Assessing the Performance of Functional Expansion Tallies in the Serpent 2 Monte Carlo Code

Abstract

Serpent has recently been implemented with a novel functional expansion tally (FET) as an alternative to tallying events occurring in specifically defined bins. Test calculations were performed to assess the potential of FETs as a replacement to bin-based tallies espe- cially in the context of transferring fuel rod power distributions to 1.5D fuel performance codes in coupled neutronics and fuel thermo-mechanics calculations. Power distributions obtained with FETs were compared to fine resolution reference results obtained with bin- based tallies using a significant amount of simulated neutron histories, with the aim of confirming whether FETs can capture the power distributions accurately in selected cases consisting of a fresh UO2 fuel rod with and without burnable absorber and of a rod located next to a partly inserted control rod. Our results showed that even low-order FETs can provide accurate axial and radial power distributions in the case of a fresh UO2 rod, but even relatively high-order FETs experience problems in capturing the power profiles cor- rectly in the more demanding cases with steep gradients in the flux shape. In these cases the FETs provided power distributions that visibly oscillate around the reference solution, even producing negative values at low power levels. Thus, it was concluded that FETs need to be used carefully especially when used to provide input data to external solvers.