SlideShare a Scribd company logo
1 of 14
Download to read offline
Influence of Dense Granular Columns on the Performance
of Level and Gently Sloping Liquefiable Sites
Mahir Badanagki, S.M.ASCE1
; Shideh Dashti, M.ASCE2
; and Peter Kirkwood3
Abstract: Dense granular columns are often used as a liquefaction mitigation measure to (1) enhance drainage; (2) provide shear reinforce-
ment; and (3) densify and increase lateral stresses in the surrounding soil during installation. However, the independent influence and con-
tribution of these mitigation mechanisms on the excess pore pressures, accelerations (or shear stresses), and lateral and vertical deformations
are not sufficiently understood to facilitate a reliable design. This paper presents the results of a series of dynamic centrifuge tests to fun-
damentally evaluate the influence of dense granular columns on the seismic performance of level and gently sloped sites, including a lique-
fiable layer of clean sand. Specific consideration was given to the relative importance of enhanced drainage and shear reinforcement. Granular
columns with greater area replacement ratios (Ar), for example Ar greater than about 20%, were shown to be highly effective in reducing the
seismic settlement and lateral deformations in gentle slopes, owing primarily to the expedited dissipation of excess pore water pressures. The
influence of granular columns on accelerations (and therefore, the shear stress demand) in the surrounding soil depended on the column’s Ar
and drainage capacity. Increasing Ar from 0 to 10% was shown to reduce the accelerations across a range of frequencies in the surrounding
soil due to the shear reinforcement effect alone. However, enhanced drainage simultaneously increased the rate of excess pore pressure
dissipation, helping the surrounding soil regain more quickly its shear strength and stiffness. At short drainage distances or higher Ar values
(for example, 20%), this could notably amplify the acceleration and shear stress demand on soil, particularly at greater frequencies that
influence PGA. The experimental insight presented in this paper aims to improve our understanding of the mechanics of liquefaction
and lateral spreading mitigation with granular columns, and it may be used to validate the numerical models used in their design.
DOI: 10.1061/(ASCE)GT.1943-5606.0001937. © 2018 American Society of Civil Engineers.
Author keywords: Soil liquefaction; Granular columns; Drains; Centrifuge modeling; Lateral spreading; Site performance.
Background and Introduction
Past earthquakes have provided many examples of damage to
important geotechnical structures such as slopes, retaining walls,
and embankments caused by soil liquefaction. For example, exten-
sive lateral slope deformations were reported by Seed (1987) and
Tokimatsu and Asaka (1998) during the earthquakes in Niigata,
Japan, in 1964; Loma Prieta, California, in 1989; and Kobe, Japan,
in 1995. Mitigation techniques are often warranted to prevent ex-
cessive lateral deformations in slopes founded on liquefiable depos-
its and the subsequent damage to infrastructure. A reliable and
performance-based design of liquefaction remediation techniques
for slopes requires a clear understanding of the influence of various
mitigation mechanisms that control performance.
Bartlett and Youd (1992) characterized liquefaction-induced lat-
eral spreading in mild (0.3–5%) slopes underlain by loose,
saturated granular soils. Despite the low angle of these slopes,
the lateral deformations produced during earthquake loading could
still exceed several meters, leading to extensive damage. Many
ground-improvement techniques such as densification, reinforce-
ment, and methods that enhance drainage can be used to reduce
the risk of liquefaction and its associated ground deformations.
In particular, the installation of dense granular columns made of
gravel or stone is an attractive mitigation method for slopes and
embankments (Seed and Booker 1977). Depending on its installa-
tion procedure, this method is believed to mitigate the soil lique-
faction hazard and its consequences through a combination of
densification, increased lateral earth pressures, shear reinforcement,
and enhanced drainage (Baez 1995; INA 2001; Rayamajhi et al.
2016a).
Previous case histories have generally demonstrated a success-
ful performance of different types of compacted granular columns
in loose, saturated cohesionless soils (e.g., Mitchell et al. 1995;
Mitchell 1986, 1988; Adalier 1996; Baez 1996; Boulanger et al.
1998; Koelling and Dickenson 1998; ISSMGE 2001). However,
although valuable insights can and must be drawn from case his-
tories, the influence and relative importance of each mechanism of
mitigation on the performance of the site and slope cannot be re-
liably evaluated from case histories alone in a systematic manner,
as is necessary for a reliable, performance-based mitigation design.
Seed and Booker (1977) introduced an analytical method for the
design of drains based on radial consolidation or excess pore pres-
sure (Δu) dissipation that is commonly used in practice. They rec-
ommended the use of gravel or stone columns with a hydraulic
conductivity at least two orders of magnitude greater than that of
the surrounding soil, to avoid significant Δu generation in the
drains. This study had a number of limitations, including (1) the
assumption of an infinite granular column permeability and ignor-
ing the well resistance or clogging potential; (2) the assumption of a
1
Graduate Research Assistant, Dept. of Civil, Environmental, and
Architectural Engineering, Univ. of Colorado Boulder, Boulder, CO 80309.
Email: mahir.badanagki@colorado.edu
2
Associate Professor, Dept. of Civil, Environmental and Architectural
Engineering, Univ. of Colorado Boulder, Boulder, CO 80309 (corresponding
author). Email: shideh.dashti@colorado.edu
3
Research Associate, Dept. of Civil, Environmental, and Architectural
Engineering, Univ. of Colorado Boulder, Boulder, CO 80309. Email: peter
.kirkwood@colorado.edu
Note. This manuscript was submitted on May 19, 2017; approved on
March 29, 2018; published online on July 6, 2018. Discussion period open
until December 6, 2018; separate discussions must be submitted for indi-
vidual papers. This paper is part of the Journal of Geotechnical and
Geoenvironmental Engineering, © ASCE, ISSN 1090-0241.
