A Heat Transport Analytical Benchmark Problem for Sensitivity Analysis, Uncertainty Quantification, and Predictive Modeling. Ⅱ: Best Estimate Predictions with Reduced Uncertainties

摘要

An accompanying paper presents a benchmark problem that models the heat conduction process in the radial direction of a heated rod, coupled to convective heat transport along the rod's vertical direction. The temperature ranges, material properties and geometry of the rod are typical of the LBE-cooled SMR designed2 by Gen4Energy Inc., LLC, and have been selected to enable the possible design of an experimental facility aimed at providing validation data for computer codes. To design such an experimental facility, however, it is necessary to optimize the placement of measuring devices (thermocouples, flow meters), to quantify the impact of computational and experimental uncertainties on the design, and to predict the impact of measurements on the subsequent calibration of reactor design codes. This paper will illustrate the use of the sensitivities computed in the accompanying Part 1 for quantifying the uncertainties produced by model parameter uncertainties in the computed temperatures within the rod and fluid, and the subsequent use of these uncertainties, along with measurements and their uncertainties, for obtaining best estimate predictions, with reduced predicted uncertainties, for the measured temperatures.