Quantifying the Contribution of Aleatory and Epistemic Uncertainties in the Estimation of Probability of Failure of Primary Piping

Abstract

The impact of aleatory (random) and epistemic (lack of knowledge) uncertainty on the probability of failure is typically quantified through a two-staged nested simulation approach. Addressing the regulatory concern over the confidence of the results has thus far had limited discussion in the literature. This paper shows how the two-staged approach can be used for modeling parameter uncertainty and analyzing sensitivity, which are unrelated to the confidence in the probability of failure. The results demonstrate how the uncertainty in the overall probability of failure depends only on the sampling error from simulation, not the input variable uncertainty or their separation.