© ASCE 04018065-1 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
level ground condition (no slope); and (3) not considering shear
reinforcement in the gravel or stone columns. In practice, however,
engineers sometimes rely on the shear reinforcement provided by
granular columns and use design procedures (based on shear strain
compatibility) to reduce the cyclic stress ratios induced in the sur-
rounding soil (Rayamajhi et al. 2016a).
Rayamajhi et al. (2016a) numerically evaluated the contribution
of the shear reinforcement provided by granular columns in isola-
tion, with and without the generation and redistribution of excess
pore pressures. Their nonlinear, coupled, three-dimensional (3D)
finite-element analyses indicated that granular columns experience
shear strain deformations that are not compatible with the surround-
ing soil. As a result, the reduction in the cyclic stress ratios induced
in the treated soil was notably less than that predicted by the
conventional design approaches that assume strain compatibility.
They subsequently provided a modified design approach to esti-
mate the reduction in cyclic stress ratios provided by dense granular
columns.
Elgamal et al. (2009) and Asgari et al. (2013) performed a
numerical parametric study with a unit cell and showed that granu-
lar columns could help reduce lateral displacements in treated
soil due to a combination of drainage and shear reinforcement.
Rayamajhi et al. (2016b) numerically showed that granular col-
umns can be effective in reducing lateral spreading through shear
reinforcement, even if liquefaction or generation of large excess
pore pressures is not prevented. They also numerically identified
the surface pressure, length and diameter of the drains, area replace-
ment ratios, liquefiable soil depth, hydraulic conductivity, and
slope angle as some of the key parameters influencing the lateral
displacement. However, the results presented in these numerical
studies were not validated with rigorous physical model studies
or case history observations.
Although limited in quantity and scope, a number of physical
model studies (for example, 1g shake table and centrifuge tests)
have been conducted to evaluate the influence of different types
of drains on site response and lateral spreading. Due to the diffi-
culties of drain installation in flight, the centrifuge experiments fo-
cused primarily on the reinforcement and drainage mechanisms of
mitigation, as opposed to installation-induced densification or an
accurate representation of the increase in lateral earth pressures
in the surrounding soil, which may not be reliable in all cases, even
in the field. As an example, Adalier et al. (2003) performed cen-
trifuge experiments with stone columns in a uniform and level de-
posit of saturated, loose silt with an overlying model of a rigid
footing. This study showed that stone columns can be effective
in reducing the foundation settlement by reducing shear strains
and generation of excess pore pressures in the underlying soil. They
did not, however, evaluate the influence of granular columns and
their properties on the performance of sites with more realistic
layering and geometry and under different earthquake motions.
Centrifuge tests were more recently conducted by Howell et al.
(2012) to evaluate the performance of liquefiable slopes mitigated
with prefabricated vertical drains (PVDs). This method did not in-
troduce notable shear reinforcement and relied primarily on the
drainage mechanism of mitigation. The study showed that PVDs
could be effective in expediting the dissipation of excess pore pres-
sures and reducing the resulting lateral slope deformations, depend-
ing on the characteristics of the earthquake motion. The combined
or independent influence of shear reinforcement and enhanced
drainage in slopes has not been experimentally evaluated. It is
not clear, for example, whether the added shear reinforcement pro-
vided by stiffer drains (e.g., dense granular columns in comparison
with PVDs) would reduce the seismic demand or the deformations
in slopes.
Dashti et al. (2010) performed a series of centrifuge experiments
to identify the dominant mechanisms of settlement on layered
liquefiable soil deposits. The study classified the primary volumet-
ric settlement mechanisms active in the free-field as (1) sedimenta-
tion or solidification (εp-SED), which occurs after significant
strength loss, as settling particles from the uppermost part of the
liquefiable sand accumulate at the bottom to form a solidified zone.
This zone increases in thickness with time and concurrently con-
solidates under its own weight; (2) reconsolidation (εp-CON), which
occurs as the excess pore pressures dissipate and soil’s effective
stress regains its initial value; and (3) partial drainage and loss
of water during cyclic loading (εp-DR) that occur according to
the 3D transient hydraulic gradients. Volumetric deformations
caused by partial drainage can be notable, as they occur while
the hydraulic gradients are kept at their peak during shaking.
Drainage-enhancing granular columns, such as those investigated
in this study, locally reduce the duration, or the amplitude, of the
elevated excess pore pressures. This, along with the altered shear
stress and strain profiles, likely reduces the contribution of sedi-
mentation to the volumetric strains. However, the increased rate
of drainage likely increases the volumetric strains due to partial
drainage. The net effect of granular columns on the competing
mechanisms leading to volumetric strains is uncertain and requires
investigation.
This paper presents the results of three centrifuge experiments
that systematically evaluated the influence of dense granular col-
umns and their various mitigation mechanisms and properties on
the seismic performance of level and gently sloping sites underlain
by a layered liquefiable deposit consisting mostly of clean sand.
