A full core homogenization approach using serpent as a cross section generation tool for the OSCAR-4 code system

OSCAR-4 is a nodal diffusion based deterministic code system used for research reactor operational support. The system is used to perform routine core follow and reload calculations for each reactor cycle. In these small heterogeneous cores, typically encountered in research reactors, homogenization and equivalence theory play an important role. The effect of the environment in which an assembly's homogenized parameters are generated, has a large impact on the overall accuracy of the core simulation. Thus, generating few-group equivalence parameters, such as face discontinuity factors, in an infinite lattice or approximate colorset environment, is often not sufficient to capture the true assembly environment in the full core simulation. Currently, the OSCAR-4 system uses a low-order interface current, collision probability based lattice code, HEADE, to generate homogenized cross sections and discontinuity factors for the diffusion core simulator, MGRAC. HEADE suffers from some geometric limitations as well as the low-order coupling between cells in the geometry, limiting the code to small colorsets and resulting in the incorrect treatment of anisotropic effects in the lattice calculation. To overcome the shortcomings of HEADE, a link between the Monte Carlo code Serpent and the OSCAR-4 code system was developed. With this new development, Serpent can be used as a lattice code in the OSCAR-4 code system and without the geometric limitations present in HEADE, large colorsets or full core models can be used to generate equivalence parameters for the core simulator. In the work presented here, Serpent was used to produce few-group equivalence parameters for research reactor applications using 2D full core models. These homogenized parameters were then used in the core simulator to follow reactor operations and model various experiments such as control rod calibration experiments. The results of the calculations are compared to plant data obtained from experiments done during reactor operation. (author)