Séminaire : Alain Rassineux

Voxel models are a robust and effective method for accurately representing complex geometries at a high level of detail, even when other techniques such as surface representation may fall short. These models are typically obtained through segmented microcomputed tomography, and are particularly advantageous in finite element simulation due to their ability to provide conformal meshes of multiple volumes in contact. However, the jagged nature of voxel grids can introduce spurious stresses, particularly when studying phenomena such as damage. To effectively tackle this challenge, we propose an approach based on smoothed voxel-based technology, which allows for the creation of realistic representative volume elements of 3D woven composites suitable for mesoscale analysis. This approach allows for the creation of conformal tetrahedral meshes of the tows embedded in the matrix, while controlling fiber volume fraction. In a final step, the mesh can be coarsened to significantly reduce the computational cost of finite element analysis. This technique is highly versatile and can be applied to any representative volume element, providing an effective and efficient method for accurately simulating complex composite materials.

To demonstrate the effectiveness of smoothed voxel techniques in generating finite element models for highly complex models, we propose the first steps of an approach to create an accurate hexahedral mesh of a composite structure produced using the filament winding process. This technique can overcome the challenges of generating a mesh for such structures due to their intricate geometries and varying fiber orientation.

 

References

Rassineux, A. (2021). Robust conformal adaptive meshing of complex textile composites unit cells. Composite Structures, 114740, DOI: 10.1016/j.compstruct.2021.114740

Fourrier, G., Rassineux, A., Leroy, F. H., Hirsekorn, M., Fagiano, C., & Baranger, E. (2023). Automated conformal meshgenerationchainforwovencompositesbasedonCT-scanimageswithlowcontrasts.CompositeStructures, 116673. DOI: 10.1016/j.compstruct.2023.116673