NAGRA Project Order 15’228 EBS – Task Force: Homogenization Analysis of Bentonite & Gas Transport in Bentonite

The Task 7 of the EBS task force is devoted to the modelling of gas transport phenomena in bentonite. This report describes the second phase of Task 7 that concerns the finite element modelling of some specific tests performed on granular bentonite to assess the capability of the existing HM models to describe the physical phenomena involved in the gas migration process. The material under investigation is National Standard WP2 granular Na-bentonite from Wyoming (USA). Following Romero and Gonzalez-Blanco (2017), WP2 properties are assumed similar to Nabentonite MX-80 which presents an equivalent montmorillonite content. In this study, the microstructural investigation of Romero and Gonzalez-Blanco (2017) were compared with the large amount of experimental data available for MX-80 bentonite to obtain a better understanding of the water retention and hydraulic behaviour of the material. The experimental data base consists mainly of oedometric tests performed in the framework of Task 7 (Romero and Gonzalez-Blanco, 2017). This experimental program involves the microstructural characterization of the granular bentonite, the preparation of saturated bentonite specimens, and the execution of gas injection tests in oedometric conditions. Complementary information on the material characterization can be derived from previous studies. The proposed modelling approach is based on the Finite Element Code Lagamine originally developed at the University of Liege (Charlier 1987; Charlier et al., 2001) and extended with the contribution of other research institutions. The code can address multi-physical problems (THMC) including many non-linear constitutive approaches and some advanced features dedicated to geomechanical problems (large displacement formulation, coupled contact elements, strain localization…). The modelling activity is based on a detailed analysis of the microstructure of the analysed granular bentonite to define its water retention behaviour and to calibrate the hydraulic and constitutive models. A series of analysis were performed with the aim of identifying the uncertainties of the modelling process. The Monte Carlo sensitivity analysist tool was employed to identify the impact of each modelling assumption on the accuracy of the model output and the modelling uncertainties. The analyses were conducted with growing degree of complexity: the first step was a simple hydraulic analysis of a homogeneous specimen whereas the following steps involved hydromechanical processes.