Verification of a Reactor Physics Calculation Scheme for the Crocus Reactor

Abstract

CROCUS is a zero power (100 W) reactor of the Laboratory for Reactor Physics and Systems Behavior (LRS) at the Swiss Federal Institute of Technology, Lausanne (EPFL). It is used for teaching and research purposes. In the CROCUS Safety Analysis Report (SAR), its modeling has relied on diffusion theory and point kinetics for the neutronic analysis and simplified thermal hydraulics models for accident analysis. Recently, an effort has started within the LRS to improve its modeling capabilities, the long term goal being to update the CROCUS SAR for improved operational flexibility.The present work is focused on the static neutron analysis of CROCUS through the developmentof a 3D nodal simulator (e.g. PARCS) model of the reactor and its verification against reference solutions provided by either experiments or whole-core Monte Carlo calculations using the MCNP5 and SERPENT codes. The quantities of interest for the verification of the model are the keff, and the control rod worths.Nodal core simulators are typically used in the industry for modeling of Light Water Reactors (LWR). The set of homogenized macroscopic cross-sections needed by the core simulator, referred in this work as nuclear data library, is generated by the SERPENT Monte Carlo code. An innovative homogenization approach to generate the nuclear data library is considered due to the irregular radial geometry of the CROCUS reactor.PARCS keff predictions are within 400 pcm of the SERPENT/MCNP5 results, which in turn deviates by about 200 pcm from the experimental values. The latter deviation is covered by the uncertainty due to the nuclear data in the keff prediction (about 500 pcm). PARCS control rod worth prediction underestimates the reference solutions by about 30 pcm pointing towards necessary improvement in the homogenization procedure for control rods.