Quantification of design margins and safety factors based on the prediction uncertainty in tritium production rate from fusion integral experiments of the USDOE/JAERI collaborative program on fusion blanket neutronics

Abstract

Abstract Several fusion integral experiments were performed within a collaboration between the USA and Japan on fusion breeder neutronics aimed at verifying the prediction accuracy of key neutronics parameters in a fusion reactor blanket based on current neutron transport codes and basic nuclear databases. The focus has been on the tritium production rate (TRP) as an important design parameter to resolve the issue of tritium self-sufficiency in a fusion reactor. In this paper, the calculational and experimental uncertainties (errors) in local TPR in each experiment performed i were interpolated and propagated to estimate the prediction uncertainty ui in the line-integrated TPR and its standard deviation σi. The measured data are based on Li-glass and NE213 detectors. From the quantities ui and σi, normalized density functions (NDFs) were constructed, considering all the experiments and their associated analyses performed independently by the UCLA and JAERI. Several statistical parameters were derived, including the mean prediction uncertainties u and the possible spread ±σu around them. Design margins and safety factors were derived from these NDFs. Distinction was made between the results obtained by UCLA and JAERI and between calculational results based on the discrete ordinates and Monte Carlo methods. The prediction uncertainties, their standard deviations and the design margins and safety factors were derived for the line-integrated TPR from Li-6 T6, and Li-7 T7. These parameters were used to estimate the corresponding uncertainties and safety factor for the line-integrated TPR from natural lithium Tn.