Analysis of Structural Systems with Imprecise Uncertainties using High Dimensional Model Representation

Abstract

Uncertainties present in any structural system inherently affect the performance and design of the system. The sources of uncertainties serve the basis for delineating the types as aleatory or epistemic. The probabilistic models can be considered as the most valuable strategies to deal with aleatory uncertainties, while convex models, possibility theory, evidence theory and Bayesian probability theory can be used to deal with epistemic uncertainty. However, when only scarce datasets are available and knowledge is incomplete, a more general framework, such as probability-box, is more appropriate to describe the uncertainty. Furthermore, analysis of complex and multi-dimensional structures is expensive and time consuming when numerical techniques are used. Therefore, simulation of such structures for many realisation of uncertain input becomes a challenging task in the uncertainty analysis. In this paper, complex structural systems with imprecise uncertain input are studied and evaluated efficiently by High Dimensional Model Representation based uncertainty analysis.