PSI - Issue 80

R. Salem et al. / Procedia Structural Integrity 80 (2026) 256–268 Rania Salem/ Structural Integrity Procedia 00 (2019) 000 – 000

258

3

property differences due to FFF's printing parameters (e.g., layer height, raster width, infill pattern), violating scaling laws and continuum theories (Ramezani Dana et al., 2019). FE modeling requires complex mesoscale approaches to capture raster paths, inter-layer bonding, and voids. In fact, case studies confirm large data discrepancies in ultimate tensile strength and strain using current standards (Sola et al., 2023). Tensile variability stems fundamentally from porosity, which critically defines mechanical performance (Hebda et al., 2019). Void characteristics (size, shape, distribution) reduce modulus and strength via stress concentration and weakened adhesion; larger voids in complex geometries cause disproportionate degradation, with significant performance gains from void reduction (Gonabadi et al., 2024). Anisotropic void distribution also weakens out-of plane properties more severely. FEA-RVE modeling effectively correlates void morphology with property reductions (Tao et al., 2021). Optimizing printing parameters (e.g., layer thickness) minimizes voids and enhances bonding, improving strength (Kumar et al., 2023). Bonding relies on thermal energy, with strength critically dependent on interfacial thermal history (Sun et al., 2008). While high temperatures promote diffusion, cooling induces detrimental residual stresses (Bellehumeur et al., 2004). Consequently, multiscale homogenization is crucial for understanding porous fused deposition modeling or material extrusion additive manufacturing materials (Anoop & Senthil, 2019; Sharafi et al., 2022). Methods include using microscale RVEs based on void geometry to homogenize elastic properties via volume averaging, explicitly considering void influence at raster interfaces (Anoop & Senthil, 2019), and employing X-ray tomography-derived 3D RVEs capturing mesostructural defects for numerical homogenization. The latter specifically links defects/voids (as interfaces) to stress concentrations governing material resistance and crack initiation (Paux et al., 2023), collectively emphasizing how void structure irregularities fundamentally control mechanical response. This paper presents a numerical investigation into the effect of porosity size on the elastic properties of FFF 3D printed lattice struts. The study focuses on struts with small cross-sections (typically ~1 to 4 mm²), where filaments are aligned parallel to the strut axis. In this context, experimental analyses have shown that stiffness can depend significantly on the cross-sectional size itself, independent of the overall porosity volume fraction. To investigate this size-dependent behavior, the work models a square pile-up of filaments, capturing both the void regions within the pile-up and the filament intersection zones. A representative volume element (RVE) is developed with cylindrical filaments and idealized pores. The elastic response under uniaxial tensile loading is computed numerically and compared against classical bounds (Voigt and Mori – Tanaka) to assess the role of pore size and filament connectivity in determining the effective stiffness of the structure. A brief literature review contextualizes the importance of these mesoscale features in the mechanics of additively manufactured materials. This section presents a numerical investigation of the mechanical behavior of FFF parts through a multiscale finite element framework. At the microscale, representative volume elements (RVEs) are constructed based on void morphologies and interfacial geometries reported in the literature (Sahoo et al., 2024; Sola et al., 2023; Paux et al., 2023), which originate from high-resolution imaging techniques such as scanning electron microscopy (SEM) and X ray micro-computed tomography (micro-CT). An example of void morphologies as obtained in an ABS strut of section dimensions 1x4 mm 2 , printed with 3 adjacent filaments in each layer, is provided in Fig. 1 . The 2D image ( Fig. 1a ) has been obtained by means of X-Ray micro-tomography using an EasyTom 150 tomograph of RX Solutions, France, and the software XAct (RX Solutions). It illustrates the presence of voids at the intersection between filaments and layers. The 3D image ( Fig. 1b ) has been obtained by 3D reconstruction from the 2D CT-scans using the software Slicer (slicer.org). This 3D image further shows that these voids are present all along the deposition direction. 2. Multiscale Analysis of Porosity and Bonding Effects in FFF Parts