Thèse soutenue

Modélisation théorique du comportement chimio -hydro-mécanique couplé en argiles gonflantes insaturés

FR  |  
EN
Auteur / Autrice : Xiaoqin Lei
Direction : Ali LimamHenry Kwai-Kwan Wong
Type : Thèse de doctorat
Discipline(s) : Génie civil
Date : Soutenance le 10/09/2015
Etablissement(s) : Lyon, INSA
Ecole(s) doctorale(s) : Ecole doctorale Mécanique, Energétique, Génie Civil, Acoustique (Villeurbanne ; 2011-....)
Partenaire(s) de recherche : Laboratoire : LGCIE - Laboratoire de Génie Civil et d' Ingénierie Environnementale, EA 4126 (Villeurbanne, Rhône)
Jury : Président / Présidente : Albert Giraud
Examinateurs / Examinatrices : Ali Limam, Henry Kwai-Kwan Wong, Albert Giraud, Jean-Michel Pereira, Patrick Dangla, Antonin Fabbri, Lingzhi Xie
Rapporteurs / Rapporteuses : Albert Giraud, Jean-Michel Pereira

Résumé

FR  |  
EN

Expansive clays used in engineering practice are usually unsaturated and are sensitive to chemical composition of the in-pore solution. To analyse the complex coupled problems involved, an efficient mathematical model which can account for these chemo-hydro-mechanical behaviours has been developed. In this thesis, expansive clays are conceptualised into three-phase multi-species porous media. Based on the modified mixture theory and irreversible thermodynamics, a thermo-electro-chemo-hydro-mechanical framework has been developed. The Clausius-Duhem inequality, which governs the dissipations associated with mechanical work, phase transformation, mass transport and thermal transport, is rigorously derived. Based on this thermodynamic framework, constitutive laws for bulk liquid and salt mass transport, free and adsorbed water inter-phase mass transfer, and the chemo-elastic-plastic deformations of soil skeleton have been developed. The model has been implemented into the FEM software Bil and validated by simulating available experimental data on Boom Clays. What’s more, the salt infiltration process into an unsaturated expansive clay layer has been simulated to illustrate the applicability of the model.