Mixed epistemic-aleatory uncertainty quantification using reduced dimensional polynomial chaos and parametric ANOVA

Abstract

In this paper, a reduced dimensional PC approach is presented to model the impact of both aleatory (random) and epistemic (ignorance based) uncertainty on the response of distributed transmission line networks. The key feature of this approach is the development of a parameterized analysis of variance (PANOVA) strategy to identify which of the aleatory dimensions have minimal impact on the response of the network over the entire multidimensional support of the epistemic dimensions. By removing these statistically insignificant dimensions, a highly compact PC representation of the response can be developed to capture the mixed epistemic-aleatory effects.