initiation criteria that perform best are QUADS and MAXS. In the case of 2015 and 7752 adhesives, the QUADS criterion results in numerical values extremely close to the experimental ones. 3.3. Effect of the damage propagation criterion To analyze the effect of different propagation laws (triangular and exponential), three values for  were also considered (0.5, 1 and 2). The initiation criterion used for this study was QUADS. Fig. 5 shows the P m comparison between the linear law and experimental data. For the AV138, when  =1 or 2, the numerical P m were very close to the experimental ones, with a maximum variation of 11.4% and 14.5%, respectively, for  =3.43°. However, if  =0.5, the numerical values differ quite significantly from the experimental ones, with a maximum deviation of -