Prediction of ground surface settlement induced by twin tunnels

In recent years traffic congestion in big cities has increased due to urban development and transportation problems. This might lead to the use of underground transportation spaces such as tunnels. A megalopolis as Istanbul that due to the high request of transportation service have to continuously develop principally the metro system and also the links between the two continents. The Eurasia tunnel project is among the most significant investment regarding solutions of traffic problems, it is designed to contribute towards alleviating current pressure on the two bridges, providing as another road crossing of the Bosphorus. With this project, Asian and European sides are connected with a road tunnel crossing under the Bosphorus and allowed the link of Kazlıçeşme to Göztepe, where the vehicle traffic is intense in Istanbul. The first part of the Eurasia Tunnel comprises of two open U-section box with a ramp leading down into to the twin cut and cover tunnel on the European side. These two tunnels, which are 430 meters long, are then connected with a double-deck cut and cover tunnel, which runs 80 meters up to the ventilation shaft on the European side. The ventilation shaft is the initial point for the 3,34 km double-deck TBM tunnel, which is excavated by a mixshield tunnel boring machine designed to adapt to the complex ground conditions expected along the route including different rock and soil types. The vertical alignment of the TBM tunnel that links the two sides, descends at a slope of 4% to the deepest point located 95 m below sea level. The TBM section then rises at a slope of 4% to end at the Asian ventilation shaft north of Haydarpaşa container terminal. From the Asian ventilation shaft, the tunnel continues as a double deck New Austrian Tunneling Method (NATM) tunnel for approximately 340 m. It then splits into two single deck NATM tunnels for a further 670 m. The excavation that links the TBM excavation with the road is characterized from passing from double deck to twin tunnel by NATM excavation, after that, it becomes a cut and cover boxes. In congested urban areas, there is now a scarcity of available land resulting in the inevitable implementation of underground highways and the construction of underground metro tunnels, it is also inevitable that infrastructures will be constructed adjacent to the each other in large cities. Generally, underground transport systems consist of a double tunnel construction, also known as twin tunnels built close by. Twin tunnel solutions provide significant advantages such as reducing tunnel diameters and ground movements due to tunnel construction. The construction of tunnels in urban areas may have undesirable effects on the ground surface. Today, therefore, the main subject of tunneling activities in urban areas is the interaction between existing structures and tunnels. Surface settlements due to tunnel excavations affect the continuity and stability of existing structures near the excavation area and thus become one of the most important problems in the excavation process. The estimation of the size and distribution of surface settlements induced by tunnel construction is critical in the tunnel design process. Because these estimates are used to predict potential damages, to estimate the tolerance of neighboring structures to stimulated deformations and to design protective measures when needed. Surface settlements can be estimated using empirical methods, analytical methods and numerical methods. This thesis has been focused on the single deck twin NATM tunnels, whose construction started in Km: 9+900.00 and continued for 670 meters, within the scope of the Eurasia Tunnel Project built for the solution of the urban mobility problem in Istanbul. In this part of the project, the overburden thickness varies above the twin tunnels and they have been constructed under diagonal configuration. The lateral distance between the two tunnels' center-lines is 25.5 meters and the vertical distance is 3 meter. Surface settlements induced by twin tunnel excavation were examined on a section which is very close to the operating room of an eye clinic. In this section under a busy local road, the groundwater level is very close to the surface and the tunnels pass through the medium - weak strength sandstone. The aim of this thesis is to compare the results obtained with the numerical, empirical and modified empirical method with the real transverse surface settlement induced by twin tunnels. First, surface maximum settlement results were obtained for each tunnel at different depths using the Herzog (1985) model developed to estimate the maximum settlement value on a tunnel excavated by NATM. The transverse surface settlement pattern triggered by twin tunnel construction was achieved by the superposition of the settlement troughs obtained by using the empirical Gaussian distribution curve for two separate tunnels. With this approach, the effect of support on surface settlements and any interaction between the two tunnels were unfortunately ignored. As a further alternative to the empirical method, the modified Herzog model was used in which the effect of the pipe roofing method on surface settlements was taken into account. With this model, the effect of the diameter and longitudinal overlapping lengths of the pipe roofing system was neglected. Maximum surface settlement values based on modified Herzog model were calculated for each tunnel and vertical displacement profile induced by twin tunnels was obtained by superposition method. Numerical analyzes were performed by using the PLAXIS 2D finite element program. A soil model has been formed using the data obtained as a result of field and laboratory tests. The vertical displacement distribution on the ground surface due to twin tunnel excavation was examined by using the ''Mohr-Coulomb Model'' which can model the elastoplastic behavior of the soil. As a result of empirical and numerical analyzes, the maximum surface settlements due to the twin tunnel construction were compared with the measured maximum surface settlement in the field. According to the Herzog model, the maximum vertical displacements on the surface formed by the passage of each tunnel are approximately 7 mm, and the maximum settlement value obtained by the superposition of the settlement curves of the twin tunnels is 11 mm. With the effect of the Herzog model ignoring the effect of the reinforcement elements, 38% higher maximum surface settlement value than the measured data in the field was calculated. The estimated maximum surface settlement amount based on the modified Herzog model is approximately 4 mm for each tunnel. It is concluded that the maximum surface settlement value which is likely to occur by the passing of the twin tunnels is 7 mm by means of the superposition approach. Modified Herzog model, which takes into account the effect of the reinforcement elements in the calculations, is the solution that gave the most similar result to the measured data in the field. By applying the modified Herzog model, 12.5% higher maximum surface settlement value than the measured data in the field was obtained. The settlement trough, which is expected to occur over twin tunnels excavation, located on the center of twin tunnels. Because the superposition approach assumes that during twin tunnel excavation, the stresses caused by the excavation of each tunnel don't affect the excavation of the other. As a result of the numerical analysis carried out with the PLAXIS 2D program using the finite element method, the maximum vertical displacement occurring on the ground surface with the passage of the first tunnel is 3 mm, and at the end of the passage of the second tunnel is 4.7 mm. The calculated settlement amount was 41% less than the real settlement value and also less conformity result was obtained by numerical method compared to the other methods. However, parametric analyzes were performed to examine the sensitivity of twin tunnel behavior to the characteristics of the geological unit through which the tunnels pass. Young's modulus (E), internal friction angle (φ), Poisson's ratio (ν) and cohesion (c) properties of the sandstone layer which the tunnels pass are investigated. It has been observed that Young's modulus of the rock mass has an effective role on tunnel behavior, the amount of the surface settlement has increased with the decrease of the stiffness and the other parameters have no significant effect on the settlement trough. Within the scope of the parametric study, the effect of the ground relaxation coefficient (β), which takes into account the three-dimensional effects of displacements caused by NATM in two dimensional analysis, on the settlement trough has been examined. It was concluded that the increase in the value of the ground relaxation coefficient had no significant effect on the amount of the surface settlement. Finally, when the field observation results were compared with the predicted vertical surface displacements in the finite element method, the 41% smaller results were obtained at the finite element method. The back analysis was carried out to investigate the this difference. The parameters which correspond to the maximum surface settlement conditions in the field were estimated by using the approximate changes of the parameters in the pre-analyzed model, which were obtained as a result of the parametric study.