First steps towards productionizing probabilistic radwaste characterization

Eric Laloy, Bart Rogiers, An Bielen, Alessandro Borella, Giada Gandolfo, Luigi Lepore, Giuseppe A. Marzo, Nadia Cherubini, Bertrand Perot, Quentin Ducasse, Cyrille Eleon, Sven Boden

Research outputpeer-review

Abstract

We present the findings of the probabilistic radiological characterization exercise conducted within the EU project MICADO “Measurement and Instrumentation for Cleaning and Decommissioning Operations”. A Bayesian inversion approach that accounts for uncertainty in the measurement efficiencies is used to interpret combinations of (i) segmented gamma scanning (SGS) spectrometry, (ii) passive neutron coincidence counting (PNCC) and (iii) active neutron interrogation (AN), in a fully virtual experiment. The considered Bayesian approach treats uncertainty in the measurement efficiencies by doing multilinear interpolation between reference efficiencies representing potential “end-member” waste matrices with respect to both composition and density, with the end-member proportions being jointly inferred with the other unknowns. The performance of the approach in terms of efficiency and accuracy is explored for two virtual case studies of increasing complexity, that are based on common, real waste packages. The used Bayesian approach appears to be fast and rather accurate for the first considered waste package. With respect to the second waste package which has a substantially more complex and heterogeneous matrix structure, some biases are noticed in the derived posterior mass distributions of the nuclides of interest. We thus discuss possible causes and solutions for these discrepancies. In addition, we devise an R package that wraps the probabilistic models in an HTTP API so that the user can send HTTP requests to a remote server that runs the computations and returns the obtained results. This should allow using the approach in a production environment.

Original languageEnglish
Article number113257
Number of pages17
JournalNuclear Engineering and Design
Volume424
DOIs
StatePublished - Aug 2024

Funding

This work received funding by the EU project MICADO Consortium . 2018: \u201CHorizon 2020, Nfrp 2018-10, Proposal 847641: Measurement and Instrumentation for Cleaning and Decommissioning Operations\u201D. An example code of the proposed approach will be available online at https://github.com/rogiersbart/micado.dap .

FundersFunder number
EC - European Commission847641

    ASJC Scopus subject areas

    • Nuclear and High Energy Physics
    • General Materials Science
    • Nuclear Energy and Engineering
    • Safety, Risk, Reliability and Quality
    • Waste Management and Disposal
    • Mechanical Engineering

    Cite this