Communication Dans Un Congrès Année : 2017

TUComp framework for modeling and analysis of composite materials based on X-ray microtomography

Résumé

Hereby presented is a TUComp suite of Python programs conceived to help in modeling and analysis of composite materials based on X-ray microtomographic scans. The approach combines existing libraries like WEKA learning algorithms library, Google’s TensorFlow for deep learning and Blender for 3D modeling with in-house developed algorithms based on implicit and dual kriging. The main applications of the system reach from the analysis of short fiber degeneration and identification of defects in manufacture with Resin Transfer Molding (RTM) to the automatic reconstruction of the realistic geometry of fibrous woven reinforcements to predict their permeability and behavior during compression.

Fichier principal
Vignette du fichier
amCSMA2017log1.pdf (2.64 Mo) Télécharger le fichier
Format typeAnnex_abstract
Licence
Loading...

Dates et versions

hal-01908918 , version 1 (30-10-2018)

Licence

Identifiants

  • HAL Id : hal-01908918 , version 1

Citer

Anna Madra, Francois Trochu, Piotr Breitkopf. TUComp framework for modeling and analysis of composite materials based on X-ray microtomography. 13e colloque national en calcul des structures, Université Paris-Saclay, May 2017, Giens, Var, France. ⟨hal-01908918⟩
138 Consultations
214 Téléchargements

Partager

  • More