Determination of correction factors for the isotopic composition in Burnup credit applications taking into account the depletion codes validation
Congress title :ICNC 2007
Congress location :Saint-Petersbourg
Congress date :28/05/2007
In order to ensure a safe evaluation of the Burnup Credit in criticality studies, it is necessary to guaranty that, for a given burnup, the depletion calculations lead to a conservative inventory of the irradiated fuel. Thus, appropriate values of the parameters that characterize the fuel irradiation (e.g. specific power, control rods presence
) have to be defined, and the depletion calculations have to be validated with post irradiation experiments. This validation can be used to determine correction factors that should be applied, if needed, to the calculated concentrations of each actinides and fission products considered in the method. The DARWIN 2.0 package is the depletion code usually used at IRSN for Burnup Credit applications. This code has been developed and validated by CEA using an experimental basis. This validation relies on comparisons between calculated and measured values of the concentrations, and is used to measure computational error for the isotopes of interest concentrations. This paper presents the study carried out by IRSN to investigate a conservative method using the CEA experimental database to estimate the correction factors for the depletion calculations. These ones are calculated in this work by means of a statistical analysis of computational error. Observed errors and experimental uncertainties are combined in a conservative manner into a correction factor for each isotope: fissile isotopes have to be corrected by a factor greater than 1, whereas absorbing isotopes have to be corrected by a factor lower than 1. After correction, isotopic concentrations may be used in criticality calculations. Different methodologies have been investigated to construct predictive models of probable extreme values of the relative bias between calculations and measures, called (C-E)/E, for a given isotope. That have been achieved by considering (C-E)/E as a function of burnup. First, a simple linear model (function of burnup) with prediction boundaries has been calculated for each isotope but it revealed its limits, as the linearity hypothesis is not significant for every isotope. Then, a kernel-smoothing model has been investigated in order to compute prediction intervals as a function of those biases and their uncertainty to make possible the correction of each isotope. This results in the computation of different sets of correction factors, depending on the methodology used. Those sets have been compared and evaluated, especially by measuring the influence of isotopic corrections on reactivity and to identify a possible lack of post irradiation experiments for some isotopes. The effect of each set of correction factors on the Burnup Credit gain has been evaluated for different configurations of storage and transport using the French criticality package CRISTAL V1.