The data obtained from these experiments facilitated (1) the mecha-
nistic evaluation of the influence of granular columns and their dif-
ferent properties (for example, the area replacement ratio, shear
reinforcement, and enhanced drainage) on the site and slope per-
formance in terms of the acceleration, excess pore pressure, settle-
ment, and lateral spread; and (2) the provision of data for the
calibration and validation of advanced numerical models in the
future. Following validation, such models can be more reliably
used in the design or development of design guidelines for the
mitigation of lateral spreading hazards using dense granular
columns.
Experimental Procedure
Centrifuge Testing Plan
A series of three centrifuge experiments comprising five separate
models were designed and conducted at the University of Colorado
(CU) Boulder’s 400g-t (5.5-m-radius) centrifuge facility. In these
tests, the primary goal was to systematically evaluate the influence
and relative importance of various granular column parameters on
the performance of a level site and a gentle slope underlain by a
layered liquefiable deposit during one-dimensional horizontal
earthquake loading. Fig. 1 shows the detailed plan and elevation-
view drawings of the model tests and their instrumentation layouts.
The dimensions are presented at both prototype and model scales
following the accepted scaling relations (Tan and Scott 1985).
For all tests, as shown in Fig. 1, a dense layer (8 m thick in the
prototype scale) of Ottawa sand F65 was dry pluviated to attain a
relative density (Dr) of approximately 90% at the bottom of a
flexible-shear-beam (FSB) container constructed of aluminum and
rubber at CU. Subsequently, a loose layer of Ottawa sand (8 m
thick) with a Dr ≈ 40% was pluviated as the liquefiable material.
This layer was subsequently overlaid by a 0.5-m-thick layer of
© ASCE 04018065-2 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
Silica silt (Sil-Co-Sil 120) to restrict the rapid excess pore pressure
dissipation vertically and to represent a case with void redistribu-
tion, followed by a 1.5-m-thick layer of coarse Monterrey sand 0/30
with Dr ≈ 90%, to create a nonliquefiable, thin crust. For all the
treated models (Models 0, 1, 3, and 4), the granular columns with
diameter = 1.75 m (in prototype scale) were placed vertically at the
bottom of the container prior to sand pluviation to avoid localized
densification during their installation. This was done to isolate the
influence of drains from ground densification and to keep the den-
sity of the surrounding soil controlled and uniform within each
layer.
Test 1 (Model 0) examined the behavior of a single draining
granular column in level ground, before adding the complications
associated with multiple drains and a slope. Vertical arrays of ac-
celerometers (Accs), pore pressure transducers (PPTs), and linear
variable differential transformers (LVDTs) were placed at three dif-
ferent radial distances from the single drain to track the wave propa-
gation, pore water pressure generation and dissipation, and
volumetric deformations throughout the soil profiles, as shown
in Fig. 1.
In Tests 2 and 3, all soil models were constructed of two
symmetric slopes on the two sides of the container separated by
an open channel, as shown in Fig. 1. In Test 2, one side of the slope
(Model 1) was treated with a grid of 1.75-m-diameter dense granu-
lar columns encased in geotextile filters (to avoid clogging) sepa-
rated by 3.5 m (from center to center) with an Ar of 20%. The Ar is
ratio of the area of granular columns to the total treatment area
(Baez and Martin 1993). The other side of the slope (Model 2)
in Test 2 was left untreated to evaluate the effectiveness of the
granular columns as a mitigation technique. In Test 3, both sides
of the slope (Models 3 and 4) were treated with 1.75-m-diameter
dense granular columns separated by 5 m (from center to center)
with Ar ¼ 10%. In Model 3, the granular columns were encased in
geotextile filters, similarly to Model 1, whereas in Model 4 they
were encased in geotextile and a thin (0.2-mm-thick) latex mem-
brane to evaluate the influence of stiffness alone without drainage.
The groundwater table was at the soil surface in all tests (the highest
elevation in the cases with slope), and all model specimens were
spun to a centrifugal acceleration of 70 g prior to the application of
earthquake motions.
Properties of Sand, Silt, and Gravel
The response of Ottawa sand F65 at various relative densities was
evaluated under monotonic and cyclic, drained and undrained tri-
axial tests prior to the centrifuge experiments presented here (de-
tailed by Ramirez et al. 2017). The engineering properties of this
sand were measured by the authors as follows: maximum void ratio
ðemaxÞ ¼ 0.81; minimum void ratio ðeminÞ ¼ 0.53; coefficient of
uniformity ðCuÞ ¼ 1.56; specific gravity ðGsÞ ¼ 2.65; and hy-
draulic conductivity ðkÞ ¼ 1.19 × 10−2
to 1.41 × 10−2
cm=s at
1g with water (which were the same at 70g with viscous fluid
70 times more viscous than water) for the relative densities (Dr)
of 90 and 40%, respectively. The critical friction angle (ϕcs) of this
sand was estimated at approximately 31° (Ramirez et al. 2017).
Coarse Monterey 0/30 sand was selected for the thin, high-
permeability, dense surface layer. The properties of Monterey sand,
obtained from previous studies, were emax ¼ 0.84; emin ¼ 0.54;
Cu ¼ 1.3; k ¼ 5.29 × 10−2 cm=s; and Gs ¼ 2.64 (Dashti et al.
