M.R. Sternat, W.S. Charlton, T.F. Nichols , “Monte-Carlo Burnup Calculation Uncertainty Quantification and Propagation Determination”, 52nd Annual Meeting of the Institute for Nuclear Materials Management (INMM), Palm Desert, CA, July 17-21, 2011.
Reactor burnup or depletion codes are used thoroughly in the fields of nuclear forensics and nuclear safeguards. Two common codes include MONTEBURNS and MCNPX/CINDER. These are Monte-Carlo depletion routines utilizing MCNP for neutron transport calculations and either ORIGEN or CINDER for burnup calculations. Uncertainties exist in the MCNP steps, but this information is not passed to the depletion calculations or saved. To quantify this transport uncertainty and determine how it propagates between burnup steps, a statistical analysis of a multiple repeated depletion runs is performed. The reactor model chosen is the Oak Ridge Research Reactor (ORR) in a single assembly, infinite lattice configuration. This model was burned for a 25.5 day cycle broken down into three steps. The output isotopics as well as effective multiplication factor (k-effective) were tabulated and histograms were created at each burnup step using the Scott Method to determine the bin width. The distributions for each code are a statistical benchmark and comparisons made. It was expected that the gram quantities and k-effective histograms would produce normally distributed results since they were produced from a Monte-Carlo routine, but some of the results do not. Statistical analyses are performed using the χ2 test against a normal distribution for the k-effective results and several isotopes including Cs-134, Cs-137, U-235, U-238, Np-237, Pu-238, Pu-239, and Pu-240.