Numerical study on the effect of tunnel on liquefiable soil on the surface settlement

Abstract

Earthquakes are among the most devastating natural hazards, often causing widespread surface and underground infrastructure damage. In seismically active regions, such as Türkiye, the protection and resilience of underground structures such as lifeline systems (including water, gas, and communication pipelines), tunnels, and manholes are paramount. Damage to these systems during earthquakes can severely disrupt emergency response, delay recovery operations, and lead to long-term socio-economic impacts. Recent seismic events, particularly the February 6, 2023, Kahramanmaraş earthquake sequence, have demonstrated the vulnerability of buried infrastructure under complex fault rupture scenarios and prolonged strong shaking. These observations have intensified the need for advanced engineering approaches to assess and mitigate seismic risks to underground systems. This thesis focuses on the dynamic behavior of structural systems embedded in soils with liquefaction potential, emphasizing nonlinear soil behavior and soil–structure interaction under strong ground motion. As essential components of urban infrastructure, buried tunnels must remain operational after major earthquakes, especially in seismically active regions like Türkiye. Recent destructive events, including the February 6, 2023, Kahramanmaraş earthquakes, have highlighted the necessity of accurate modeling approaches to consider the complexities of earthquake and soil response, particularly under conditions involving multi-fault ruptures and long shaking durations. This study employs a series of numerical modeling approachs using the Finite Element Method (FEM), consuming the OpenSEES platform as a pre- and post-processing environment. Initially, a free-field analysis was conducted without a structure to meassure the soil's independent seismic response. Within this framework, a 2D fully coupled effective stress analysis was developed. Subsequently, the tunnel was incorporated into the model, and dynamic analyses were repeated. A two-dimensional plane strain model was developed to represent a typical buried tunnel configuration. The soil media is 100 meters long and 30 meters deep, and contains a square tunnel with a dimension of 4 meters, buried 2 meters below the surface. The groundwater table was accepted to be at the ground level. The model was meshed using 9-node quadratic quadrilateral elements with pore pressure capabilities (QuadUP), and a 1-meter element size was selected to ensure sufficient numerical resolution without excessive computational cost. The PDMY02 (Pressure-Dependent Multi-Yield 02) constitutive model was used to model the complex behavior of the soil during seismic excitation. This model records fundamental features of soil performance, including excess pore water pressure generation, displacements in both horizontal and vertical directions, and shear strains. Model parameters were validated to ensure accuracy before application to the fullscale system. Nevada sand with various relative densities was utilized during this thesis to study the effect of Dr on the dynamic behavior of sands. Various seismic inputs for the dynamic analyses were the Antakya and Arsuz ground motion records from the 2023 Kahramanmaraş earthquake, which is characterized by strong intensity and complex rupture patterns, and Sakarya and Izmit records from the 1999 Kocaeli earthquake. The inputs were horizontally applied as shear waves at the model's base to study the effect of maximum acceleration and loading time on dynamic effects. This study compared results from two cases, free field and soil tunnel system, to understand how the structure influences ground response. It examined how relative density and input motion characteristics impact the seismic response of buried tunnel systems. Relative density significantly affects the stiffness and strength of granular soils, impacting how underground structures behave during earthquakes. This study examined how altering soil density in numerical models influences ground response and tunnel system behavior. Understanding relative density is crucial for predicting the liquefaction potential of soil and its influence on tunnel deformation and stability. The study examined how different seismic records' characteristics, such as amplitude, frequency, and duration, affect soil-tunnel interaction. Results indicated that variations in input motion significantly impact soil displacements, pore pressure development, and tunnel structure stresses.