X-ray CT based multi-layer unit cell modeling of carbon fiber-reinforced textile composites: Segmentation, meshing and elastic property homogenization. (15th October 2022)

• Record Type: Journal Article
• Authors: Sinchuk, Yuriy; Shishkina, Oxana; Gueguen, Mikael; Signor, Loic; Nadot-Martin, Carole; Trumel, Herve; Van Paepegem, Wim
• Is Part Of: Composite structures. Volume 298(2022)
• Publication Date: 2022-10-15
• Subjects: Material modeling -- X-ray computed tomography -- Textile composites -- Finite element analysis (FEA) -- Deep learning; Composite construction -- Periodicals; Composites -- Périodiques; 624.18
• DOI: 10.1016/j.compstruct.2022.116003
• Languages: English
• ISSNs: 0263-8223

Abstract

Graphical abstract: Abstract: Generation of realistic finite element method (FEM) geometry of a textile composite material at tow scale remains a challenging stage of material modeling. In this paper, a FE model generation approach is introduced, based on micro-Computed Tomography (µ-CT) images of a carbon fiber reinforced textile composite with multiple layer orientations. For the specified laminate layup made of plain weave fabric layers, a tow instance segmentation method based on deep learning was proposed. The fiber local orientation field was extracted from the tow centerlines approximation. For the image-based finite element meshing, a material interface reconstruction algorithm was proposed. The algorithm leads to a conformal and smooth finite element (FE) mesh with high-quality elements. FE models with voxel and tetrahedral meshes considering different approximations of the material orientation field were generated and used for numerical homogenization of the composite material's elastic properties.