Abstract
Experimental simulation to a line source has been realized at FNS, JAERI, within the USDOE/JAERI collaborative Program on Fusion Neutronics. This simulation achieved by cyclic movement of an annular Li2O test assembly relative to a stationary point source was a step forward in better simulation of the energy and angular distributions of the incident neutron source found in Tokamak plasmas. Thus, in comparison to other experiments previously performed with a stationary point source in the program, the uncertainties (that are system-dependent) in calculating important neutronics parameters, such as tritium production rate, will be more representative of those anticipated in a fusion reactor. The rectangular annular assembly used is 1.3 m x 1.3 m and 2 m long with a square cavity of 0.42 m x 0.42 m cross-section where the simulated line source is located axially at the center. There is a 1.5 cm - thick S.S. first wall followed by a 20 cm-thick Li2O zone and a 20 cm - thick Li2CO3 zone. The ends of the rectangular assembly were left open. Calculations were performed for many measured items that include tritium production rate from Li-6(T6), Li-7(T7), in-system spectrum measurements, and various activation measurements [e.g. Ni-58(n, 2n), Ni-58(n, p), Nb-93(n, 2n), Zr-90(n, 2n), Al-27(n, alpha), In-115(n, n'), and Au-197(n, gamma)]. These measurements were performed in three radial drawers inside the Li2O and Li2CO3 zones. Flux mapping with foil activation measurements were also performed in the axial direction (z = -100 cm to z = 100 cm) at the front surface of the assembly in the cavity with the annular blanket in place and comparison was made to the bare line-source case (without annular blanket). The US has used the DOT5.1 code along with RUFF in the deterministic calculations while MCNP was used in the Monte Carlo analysis. END F/B-V data was applied in this case. The corresponding codes/data used by JAERI are DOT3.5 along with FNSUNCL, MORSE-DD, and JENDL-3 cross-section data file. In this paper, the calculated-to-measured values, C/E, for the above-mentioned measured items will be given, as obtained individually by the US and JAERI. It will be shown that even with mechanically simulating a line-source, the present methods and codes can well predict the measured items under consideration without a particular difficulty in modeling. The C/E values for T6 and T7 are closer to unity (10%) than those obtained in the previous point source experiments, while the reaction rates are within 10-15% of the measured values.