Article

Dose assessment based on short-ranged computer simulation in a radioactive release event

¹ Nuclear Engineering Graduate Program, Military Institute of Engineering, Rio de Janeiro, Brazil
² Defense Engineering Graduate Program, Military Institute of Engineering, Rio de Janeiro, Brazil
³ Institute for Chemical, Biological, Radiological and Nuclear Defense (IDQBRN), Rio de Janeiro, Brazil
⁴ FAAP, Faculty of Economics, Graduate Program, São Paulo, Brazil

^* Corresponding author: andradepged@ime.mil.br

Received: 27 October 2022
Accepted: 27 February 2023

Abstract

This study describes the computational modeling of the atmospheric dispersion of radionuclides resulting from a hypothetical nuclear accident involving an off-site release of the inventory of a Small Modular Reactor (SMR). The inventory released was simulated with the aid of the SCALE software, considering 2 years of continuous operation. The radionuclide Cs-137 was selected due to its abilities to cause environmental damage. A simulation of the atmospheric dispersion of the radionuclide was performed using the codes HotSpot (analytical) and ANSYS (numerical). The distribution of concentrations and consequent equivalent total effective doses (TEDE) were estimated as a function of the distance from the point of release and the local atmospheric stability conditions. In the study, a hybrid computer simulation refers to the simultaneous application of the analytical and numerical models of the nuclear event. The results from the simulated conditions agree with the environmental Cs-137 concentration values, regardless of whether those values are from a numerical or analytical model. However, there is a disagreement about the location and distribution of these concentrations. It was also verified that the average percentage error of the results obtained in the two simulations was around 5%, considering the evaluation between 400 and 1500 m from the origin of the dispersion. For smaller distances in relation to the release point, the analytical and numerical models produce solutions that, although divergent, were shown to be complementary. The relevance of this investigation shows the importance of the convergence of isolated computing platforms in order to support decision-making.