Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : October 07, 2017 - October 09, 2017
The authors have recently developed a first-order perturbation based stochastic homogenization (FPSH) method to predict the variability in the macroscopic properties of composite materials before manufacturing. Both geometrical and physical uncertainties at the microscale were considered in the stochastic multiscale modeling, i.e., the generation of RVE (representative volume element) models. The parameterization and statistical measurement of uncertainties have also been reported. As an extension of FPSH method, this paper mainly describes the formulation to predict the probabilistic density of microscopic strain under uniform macroscopic strain and damage probability. Short fiber reinforced composite materials are considered in this study, where RVE models were generated using Meshman_ParticlePacking.
