An integrated and efficient numerical framework for uncertainty quantification: application to the NASA Langley multidisciplinary uncertainty quantification challenge

In this work, an integrated framework to deal with scarce data, aleatory and epistemic uncertainties is presented. Generally, dealing with the uncertainty requires the availability of efficient and scalable computational tools. For this reason, the proposed strategies have been implemented in an open general purpose computational framework for uncertainty quantification and management that allows for a significant reduction of the computational time required by adopting efficient techniques for uncertainty quantification and resorting to the computational power of a cluster computing. The proposed framework has been adopted to solve the NASA Langley multidisciplinary uncertainty quantification challenge. All the five subproblems have been tacked, i.e. uncertainty characterization, sensitivity analysis, uncertainty quantification, extreme case analysis and robust design. All the subproblems have been solved using different approaches based on different hypotheses and assumption in order to cross-validate the results and showing the exibility and potentiality and computational framework.