In the Äspö Hard Rock Laboratory Germany plans an in situ test on two-phase flow in a water-saturated formation. This test will concentrate on investigations related to groundwater flow, radionuclide transport and geochemistry, on two-phase flow investigations, and on the development and testing of instrumentation and methods for underground rock characterization.

As far as the migration of released radionuclides from the repository into the biosphere is concerned, two-phase flow is of importance, as water and gas govern the transport. Gases can be generated in a repository predominantly by corrosion of waste canisters or may be released from the waste, from the geological formation, or from formation water. Therefore, it is essential to improve the understanding of the two-phase flow processes in the more or less water-saturated rock formations.

In preparation for the in situ experiment, preliminary measurements of gas concentrations in the ventilation air, in the host rock, and in the formation water have been performed. The analyses of the ventilation air and of the formation water have indicated that hydrogen, methane, and carbon dioxide are released in detectable amounts. At specific locations somewhat higher gas concentrations were found; these may possibly correlate with mineralogical and stratigraphic rock properties, fissures, or tectonic fracture zones.

Since excavation-disturbed zones around underground drifts represent potential flow paths in a repository, geoelectrical measurements have been started to determine the water-bearing structures in such zones. An array with 299 electrodes has been installed around an area of 25 m x 50 m. With the array, which is one of the largest ever installed underground, a tomographic image of the area can be obtained. The applicability of the geoelectrical method was confirmed in principle by the successful identification of a known water inflow in a borehole. For quantifying the correlation of the measured rock resistivity with the water content, however, additional laboratory investigations on rock samples have yet to be performed.

The selection of the two-phase flow test field was performed primarily on the basis of the geological situation at the site and results of the thermographic work. Investigations were focussed on the location at tunnelmeter 2715. With respect to modeling, the configuration of horizontal and vertical fractures represents well defined boundary conditions for the hydraulic flow field. The possible influence of artificial fractures on the flow field could be minimized by removing the brittle front of the niche by additional smooth blasting.

In nuclear repositories in deep geological formations, gases will be present because of natural degassing of the host rock, because of corrosion of materials, and as a result of radiolysis. The migration of contaminated gases out of the repository as well as the possible pressurization of repository areas represent important issues in long-term safety analyses.

Degassing of the disposal areas in a repository is rather likely because a complete backfilling of the storage rooms is hardly achievable. Possible pathways for the gases are represented by the (moist) rock formations around the disposal areas, by the excavation-disturbed zones around drifts and rooms, as well as by the geotechnical barriers inside the repository. The two-phase flow (TPF) of water and gases in the host rocks and in the seals must therefore be considered in the safety analyses of a repository.

The appropriate methods have been tested mainly with regard to the characterization of a suitable test site in the Äspö mine for the performance of the envisaged experiment concerning the two-phase flow. The following measurements have been performed:

• measurement of the gas content of the rock mass and of the formation water as well as the distribution and propagation of the gas by taking samples from sealed boreholes and from the open test field,
• determination of the extension of the excavation-disturbed zone (EDZ) by applying a new combined geoelectric method and detection of the moisture distribution in the EDZ and in the uninfluenced rock mass, respectively,
• recording of water-bearing structures as well as their spatial distribution by measurement of the characteristic drift surface temperatures.

In order to determine the continuous gas release from the host rock, samples of the mine air were taken at 9 different locations of the access drift. To separate the influence of the diesel exhaust from the machines and cars, samples of the mine air were taken on Friday, after continuous ventilation during a five-day period, and on Monday, after the ventilation had been switched off for about 60 hours.

Additional air samples were taken at the exhaust of the mine ventilation system on Friday before switching off the ventilation, and on Monday, 0.25, 0.5, 1.0, 1.5, 2.0, and 24 hours after switching on the ventilation again. For comparison, samples of the air entering the mine were taken.

The air samples were taken with a manual pump and transferred into gas bags for transport to the laboratory to be analyzed by gas chromatography.

