A systematic study of cross section library based discrepancies in LWR criticality calculations

Discrepancies in fundamental nuclear data pose a potential source of uncertainty in neutron transport calculation. These discrepancies are often overlooked in reactor physics calculations. There is evidence that differences in the order of 1% in the multiplication factor are encountered when criticality calculations are carried out using different neutron cross section libraries. Monte Carlo transport calculation codes are especially prone to such discrepancies, since the base evaluated data can be used without modifications. This paper presents some results of a study, in which cross section library-based discrepancies in light water reactor criticality calculations were investigated in a systematic manner. Point-wise cross section libraries were generated from the ENDF/B-VI.8, JEFF-3.0, JENDL-3.3, JEF-2.2 and JENDL-3.2 evaluated nuclear data files using the NJOY-99 nuclear data processing system. The comparison calculations were carried out using the MCNP4C Monte Carlo transport calculation code. A systematic method based on the neutron balance of the system was developed in order to study the origin of the reactivity discrepancies. The energy dependence of the cross section data was taken into account by dividing the neutron flux spectrum into four energy groups. The comparison calculations cover the most typical LWR operating conditions. The basic geometry is an infinite pin-cell lattice. The main free parameter in the system is the fuel-to-moderator ratio. Several variations of the basic geometry were studied, including lattices with burnable absorber and control pins, a finite lattice with leakage and lattices with low- and high-burnup fuel pins