The impact of covariance information on criticality safety calculations in the resolved resonance energy range.

Resonance data play a significant role in the calculations of systems considered for criticality safety applications. K{sub eff}, the major parameter of interest in such a type of calculations, can be heavily dependent both on the quality of the resonance data as well as on the accuracy achieved in the processing of these data. If reasonable uncertainty values are available, in conjunction with their correlation in energy and among type of resonance parameters, one can exploit existing methodologies, based on perturbation theory, in order to evaluate their impact on the integral parameter of interest, i.e., K{sub eff} in our case, in practical applications. In this way, one could be able to judge if the uncertainty on specific quantities, e.g., covariances on resonance data, have a significant impact and, therefore, deserve a careful evaluation. This report, first, will recall the basic principles that lie behind an uncertainty evaluation and review the current situation in the field of covariance data. Then an attempt is made for defining a methodology that allows calculating covariances values for resolved resonance parameters. Finally, practical applications, of interest for criticality safety calculations, illustrate the impact of different assumptions on correlations among resolved resonance parameters