Sammendrag
Oslo is the fastest growing city in Europe. The prognosis indicate that the population growth will continue. Today 90% of the water supply to Oslo comes for Maridalsvannet at the water treatment plant at Oset. This makes the city very vulnerable to incidents that strikes either the source or the treatment plant. The municipality of Oslo is therefore in the process of building a secondary water supply. By January 1st 2028 the new water supply to Oslo will be ready. The new water treatment plant is located at Huseby under Husebyskogen. The treatment plant consists of six large caverns with cross sections varying between 20m x 24m to 26m x 43 m. In addition, we are building an assembly chamber for the TBM that is going to excavate the water supply tunnel. The treatment plant makes three different levels and is a complex system of tunnels and caverns. The volume is approximately 100 000 m3 which is all excavated in a relatively small area. It was expected to be complicated to build the treatment plan. The plan was to use heavy rock support with arches of lattice girders in combination with rock bolts and shotcrete. Similar, but smaller and less complex caverns nearby is built using the same rock support. There was also identified several weakness zones, adding to the supposed need for lattice arches. VAV found it necessary to follow-up the effect the excavation of such a large volume on such a small area could have on the stress system in the rock mass and the potential deformations in the caverns. In collaboration with Skanska, SINTEF was engaged to model the development of the stress and strain before starting the actual excavation. Inspired by the Tripod-model that SINTEF had developed and used with e.g. BaneNOR at the Folloline project, we established a surveillance program to be used for calibration and verification of the model. The model has been used to identify the most critical areas in the caverns and to document that the roof of the caverns is not as critical for stability as first expected. Rock joints were mapped in the adjacent tunnels and available surface rock. Based on the mapped joints, the most critical wedges were identified, and the rock support decided based on the most critical wedges. The simulations and the measurements have given us the confidence to downsize the rock support in the caverns. The caverns are now supported by 20 cm thick shotcrete applied in two layers and rock bolts with length between 5 and 6 m applied between the two rounds of shotcrete. With help from the numeric model, we have identified areas that will be subject to frequent visual inspection throughout the excavating process. So far, we have fund very little fracturing or spalling of shotcrete. Before applying shotcrete, the caverns are scanned, using a LIDAR scanner. The scan will be used to identify the actual jointing of the rock mass and to verify that the most critical wedges are supported sufficiently by the applied rock support. Stress and strain are monitored continuously, and the measured values are compared to the modelled values. The model is updated and calibrated with the observations made in the surveillance program. This gives us a strong basis for deciding and verifying the rock support. So far, the measurements have confirmed that the rock support is sufficient for both local and global rock stability.