2010). Silica silt (Sil-Co-Sil 120) was selected as the thin, low-
permeability cap above the liquefiable layer. The properties of silica
silt, also obtained from previous studies (Walker and Stewart 1989),
were Cu ¼ 7.3; k ¼ 3 × 10−5
cm=s; and Gs ¼ 2.65; and the critical
friction angle (ϕcs) was approximately 25°. In this study, the material
428mm 428mm
[30m] [30m]
Treated side
(Model 1)
Untreated side
(Model 2)
3°
60°
Silica silt
Monterey sand
(Dr=90%)
Ottawa sand
Dr=90%
Ottawa sand
Dr=40%
Gravel column
(Diam.=25mm)
[1.75m]
150 mm Model
10.5 m [Prototype]
0 75
5.30
Acc. (A)
PPT (P)
LVDT (D)
Latex
Shaking table
Without latex
(Model 3)
With latex
(Model 4)
Model 0
Test #1
Test #2
Test #3
177mm144mm158mm
[12.4m][10m][11m]
956mm
[26.3m]
376mm
[66.9m]
R
35
[2.5m
]
R121
[8.5m]
[17.4m]
R249
Shaking table
Porous stone
Shaking table
228mm
328mm
[16m]
[23m]
21mm
7mm
[1.5m]
[0.5m]
[8m]
115mm
[8m]
115mm
Shaking
Shaking
3°
Fig. 1. (Color) Three centrifuge experiments with their instrumentation
layout.
© ASCE 04018065-3 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
representing the granular columns was made of relatively uniform,
clean, medium gravel with Cu ¼ 1.54. It had the critical and peak
friction angles (ϕcs and ϕpeak) of 35° and 40°, respectively, at a dry
unit weight of 17 kN=m3
, obtained from isotropically consolidated
drained triaxial tests conducted by the authors. The hydraulic con-
ductivity of these columns was estimated at k ≈ 2.9 cm=s from con-
stant head permeability tests (roughly 200 times more permeable
than Ottawa sand F65). Fig. 2 shows the grain-size distribution
curves for all four materials used in this study.
Model Construction
The Ottawa and Monterey sand layers were prepared using the air
pluviation reconstitution method with the automated pluviator at
CU [Fig. 3(a)], to improve the repeatability among tests and the
uniformity within each model. The silt layer was dry pluviated from
a constant height using a #10 sieve to avoid lumps and distribute silt
across a controlled area. It was then gently compacted using a static
surcharge pressure of 5 kPa to achieve a dry unit weight of approx-
imately 16 kN=m3
with a final thickness of approximately 0.5 m in
the prototype scale.
To prepare the dense granular columns, geotextile filters were
formed to the desired column’s shape and diameter, and they were
subsequently plugged and glued at the bottom. Then the gravelly
soil was poured inside and tightly compacted in the geotextile col-
umn in three layers (previously calibrated to achieve the desired dry
unit weight of approximately 17 kN=m3
) and subsequently glued
from the top by adhesive tape, to prevent sand from entering the
columns during pluviation. For Model 4, the granular columns
were encased first in geotextile filters and then in 0.2-mm-thick
latex membranes. At their base, the membrane latex was sealed
with silicone sealant and end-cap plates. This sealing process pre-
vented any water in the soil from moving into the constructed
granular columns in Model 4 from the bottom or sides. The granu-
lar columns in Models 0, 1, and 3 were capable of drainage and
were not encased with a latex membrane. The shear modulus ratio
(Gr), ratio of the small-strain shear modulus (Gmax) of granular col-
umns to that of the surrounding loose Ottawa sand, was approxi-
mately 3. The Gmax value was calculated for sand and gravel on the
basis of the empirical procedures detailed by Seed and Idriss (1970)
and Seed et al. (1986), respectively.
The construction of dense granular columns inside geotextile
may not be practical in loose saturated sandy deposits, based on
the existing construction techniques. Nevertheless, in this experi-
mental study, the geotextile was used to prevent clogging and
preserve the drainage capability of granular columns during sub-
sequent motions. This design also provided an upper-bound con-
dition for the shear stiffness and drainage ability of these columns,
enabling an easier evaluation of the underlying mechanisms.
Fig. 2. (Color) Particle size distribution curves for all soil types used in
this study.
Fig. 3. (Color) Photographs taken (a) during air pluviation of Ottawa Sand; (b) after model completion of Test 2; (c) showing horizontal LVDT
holders and setup designed for the centrifuge experiments; and (d) after model completion of Test 3.
© ASCE 04018065-4 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
To create accurate 3° slopes that were consistent among the
models, the sand surface was vacuumed rather than scraped (to
minimize soil disturbance) with the aid of guide rails clamped
along the length of the container. After the model completion in
Tests 2 and 3, the top surface of Monterey sand was sprayed with
a sugar solution (5% by mass), lightly heated, and left overnight to
dry. This treatment kept the slopes intact during the model transport
across the lab floor to the centrifuge, but cementation effects were
removed when the sugar dissolved during saturation.
For all the treated models (Models 0, 1, 3, and 4), the granular
columns were placed vertically at the bottom of the container at
predetermined positions in a grid prior to the pluviation of sand.
Rows and vertical columns of colored sand (Ottawa) were also
placed at various locations throughout the model, to enable meas-
urement and evaluation of various modes of deformation during
model excavation. Figs. 3(b and d) show photographs of the fully
assembled models. Figs. 3(b and c) also show the setup for meas-
uring vertical and horizontal soil movements. The vertical LVDTs
measured the settlement of bearing plates placed on the soil surface.
