Technical Report NTB 98-03

Large-scale Experiment for Water and Gas Transport in Cementitious Backfill Materials (Phase 1) COLEX I

In the planned Swiss repository for low- and intermediate-level radioactive waste, the voids between the waste containers will be backfilled with a highly permeable mortar (Nagra designation: mortar M1). As well as providing mechanical stability through filling of voids and sorbing radionuclides, the mortar must divert gases formed in the repository as a result of corrosion into the neighbouring host rock. This will prevent damage which could be caused by excess pressure on the repository structures. Water transport, which is coupled to gas transport, is also of interest. The former is responsible for the migration of radionuclides.

Up till now, numerical simulations for a repository situation were carried out using transport parameters determined for small samples in the laboratory. However, the numerical simulations still had to be validated by a large-scale experiment. The investigations presented here should close this gap.  

Investigations into gas and water transport were carried out using a column (up to 5.4 m high) filled with backfill mortar. The column has a modular construction and can be sealed at the top end with a material of defined permeability (plug or top plug). The possibility to vary the material of the plug allows the influence of the more impermeable cavern lining or possible gas escape vents in the cavern roof to be investigated. A gas supply is connected to the bottom end and is used to simulate different gas generation rates from the waste.  

A total of 5 experiments were carried out in which the gas generation rate, the column height and the permeability of the plug were varied. Before the start of the experiments, the mortar in the column and the plug were saturated with water to approx. 95 %. In all the experiments, an increase in pressure with time could be observed. The higher the gas generation rate and the lower the permeability of the plug, the more quickly this occurred. At the beginning, only water flow out of the top of the column was observed. Only after the saturation of the mortar had dropped to approx. 80 % could gas be detected flowing out of the column. This gas flow was not continuous but occurred in pulsed intervals.  

The experimental investigations were then numerically simulated and the results compared with those from the experiments themselves. This demonstrated that, with the transport parameters used up to now, the experiments could be simulated, sometimes resulting in a very good correlation with the actual tests. Extensions to the numerical model are suggested, which will allow simulation of phenomena not yet implemented in the model.