The Oslofjord crossing includes 26.5km of new highway and will open in summer 2000 (Fig 1). It will replace the present low capacity ferry between Drøbak and Hurum, which connects the communities on both sides of the Oslofjord, about 50km south of Norway’s capital. Because of the hilly topography, the route includes seven bridges and six tunnels, several of them specified mainly for environmental reasons. The main tunnel is the 7.2km long Oslofjord subsea tunnel from Verpen in Hurum to Måna in Frogn.
Pre-investigations
The Oslofjord is situated in a major regional rift belt, the so-called Oslo graben, in which a total vertical displacement of about 2000m has taken place. In this area, the rock types on both sides of the fjord are Precambrian gneisses. The rift belt here consists of several faults and weakness zones, some of the major ones following the fjord over long distances.
Investigations for a possible road crossing the Oslofjord started as early as 1952. Acoustic measurements, reflection seismics and refraction seismic investigations were performed, mainly in 1979/80 and 1985/86, resulting in the present tunnel alignment, which follows a rock threshold in the fjord for approximately 2.1km. Due to the sensitive environment along the fjord, the rest of the tunnel crosses to portals well away from the shore areas on both sides.
Seismic investigations showed there were three wide channels in the rock, eroded by glaciation. In each of these channels major weakness zones were confirmed (Fig 2). At the tunnel low point, the weakness zone in the western channel, the ‘Hurum weakness zone’, showed the lowest seismic velocity of 2600m/s, with a relatively poor velocity contrast to the glacial soil deposits above. At this point, it was beneficial to reduce the rock cover to a minimum to shorten the tunnel length.
The Norwegian Public Road Authorities require 50m of rock cover as a normal minimum but allow less if favourable conditions are verified by investigations. Accordingly, two boreholes were drilled and seismic tomography was carried out, which confirmed the results from the refraction seismics. The tunnel alignment through the Hurum weakness zone was therefore located with a minimum rock cover of 32m.
The core samples were logged and tested in the laboratory. The material in the central part of the Hurum weakness zone consisted of familiar crushed rock material and clay, conditions similar to those of many such weakness zones crossing the 18 Norwegian subsea road tunnels already completed.
Facing the Hurum weakness zone
Excavation of the Oslofjord tunnel started in April 1997 from an access tunnel at Storsand, which had been included in the layout to provide early access to the subsea section and to reduce total construction time. The tunnel face from the west proceeded under the sea in the autumn of 1997. Because investigations had already revealed the Hurum weakness zone, comprehensive investigations were performed from the tunnel face, including percussion probe drilling ahead and above the tunnel face and core drilling ahead of the face. Systematic pregrouting with cement was also performed as the zone was approached.
On December 17 1997, percussion probe drilling showed the weakness zone in the upper part of the cross section contained soil materials with a mix of gravel and sand to clay under full water pressure of 120m (12 bars). To gain a first impression of the extent of the soil zone, percussion probe holes were drilled in three vertical sections: 5m to each side and in the middle of the tunnel. Where possible, the holes were drilled through the zone and into solid rock on the other side. 480m of percussion probe holes were drilled in this first investigation stage. In the upper left part of the tunnel, the zone seemed to contain clay, sand, gravel and blocks of solid rock, and water under 12 bar pressure. In the lower right part of the tunnel, the zone seemed to contain normal weakness zone crushed rock and clay with only minor water seepage.
Because of the extensive investigations ahead of the face, the excavation was stopped in fair rock at a safe distance of 17m before the zone. The rock mass at the face had low permeability, with negligible water seepage. Further investigations were started immediately to determine the extent and content of the zone.
Investigating and treating the weakness zone
A reference group composed of members of the client and the contractor and specialist external advisers was formed to advise on the investigations and the methods for preparing and crossing the soil zone. Based on information from the percussion probe holes, it was decided to drill four core holes through the soil materials in the upper part of the cross section under high water pressure. A casing with three different diameters was needed to ensure the drilling was a success. All four core holes were drilled through the zone 10-15m into solid rock on the other side. The core holes confirmed the extent and content of the zone. Fig 3 represents an extract of the information gained about the problems ahead of the face.
The upper soil zone consists of materials from rock block size to clay of local and non-local rock types, with sharp to well rounded edges. Some grains were rounded like glacifluvial material, and were obviously not normal crushed zone material as had been expected. A trench or narrow canyon had been eroded along the weakness zone by melting water at the end of the glaciation period, and filled in with loose deposits. The erosion depth was 35m, reaching below the assumed rock cover of 32m. The deposits were under full water pressure of 12 bar. Such conditions were not expected and it was possibly not feasible to pass through them with the methods stipulated in the contract.
Below the soil zone, the weakness zone consisted, as expected, of heavily jointed to crushed clay filled rock. The intersection between the upper soil zone and the lower crushed rock zone runs from the middle of the left wall to the abutment on the right side. Although pre-investigations had determined the location of the weakness zone, the directional core hole had missed the soil zone by being located a few metres too low. Knowing the extent and content of the weakness zone, the main question now was how to stabilise the zone before excavation. Two possible methods were evaluated: stabilisation by grouting and stabilisation by freezing.
Because of the water pressure, normal cement grouting was considered necessary as a first step for both methods. The results of drilling the grout holes and the grouting itself could be used to decide on the final stabilisation method. The main uncertainty was whether further grouting could confer enough stability of the soil or if freezing would be needed to obtain the required safety during excavation.