The vertical LVDT rods were free to slide horizontally across the
surface of the bearing plate, while the plate displaced with the soil
owing to the nails placed around its perimeter. The nails had a
diameter of 0.9 mm and height of 9 mm in the model scale,
and their effect on the soil response was assumed to be negligible.
The horizontal LVDTs recorded lateral displacements along an ini-
tially vertical slot in an aluminum plate. This permitted the vertical
settlement of soil without influencing horizontal displacement mea-
surements [Fig. 3(c)]. The bearing plate for horizontal LVDTs was
embedded in Monterey sand to resist the overturning moment aris-
ing from the inertia of LVDT rods. To minimize this moment, the
LVDT rod was initially positioned near the base of the slot.
After dry preparation, the models were saturated with a hydrox-
ypropyl methylcellulose solution of 64 cSt viscosity, prepared per
Stewart et al. (1998) and measured before use. The viscosity was
70 times greater than that of water, to satisfy both the diffusive and
the dynamic scaling laws (Taylor 1995). A computer-controlled sat-
uration system was designed and implemented to improve the qual-
ity and rate of saturation, similar to that proposed by Stringer and
Madabhushi (2009). Initially, the soil model was flushed with CO2
from the bottom of the container for about 1 h, after which it was
kept under constant vacuum. The fluid tank was placed on a scale,
and its vacuum level was subsequently controlled automatically to
maintain a safe and constant flow rate below that required for flow-
induced liquefaction (in this case, 19 g=minute just prior to reach-
ing the silt layer and then 2 g=min, based on Stringer and
Madabhushi 2009). Complete saturation of each model required
approximately 96 h.
Properties of Input Motions
Once they were under 70 g of centrifugal acceleration, a series
of one-dimensional horizontal earthquake motions were applied
to the base of each model using a servo-controlled hydraulic shake
table available at CU (Ketchum 1989). These motions were scaled
versions of the horizontal component of recorded earthquake
motions, selected to cover a range of characteristics in terms of am-
plitude, frequency content, and duration, thus enabling evaluation
of the impact of these properties on system performance. Table 1
provides a summary of the sequence and properties of the first three
earthquake motions applied at the base of the model container.
Fig. 4 shows the acceleration response spectra (5% damped) and
the Arias Intensity time histories of the base motions achieved in
the centrifuge (the mean and their range of variability among the
three tests). Ample time was allowed between each phase of shak-
ing to allow and monitor full excess pore pressure dissipation. The
1992 Mw 7.3 Landers Earthquake recording at the Joshua Tree sta-
tion was selected because of its longer duration and slower rate of
energy buildup relative to the other records. The 1995 Mw 6.9
Kobe Earthquake recording at the Takatori station had a more rapid
buildup of energy and a shorter duration, but it was scaled to have a
lower amplitude in comparison with the Landers event (for exam-
ple, in terms of the PGA). Finally, the 1994 Mw 6.7 Northridge
Earthquake recorded at the Newhall-WPC station was a near-fault
record with a significant velocity pulse. The three selected motions
also differed in their frequency content, as shown in Fig. 4, for the
achieved or measured base motions.
Centrifuge Experimental Results
During all tests, the data were recorded at a sampling rate of 3,500
samples per second. In the following sections, all the test results
are presented and discussed in prototype units. The responses of
Model 0 (level ground with a single drain), Model 1 (slope with
36 draining granular columns and Ar ¼ 20%), Model 2 (slope
without granular columns), Model 3 (slope with 25 granular col-
umns and Ar ¼ 10%), Model 4 (slope with 25 granular columns
and Ar ¼ 10%, encased with latex membrane) were analyzed and
compared on the basis of the recorded accelerations, pore water
pressures, and vertical and horizontal displacements, in order to
provide insight into the effects of granular columns and their vari-
ous properties on site and slope performance. Due to space limi-
tations, only selected, representative results are presented.
Table 1. Ground motion properties recorded at the base of the container during Test 3
Motion
number
Ground
motion identifier Event Station PGA (g)
Significant duration,
D5–95 (s)
Mean period,
Tm (s)
Arias intensity,
Ia (m=s)
1 Kobe 1995 Kobe Takatori 0.38 11.9 0.86 1.92
2 Joshua 1992 Landers Joshua Tree 0.58 27.6 0.66 7.54
3 Northridge 1994 Northridge Newhall-WPC 0.75 16.6 0.94 4.94
Fig. 4. (Color) Acceleration response spectra (5% damped) and Arias
Intensity time histories of the base motions recorded during the three
tests. The solid lines show the average of the three tests, and the
highlighted regions show the variation among tests.
© ASCE 04018065-5 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
In this paper, toe refers to the location of the instrument array closer
to the slope toe but still within the slope, whereas top refers to the
location of the instrument array closer to the slope top and container
boundaries.
Influence of a Single Granular Column on Site
Performance
The primary objectives of Test 1 (Model 0) were (1) to evaluate
site performance in the far field and in the vicinity of a single granu-
lar column in a level, layered soil profile in terms of accelerations,
pore pressures, and settlements; and (2) to provide a data set for
the calibration and validation of numerical models in the future,
before adding the complexities of a slope and multiple drains.
Furthermore, drain systems are often designed for a unit drain
cell surrounded by other drains on all sides. The performance of
a drain outside the unit cell framework (for example, drains at
the edge of a group or in isolation) has not been adequately studied
experimentally.
Fig. 5 shows a summary of the time histories of excess
pore pressure, vertical displacement, and horizontal acceleration
recorded at three different radial distances from the single granular
column at different depths in Model 0 during the Kobe motion.
