Activity estimation of shielded or hidden radionuclides in emergency conditions: Impact of environmental conditions

Activity Acronym:

RadShield

Authors:

H. Toivonen, H. Ramebäck, M. Granström, S. Ihantola, G Jónsson, M. Ilkov, M. Kilkki,

Abstract:

In 2017 an NKS project focused on the shield analysis, which showed that the spectrum contains enough information to determine the attenuation of the photons in a material between the source and the detector. Two approaches were studied: the step ratio- and the peak ratio methods. In the step ratio method, the height of the step just below a full energy peak originates from photon scattering to small angles, primarily from scattering in a shield. Using that information, the ratio of the step divided by the net area of the peak, is a function of the shield thickness. If a calibration is done with different thicknesses, that calibration can be used to determine the shield thickness. Moreover, with knowledge of the distance between the detector and the source and the efficiency of the detector at a reference distance, the activity of the source can be determined. The other approach, the peak ratio method, can be used for radionuclides emitting more than one gamma photon having enough difference in energies in its decay. That method uses the fact that the attenuation of the two gamma rays will be different. Again, knowledge of the distance and the efficiency of the detector gives the source activity.
The study in 2017 showed the effect of the environment on the shield analysis for the step ratio method when the information in the step just below a peak is to be used for the analysis. The present study focuses on the impact of the environment for this method. In particular, material other than the shield, including the ground, contributes to the small angle scattering and therefore to the step underneath the peaks. Such contributions will cause a bias in the method which has to be accounted for in the uncertainty calculation.
Another problem correlated to the uncertainty estimation is the fact that the activity of a shielded source is not linear with respect to some input quantities. If the uncertainty of those input quantities is large, normal uncertainty propagation will result in bad estimates for the uncertainty. One solution to this problem is to apply a Monte Carlo method, i.e. propagate the distributions. This was also done for the activity estimations in this work.