The impact of time dependant spectra on fusion blanket burn-up

Abstract

Knowledge of nuclide burn-up within tritium breeding blankets has a crucial part to play in the safety, reliability and efficiency of fusion reactors. The modelling of burn-up requires a series of neutron transport calculations which can compute the reaction rate either directly, via Monte-Carlo estimators, or by implementing the multi-group method. These reaction rates can then be directly substituted into the burn-up equations, which can calculate nuclide number densities after a specified period of burn-up. The material burn-up will change the neutron spectra and the rate of nuclear reactions. Hence, a new neutron transport calculation needs to be performed after burn-up and the sequence is repeated for several time-steps. Radiation transport calculations are computationally expensive, therefore the minimisation of reaction rate calculations via Monte-Carlo simulations is desirable. Thus, time-intervals between Monte-Carlo simulations should be as large as possible. This paper addresses the effect of neutron spectra on the burn-up of parent and daughter nuclides found in EUROFER steel and the tritium self-sufficiency time.Using a spherical reactor geometry with lithium–lead tritium breeding material, a neutron spectrum is computed at time=0 and time=2 years after a detailed depletion calculation using 1 day time intervals. These two spectra are then used to calculate reaction rates for every isotope listed within the EAF2005 database using the FISPACT code. The results show that the difference in nuclide number densities are less then 11% for all nuclides within the database and less than 4% for all fusion relevant nuclides.Using the same methodology as the first model, EUROFER parent and daughter nuclide number densities produced by each neutron spectra are compared. This study found that the change in burn-up for parent nuclides is statistically insignificant. However, the difference in daughter nuclide production is significant, especially for Rh, Ru, Re, Os, Pt and Ir where the differences range from 20% to 504%. Thus, in order to model the metallurgical properties of steels within fusion blankets over time, a multiple transport-burnup depletion code (such as used by FATI, VESTA or MONTEBURNS) must be implemented.The final part of this work studied the effect of time-step interval (used to update neutron spectra) on the tritium self-sufficiency time of a blanket. The FATI depletion code modelled the same geometry as in part 1 of the study, however time-steps ranging from 1 day to approximately 800 days were used to predict when the blanket would cease to be able to breed enough tritium to sustain the fusion reactor. The single time-step model (i.e. where a constant neutron spectrum is used for the entire simulation) underestimated the tritium self-sufficiency time of the blanket by approximately 70%. Only time-steps less than 1 month produce a self-sufficiency time which is within 5% accuracy. Hence, this work suggests that spectra time-stepping is important in the modelling of tritium production with solid breeders.