PSI - Issue 53

S. Leonardi et al. / Procedia Structural Integrity 53 (2024) 327–337

329

S. Leonardi et al. / Structural Integrity Procedia 00 (2023) 000–000

3

Fig. 1. Strategy used to design the metallic cellular architectures of this work. The latter contain through-thickness porous inclusions heteroge neously dispersed into a dense matrix. To design such architectures a four-step protocol is used. This protocol begins with the generation of the material architecture by means of the RSA algorithm and ends with the LPBF fabrication.

consists in placing, randomly and sequentially, non-overlapping inclusions of quadratic shape in a solid cell with ar bitrary geometry. For convenience, here we focus on circular inclusions embedded into a squared cell. Our prior work highlights the versatility of the algorithm that has been employed to generate a variety of particulate architectures, ex tending from closed-cell foams with spherical voids (Zerhouni et al. (2018, 2019); Tarantino et al. (2019)) to open-cell cellular solids with arbitrary geometry (Hooshmand-Ahoor et al. (2022)). It is also noted in passing that RSA algo rithm yields random particulate architectures whereby the particle inclusions can be of any constituent phase, namely either solid or air. Here, the inclusions are designed to be voided particles. This is achieved during the second step of the design protocol, which consists in converting the RSA-generated particulate architecture into a FE numerical model by means of the open source meshing software Gmsh . The FE model is then extruded and converted into STL format for LPBF additive manufacturing, see steps 3 and 4 in Figure 1. In this work, four random porous architectures were designed and studied. These are reported in Figure 2 and contain an increasing content of porosity comprised between 20 and 60 %. Each porous architecture was designed to contain 200 pores with circular cross-section embedded into a squared solid domain, see Figure 2. Specifically, pores are equisized for architectures containing up to 40 % porosity, whereas polydisperse distributions of pores are generated by the RSA algorithm to achieve the highest porosity, i.e. 60 %. This is consistent with our earlier work (Tarantino et al. (2019)). Likewise, each generated porous structure is designed to be periodic - namely all voids intersecting the cell edges are periodically reproduced on opposite edges of the cell. This in turn allows the study of these materials by means of computational homogenization (Zerhouni et al. (2019); Tarantino et al. (2019)). A detailed summary of the pore features of the RSA-generated porous architectures in Figure 2 is given in Table 1. In the interest of presentation, the designation N-Y is used to denote the numerical porous model, where N stands for numerical and Y indicate the total porosity expressed in %. Finally, it is noted in passing that the choice of

Fig. 2. Random porous architectures designed and explored in this study. The latter are numerically generated by means of the RSA algorithm and are designed to contain between 20 and 60 pct of porosity. Each porous geometry contains 200 circular pores randomly dispersed into a square cell, whereby the pores are extruded throughout the thickness prior to fabrication.