Reliability and sensitivity analysis of composite structures by an adaptive Kriging based approach

Abstract

Uncertainties widely existing in composite structures make the reliability and sensitivity analysis highly necessary. In this work, the approach based on adaptive Kriging is discussed to estimate the failure probability, local and global sensitivity of composite structures. The local sensitivity can measure how the local perturbations of the distribution parameters of input variables affect the structural reliability, while the global sensitivity can measure the contributions of input variables to the reliability over their entire uncertainty ranges. The response model such as finite element model of a composite structure is approximated by the Kriging model, which is adaptively updated by the U learning function. Then the Kriging model is combined with numerical simulation for reliability and sensitivity analysis. A composite beam with explicit response function and a composite plate is analyzed to test the accuracy and efficiency of the Kriging based approach. Then a composite radome which is used for aeronautical purposes is analyzed. The reliability of the radome is obtained, and those input parameters that are important to the radome reliability are identified. The adaptive Kriging based approach offers a viable tool for reliability evaluation of composite structures.