PSI - Issue 33

Umberto De Maio et al. / Procedia Structural Integrity 33 (2021) 954–965 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

960

7

The Young’s modulus and the Poisson’s ratio of the UHSC concrete are equal to 52,837 MPa and 0.18, respectively, whereas, the inelastic parameters required by the adopted traction-separation laws, described in Section 2.1 and 2.2 for DIM and ECM respectively, are listed in Table 1.

Table 1. Material parameters for the traction-separation law used in the wedge splitting test. max  [MPa] max  [MPa] Ic G [N/m] IIc G [N/m] 0 n K [N/mm 3 ] 0 s K [N/mm 3 ] DIM 8.6 8.6 128.61 128.61 7.0507E14 3.751E14 ECM 8.6 - 128.61 - - -

In particular, to reduce the computational effort of the analysis the cohesive interface and the breakable solid elements, used in the DIM and ECM, respectively, to simulate the crack nucleation and propagation, are located only in a rectangular region, called damage zone (see Fig. 3a), ahead the pre-crack. In this zone the following mesh topologies have been considered: (i) a free-triangular Delaunay tessellation for the DIM (Fig. 3b), consisting of three node triangular elements for the bulk and four-node zero thickness elements for the embedded interfaces, and (ii) a mapped mesh for the ECM (Fig. 3c), composed by eight-node constant-stress user-defined solid elements enriched with the embedded crack formulation. In the remaining part of the analyzed specimen a free quadrilateral mesh is employed for both models. A Newton-Raphson solution algorithm with quasistatic displacement-controlled loading conditions, has been used to conduct the following DIM analyses, whereas for the ECM simulations, a finite element solver with explicit time integration has been used, adopting a suitable loading rate to prevent the appearance of appreciable inertial effects.

Fig. 4. Comparison between the load versus CMOD curves obtained by the DIM and ECM simulations.

The comparison between the structural responses, in terms of load versus crack mouth opening displacement (CMOD) curves, predicted by the adopted models and by a reference model realized for comparison purposes, are reported in Fig. 4. Such a reference model, already used in (De Maio et al., 2020b), whose validity is assured by the prescribed symmetric conditions, consists in a set of cohesive elements inserted along the vertical line coinciding with the self-similar growth direction for the preexisting crack. The structural response predicted by the DIM approach is slightly stronger respect to the reference numerical model at both peak and post-peak stages due to the artificial toughening effect induced by the used unstructured mesh. On the other hand, the loading curve predicted by the ECM