PSI - Issue 72

H.G.E. da Silva et al. / Procedia Structural Integrity 72 (2025) 26–33

29

a)

b)

Fig. 2. Compression test (a) and 3PB test setup (b).

2.3. Numerical modelling The two-dimensional specimens were modelled as a deformable solid, while the supports and punch were modelled as rigid discrete elements. The specimen was divided into five partitions, representing the two skins, the two layers of adhesive and the core. The skins were also partitioned into eight plies to individually model each ply orientation. For the laminate materials, elastic behavior was used. Autodesk’s Helius C omposite program was used to determine the engineering constants and the rupture data for the Fail stress sub-option. The adhesive was created as a cohesive layer with a triangular traction-separation law. The core and the carbon and glass fibers were created as homogeneous solids. A static analysis was carried out with geometric non-linearities. The interaction between components was defined to guarantee support without friction. A surface-to-surface interaction was created between the punches and the skin faces. The supports were assumed to be clamped. The loading punches were assumed to have a vertical displacement and restricted movement in horizontal axis. The skins and core were assigned a type of element from the plane strain family, labelled CPE4R. The adhesive was assigned cohesive elements with the designation COH2D4. To optimize the mesh, bias effects were used in the models to provide a finer mesh at critical regions (adhesive layer and in the punch contacting regions with the specimen). The adhesive was modelled using cohesive elements with a triangular law, including the quadratic stress (QUADS) damage initiation criterion, and a linear energy criterion to assess failure of the elements. The plastic behavior of the core was of the crushable foam type with isotropic hardening, using compression ratio values of 1 and plastic Poisson of 0. The Tsai-Wu criterion [2], in plane stress, was employed to predict ply failure. To apply the model, tensile ( σ ut1 and σ ut3 ) and compressive ( σ uc1 and σ uc3 ) strengths, both in fiber and transverse direction, as well as shear strength ( τ u13 ) values, were taken from the manufacturer’s information. 3. Results 3.1. Experimental results The area ( A ), maximum load ( P m ), compressive strength (  ) and modulus of the foam ( E c ), corresponding to the first maximum, and the elastic slope ( S ), are shown in Table 4.

Table 4. Compression test results.

Specimen

A [mm 2 ]

m [N]

s [MPa]

S [N/mm]

E c [MPa]

P

1 2 3

1481 1485 1501

1847 1906 1889

1.248 1.284 1.258 1.263 0.019

5181 5418 5480

34.96 36.49 36.36 35.94 0.85

Average value Std. deviation

- -

- -

- -

Table 5 presents the summary of 3PB results for configurations 1 and 2 (  is the shear stress in the core,  f is the flexure stress in the skins, and s 1 is the elastic stiffness of the specimen). The specimens in configuration 2 have in general lower deviation values than the specimens in configuration 1. It can be concluded that, by carrying out the 3PB test, configuration 2 has marginally better properties than configuration 1. Thus, it can be concluded that the orientation of the glass fiber plies is of little significance in the overall behavior of the structure.