Levee Characterization by Means of Data Fusion of In-Situ Geophysical and Geotechnical Information
Résumé
To prevent catastrophic consequences imputed to levee breakage, the identification of the subsoil constitutive materials, as well as the detection of interfaces and anomalies are crucial. During investigation campaigns, complementary geophysical and geotechnical methods are usually used. These two sets of methods provide data with very different spatial scales, levels of incompleteness, uncertainty and inaccuracy. The effective combination of these two different types of data is challenging. In this work, we propose the fusion of geophysical and geotechnical data by means of Belief Functions, implementing Smets and Dempster-Shafer rules. We demonstrate our approach on a real case study, at Saint-Clément-des-Levées (France) including one geophysical method (electrical resistivity tomography) and two geotechnical ones (cone penetrometer and grain size analysis after drilling). This work also aims at supplying a better visual representation of a levee section with related degrees of belief. Our results highlight the ability of our fusion methodology to properly characterize the considered materials as well as to indicate the positions of the interfaces and the associated levels of confidence.