Tritium breeding ratio optimization in simple multi-layer blanket with genetic algorithm

Abstract

Investigation on tritium production via fast neutrons generated by magnetically confined plasma in a tokamak device is conducted for optimization of tritium breeding ratio (TBR). The blanket is configured as a 1-dimensional multiple layer of materials for the wall, moderation, reflection, and tritium productions, and genetic algorithm is adopted to select the optimal material on each order and location. The configuration selected in the process is evaluated in aspect of tritium production to incoming neutron ratio. To construct the algorithm, the ratio of tritium production is parametrized by neutron energies from 0 to 14 MeV for the materials with Geant4 Monte Carlo simulation toolkit, and the simulated tritium in the algorithm are evaluated for the selection of parent for the next generation. Result configuration from the algorithm is put back to the Geant4 simulation for verification, and TBR is evaluated with blanket designs in other facilities.