Seismic performance of tunnel in liquefiable soil

Seismic performance evaluation of structures below the ground surface such as tunnels, manholes, pipelines, subway stations, deep foundation and lifeline systems is an important issue in growing modern urban areas. Therefore, needs for developing reliable tools and methods to decrease damage risk is a priority especially in high seismic risk regions. Evaluation of risk for designing an underground structure can be accomplished by gaining insight into the ground response during strong ground motion. During past two decades, uplift of underground structures buried in liquefiable site has been come to attention between scholars. In this thesis, the emphasize was to evaluate the uplift of tunnel due to geotechnical aspects rather than structural. In second chapter, some background information available in literature related to underground structures in liquefiable soils were given. Effect of different parameters such as buried depth, wide to buried depth ratio of tunnel, direction of ground motion propagation, permeability, friction angle and dilation angle of soil on tunnel behavior in liquefiable sand were explored. It should be noted that, current understanding on the influences of these parameters on the seismic behavior of tunnels is restricted due to lack of experiments or real case data. Present numerical methods are based on assumptions. Therefore, for the rationality and validation of numerical analyses can be established by carrying out experimental studies and numerical simulations cautiously. In fifth chapter, the procedure of finite element modeling for implementation in Opensees platform related to tunnel buried in saturated sand explained in detail. The most important issues that may occur during modeling such as chosen time step and adequate element mesh size clarified. Different available and practical boundary conditions procedure in geotechnical computation are described. However, in this research only absorbent boundary condition used and the effect of different boundary conditions on soil-tunnel system behavior was neglected. The numerical outline for free-field analyses consists more attention to the boundary condition to reduce the reflection of energy from far-field lateral sides of the domain and to represent structural and hydraulic interactions at the soil-tunnel interface. In the analyses procedure gravity analyses performed following by dynamic analysis to evaluate tunnel uplift behavior. The model analyzed with different ground motion amplitudes to investigate the effect of peak ground acceleration on tunnel uplift behavior. Additional evaluation performed for post-liquefaction behavior of tunnel in cohesionless saturated loose sand. Chapter 6 describes characteristics of the site response model used to perform the parametric analyses as described in Chapter 5. Single element level calibrations procedure was enlightened which were conducted for both cyclic and monotonic under drained and undrained conditions. The soil constitutive parameters calibrated against a commonly used semi empirical liquefaction triggering routine. As procedure of this constitutive model calibration, a new way was used for assuring that shear modulus reduction curve is compatible with soil shear stiffness. An extensive soil constitutive model calibration was carried out for PressureDependentMultiYield model (PDMY02) in Opensees. The aim was to capture more accurate response of the soil to be implemented in nonlinear 2D site response analyses for real liquefiable profile case studies. Calibration was only conducted for loose to dense sand with different blow counts of SPT test reported by researchers. Numerical evaluation was conducted with Single element simulation for monotonic and cyclic simple shear test. Different initial confining vertical pressures was considered to simulate soil behavior at different depths. The simulation results was compared with EPRI1993 recommended modulus reduction for sand. The result of calibrated parameters were in agreement with EPRI1993 recommended modulus reduction. Furthermore, the calibrated constitutive model parameters were used to validate the 2D site response of Wildlife Array site and Kobe port island site as real case studies. For Wildlife Array site, downhole array record of Superstition Hills earthquake was used as an input motion at the base of the model. For Kobe Port Island station site, the downhole array record at depth 32m of 1995 Kobe earthquake was used as an input motion. The acceleration, velocity and displacement time histories at ground surface level from simulations were in good agreement with recorded data at each site. In the second part of chapter 6, the calibrated model parameters are then verified against real case solutions for the linear behavior, and validated against down-hole array recording for Superstition Hills earthquake at the Wildlife Liquefaction Array (WLA) site and Hyogoken-Nanbu, Kobe earthquake at Kobe Port Island site. a series of parametric 2D site response analyses were performed to validate the soil parameters and to study geotechnical parameter that are typically unavailable in the empirical data. The resulting surface acceleration spectra are compared in a way that is comparable to the empirical method. The differences might be attributable to the inability of the 1D site response analyses to capture 2D and 3D effect such as surface wave and basin effect, or possibly shortcomings the in the 1D model's ability to faithfully represent all salient aspects of 1D wave propagation under liquefaction. Effect of different soil constitutive model parameters on uplift and settlement of soil structure system, effect of geometric of tunnel and soil domain on uplift and settlement of tunnel-soil system such as buried depth of tunnel and effect of height to wide ratio of tunnel and effect of ground motion characteristic's parameters such as frequency contents of motion, amplitude of ground motion and duration are the main parameters effecting behavior of tunnel. On the other hand, investigation the effect of geometry and ground motion characteristics, only frequency content and acceleration amplitude, on tunnel-soil system in liquefiable soil have been widely evaluated in literature due to simpler procedure for implementation. In this research, emphasize was on evaluating effect of combined horizontal and vertical excitation, post-liquefaction phase on uplift and settlement of tunnel-soil system and effect of different relative density on tunnel uplift. In terms of evaluating the effect of horizontal and vertical excitation combination, for vertical to horizontal maximum amplitude ratio higher than unity (V/H>1) an increase in uplift displacement of tunnel was observed. On the other hand, combination of horizontal and vertical excitation with V/H less than 1 will reduces the uplift displacement as it damps out the effect of horizontal excitation. Sand with higher relative density did not liquefied due to generation of less pore water pressure. Therefore, dissipation begins earlier compare to sand with smaller relative density. The rate of uplift and settlement displacement of tunnel increases as relative density decreases. For evaluation of post-liquefaction consolidation settlement of tunnel numerical simulations using solid-fluid fully coupled effective stress were conducted in OpenSees open source code platform. horizontal components of 1995 Kobe earthquake records was selected and first 40s of records only considered in analyses without any change in frequency contents of motion. The soil assumed as medium dense Nevada sand with relative density Dr=57% overlaying dense sand with relative density Dr=74%. Extracted results for Far-field, near-field and center nodes of domain. To evaluate the tunnel behavior with the liquefiable soil, critical points within the soil domain are considered horizontally and vertically. Both uplift of tunnel near-field and settlement in far-field in sand with higher relative density are smaller as relative density increases.