Re-estimation of nuclear data and reliable covariances using integral experiments. application to jeff3 library.

Overview & Synopsis

Differential measurements of Nuclear Data (ND) are not enough accurate in order to meet the target-accuracy required by reactor design calculation. Moreover, the uncertainties associated with the various ND evaluations are inconsistent. Therefore, integral experiments are needed to improve nuclear data files and to derive reliable covariance matrices.Experimental validation of international ND files is generally based on critical Keff measurements, using mainly the International Handbook of Benchmark Experiments ICSBEP. Although it is satisfactory for Criticality-Safety calculation codes, this validation is insufficient for Reactor Physics; indeed, nuclear data of minor actinides (238-241-242Pu, 237Np, 241-243Am, 244-245Cm) have to be validated, particularly for fuel depletion and cycle length calculations. Therefore, spent fuel chemical assays also have to be considered in the validation process.The international library JEFF3.1.1 is currently used in France both in the safety-criticality package CRISTAL and in the EDF and AREVA LWR calculation packages. Therefore, it is essential to validate the main evaluations of this library and to determine the associated realistic uncertainties.In order to improve the JEFF3.1.1 library and to obtain reliable covariance matrices, targeted experiments are required. Critical experiments such as ICSBEP benchmarks are useful; however the rigorous uncoupling between capture, fission and multiplicity data can be met only through the addition of Post-irradiation experiments (PIE). From the French experimental database including critical LWR cores and PWR spent fuel chemical analyses, using the ND re-estimation method, this study provides the ND improvement trends as well as the JEFF3.1.1 covariance matrices for the main nuclides.