Neutronic predesign tool for fusion power reactors system assessment

Abstract SYCOMORE, a fusion reactor system code based on a modular approach, is under development at CEA. In this framework, this paper describes a methodology developed to build the neutronic module of SYCOMORE. This neutronic module aims to evaluate main neutronic parameters characterising a fusion reactor (tokamak): tritium breeding ratio, multiplication factor, nuclear heating as a function of the reactor main geometrical parameters (major radius, elongation, etc.), of the radial build, Li enrichment, blanket and shield thickness, etc. It is based on calculations carried out with APOLLO2 and TRIPOLI-4 CEA transport code on simplified 1D and 2D neutronic models. These models are validated versus a more detailed 3D Monte-Carlo model (using TRIPOLI-4). To ease the integration of this neutronic module in SYCOMORE and provide results instantly, a surrogate model that replicates the 1D and 2D neutronic model results was used. Among the different surrogate models types (polynomial interpolation, responses functions, interpolating by Kriging, artificial neural network, etc.) the neural networks were selected for their efficiency and flexibility. The methodology described in this paper to build SYCOMORE neutronic module is devoted to HCLL blanket, but it could be applied to any breeder blanket concept provided that appropriate validation could be carried out.