It was found that the oxygen and nitrogen concentrations were comparable with the atmospheric concentrations. Only carbon dioxide and methane were measured in concentrations significantly above the detection limit of the gas chromatograph.

The carbon dioxide concentration during ventilation was between 350 and 925 vpm
(1 vpm = 1 cm³ per m³). Without ventilation for 60 hours it increased to a level between 440 and 986 vpm.

Methane at a concentration between 0 and 3 vpm was measured during ventilation. Without ventilation, the methane concentration was between 0 and 32 vpm.

From the results of the carbon dioxide and methane concentration measurements, it can be concluded that a correlation between both components exists, because methane can be oxidized by microbes if oxygen is present.

The measurements at the exhaust indicated that the carbon dioxide concentration in the air entering the mine ranges between 350 and 380 vpm, whereas the methane concentration is below the detection limit. The air leaving the mine after continuous ventilation and an active mine period of 5 days shows a carbon dioxide content of about 380 vpm and a methane content of about 0.3 vpm. After a period of 60 hours without any ventilation, the concentration of carbon dioxide was about 500 vpm (an increase by 120 vpm), and that of methane was about 10 vpm. With ventilation both concentrations decreased, and after 24 hours they were almost at the same level as that observed during continuous ventilation.

The measurements indicate that during ventilation the operation of cars and machines does not significantly affect the concentration of carbon dioxide and methane in the mine air.

For estimating the gas release rates from the host rock the following assumptions were made:

This specific rate may be higher in some regions by a factor of at least 3 to 10. A correlation between the gas release and the mineralogy, stratigraphy, fissures, or tectonically fractured zones is rather likely.

The following gases have been identified:

1. Hydrogen. Up to 5.9 µl per g of water were released at 20°C, and up to 5.6 µl per g of water at 75°C.

2. Methane. Up to 2 µl per g of water were released at 20°C and, in addition up to 2 µl per g at 75°C

Nitrogen was not found above the detection limit of about 0.0001 µl per g of water. That means that the formation water contains no significant amounts of dissolved nitrogen.

Other gases like helium or further hydrocarbons were not found above the detection limit of about 0.0001 µl per g of water.

The excavation-disturbed zone (EDZ) around underground drifts with its increased permeability represents potential flow paths for water into a repository. Since the extension and hydraulic properties of this zone affects the long-term safety of a repository, a more detailed investigation of its properties and their spatial and temporal variations is required.

The water content of a rock can be estimated through a correlation with its electrical resistivity. In Germany this method has been applied for many years for the quantification of the brine or water content in rock salt, anhydrite and granite. In the field the electrical resistivity of the rock is determined with different 4-point configurations. By means of two electrodes (A-B) and a known voltage, a current is induced in the rock. At two other electrodes (M-N) the voltage is measured. From the measured voltage, the measured current, and the positions of the electrodes involved, an apparent resistivity is calculated. From many single measurements the resistivity distribution in the rock can be calculated.

To determine the properties of the excavation-disturbed zone around some drifts at the Äspö HRL, a geoelectric array of 299 single electrodes has been installed in three drifts and one borehole around the area of investigation. Figure 1 shows the distribution of the electrodes in the different tunnels. The electrodes have a mutual spacing of 0.5 m in the drifts and 1.0 m in the borehole. For the successful application of the resistivity tomography it is essential to cover the fourth side of the area with a borehole. Since many single measurements are necessary to determine the resistivity distribution in this area between the ZEDEX- and the Demo-tunnel, an automatic recording unit is installed in a container at the end of the Demo-tunnel. With the use of a computer, any four of the 299 electrodes can be addressed for a measurement. This computer is connected to a telephone line by a modem and can be controlled from Braunschweig / Germany. The data are transferred by this connection too. Repeated measurements allow the assessment of the temporal variations in the resistivity distribution.

