Water in the surrounding rock mass flows into the tunnel via naturally occurring joints and via cracks caused by the blasting used to excavate the tunnel. The most common method in Sweden to reduce or prevent leakage problems are first and foremost the use of grouting. However, experience shows that despite extensive pre-grouting and supplementary post-grouting, it is difficult to seal the rock mass so that drips and moisture are completely eliminated. Although the water itself causes degradation of the tunnel, the degradation process increases dramatically when the water is exposed to freezing temperatures. Water expands during freezing and due to water migration, which occurs in rock in a similar way as in soil, the ice causes frost shattering of the interface between rock and shotcrete and also to the shotcrete and the rock itself. This can damage the main load-bearing system. The ice formation itself is a maintenance problem, as the tunnels must be kept clear of icicles, ice pillars and ice layers in the tracks or on the roads. One of the main tasks in this research project has been to identify which problems cause the most maintenance work and where and when these problems occur in the tunnel.
During the field observations carried out as part of this doctoral study, many problems with water and ice were discovered, all of which contribute to increased maintenance. Many ice problems are directly linked to frost insulated drain mats. Leakage and ice formations occur at the edge of the drains, in mat splices and when brackets for cable racks, handrails or other installations puncture a drain and it has not been properly sealed. In drains covered with shotcrete, frost shattering and cracking in the shotcrete can be a problem. Frost cycles in the tunnel cause the water to freeze and thaw alternately, allowing more water to reach the freezing area due to water migration, resulting in frost shattering of the rock and the shotcrete. If not anchored with bolts, the reinforcing effect and the stability of shotcrete in a tunnel is dependent on the adhesion to the rock surface. It is, therefore, important to take all available measures to ensure good adhesion. Poor adhesion in itself is not a degradation problem, but a void can form in the interface between rock and shotcrete as a result of poor adhesion. If this void is filled with water that cannot drain away, ice pressure can occur in the layer between rock and shotcrete. The ice pressure can cause cracking and degradation of the shotcrete if the pressure exceeds the tensile strength of the adjacent material. In some of the reported fall-outs of rock and shotcrete, an ice layer was discovered between the rock surface and the edges of the remaining shotcrete layer. Therefore, frost shattering is a likely cause of the fall-outs. Many frost cycles combined with water leakage can cause frost shattering. The field measurements conducted as a part of the doctoral study have shown that most frost cycles do not occur closest to the tunnel entrances, but instead about 100 to 200 m into the longer tunnels. The results from the laboratory tests performed as part of the doctoral study showed that the adhesive strength between rock and shotcrete decreased significantly when the test panels were subjected to freeze-thaw cycles. Furthermore, more of the micro seismic events (AE - acoustic emission monitoring) occurred in the test panels that had access to water during freezing. Therefor, maintenance personnel and inspectors should pay particular attention to water leakage in sections that have an increased number of frost cycles, to avoid future problems with frost shattering of rock or shotcrete.
In the longer tunnels studied in this work, a greater number of ice formations occurred in the inner parts of the tunnel, than close to the entrances. The rock mass emits heat, which heats up the cold outside air that enters the tunnel. Due to the heat transfer from the rock mass, leakage points located further along the tunnels can remain unfrozen. A leak that is closer to the tunnel entrances in the longer tunnels or a leak in a shorter tunnel are exposed to higher freezing rates. The entire rock mass freezes and the leak ‘freezes dry’, that is, ice forms in the water-bearing fracture, preventing further water leakage.
Where and when ice problems occur along a tunnel depends on many factors. Besides the obvious water leakage, the length of frost penetration into the tunnel is the main reason for where and when ice problems occur. The predominant cause of frost penetration in most of the tunnels is the thermally induced airflow. In the longer tunnels, the inclination of the tunnel affects frost penetration the most. The field observations showed that there was a difference in where and when leakage points appear during the year and also in terms of variation in the amount of leakage water. There was also a variation over different years. The conclusions of the field observations are that it is difficult to estimate where the insulated drain mats should be located along a tunnel. Based on experience from this survey, the location of the drains should be determined only after several inspections and especially after a winter period, when the main problems with ice formation occur. Previous perception regarding ice problems have been that ice formation only occurs at the tunnel entrances and in the outer parts of the tunnel. A proposed measure has, therefore, been to cover the first 300 m from each entrance with frost insulated drains to try to completely eliminate the ice problems. However, this is not an effective solution to the problem. The insulation not only prevents the cold from reaching the leakage point, but it also prevents the rock mass from emitting heat that warms up the cold outside air entering the tunnel. Thus, the frost can penetrate further into the tunnel and the problems with ice formation are only moved further into the tunnel. As the amount and location of the frost insulation affects frost penetration, the dimensioning of insulation must, therefore, be carried out in several iterations, where each new distribution of insulation along the tunnel is calculated separately.
For the tunnels that have been studied as part of this doctoral study, the following has emerged. The central and southern parts of Sweden have shorter cooling periods and the tunnels are exposed to many temperature fluctuations around 0°C during the winter. The frost does not have time to penetrate as far here as in the tunnels in the northern parts of Sweden. Therefore, more ice problems arise around the entrances of the tunnels in the southern parts of Sweden than for those in the northern parts. For northern parts of Sweden, the problem of growing ice formations in sections near the tunnel entrance usually occurs only during the autumn and spring, but not in winter. The field observations showed that the problems with ice growth and temperature fluctuations around 0°C occur further along the longer tunnels in the northern parts of Sweden. This is because the temperature of the tunnel air is higher due to heat transfer from the rock mass. For shorter tunnels that adopt the same temperatures as the outside air, ice formations can occur along the entire length of the tunnel in the sections that have leakage problems. The Swedish Transport Administration’s regulations are currently being updated and the observations and measurements carried out in this doctoral work are now being used to evaluate new requirements regarding frost penetration in tunnels.