Fusion Science and Technology / Volume 13 / Number 4 / May 1988 / Pages 616-643
Technical Paper / Tritium System / dx.doi.org/10.13182/FST88-A25138
Articles are hosted by Taylor and Francis Online.
Several integral experiments on tritium breeding were jointly performed at the Fusion Neutronics Source (FNS) facility at the Japan Atomic Energy Research Institute (JAERI), in connection with the U.S./JAERI Collaborative Program on Fusion Breeder Neutronics. Tritium production rates from 6Li (T6) and 7Li (T7) were measured at several locations in an Li2O assembly (D = 60 cm, L = 60 cm) embedded in the concrete wall of a 5- × 5- × 4.5-m room (reference experiment). JAERI has also performed independent benchmark experiments with the Li2O assembly located in a large room of negligible room-return neutrons. In the reference experiment, large discrepancies in T6 were found at the front locations in the Li2O assembly. At middle locations, the calculated-to-experimental (C/E) values for T6 are ∼1.2 (U.S.) and ∼1.1 (JAERI). The C/E values for T7 are ∼1.18 (U.S.) and 1.05 (JAERI). To assess the contribution to the uncertainty in predicting T6 and T7 that results from the current uncertainties in the nuclear data base, an extensive two-dimensional cross-section sensitivity/uncertainty analysis was performed. For that purpose, the FORSS module, and the VIP and DOT 4.3 codes were used along with the PUFF-2 covariance code. Two systems were considered for the analysis: the benchmark system and the reference system. The models used simulate the geometrical details and source conditions for the experiments. After coupling the sensitivity profiles with the cross-section uncertainty information (ENDF/B-V, file 33), it was found that the standard deviations in T6 are 2.0 to 3.5%. In the reference system, the uncertainties in T6 at front locations due to data uncertainties were found to be very small (∼0.3%). The large discrepancies at these locations between the calculation and measurements were attributed to inaccuracy in modeling and predicting the room-return component of incident neutrons. The uncertainties in T7 due to the uncertainties in nuclear data were found to be 3 to 6%, with the largest values at back locations. The discrepancies with experimental values were attributed to the inaccuracy in the 7Li(n,n′α)t cross section, which requires further evaluation.