Speaker
Description
Authors (affiliation): Paul Malfrait (IRSN, France), Jérôme Bobin (CEA, France), Anne de Vismes-Ott (IRSN, France).
In the context of environment surveillance, more precisely in the framework of the air monitoring, we aim at (i) reducing the time between the sampling and the detection of radionuclides in the samples and (ii) get a precise estimation of the activity in the sample we analyse (even at low-level). To achieve this, we focus on a full spectrum analysis algorithm on gamma-ray spectrum obtained on HPGe detectors.
The full spectrum analysis is proven to perform better than the usual peak-based analysis as viewed in Xu et al. (2020, 2022), lowering the detection limits and estimating the activity of the radionuclide which activity is at the mBq level. In Malfrait et al. (2022) we achieved to detect the activity of low-level radionuclide earlier thanks to temporal analysis of the gamma-ray spectra. In a nutshell, we split the one-week measurement in multiple short duration ones. This allows us to estimate the activity of the radionuclides earlier than before as the analysis is carried out during the measurement. The model uses the joint analysis of the different time segments and the decay model to estimate the activity in the sample.
Speeding up gamma-ray spectrum analysis mandates processing increasingly smaller and therefore more numerous time intervals at the cost of dramatically increasing the computation time. To that purpose, and building upon Malfrait et al. (2022), we first introduce an online temporal full spectrum analysis algorithm to process gamma-ray spectra. Such a procedure allows to update the radionuclides' activity each time a new measurement is available. We show that the proposed online algorithm allows for a faster processing of the sample. More precisely, the detection of Cs-137 at trace level (few µBq/m3, namely few millibecquerel per sample) can be reached one day after the sample has been collected, much faster than the 8-day delay in routine procedure and 4 days with the method proposed in Xu et al. 2020, 2022).