Liquefaction, defined as an excess pore pressure ratio (ru ¼
Δu=σ0
vo) of 1.0, was observed quickly in the loose sand layer. Large
excess pore pressures were also generated in the lower dense layer
of Ottawa sand, causing liquefaction during the first motion but at a
slightly slower pace. The accelerations generally showed large,
high-frequency spikes due to dilation within dense Ottawa sand.
These were attenuated in loose sand.
The pore pressure recordings showed that a single granular col-
umn was unable to prevent liquefaction even at a radial distance of
2.5 m, a result that was not surprising given that the drain spacing
was effectively infinite. However, the rate of excess pore pressure
dissipation was increased after shaking ceased, leading to a faster
rate of dissipation near the granular column, particularly at the
lower depths. The net dissipation of excess pore pressures was
slower at higher elevations because of the upward flow from lower
elevations. The influence of a single granular column considered in
this study (in terms of its effects on pore pressures) appeared to fall
within a radius of 2.5–8.5 m. Further, the ru values greater than 1.0
at the top of the liquefiable layer may indicate the possible forma-
tion of a water film below the silt interface and/or the slight settle-
ment of PPTs with respect to the surrounding soil at that location.
The slight movement of those PPTs with respect to colored sand
was confirmed during the excavation.
The dynamic total head isochrones were used to show the di-
rection and magnitude of transient hydraulic gradients at different
times and the resulting flow tendencies around the drain. Fig. 6
shows a comparison of the dynamic total head isochrones obtained
from the PPT recordings at three different radial distances from
the granular column center during the Kobe and Joshua motions.
At different radial distances, the isochrones indicated that liquefac-
tion (ru ¼ 1.0) was achieved quickly within the loose layer
of Ottawa sand during all motions. The hydraulic gradients were
generally formed upward from the dense layer toward the surface.
Fig. 5. (Color) Acceleration, excess pore pressure, and vertical displacement time histories during the first motion (Kobe) at three radial distances
from the single granular column in Test 1.
© ASCE 04018065-6 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
The presence of a drain led to a radially inward and upward flow
pattern in its vicinity, which slightly reduced the duration of large
excess pore pressures near the drain, particularly in deeper layers,
in comparison with the far-field (for example, a radius of 17.4 m in
this experiment) that experienced only vertically upward flow.
More rapid drainage, particularly for the deeper liquefiable layers,
would be expected to reduce the duration of large lateral spreads in
the slopes and benefit their performance; this is explored in the next
section.
The acceleration time histories showed spikes at large strains
after significant pore pressure generation and softening of the soil.
The amplitude of the acceleration spikes increased as the shaking
intensity and soil density increased in the subsequent events (not
shown, for brevity), which may be attributed to increased soil di-
lation and restiffening (Dashti et al. 2010). The relative density of
the loose sand layer increased from approximately 40% prior to
Kobe to about 60% prior to the Northridge motion on the basis
of far-field LVDT recordings, assuming uniform settlement across
the far-field.
Fig. 7 shows the time-frequency Stockwell spectra (Stockwell
et al. 1996; Kramer et al. 2016) of the transverse accelerations at the
container base as well as the bottom, middle, and top of the lique-
fiable layer during the Kobe motion. These plots enable visualiza-
tion of the spectral changes in acceleration over time, which is more
appropriate and insightful than a Fourier transform in a highly non-
linear and nonstationary system. In this figure, the color scale is
constant across all plots to facilitate comparison, and the peak
ground acceleration (PGA) measured during the response is shown
in the top right corner of each plot. Significant deamplification of
accelerations occurred from the base toward the surface due to ex-
cessive softening, particularly in loose Ottawa sand.
Within the loose Ottawa sand layer, the acceleration’s frequency
content varied with distance from the granular column. Fig. 5
shows that the excess pore pressures developed in this layer were
relatively unaffected by the column for the duration of strong shak-
ing. Therefore, enhanced drainage was likely not primarily respon-
sible for this change in acceleration’s frequency content. It is
proposed that the shear stiffness increment due to the presence
of a gravel column shifted the site’s modal frequencies as a function
of the radius (R). Owing to the increased shear stiffness, the lower
frequency shear waves (less than approximately 1 Hz) experienced
deamplification in soil closer to the column in comparison with the
far-field. The frequency bandwidth of the deamplification may
have been more extensive in the absence of enhanced drainage.
However, the high-frequency accelerations (around 4–10 Hz), were
amplified near the granular column in comparison with the far-field
due to a slightly faster dissipation of excess pore pressures, even
during shaking, and hence the instances of recovery in the soil shear
stiffness (and the reduction in damping). This hypothesis is tested
in subsequent sections in which the effect of drainage is separated
from that of reinforcement.
When used on a gently sloping site (explored in the next sec-
tions), the increased shear stiffness close to the granular column
may help limit the lateral spreading. However, designers should
Fig. 6. (Color) Dynamic total head isochrones at three different radial distances from the granular column during the Kobe and Joshua motions.
Fig. 7. (Color) Time frequency (Stockwell spectra) of transverse
accelerations at the bottom, middle, and top of liquefiable layer and
container base at three different radial distances from the granular
column during the Kobe motion. (The PGA value is noted at the
top right corner of each figure.)
