Optimization of a calculation scheme through the parametric study of effective nuclear cross sections and application to the estimate of neutronic parameters of the ASTRID fast nuclear reactor

This thesis presents the project for the optimization of the APOLLO3® neutronic calculation scheme applied to the 4th generation fast neutron reactor ASTRID. APOLLO3® is the new multipurpose neutronic platform developed by the CEA. It incorporates many of the previous generation codes used in the French reactor core design supply chain. Like all deterministic codes, APOLLO3® solves the neutron transport equation with a discretization of the variables of interest: multi-group method for the energy, discrete ordinates and spherical harmonics for the angular variable, collision probabilities and characteristics methods for the spatial variable. The resolution of the transport equation handles useful quantities such as the neutron flux and multiplication factor, fission rates and cross sections to understand the physical behaviour of the reactor core. Currently it is not possible to use deterministic codes to simulate an entire reactor with a heterogeneous 3D geometry and a fine energy description, so to simplify the study of complete neutron field at core level, the calculation scheme is divided into two phases: lattice and core calculation. The main purpose of this work is to find an optimal degree of approximations of the calculation scheme for the evaluation of a desired physical effect and of the user constraints. In order to reach this optimum, several studies have been carried out with different levels of approximations. The results have been benchmarked with the ones obtained using the stochastic code TRIPOLI4®, used as a reference and to ensure a good accuracy. Furthermore, several sensitivity studies have been carried out to understand how the different approximations affect the macroscopic cross sections evaluation, because these dependences are not yet fully understood.