skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Application of Sensitivity and Uncertainty Analysis Methods to a Validation Study for Weapons-Grade Mixed-Oxide Fuel

Conference ·
OSTI ID:788580
 [1];  [1]
  1. Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)

At the Oak Ridge National Laboratory (ORNL), sensitivity and uncertainty (S/U) analysis methods and a Generalized Linear Least-Squares Methodology (GLLSM) have been developed to quantitatively determine the similarity or lack thereof between critical benchmark experiments and an application of interest. The S/U and GLLSM methods provide a mathematical approach, which is less judgment based relative to traditional validation procedures, to assess system similarity and estimate the calculational bias and uncertainty for an application of interest. The objective of this paper is to gain experience with the S/U and GLLSM methods by revisiting a criticality safety evaluation and associated traditional validation for the shipment of weapons-grade (WG) MOX fuel in the MO-1 transportation package. In the original validation, critical experiments were selected based on a qualitative assessment of the MO-1 and MOX contents relative to the available experiments. Subsequently, traditional trending analyses were used to estimate the Δk bias and associated uncertainty. In this paper, the S/U and GLLSM procedures are used to re-evaluate the suite of critical experiments associated with the original MO-1 evaluation. Using the S/U procedures developed at ORNL, critical experiments that are similar to the undamaged and damaged MO-1 package are identified based on sensitivity and uncertainty analyses of the criticals and the MO-1 package configurations. Based on the trending analyses developed for the S/U and GLLSM procedures, the Δk bias and uncertainty for the most reactive MO-1 package configurations are estimated and used to calculate an upper subcritical limit (USL) for the MO-1 evaluation. The calculated bias and uncertainty from the S/U and GLLSM analyses lead to a calculational USL that supports the original validation study for the MO-1.

Research Organization:
Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA), Nuclear Criticality Safety Program (NCSP)
DOE Contract Number:
AC05-00OR22725
OSTI ID:
788580
Report Number(s):
P01-111333; TRN: US0200514
Resource Relation:
Conference: 2001 ANS Embedded Topical Meeting on Practical Implementation of Nuclear Criticality Safety, Reno, NV (United States), 11-15 Nov 2001; Other Information: PBD: 20 Jul 2001
Country of Publication:
United States
Language:
English