Performance of a Monte-Carlo solution for the H.B. Robinson-2 Pressure Vessel Dosimetry Benchmark

Abstract

At the Paul Scherrer Institut, a methodology for PWR fast neutron fluence estimations, based on the Monte-Carlo particle transport code MCNPX with general-purpose neutron data libraries and using neutron source data from deterministic 3-D core-follow calculations, has been developed. The methodology has been validated on the basis of experimental data related to the fluence at the inner surface of a Swiss PWR Reactor Pressure Vessel. In this technical note, a first objective is to enlarge the validation basis of the methodology as well as to extend it for applications to RPV outer-surfaces. To that aim, a preliminary analysis with the MCNPX-2.4.0 code, along with the JEFF-3.1 continuous-energy neutron data library, of the “H.B. Robinson-2 Pressure Vessel Benchmark”, providing in-vessel and out-vessel experimental dosimetry data, is presented. In addition, considering that the available original solutions of the benchmark employed deterministic transport methods with associated libraries, a second objective of this technical note is to assess the progress achieved for this type of problems when applying modern Monte-Carlo based methods. The results show that for the 237Np and 238U fission dosimeters, which were of primary interest in the given study, a non-negligible improvement is seen in the MCNPX solution, indicating thus a rather good performance of the employed Monte-Carlo method and providing thereby, additional confidence for the overall PSI fluence methodology. For other high-energy dosimeters, the presented new results do not show yet any significant accuracy improvement versus previously reported results. This can however not be confirmed before additional studies, e.g., with focus on improvements of the involved modelling approximations and the statistical precision of the MCNPX calculations, be carried out. Similarly, investigations of neutron cross-section and dosimetry libraries effects remain to be addressed. These further studies are however not included here since at this stage, the principal aim was mainly to model and analyse this benchmark at most consistent manner with the previous solutions using a continuous-energy Monte-Carlo based method.