Systematic pregrouting started from the bottom of the face by drilling 30m long holes with the normal drilling jumbo through the crushed rock zone. Here, cement grout successfully ‘filled up’ the zone. In the soil zone above, cement grouting gradually reduced the water leakage to 10-100 litres/min/drill hole. Drilling problems made it impossible to drill more than 21-25m into the soil zone. However, because of the bypass tunnel, it was possible to carry out supplementary grouting from the ‘back side’ of the zone, where the main tunnel had been excavated up to approximately 20m from the zone.
At the beginning of June 1998, 700 tonnes of cement grout, including 25 tonnes of micro-silica and 40 tonnes of sand, had been pumped into the weakness zone from both sides. As a trial, 11m3 of concrete was pumped in but was discontinued for practical reasons.
Continuous drilling problems indicated that the soil zone would be difficult to stabilise by grouting. Large volumes of grout material would be needed, and it would be difficult to control where the grout went in the zone as permeable channels appeared to be present. In particular, there was concern that it could be difficult to control the result of the grouting. The zone would not only have to become reasonably tight but it would also be necessary to have a minimum stand-up time after exposure by excavation to allow support.
It was therefore decided that freezing was necessary to stabilise the zone before excavation, and no further grouting was performed. The decision was based primarily on the assumption that stabilisation by freezing would be easier to control. It was assumed that the bypass tunnel had bought enough time to complete the time consuming activities of drilling freeze holes, the freezing itself and final excavation and support.
Stabilisation by freezing
The freezing operation was mobilised by sub-contractor Geofrost on the west side of the zone, and all the freeze holes were drilled from this side. Excavation through the frozen zone was performed from the east side, taking advantage of the available access to the zone from both sides. The freezing liquid was salt water. Nitrogen was considered but was discarded for safety reasons. The design of the freezing operation was based on the assumption that sufficient stability could be gained with one row of freeze pipes. In the soil zone, the freeze holes were drilled with a spacing of 1m. The crushed rock zone had higher compaction and less water, and was easier to freeze, so here spacing was increased to 2m.
The tunnel cross section had to be extended to a width of 20m to gain enough space for the freeze holes. Long percussion probe holes were drilled above the tunnel to ensure sufficient rock cover in this area. An extensometer was installed in the tunnel roof and showed less than 2mm vertical displacement as a result of the extension of the cross section.
Parallel to the freeze holes, tests were performed on samples of material from the soil zone. Because of the water pressure, sand and clay were flushed out of the drill holes during drilling (Fig 4), which resulted in separation of the different fractions, making the real composition of the soil material difficult to determine. The design of the freezing operation was therefore based on the sand material as an assumed worst case.
Testing performed at SINTEF at the Norwegian University of Science and Technology, Trondheim, showed that, for the sand layer containing seawater, a temperature of -28°C was needed to penetrate to 2-3m for sufficient stability during excavation. To achieve this, a second row of freeze pipes was required 2-3m inside the first row in the soil zone. The diameter of the freeze holes, including casing, is 165mm. Despite the 700 tonnes of grout, the freeze holes were exposed to a full water pressure of 12 bars in the soil zone. To make the drilling safe, it was necessary to drill through a ball valve connected to a safety pipe grouted into the rock. In the crushed rock zone, drilling progressed quickly, with no problems, but it was very time consuming. It took about a year to drill a total of 105 holes, including replacing holes whose deviation was too large. The final design consists of 81 freeze holes and 15 temperature control holes (Fig 5).
Excavation of the remaining 40m started in August 1999 after about 12 weeks of freezing. The temperature had then reached -25°C in the soil zone. At the end of September, when driving through the weakness zone, the required temperature of -28°C was obtained. Excavation is performed by drilling and blasting in full cross section in rounds of 1.9m length in the soil zone. A layer of 300-400mm of fibre reinforced sprayed concrete, with alkali-free additives for fast build-up of thickness, is applied as immediate support directly on the exposed frozen ground. A 1-1.2m thick concrete ring lining (Fig 6) is cast and cured about one day before excavation of the next section.
The freezing will be turned off after excavation has progressed through the zone, leaving the concrete lining as permanent support. The concrete lining is designed to sustain a total pressure of 14 bars, estimated as full water pressure plus some soil pressure.
Excavation has shown that the investigations of the zone performed from the tunnel face gave a very good picture of the actual conditions, which include glacifluvial materials overlying a thin layer of moraine deposits. The intersection between the soil zone and the crushed rock zone is in the middle of the tunnel face as expected. Cement grout does not seem to have infiltrated the soil masses, but is found in channels. This confirms that freezing was necessary.
Each excavation round and support takes about a week, and crossing the frozen zone is estimated at about 15 weeks. Tunnelling through the Hurum weakness zone is expected to be completed at the end of 1999. Due to the problems with the soil zone and the need for two rows of freeze holes, this is significantly longer than anticipated. Such delays can have damaging effects on the schedule for most projects, but because of the bypass tunnel, tunnel breakthrough was achieved below the eastern wide channel on February 4 1999, six weeks ahead of schedule. Completion work is thus proceeding in parallel with freezing, excavation and support of the remaining 40m of tunnel, and it is expected that the commissioning date will not be affected.
The detailed requirements for procedures for pre-investigations to verify rock cover at critical low points of subsea tunnels are under review. Refraction seismics are still considered to be an invaluable tool for such investigations. It is not the results of these investigations that have been misleading but the interpretation of the results in rock engineering terms and the procedures for verification by directional core hole drilling that need improvement. As a first principle, probe drilling ahead and above the face during construction is the only real assurance against catastrophe when unexpected conditions occur in a subsea tunnel.
Related Files
Concrete Lining Cross Section
The Bypass Tunnel
Alignment of the Oslofjord Crossing
Cross Section of the Weakness Zone
Geological Profile
Expanded Cross Section