Exploring Dyke Geometry Effects on Tunnel Stability Using Particle Flow Code

Istanbul is a unique city spanning two continents and one of the world's most populous cities. There is a significantly increasing demand for the construction of deep engineering structures, particularly transportation and infrastructure tunnels because of a population exceeding 16 million. Numerous projects, including intercontinental strait crossings, will be implemented over the coming decades to meet the city's needs. Istanbul’s complex geological structure, rugged topography, and earthquake risk make construction efforts more challenging. Generally, deep engineering structures in Istanbul are primarily constructed within the classic and carbonate rocks, which are part of the Istanbul Paleozoic Sequence. These sedimentary rocks are intersected by dikes composed of andesite, diabase, and dacite, which exhibit varying geometries. Due to the different engineering properties of these rock units, which have different geological origins, their behavior in deep underground excavation openings also varies. Engineering problems such as rock bursts, water inflows, and TBM (Tunnel Boring Machine) jamming arise during tunnel construction, leading to both cost overruns and time delay. In this study, the effect of dyke geometries on tunnel stability was analyzed using ITASCA’s Particle Flow Code (PFC), which employs the discrete element method (DEM). Models were calibrated based on field experiences and laboratory data of the host rocks and dykes obtained during tunnel construction on the Anatolian side of Istanbul. The uniaxial compressive strength and elasticity modulus values of the host rocks and dykes are 28 MPa and 46 MPa, 12 GPa, and 16 GPa, respectively. Initially, a base two-dimensional model with dimensions of 4 by 4 meters was calibrated using the flat-joint model, and upscaled to 80 by 60 meters to represent the tunnel environment. Additionally, models featuring singular and dual dyke geometries with different orientations were utilized in an environment containing a tunnel opening with a radius of 4 meters. Analyses were carried out on a total of 18 models, which included seven single dyke geometries with a dip angle varying between 0 and 90 degrees in 15-degree increments, six geometries featuring a secondary dyke with a horizontal dip angle of 0 degrees intersected by another dyke at 15-degree increments, and five models with two dykes positioned at dip angles of 15, 30, 45, 60, and 75 degrees relative to each other. Each model was run for up to 100,000 cycles under gravitational loading conditions, and data on force chains, fragmentations, stress, and strain values were collected using measurement spheres positioned at 0, 90, 180, and 270 degrees around the tunnel opening. According to the results obtained from the models, the highest deformation in the tunnel section was observed at 60 degrees in single dyke geometries, while in double dyke geometries, it was observed at 0-30, 0-45, and 0-60 degrees.