Skip to Main content Skip to Navigation
Theses

Combinaison de la géostatistique et des simulations à base physique – application à la caractérisation de panaches de contaminants

Abstract : A major issue concerning the remediation of industrial plants is the characterization of contamination within soils and groundwater. Two classical approaches are combined in this work: geostatistics and numerical simulations of flow and transport. It aims at improving the estimates of radionuclide activities by taking into account physically-based information into a mathematical model. Two methods are introduced in this work. They are both based on a large number of simulations of contaminant plumes, taking into account the uncertainties on input parameters. The first method is kriging with an external drift, in which the drift is the average of the simulations. The second method is kriging with a non-stationary empirical variogram computed from the set of simulations, which models the spatial structure of the phenomenon under study. The performances of the methods are first assessed on a synthetic test case, which deals with the migration of a tritium plume in the unsaturated zone of a sandy aquifer. Then, they are assessed on two datasets collected around Chernobyl nuclear plant. The results underline the benefit of combining geostatistics and numerical simulations, especially when only few observations are available.
Document type :
Theses
Complete list of metadata

https://pastel.archives-ouvertes.fr/tel-03135798
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, February 9, 2021 - 11:38:09 AM
Last modification on : Wednesday, November 17, 2021 - 12:31:40 PM
Long-term archiving on: : Monday, May 10, 2021 - 6:32:42 PM

File

2020UPSLM047_diffusion.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03135798, version 1

Citation

Léa Pannecoucke. Combinaison de la géostatistique et des simulations à base physique – application à la caractérisation de panaches de contaminants. Ingénierie de l'environnement. Université Paris sciences et lettres, 2020. Français. ⟨NNT : 2020UPSLM047⟩. ⟨tel-03135798⟩

Share

Metrics

Record views

180

Files downloads

137