Since the measurements along the profiles in the different tunnels do not cover the whole area between the tunnels, tomographic techniques have been applied to determine the more regional resistivity distribution. In contrast to the measurements along profiles in the different tunnels, the pairs of electrodes involved (A-B; M-N) can be located anywhere on the array. All the readings are inserted into a general inversion algorithm, which is almost the same process as for the well known seismic tomography. The result is a picture of the resistivity distribution in the whole area between the ZEDEX- and the Demo tunnel, but with lower resolution (figure 2; right).

Investigations on site evaluation for the planned two-phase flow (TPF) experiment in a fracture/matrix system were focused on the niche at the outer drift loop at tunnelmeter 2715 (Figure 3). Geological and thermographic methods were used for the test site investigation.

The test site was selected under consideration of the following criteria:

• an undisturbed and hydraulically representative rock mass,
• precise knowledge of predominating fracture systems,
• sufficient distance from existing or planned in situ experiments, to avoid undesirable influences.

From a macroscopic point of view, the major fracture systems as well as the matrix areas are saturated with water.

Although the geology is in reality more complex, the figure 4 (right) gives an impression of the major hydraulic pathways and the intersection of hydraulic structures. The front is characterized by subvertical and horizontal fractures which divide the matrix of the ÄSPÖ diorite into different blocks. The area between the major discontinuities is considered to be suited for the design of the planned two-phase flow experiment. Excavation of this area by smooth blasting prior to the instrumentation of the test area is planned.

The temperature distribution on rock surfaces in the niche at meter 2715 was determined by means of thermographic measurements. Figure 4 (left) shows the geometry of the scanned area and the measured overall temperature distribution in the niche. The front face of the niche (5.9x3.6 m), the south wall and the north wall (each 3.6x2.2 m), including the transition zone to the roof, were scanned.

The measured temperature at the side walls of the niche is in the range between 11.7 and 12.5°C. In relation to the resolution of the thermographic measurement system (+/- 0.1°C), the temperature distribution is considered to be homogeneous. ‘Warmer’ areas (mean temperature plus 0.2°C) are measured on the left side of the center and at the north wall near the floor. The only ‘cold’ area (mean temperature minus 0.2°C) was found in the upper part of the north wall.

In Figure 4 (right) the temperature distribution of the front face is illustrated. The deviation of the color at the edges from mean values of the image is caused by the detection limits of the scanner. Only some correlations are given between temperature distribution and results of the geological mapping. Because no specific anomalies were measured, the temperature distribution can be considered as homogeneous. This leads to the interpretation that the front face is totally water-saturated.

The distribution of the dissolved gases and their release are very heterogeneous and vary by a factor of at least 10 within the mine. The possible correlation of higher gas contents with mineralogy, stratigraphy, existing fissures, or tectonically fractured zones has not yet been investigated. Further investigations concerning this subject are considered necessary.

As far as two-phase flow experiments in granite are concerned, this investigation shows that gases stored in the granite or dissolved in the formation water have to be taken into account.

The results of the geoelectric measurements obtained so far show that the conventional method of resistivity mapping as well as the tomographic method are applicable for estimating the water distribution in large areas and around drifts. Especially the temporal variations can be monitored very well by repeated measurements. Quantitative correlation of the resistivity with the water content is possible, but the accuracy depends on the validity of the different assumptions, for instance, 2-dimensionality, unfractured rock etc.

The geological mapping in the niche at tunnelmeter 2715 indicates different water-bearing structures. The rock mass temperature distribution was found to be homogeneous. The recent explanation of these two diverging results is that the rock is totally water-saturated, and that evaporation therefore does not affect the temperature distribution of matrix areas.

Under consideration of the preliminary selection criteria, the results of the geological mapping lead to the conclusion that the location at tunnelmeter 2715 will be appropriate for the performance of the planned two-phase flow experiment. Especially with respect to the modeling part, the configuration of vertical and horizontal fractures in the near field of the niche guarantees well defined boundary conditions of the hydraulic flow field. The possible influence of artificial fractures on the flow field can be minimized by removal of the damaged front of the niche by additional smooth blasting.

The report is funded by the German Federal Ministry for Education, Science, Research, and Technology (BMBF) within the contract 02E 8936.