© ASCE 04018065-7 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
be aware of a consequential increase in PGA. The PGA of the sur-
face motion (controlled by higher-frequency vibrations) and hence
the cyclic stress ratio (commonly used in liquefaction-triggering
analyses as the measure of seismic demand) were increased by ap-
proximately threefold when the radial distance (R) to the drain was
decreased from 17.4 m (PGA ¼ 0.21 g) to 2.5 m (PGA ¼ 0.61 g),
during the Kobe motion. The adverse influence of the granular col-
umns on surface PGA and the cyclic demand must be considered in
design.
The readings from vertical LVDTs provided a measure of set-
tlement (positive readings) at the top and bottom of the loose sand
layer during different shaking events at three different radial dis-
tances from the drain center, as shown in Fig. 5, during the Kobe
motion. Generally, significant settlements occurred both in the
dense and loose layers of Ottawa sand during shaking due to sed-
imentation and partial drainage (εp-SED and εp-DR), followed by a
smaller contribution from post-shaking consolidation (εp-CON). As
the distance to the drain decreased, settlements within both loose
and dense layers of Ottawa sand generally reduced. Even though
the drain tended to amplify volumetric strains due directly to partial
drainage (εp-DR), it was shown to reduce the net volumetric settle-
ments at the surface by limiting the duration of large excess pore
pressures and therefore the extent of sedimentation (εp-SED). The
patterns were similar during other ground motions.
Influence of Granular Columns and Area Replacement
Ratio on Slope Performance
In this section, the effectiveness of granular columns is assessed in
reducing the extent of excess pore pressure generation, accelera-
tions, and lateral spreading in a gently sloping site. The influence
and relative importance of the area replacement ratio (Ar) of granu-
lar columns was assessed experimentally. Models 2, 3, and 1 con-
tained a gentle slope treated with granular columns that had
Ar ¼ 0% (no drains); Ar ¼ 10%; and Ar ¼ 20%, respectively.
Fig. 8 shows a comparison of the dynamic total head isochrones
for the Kobe motion, which were obtained from the PPT recordings
for Models 1 through 3 at the depths shown in Fig. 1. These
isochrones show that liquefaction was achieved quickly within
the loose layer of Ottawa sand in Models 2 and 3 with the Ar of
0 and 10%, respectively. Large excess pore pressures were mea-
sured in the dense layer of Ottawa sand, but liquefaction was typ-
ically not reached in this layer in the presence of a gentle slope
(even for the untreated case of Ar ¼ 0%). Importantly, the greater
Ar of 20% in Model 1 successfully reduced the extent and duration
of large excess pore pressures in all layers in comparison with the
other two models. Model 3 (Ar ¼ 10%) also appeared to slightly
increase the rate of excess pore pressure dissipation in comparison
with the untreated Model 2 (Ar ¼ 0%), but not as well as Model 1.
These patterns were consistent during different motions.
Fig. 9 shows a comparison of the excess pore pressure ratio (ru)
time histories at various depths together with the vertical and hori-
zontal displacements recorded at the top and toe of the slope during
the three shaking events in the three models. The container base
acceleration time histories are also provided for comparison.
The horizontal displacement is considered positive in the fall line
or downslope direction for all models. The rate of excess pore pres-
sure buildup at all depths was generally decreased and the rate of
post-
shaking dissipation increased when the Ar increased from 0 to 10%
and to 20%, as expected. Similar to Model 0, the fastest rate of
dissipation was observed at lower elevations, due to an upward flow
tendency, which is also shown in Fig. 8.
The time-frequency response (Stockwell spectra) of transverse
accelerations recorded in the three models during the Kobe motion
is shown in Fig. 10. The figure shows that the amplitude of soil
surface accelerations generally tended to decrease at lower frequen-
cies over an extended period of time, as the Ar increased from
0 to 10% (the left two plots on the top row of the figure). On the
other hand, the acceleration amplitudes increased considerably
when Ar was increased from 10 to 20% (the right two columns in
the figure), particularly at frequencies greater than 0.7 Hz during
and after strong shaking. This led to a notable increase in PGA in
the top half of the liquefiable layer. This response may be explained
by the counteracting effects of shear reinforcement and enhanced
drainage provided by granular columns. For both treated Models 3
(Ar ¼ 10%) and 1 (Ar ¼ 20%), the added shear stiffness of the col-
umns alone was expected to reduce the amplitude of accelerations
and shear stresses in the surrounding liquefiable soil, particularly at
lower frequencies. In the experiments, an Ar of 10% provided shear
reinforcement, resulting in a slight reduction of accelerations in
comparison with the unmitigated slope (Ar ¼ 0%) at frequencies
less than around 1 Hz. This model also enhanced the drainage rate,
but not sufficiently to noticeably amplify accelerations at higher
frequencies. At the accelerometer locations (midway between the
columns), the accelerations were in this case only marginally am-
plified or preserved at higher frequencies.
For the greater Ar of 20%, the drainage rates were increased
sufficiently to prevent soil liquefaction in most cases. Therefore,
the average soil stiffness was increased and the damping decreased
in comparison with Models 2 (Ar ¼ 0%) and Model 3 (Ar ¼ 10%).
This impacted the propagation and amplification of accelerations,
particularly above the dense Ottawa sand layer. The result was
more intense accelerations measured near the soil surface (the top
of loose Ottawa sand) across a range of frequencies, concentrated
around the strain-compatible fundamental frequency of the
treated slope (approximately 0.7–0.8 Hz as determined from the
Stockwell spectra) due to the combined effect of reinforcement
and drainage. These observations are in line with those in Model
0 and were consistent during various motions, despite their varia-
tions in intensity, frequency content, and duration. These effects on
the amplitude and frequency content of the motions near granular
Fig. 8. (Color) Dynamic total head isochrones recorded in sloped sites
with granular column treatment of Ar ¼ 0, 10, and 20%, during the
Kobe motion.
