Probabilistic defect analysis of fiber reinforced composites using kriging and support vector machine based surrogates

In this study a framework for analyzing defects and their effect on the performance of composites, using wavy fiber composites as a test case, is presented. The focus is on predicting the composite strength and failure mode in the presence of through thickness waviness. The framework utilizes surrogate models, derived from a high fidelity validated Finite Element analysis, to perform Monte Carlo simulations. The final analysis product that is developed is versatile enough to solve the forward and inverse problem, while considering partial or uncertain data as input.