Prediction and examination of the impact of the raster angle on the orthotropic elastic response of 3D-printed objects using a novel homogenization strategy based on the real clustering of RVEs

Abstract

Abstract The distinct process of layer-by-layer 3D printing and the differential cooling of the 3D-printed component generates a low adhesive force between the layers, defects, and gaps interlayers, which results in an anisotropy distribution of mechanical and physical properties. Computational models, such as homogenization via green functions or homogenization based on ideal geometry-material models, are required as a first approximation of the elastic response of 3D-printed components, but have received less attention than experimental methods. This research establishes a novel multi-scale method for accurately predicting the mechanical behavior of 3D-printed components. At the micro and mesoscale, the micro-mechanical analysis of a representative volume element (RVE) corresponding to the real morphologies is acquired by clustering the microscopic observations of the internal structure of the 3D-printed samples using K-means algorithms, then used to generate micro-mechanical models in a function of the raster angle to compute the effective orthotropic constants, and these outputs are used to generate macro-scale numerical models that simulate the mechanical behavior under a tensile stress of the 3D-printed samples Compared to the idealized antecedent models and confirmed by experimental data, this methodology yields results that are consistent with reality. In conclusion, the incorporation of clustering approaches applying K-means algorithms into the homogenization procedure yields accurate prediction results that match the experimental elastic response of 3D-printed components. This study presents reliable prediction laws that enable the designer to conduct a faster iterative analysis and choose the optimal printing process parameters based on FE analysis to produce high-quality 3D FFF-printed components.