The US Nuclear Regulatory Commission (US NRC) and the Electric Power Research Institute (EPRI) have jointly funded the development of the probabilistic analysis tool xLPR (extremely Low Probability of Rupture) version 2.0 for calculating the probability of a leak before break for welded pipes . The acceptance testing for the program was completed in July of 2016. This allowed the release of the code to the development team. The released version of xLPR v 2.0 was used to run probabilistic simulations on selected scenarios and then perform uncertainty and sensitivity analysis on selected results. Uncertainty Analysis (UA) informs on how much the uncertainty affect the outputs while Sensitivity Analysis (SA) ranks the uncertain parameters in term of importance (i.e., influence over uncertainty) for selected outputs.
In this paper, we present the methodology used to perform Sensitivity Analysis using three regression techniques as well as scatterplots for graphical study. The results generated are consistent both with the experts’ expectations and our understanding of the equations used in the sub-models.