3D stochastic computational homogenization model for carbon fiber reinforced CNT/epoxy composites with spatially random properties

In this paper, a 3D stochastic computational homogenization model for carbon fiber-reinforced (CFR) CNT/epoxy matrix composites was presented. Stochastic waviness, agglomeration and orientation of CNT fillers cause random spatial variations of the elasticity tensor of the CNT/epoxy matrix within a microscale RVE, resulting in probabilistic variations of the effective homogenized stiffness of the RVE. The present computational homogenization model is based on two-scale asymptotic homogenization theory. An ensemble average of the multiple homogenized effective stiffnesses was obtained for structural analysis of the macroscale CFR CNT/epoxy composite materials. From the proposed model, we could observe significant effects of the CNT alignment orientations, agglomeration and waviness on the effective stiffnesses. Effective stiffness changes of a microscale RVE caused by the nanoscale uncertainties were investigated. The proposed multiscale modeling method and approach will be a basis for hierarchical multiscale material design of nanocomposite materials.