PSI - Issue 47

Rachid Bensaada et al. / Procedia Structural Integrity 47 (2023) 503–512 R. Bensaada et al. / Structural Integrity Procedia 00 (2019) 000 – 000

509

7

The results in terms of iso-values are given by Fig. 6 a) for the failure of fibers and Fig. 6b) for the failure of the matrix by using the Hashin damage model. These figures show that the damage is fully achieved in both cases. The results obtained in terms of load - displacement curve from finite element analysis is superimposed to the experimental result as illustrated by Fig. 7 . A good reproducibility of the experimental results is obtained by finite element analysis. The Hashin damage model is efficient to capture strain localization and the resulting damage. This step was necessary for the comprehension of the local phenomena of composite damage and could help to complete the out-of-plane testing of the composite and its modelling in this framework. 3.2 Adhesive testing and numerical validation The results obtained under pure tensile (0°) loading of the adhesive show a quasi-brittle response of this later. Its behavior could be modeled by a bi-linear cohesive zone model. Its complete formulation is given in (Bensaada et al. 2018) and the identified parameters are given in Table 3. This Arcan test is 3D fully simulated by considering the beaks (Fig. 8a), The cohesive zone meshed by COH3D8 elements, while the substrate surface is meshed by C3D8R elements and beaks zone by C3D10 elements.Damage evolutionwithin the cohesive layer that represents the adhesive is given in Fig. 8b. The comparison between experimental results obtained by DIC post-treatment and numerical simulation is given in Fig. 9.

Table 3. Cohesive zone model parameters Ts (MPa) G f (kJ/m²)

Lc (mm)

7.25

3.75

0.4

a)

b)

Fig. 8. a) Finite element model, b) Damage evolution within cohesive elements

Fig. 9. Tensile response of Permabond TA4610 vs.CZM finite element simulation