PSI - Issue 79

Osman Bayrak et al. / Procedia Structural Integrity 79 (2026) 413–420

417

3. Results and discussion Interfacial interaction was defined with the glue contact at the beginning of the porosity induction process. At the end of the process, some of the opposing nodes in the interfaces were debonded from each other, leaving a porosity effect in the interfaces (see Figure 3). The same process was applied, and the same porosity effect was observed for both tensile and bending test models. As stated in the modelling details section, there is no statistical data in literature regarding the dispersion of the pores. Also, this study is not concerned with converging fracture toughness values reported in experimental studies in literature but concerned with simulating the reported fracture toughening mechanisms with the effect of dispersions of GNPs and of the porosities that sits in the interfaces. Therefore, dispersion pattern of pores in this study was deemed acceptable with respect to the microstructural studies by (Bódis et al., 2019; Michálková et al., 2014).

Figure 3 Porosity configuration in the interfaces before (a) and after (b) the porosity induction process. All interfaces were fully bonded (red and yellow colour lines) before the process. After a uniaxial loading and an unloading, some of the contacts were deactivated, leading to sparse debondings. Therefore, porosities were introduced.

As discussed in the introduction section, some toughening mechanisms are commonly reported in literature; crack deflection, crack branching, crack bridging, crack pull-out, and interface separation (see Introduction section for the references). (Bódis et al., 2019) stated that sparsely located pores are commonly found at the interfaces, promoting the GNP pull-out. Also, (Cui et al., 2022) argued that the GNP pull-out is the most “dominant” fracture toughening mechanism; crack bridging and interface separation follow. While the existing literature on the field of numerical modelling of graphene-Si3N4 nanocomposites (Chen et al., 2021, 2022) successfully simulate the crack bridging and crack deflection mechanisms, there was still a need for the other critical mechanisms to be simulated. Post-processing of the FE analyses of the tensile and bending test models in this study showed that all major toughening mechanisms can be successfully simulated with microstructure-informed FE models. Results of the FE