Visual observation of freezing and thawing processes in 3d printed fracture replicas

In the German site selection process for a nuclear waste repository, crystalline rock is one of three different types of host rock that are currently considered. As contact of groundwater with the waste canisters poses one of the main threats to such a repository, knowledge about groundwater flow in the general area is essential for a safety assessment. By law, the safety of a nuclear waste repository in Germany must be ensured for at least one million years. During this time, several ice ages are very likely to occur. They are expected to cause permafrost conditions in the underground at any conceivable location for a repository and are to be considered in the safety assessment as they are accompanied by considerable changes in groundwater flow due to freezing.

Freezing of water in a classic porous medium does not result in an instantaneous phase change of the whole pore water, though. Within a certain temperature range, an increasing volume of ice builds up in the pore space with falling temperatures. The referring Soil Freezing Characteristic Curve (SFCC) relates the degree of water/ice saturation with temperature thereby providing a key parameter for the temperature-dependent relative permeability. While these constitutive relations have already been investigated for classic porous media, hardly any information is available yet for fracture flow in crystalline rock. A new methodology has thus been developed for measuring the temperature-dependent relative permeability in fractures.

Based on a digital representation, a transparent fracture replica with a size of 7 by 10cm has been 3d printed. As groundwater flow in granite mainly occurs within the fractures, the influence of the rock matrix on such processes is neglected by the new method. To detect the formation of ice inside the fracture replica, a LED light source was placed underneath the fracture replica and freezing as well as melting processes were observed by a camera positioned above the fracture. The whole experimental setup was placed inside a climate chamber to test the influence of different temperatures. Before the tests, all components were tempered for at least 3 days before the system got completely flooded. The distinction between water as well as ice who are both transparent in their natural state was rendered possible by using a 0,05% methylene-blue solution. In a liquid state it is dark blue while being transparent in solid state. Pretesting ensured that the added methylene-blue has only a negligible effect on the freezing behavior. In parallel to observing visually the freezing and thawing in the fracture, the related effective permeability was determined by measuring the outflow rate at a constant inflow pressure.  

The obtained images were segmented using Matlab for evaluating the ratio between ice and liquid water. In combination with the performed flow tests, the relation between relative permeability and ice content has been determined for the printed fracture replica. Further tests with variations of the replica are envisioned.