INTEGRAL PARAMETER SENSITIVITY, UNCERTAINTY, AND REPRESENTATIVITY OF A FAST CRITICAL EXPERIMENT WITH REGARD TO A GEN-IV FAST REACTOR CONCEPT

摘要

A criticality benchmark experiment, ZPPR-15 Assembly B, is selected for analysis to reduce uncertainties on a Generation IV reactor design, ABTR, via a data assimilation method. The ABTR is modeled with fuel compositions reflecting both startup and equilibrium recycle cores. The ZPPR experiment is modeled as built and then computationally modified to include minor actinide targets to give information on their reaction rates. All four models are created in the REBUS code system for fast reactor analysis. The Efficient Subspace Method is used for a sensitivity and uncertainty analysis of several integral parameters, selected reaction rates at various spatial nodes, and disposal heat load. The similarity of the benchmark experiment to the reactor design concept is quantitatively compared using representativity. Finally, uncertainty and sensitivity information from the benchmark models is used to compute posteriori uncertainty information for the design concept via a data assimilation technique.