A comparative study on stochastic multiscale analysis of composite laminates

Abstract

The progressive damage analysis of fiber composites is performed using deterministic field variables precluding the consideration of natural variability in material manufacturing. Even with high quality control techniques employed by manufacturers, there are sources of variability in the material at different spatial scales. The natural variability at microscale within the material is modeled using (1) random field approach, (2) incorporating spatially varying micro-defects. In both the approach same sampling-based non-intrusive method was adopted to incorporate aleatory uncertainty in a concurrent multiscale modeling framework. A comparison between the approaches to predict ultimate strength and stiffness of laminates is presented. Both unnotched and open-hole quasi-isotropic laminates were analyzed for validation.