Engineering geological properties of Jurong formation for underground cavern excavation in Singapore
U Kar Winn
Date of Issue2019-01-31
School of Civil and Environmental Engineering
The excavation of underground hydrocarbon storage rock cavern project had recently been completed in Singapore’s sedimentary Jurong Formation. Geologically speaking, it belongs to Ayer Chawan Facies based on constituted rocks, location and stratigraphic position. Based on the field relationship and petrographic studies, it is very likely that the pyroclastic lava intruded into the muddy sediments of Ayer Chawan Facies while they were deposited in a wet and unconsolidated condition. As a result, peperitic textures were generated. Subsequent thermal metamorphism made the rocks harder than the similar type of typical sedimentary rocks elsewhere in the world. The LA ICP-MS U-Pb zircon age of pyroclastic rocks (lithic tuff) of the Ayer Chawan Facies is determined to be 240.6 ± 1.2 Ma (Middle Triassic) in this study. It is in the same age range assigned after age determination from previous finding on collected fossil species. The engineering geological properties of intact rocks were measured in different stages before and during the excavation work. Joint analysis was carried out with 350 numbers of representative joints, bedding and fault information, which have been recorded from the geological mapping in one storage cavern. Joint sets were analyzed by DIPS program and 1 bedding set associated with 4 joint sets were identified. The strength properties of intact rocks in laboratory such as point load strength, unconfined compressive strength, triaxial compressive strength, Brazilian tensile strength and shear strength of natural rock joints were determined. The hydrogeological properties of rock mass was analysed with the records of water ingress rate versus pressure measured from the observatory probe holes, water bearing zones and water recharging holes in water curtain galleries. The calculated k value is in the order of 1x10-6 m/s and the average value is 1.73x10-6 m/s. Detailed geology mapping was conducted on every excavated rock faces to evaluate the rock mass quality using Q and RMR systems and to determine qualitative Geological Strength Index (GSI). The quantitative GSI was also calculated by four different approaches and all the results are approximately in the same range from 45 to 55. The quantitative and qualitative GSI varied within a range of ±10. Afterward, the strength and deformation properties of jointed rock mass were determined based on the intact rock properties with the input of measured GSI and excavation disturbance factor. In this study, an intelligent approach based on the Mamdani inference mechanism was employed to predict UCS and quantitative GSI determination of sedimentary rock mass. The attained results are evaluated with those from statistical model using coefficient of correlation (R2) and root mean square error (RMSE). The joint sets analysis results were applied to the wedge analyses in full cavern excavation case to evaluate the possibility of formation of wedges due to bench excavation. The results shows that no additional wedges can be formed with investigated joint sets due to steep dip angle of joints and faults. 2 D finite element program for discontinuum (intact rock parameters) and continuum analyses (rock mass parameters) and 3D finite element program for continuum analysis were conducted to evaluate the cavern stability in subsequent excavation stages. The calculated vertical displacements were in the comparable range with measured values. The safe strength factors after full face of excavation were recorded from all analyses. The calculated rock bolts and shotcrete strength are much lower than the design capacity of them. The high horizontal to vertical stress ratio measured in the project also provided tunnel crown stability condition. Having competent sub-horizontal bedded layers intersected with sub-vertical joints, flat-roofed tunnel profiles were occasionally observed in the project. The stability of such tunnels were evaluated by analytical (Voussoir beam algorithm), empirical (stability graph) and numerical approaches. The abutment relaxation is unlikely to be occurred in the project since the sufficient pillar width between adjacent tunnels is designed to balance the overstressing caused by the combination of rock mass strength and in situ stresses.