Sensitivity Analysis and Bayesian Calibration for 2014 Sandia Verification and Validation Challenge Problem

Abstract

As the wide use of modeling and simulation in many fields, the credibility of model is of great importance to predict the model response. In this paper, taking the 2014 Sandia V&V Challenge problem for example, some detail works were done. Firstly, a parametric model was established and influential parameters were selected based on global sensitivity analysis. Bayesian updating was conducted and the uncertain parameters were identified. Uncertainty propagation was investigated and all kinds of uncertainties were quantified based on the corresponding experimental results, and the credibility of model was assessed. Then, a relationship between displacement and strain was constructed and responses were predicted based on the validated model.