Measurements by activation foils and comparative computations by MCNP code

Systematic study of the radioactive waste minimisation problem is subject of the SPHINX project. Its idea is that burning or transmutation of the waste inventory problematic part will be realized in a nuclear reactor the fuel of which is in the form of liquid fluorides. In frame of the project, several experiments have been performed with so-called inserted experimental channel. The channel was filled up by the fluorides mixture, surrounded by six fuel assemblies with moderator and placed into LR-0 reactor vessel. This formation was brought to critical state and measurement with activation foil detectors were carried out at selected positions of the inserted channel. Main aim of the measurements was to determine reaction rates for the detectors mentioned. For experiment evaluation, comparative computations were accomplished by code MCNP4a. The results obtained show that very often, computed values of reaction rates differ substantially from the values that were obtained from the experiment. This contribution deals with analysis of the reasons of these differences from the point of view of computations by Monte Carlo method. The analysis of concrete cases shows that the inaccuracy of reaction rate computed is caused mostly by three circumstances:-space region that is occupied by detector is relatively very small;- microscopic effective cross-section R(E) of the reaction changes strongly with energy just in the energy interval that gives the greatest contribution to the reaction; - in the energy interval that gives the greatest contribution to reaction rate, the error of the computed neutron flux is great. These circumstances evoke that the computation of reaction rate with casual accuracy submits extreme demands on computing time. (Author)