Multi-scale Progressive Failure Modeling: From Nano-structured Carbon Fibers to Textile Composites

Abstract

A multi-scale predictive analysis methodology based on the recently developed 3-D anisotropic solid mosaic chain approach is described. It is applied, in a hierarchical fashion, to (1) nanostructured carbon fibers, (2) microstructured unidirectional composites reinforced with carbon fibers, and (3) meso-structured 2-D woven and 3-D woven composites reinforced with carbon fiber tows. This approach enables for a unified modeling of complex composite structures at nano-, micro-, and meso-levels with the use of 3-D mosaic chain assemblies of anisotropic bricks having, generally, distinct mechanical properties and bonded together by realistic internal boundary conditions. The bricks are, firstly, assembled in series thus forming individual mosaic chains. Those chains are then assembled in parallel thus forming blocks of the reinforcement chains separated by the matrix chains and/or by the interphase material chains. The obtained elementary 3-D blocks of chains can then be assembled in series and/or in parallel thus resulting in more complex hierarchical spatial composite models that approach the actual composite architecture. Numerical examples illustrate simulations of progressive failure processes and predictions of ultimate strains/stresses under quasi-static tensile loading conditions. The obtained strength properties of carbon fibers are compared to experimental data; this step allows one to further calibrate the structural models and perform fine-tuning of the nanoscale input data. The modeling results obtained for homogenized carbon fibers form the necessary set of input data for the further modeling of curved unidirectional composite tows. The generated effective elastic and strength properties of such composite tows are then used in modeling 2-D plain weave and non-crimp 3-D orthogonal weave reinforced composite unit cells. Their predicted effective strength properties are compared with experimental data available from the macroscale specimen tests. The observed disagreements between theoretical and experimental results are analyzed and further structural model improvements suggested. The gained experience emphasizes an importance of using multiple staggered unit cells for each modeled composite type in order to achieve close agreement with experimental strength data.