New approaches for interlaboratory comparisons analysis using dark uncertainty applied to radioactive materials.

In order to further improve the management of contaminated materials in nuclear facilities subject to a decommissioning programme, as well as during post-accidental site remediation and clearance, the definition and selection of the most appropriate intervention scenarios producing well-characterized radioactive waste for which storage and disposal routes are clearly identified is needed. As a step towards this accomplishment, we propose a methodology for the organization and analysis of coordinated interlaboratory comparisons (ILC) for the performance assessment and the uncertainty evaluation of available measurement techniques (methods and tools) of radioactive materials. This methodology is new for this type of comparison and demonstrated on the BR3 (Belgian Reactor 3, Belgian Nuclear Research Centre, Mol) case study from the H2020 INSIDER project (2017-2021), for which barium 133, cobalt 60 and europium 152 are analysed with gamma spectroscopy in ILC, based either on irradiated concrete from the BR3 bioshield or from spiked concrete certified reference material (CRM). On one hand, we show the advantage of organizing ILC on CRM for a more reliable uncertainty evaluation taking bias into account following ISO 21748:2017. But using CRM may be impossible due to their scarcity or too costly for performance assessment thus limiting the use of CRM in ILC in practice. On the other hand, we show that for performance evaluation and monitoring, ILC can be alternately performed on reference materials provided that laboratories' uncertainties are reported and the most appropriate analysis of data is performed using dark uncertainty (excess variance) in the presence of inconsistent data.