Numerical Modeling of Structure-Soil-Structure Interaction on Liquefiable Soils and Effects of Mitigation in Urban Settings

Hwang, Yu-Wei

Graduate Thesis Or Dissertation

Numerical Modeling of Structure-Soil-Structure Interaction on Liquefiable Soils and Effects of Mitigation in Urban Settings Public Deposited

Analytics

Download PDF

Citations

Citeable URL: https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/k643b229x

Abstract

Soil liquefaction has caused substantial foundation and structural damage in urban environments during prior earthquakes. To improve the seismic performance of structures founded on liquefiable soil deposits, a number of mitigation techniques have been studied over the past years, including ground densification, replacement, reinforcement, and enhanced drainage. The state-of-practice for designing many of these mitigation approaches is commonly based on empirical evaluation of liquefaction triggering and ground settlement in free-field conditions, ignoring seismic soil-structure interaction (SSI) near building structures. Additionally, well-documented case histories of the performance of liquefaction mitigation techniques are rare, particularly in urban settings, where structure-soil-structure interaction (SSSI) could affect the settlement and tilt patterns as well as damage potential of both treated and untreated adjacent structures. This lack of information (about the treatment location and size, foundation type, characteristics of individual structures and the cluster, or ground shaking) limits our understanding and ability to guide future mitigation design for structures, particularly in urban settings.

In the first phase of this study, we used three-dimensional (3D), fully-coupled, nonlinear finite-element analyses that identified the following critical ingredients to predicting the general trends in foundation settlement and tilt compared to centrifuge recordings: a comprehensive calibration of the soil constitutive model parameters, use of higher order elements, and a sufficiently large domain size in 3D. Subsequently, a limited numerical sensitivity study was performed to evaluate the seismic performance of multiple structures in terms of foundation settlement, tilt, and accelerations transferred to the superstructure. In this dissertation, we refer to seismic interaction near two adjacent buildings as “SSSI₂”, distinct from seismic soil-structure interaction (SSI) near one isolated structure, while SSSI₃₊ is used for interactions among multiple adjacent buildings (≥ 3) in a city block with soil. For the conditions considered, the numerical results identified the key parameters affecting SSSI₂and SSSI₃₊ as: (i) spacing between the two foundations (S) in relation to their width (W); (ii) the contact stress and geometry of the neighboring foundation-structure system; and (iii) building location and arrangement in 3D.

In the second stage of this research, 3D, fully-coupled, nonlinear finite element analyses, validated with centrifuge experimental results, were first used to evaluate the impact of mitigation mechanism (e.g., densification, enhanced drainage, or their combination) on SSI and SSSI₂on a level and layered, liquefiable deposit. Combined mitigation (i.e., densification with enhanced drainage) was an effective strategy to curb foundation’s permanent settlement and tilt near acceptable levels, compared to the models involving densification or drainage alone. For the conditions evaluated, the presence of mitigation method under one structure is shown to notably amplify foundation tilt as well as accelerations and damage within the superstructure of even an adjacent, untreated structure.at shorter spacings (S < W/3).

Later, the role of building properties and spacing as well as the location, geometry, and symmetry of ground densification on engineering demand parameters of interest were further explored numerically. The experimental and numerical simulations indicated that, for the conditions considered, the building properties and spacing, as well as the location, dimensions, and symmetry of mitigation, strongly affected the seismic performance of treated and neighboring untreated structures in terms of settlement, tilt, and deformations within the superstructure. For the spacings considered, ground densification effectively reduced the permanent settlement of two adjacent mitigated structure(s), particularly when densification depth covered the full depth of critical layer. However, the combination of SSSI₂and ground densification only below one in a pair structures notably amplified asymmetrical deformations below the foundations (hence, permanent tilt) as well as column strains, particularly for an unmitigated neighbor. This effect was strongest when closely spaced and when densification width approached building spacing. Finally, in the third stage of this research, a comprehensive and statistically-designed (through quasi-Monte Carlo sampling) numerical parametric study was pursued to develop a probabilistic predictive model for foundation’s permanent average settlement on liquefiable soils improved with ground densification. This part of the study focused on one isolated structure experiencing SSI. The nonlinear regression with lasso-type regularization identified the primary predictors of foundation’s settlement as: the cumulative absolute velocity (CAV) of the outcropping rock motion; total thickness of the soil deposit above bedrock and cumulative thickness of the critical liquefiable layer(s); foundation’s bearing pressure, size, and embedment depth; structure’s total height; the achieved density and size of ground improvement, and thickness of the remaining undensified susceptible soils within the foundation’s influence zone (≈1.5 times foundation width). The numerical database and the first of its kind predictive model aimed to guide the ground densification design that improves the performance of the soil-foundation-structure system on liquefiable soil deposits, setting the stage for future considerations of SSSI₂and SSSI₃₊.

Creator