Evaluation of the predictive capability of coupled thermo-hydro-mechanical models for a heated bentonite/clay system (HE-E) in the Mont Terri Rock Laboratory

Abstract

Process understanding and parameter identification using numerical methods based on experimental findings are key aspects of the international cooperative project DECOVALEX (DEvelopment of COupled models and their VALidation against Experiments http://www.decovalex.org). Comparing the long-term predictions from numerical models against experimental results increases confidence in the site selection and site evaluation process for a radioactive waste repository in deep geological formations. In the present phase of the project, DECOVALEX2015, eight research teams have developed and applied models for simulating the HE-E in situ heater experiment in the Opalinus Clay in the Mont Terri Rock Laboratory in Switzerland. The modelling task was divided into two study stages, related to prediction and interpretation of the experiment. A blind prediction of the HE-E experiment was performed based on calibrated parameter values for both the Opalinus Clay, which were derived from the modelling of another in situ experiment (HE-D experiment in the Mont Terri Rock Laboratory), and calibrated parameters for MX80 granular bentonite and a sand/bentonite mixture, which were derived from modelling of laboratory column tests. After publication of the HE-E experimental data, additional functions for coupled processes were analysed and considered in the different models. Moreover, parameter values were varied to interpret the measured temperature, relative humidity and pore pressure evolution. Generally, the temperature field can be well reproduced and is mainly controlled by thermal conductivity in the heat conduction process; the thermal conductivities of buffer materials and Opalinus Clay strongly depend on the degree of water saturation. The distribution of relative humidity is acceptable as it is reproduced by using both the Richards’ flow model and the multiphase flow model. Important here is to consider the vapour diffusion process. The analysis of the predictive and interpretative modelling confirms that the main processes in the system have been understood at least for the short-term experimental duration and captured using the models developed and associated parameters with respect to the thermal and hydraulic aspects in the high-level nuclear waste disposal in clay formations. The additional experimental results will help to increase confidence in the THM models and in process understanding.