论文和技术资料

A thermo-hydraulically (TH) coupled code for modelling the water uptake of compacted bentonite has been developed and presented by (Kröhn and Fromme 2023). This code is internally called “COMSOL-VIPER” as it is based on the balance equations developed for the experimental code VIPER (Kröhn 2011). It’s purpose is to be employed in the context of the safety assessment for deep geological repositories (DGR) for radioactive waste. In the past, water uptake experiments have covered only the case that always as much water is provided for uptake as the bentonite is capable of taking up (unimpeded access to water UA). Boundary conditions in COMSOL-VIPER are therefore designed to meet this condition. However, experiments at GRS have shown that the uptake dynamics are quite different at a limited water supply rate (LWSR) (Noseck et al. 2018) i.e. if less water is provided than the bentonite could potentially take up. In effect, this means that the rate at which water is supplied from outside, for instance from a low-permeable rock, is the same as the diffusive flux of vapour inside the bentonite directly at the bentonite-rock contact. This leads to a slow but steady-increase of hydrated water at the inflow boundary and due to the adsorption isotherm also to an increase of the relative humidity in the pore atmosphere. The gradient of the relative humidity that drives vapour diffusion must remain constant, though. Accordingly, LSWR-conditions end when a relative humidity of 100 % is reached at the boundary which represents the well-tried UA-conditions. LSWR-conditions can thus be described by a Dirichlet boundary condition where the as-cribed value for the relative humidity must be freshly calculated and updated for each new time step. More details can be found in (Kröhn 2019). Since the balance equations derived for the hydraulic part of the VIPER-code are rather special, the COMSOL-VIPER code has been programmed using the COMSOL Physics Builder. Coupling of the resulting VIPER interface for the hydraulics and the Heat transfer interface proved to be challenging. The exchange of values for primary variables between the two interfaces unfortunately calls for a manual assignment for every new model. Addressing the LSWR-condition requires reading and processing humidity values from nodes, which need to be updated for each calculation step. The automatic location selec-tion for the measurement points as well as the update of the values in each calculation step are the most challenging parts of the implementation of the LSWR-condition. Therefore the LSWR-condition is firstly programmed for 1D cases to test the methodology. It is intended to verify the new boundary condition by applying it to the water uptake tests performed at GRS. If successful, an extension of the implementation for 2D and 3D cases is envisaged.