Performance assessment of a 3-D steady-state and spatial kinetics model for the CROCUS reactor

This dissertation covers both experimental and numerical neutronics studies to evaluate the adequacy of the Serpent/PARCS code sequence for modeling the steady-state and kinetics behavior of the CROCUS reactor. The reactor presents design characteristics that raise questions about the acceptability of diffusion theory for its modeling. The PARCS model of CROCUS has been developed considering several potential sources of biases. More precisely, albedo boundary conditions were used to limit the axial geometry to the grid plates where diffusion theory may lead to inaccuracies due to the presence of Cadmium layers. Proper treatment of scattering anisotropies through in-scatter correction of diffusion coefficients were also fundamental for producing accurate eigenvalues in the CROCUS reactor. A parametric study has been conducted to evaluate transport effects and the impact of energy discretization on eigenvalue and pin power distribution. Steady-state and time-dependent experimental data has been obtained from CROCUS with the purpose of validating the computational scheme. A comprehensive evaluation of experimental uncertainties provided support for the generation of reliable experimental data. Particular focus was placed upon the development of transient experiments that involve local perturbations of the flux. Delayed neutron effects were not captured in these transients because of the tightly coupled nature of the reactor. The comparison of PARCS simulations against experimental data indicated that control rod reactivity worth is predicted within (43)%. PARCS radial fission rate distributions are in considerable disagreement with experimental data for the outer core region, where differences are as large as 15%. This was attributed to the fact that PARCS does not allow using adaptable mesh sizes in the radial plane, which results in a mismatch between the mesh and explicit pins of the outer core region. However, from a safety viewpoint, these biases are conservative and are located in the outer core region where the power is low. PARCS axial fission rate profiles agree within 1% with experimental data for the bottom and mid regions for the core. On the other hand, larger deviations of about 20% were encountered for the top region, which are attributed to transport effects near the water/air interface. Finally, the investigation on neutron kinetic effects verified that the PARCS code is capable of modeling the transient experiments with spatial effects in the CROCUS reactor, where maximum differences are in the order of 5%. Overall, the Serpent/PARCS scheme shows satisfactory performance for modeling the CROCUS reactor, except for the estimation of radial reaction rate profiles, where biases were attributed to the impossibility of adapting the mesh size to match the fuel pitch of both fuel zones.