Present-day precision software makes it possible to create a geometrical model of a research reactor (RR) with practically any level of detail. However, the criteria for choosing this level of detail have not been stated yet. This paper offers a justification of spatial partitioning for calculations of power density and burn-up distribution in research reactors (RR) using Monte-Carlo code. The results of maximum power density calculations depend heavily on the selection of the size of the region that is considered to be a "hot" point. When neutronic problems were solved using diffusion finite-difference programs, a most detailed spatial grid (1 cm or less) was necessary to minimize the error of the calculation algorithm itself. When using the Monte-Carlo model, on the one hand, the exact value of an average functional in the selected region can be obtained for the region of any size (including large ones), on the other hand, to determine the functional we can choose a region of any form and rather small size. The necessity to reduce the grid to get the exact solution disappears, and another problem arises, which is connected with accumulating statistical data in small regions. Nevertheless, in many cases, RR calculations using Monte-Carlo codes demonstrate commitment to calculate power density distribution in maximum detail. At the same time, the possibility of using less detailed partitioning, which gives practically the same results in a considerably shorter time, is not investigated.

Key words
Monte-Carlo method, research reactor (RR) model, RR neutronic characteristics, "hot" point, RR burn-up, power density distribution, spatial partitioning, uranium fuel

Article Text (PDF, in Russian)


UDC 621.039.54+621.039.51...17

Problems of Atomic Science and Technology. Series: Nuclear and Reactor Constants", issue 3:3, 2014