Date of Award

Spring 1-1-2019

Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Shideh Dashti

Second Advisor

Abbie Liel

Third Advisor

Dobroslav Znidarcic

Fourth Advisor

J. Antonio H. Carraro

Fifth Advisor

Yida Zhang


Soil liquefaction and ground failure have been a major source of damage to slopes, embankments, and structures during previous earthquakes. The risk of liquefaction and associated ground deformations can be reduced by various forms of improvement, including the granular column technique. Granular columns are known to be cost effective and environmentally friendly, and they have been in use since the 1970’s to mitigate the liquefaction hazard. Granular columns mitigate the consequences of soil liquefaction through a combination of: 1) installation-induced ground densification, 2) increase in lateral effective stresses in the surrounding soil, 3) increase in shear stiffness, and 4) enhanced drainage. These mechanisms aim to reduce the liquefaction potential of the improved soil or the resulting deformations. However, the independent influence and contribution of these mitigation mechanisms (in particular shear stiffness and drainage) on excess pore pressures, accelerations, and lateral and vertical deformations experienced in level and gently sloped sites are not sufficiently understood to facilitate their reliable performance-based design. In addition, the net influence of granular columns on competing mechanisms leading to lateral and vertical deformations with or without a structure is uncertain and requires further investigation, particularly in the presence of stratigraphic variations and layering expected in natural deposits.

In this dissertation, centrifuge experimental results are presented to investigate the influence of dense granular columns and their properties on the performance of layered liquefiable deposits with stratigraphic variations with and without a structure. The first set of experiments enabled investigation of granular columns in level sites with uniformly layered liquefiable deposits as well as gently sloping sites with a slight variation in the thickness of the liquefiable layer. No structure was present in these tests. The spacing and drainage capacity of columns were varied. In the second set of experiments, the same granular columns were evaluated in non-uniform liquefiable deposits with a gentle surface slope and a level surface. The presence of a shallow-founded structure and its seismic interaction with soil was also evaluated in terms of the performance of the soil-mitigation-foundation-structure system.

The experiments showed that granular columns can be effective in reducing the lateral and vertical deformations in gently sloped sites only if closely spaced (area replacement ratios exceeding about 20%) and able to enhance drainage. Hence, it is critical in such conditions to avoid clogging in subsequent events. Test results also showed that a slight variation in the liquefiable layer thickness can produce large permanent lateral ground deformations, even in the absence of a surface slope or a structure, which could damage utilities and lifelines. Use of granular columns below the foundation could effectively reduce the magnitude of void redistribution and shear strain localization underneath the sand-silt interface, hence reducing net settlements, rotations, and lateral displacements. However, granular columns could transfer greater accelerations to and seismic deformations in the superstructure, depending on the amplitude and frequency content of the motion in relation to the structure’s modal frequencies.

This dissertation provides a comprehensive experimental database that aims to improve our fundamental understanding of deformations in layered liquefiable deposits of varying stratigraphy, when unmitigated and when mitigated with granular columns. Overall, the results point to the importance of considering even slight variations in surface slope, liquefiable layer thickness, as well as seismic soil-foundation-structure interaction when designing mitigation strategies in the context of system performance. However, additional physical and numerical modeling with a range