Bentonite Resaturation & Swelling

QPAC is being used to model coupled thermal, hydrological and mechanical (THM) processes in bentonite buffers used in several designs of geological repositories for radioactive waste.

For example, QPAC was used within the EC THERESA project to model a thermo-hydro-mechanical laboratory experiment in which bentonite was resaturated, with resultant swelling, in a temperature gradient. Representing such a system requires consideration of multi-phase flow of water, water vapour and air, in conjunction and fully coupled with advective and diffusive transport of heat.  The most complex aspect of the problem however is the treatment of the hydro-mechanical interaction of water and bentonite, resulting in significant changes in mechanical behaviour and hydraulic properties as the bentonite is resaturated.

Bentonite is a complex material, and modelling its behaviour is a challenging problem. Good comparisons were obtained between QPAC calculations and experimental data for the evolution of temperature, saturation, density, relative humidity, stresses and pressures (see figures).

Process model developments were at the heart of the QPAC contribution to the THERESA project, and latterly included a more sophisticated, complete and internally consistent representation of bentonite for comparison with the standard THM approach. This new process model explicitly considers the disposition and physical form of water in the bentonite-water-air system, and uses this micro-scale knowledge to model macro-scale emergent phenomena such as swelling and porosity structure changes. The main advantage of this process model is that it uses a smaller set of free parameters than previously required and also gives information on bentonite structure relevant for coupled reactive species transport calculations. Hence it is more suitable for fully coupled THMC bentonite calculations than more traditional approaches.

This project provides a good example of how complex non-linear systems and novel process systems can be modelled effectively using QPAC technology.