Automated conformal mesh generation chain for woven composites based on CT-scan images with low contrasts

A novel approach to generate conformal meshes of woven composites at the mesoscopic scale from low-contrast CT-scan images with reconstruction artifacts is presented. The segmentation of the CT-scans relies on texture analysis of the tomography to differentiate the orientations of warp and weft yarns. The yarn paths are tracked through the scanned volume starting from the relative positions of the yarns on the external faces of the scanned volume. These paths provide seeds for a growing algorithm that labels the yarns in the initial voxelized mesh. The final conformal mesh is generated using smoothing and coarsening algorithms. The presented approach is mostly automated and allows generating quality controlled meshes suitable for finite element analyses with limited user intervention. Its flexibility is demonstrated by an application to a woven composite with manufacturing defects (fiber disorientation and yarn splitting).