The Bolu Tunnels lie along Trans European Motorway (TEM) which is connecting Eastern Europe with the Middle East. The tunnels are approximately 3.0 km long, 40 m apart and have excavated cross sections more than 200 m2. In construction, New Austrian Tunneling Method (NATM) was used in soft ground. Due to the challenging ground conditions, many problems have been encountered during tunnelling. To solve these problems special construction techniques were adapted. To simulate and demonstrate the effectiveness of these construction techniques, 2D and 3D Finite Element Methods are utilized in this study. Through comparison between 2D and 3D modelling of advance of Bolu Tunnels, respective merits of these two approaches are investigated and the conditions under which shortcomings of the 2D approach become serious are identified.

The influence of the surface excavation and loading on the existing tunnels has been investigated using a Finite Element Method program, Plaxis 3D Tunnel. A parametric study has been carried out where the parameters were the depth of the surface excavation, the eccentricity of the excavation with respect to the tunnel axis, the height of the embankment fill and the stiffness of the soil. It is found that, excavations over the existing tunnels have a negative effect on the tunnel lining capacity since the unloading due the excavation reduces the normal forces and increases the bending moments. On the other hand, it is found that surface loading within the limits considered in this study is not critical in terms of the tunnel stability due to the increase in normal forces and decrease in bending moments.

The aim of the study is to develop new analytical fragility curves for the vulnerability assessment of tunnels based on actual damage data of tunnels obtained from past earthquakes. For this purpose, additional important damage data belonging to Bolu Tunnels, Turkey was utilized as a case study. Bolu Tunnels constitute a very interesting case from the earthquake hazard point of view, since two major earthquakes, 17 August 1999 Marmara and 12 November 1999 D&uuml / zce, occurred during the construction of the tunnels. The August 17, 1999 earthquake was reported to have had minimal impact on the Bolu Tunnels. However, the November 12, 1999 earthquake caused some sections of both tunnels to collapse. The remaining sections of the tunnels survived with various damage states which were subsequently documented in detail. This valuable damage data was thoroughly utilized in this study. To develop analytical fragility curves, the methodology described by Argyroudis et al. (2007) was followed. Seismic response of the Tunnels was assessed using analytical, pseudo-static and full-dynamic approaches. In this way, it was possible to make comparisons regarding the dynamic analysis methods of tunnels to predict the seismically induced damage. Compared to the pseudo-static and full-dynamic methods, the predictive capability of the analytical method is found to be relatively low due to limitations inherent to this method. The pseudo-static and full-dynamic solution results attained appear to be closer to each other and better represented the recorded damage states in general. Still, however, the predictive capability of the pseudo-static approach was observed to be limited for particular cases with reference to the full-dynamic method, especially for the sections with increasingly difficult ground conditions. The final goal of this study is the improvement of damage indexes corresponding to the defined damage states which were proposed by Argyroudis et al. (2005) based on the previous experience of damages in tunnels and engineering judgment. These damage indexes were modified in accordance with the findings from the dynamic analyses and actual damage data documented from Bolu Tunnels following the D&uuml / zce earthquake. Three damage states were utilized to quantify the damage in this study.

A completed part of the Bolu Tunnel at Elmalik side collapsed during the 1999 D&uuml / zce earthquake. In order to by-pass the collapsed section, a new tunnel route was determined. 474 meters of the new route, including two portals and double tubing, crossed through the weak to very weak rock units with intersecting fault gouge, excavated from Elmalik side. In this study, the characteristics of the rock masses and support classes are determined for new route of the Elmalik Side. Then, during the tunnel excavation, the deformations of temporary and permanent support systems were precisely measured and recorded. The support system properties as determined from NATM were analyzed by two dimensional convergence confinement method using the numerical RocSupport software. As a result of this study, for weak ground tunneling, duration of primary support installation should be kept at minimum. Besides that, temporary support measures such as forepoling, face sealing and temporary invert have an important role in controlling deformations before the primary support installation. With the application of temporary supports, loading on the permanent support, and hence the final deformation of the excavation, was found to be reduced significantly. Application of rigid lining was found to be necessary in order to prevent long-term deformations in weak ground tunnels, even though it is contradictory to the NATM philosophy.

In many tunnel designs, reinforced concrete tunnel lining design is selected based on construction requirements rather than design loads. A constant cross-section is typically used along a tunnel even if the design loads change from one location to another, especially for tunnels constructed by tunnel boring machines (TBM). Factor of safety against failure is not constant along the length of tunnel and is typically high at shallow depth regions. Factor of safety during a rare event is usually much less than the ones set for service load states. Rare events such as earthquake, train derailment, explosion and long duration fires do not happen daily and generally a minor reparable damage is targeted at the structure during those types of events. The focus of this study is to analytically investigate structural fire safety of reinforced concrete circular tunnel linings in terms of reduction in service load safety and to develop recommendations for preliminary assessment of structural fire endurance of circular tunnel linings. Analytical methods accounting for thermal non-linearity, material degradation, tunnel lining-ground interaction and fire time stages are available to assess the structural fire safety of the concrete tunnel linings. Analytical results are determined to be in good agreement with tunnel key segment hydrocarbon fire test.

The use of sprayed concrete lining is common in tunneling practice since it allows the application of non-circular tunnel sections and complex tunnel intersections. Low capital cost of construction equipment is also an important factor for the selection of the sprayed concrete lining. In general the use of sprayed concrete lining is referred as New Austrian Tunneling Method (NATM). Depending on the requirements regarding tunnel heading stability and limitations on tunneling induced soil displacements, tunnel cross sections often advanced by different construction sequences and round lengths in NATM. For the purpose of assessing the effects of excavation sequence, round length, soil stiffness and tunnel depth on surface settlements and on tunnel lining forces, a parametric study has been carried out, considering short-term and long-term soil response. Three dimensional finite element analysis are performed to model the excavation sequence and stress distribution around the tunnel lining during excavation. The parameters used in the parametric study can be listed as: tunnel diameter, tunnel depth, round length and soil stiffness. Existing analytical and empirical solutions, which are used for prediction of ground subsidence due to tunneling and forces on tunnel lining, are also reviewed in this study / and their predictions are compared with the results obtained from numerical analysis. This comparison also provides an opportunity to evaluate the performance of the existing efforts. The variations between the results obtained from different methods are discussed and it is concluded that the limitations of the existing methods are the primary reason of the variations between results.

Contrary to the linings of conventionally driven tunnels, the linings of tunnels bored by tunnel boring machines (TBMs) consist of precast concrete segments which are articulated or coupled at the longitudinal and circumferential joints. There are several analytical and numerical structural analysis methods proposed for the design of TBM segmental linings. In this thesis study, different calculation methods including elastic equation method and two dimensional (2D) and three dimensional (3D) beam &ndash / spring methods are compared and discussed. This study shows that in addition to the characteristics of concrete segments, the mechanical and geometrical properties of longitudinal and circumferential joints have significant effects on the structural behavior of segmental lining.