© ASCE 04018065-8 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
columns must be considered in evaluating liquefaction triggering
and consequences as well as the design of overlying structures.
Fig. 9 shows the surface settlements at the top and bottom of the
slope for each of the three models. These settlements were domi-
nated primarily by volumetric strains and to a smaller degree by
deviatoric or shear strains, due to the slope’s small angle. For ex-
ample, the settlements recorded at the top of the slope in Model 2
were quite similar to those of a similar profile with level ground in
Model 0. The settlements recorded at the top of the slopes generally
decreased with an increasing Ar due to the reduction in the extent
and duration of large excess pore pressures and hence the reduction
in volumetric strains caused by sedimentation and deviatoric strains
caused by the existing static and dynamic shear stresses. These fac-
tors were more significant than the increase in volumetric strains
due to partial drainage. At the toe of the slope, the settlements were
more strongly affected by the deviatoric strains and rotational
movements from the slope above, at times causing bulging or heave
in the central channel and close to the channel. Larger deviatoric
displacements downslope, for example, in Model 2 (Ar ¼ 0%), led
to smaller net settlements at the toe of the slope in comparison with
the top, or with other models, during the first motion (Kobe) due to
the bulging effect induced by rotational movements toward the cen-
tral channel. The settlements generally reduced during the sub-
sequent motions due to densification in the prior motions as
well as the cumulative reduction in slope’s angle.
Lateral displacements were recorded in the slopes with the
LVDT setup shown in Fig. 3(c) and the locations shown in Fig. 1.
It is noteworthy that these measurements were sensitive to the
rotation of the slotted aluminum plate and in many cases malfunc-
tioned. However, in comparison with the plates placed
perpendicular to the slope in previous experimental studies, the
horizontal LVDT measurement system had a far smaller impact
on the response of the slope and the development of lateral spread-
ing. The results from horizontal LVDTs that were subject to
Fig. 9. (Color) Base acceleration, excess pore pressure ratio, vertical and horizontal displacement time histories recorded in sloped sites with granular
column treatment of Ar ¼ 0, 10, and 20%, during the Kobe, Joshua, and Northridge motions.
© ASCE 04018065-9 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
excessive rotation are not shown in Fig. 9. The successful record-
ings as well as measurements obtained during model excavation
highlight the effectiveness of the granular columns, particularly
with Ar ¼ 20%, in reducing permanent lateral spreading, which
is a critical measure of slope performance. This was due partly
to the added shear reinforcement and partly to the faster dissipation
of excess pore pressures. The contribution of each mechanism is
subsequently evaluated for one case.
Lateral displacements were generally reduced as the Ar in-
creased. These lateral deformations, governed by deviatoric strains
under the existing static and dynamic shear stresses, were in all
cases greatest at the toe (shown in the second row in Fig. 9) of
the slope in comparison with the top (shown in the first row in
Fig. 9), and could be unacceptable even with Ar ¼ 20% (that is,
on the order of 15 cm or more during the first motion). The per-
manent lateral displacements reduced significantly after the first
motion (Kobe) in the treated slopes (both Ar ¼ 10% and
Ar ¼ 20%), whereas the large lateral deformations continued in
the untreated model (Ar ¼ 0%). However, the slope angle was re-
duced significantly in the untreated Model 2 after the first two mo-
tions (from 3° to about 0.6° during the Kobe and to about
0° during the Joshua, calculated roughly with the two vertical
LVDT recordings on the soil surface), which indicated negligible
static shear stresses driving the permanent lateral deformations be-
yond this point. Following the first two motions, the lateral dis-
placement recordings in the untreated Model 2 were therefore less
reliable and were strongly affected by the rotation of the aluminum
plate (hence, not presented during the Northridge). Nevertheless,
the recordings indicated excessive lateral spread potential in an
untreated slope.
Fig. 11 shows the cumulative settlements (greatest at the top of
the slope), lateral displacements (greatest at the toe of the slope),
and slope angles (calculated roughly from the two vertical LVDT
recordings on the surface of the Monterey sand), as recorded during
the application of three consecutive motions, when available. The
results generally support the conclusion that using granular col-
umns reduces the volumetric settlements and deviatoric lateral
displacements in gentle slopes. The magnitude of this reduction
was shown to depend strongly on Ar, the characteristics of shaking,
the initial preshaking soil properties, and the slope geometry. The
use of granular columns was highly effective in reducing the mag-
nitude of lateral spreading. The vertical settlements were also re-
duced, but additional mitigation measures may be required to limit
the deformations for structures or other sensitive infrastructure.
Relative Influence of Shear Stiffness and Drainage on
Slope Performance
The influence and relative importance of mitigation arising from
the mechanisms of shear reinforcement and enhanced drainage
in granular columns were evaluated experimentally by comparing
the responses of Models 3 and 4. Both models contained a gentle
slope treated with granular columns at Ar ¼ 10%. In Model 3, the
granular columns were encased only in geotextile to avoid clog-
ging, combining the drainage and shear reinforcement mechanisms.
In Model 4, the dense granular columns were encased in both a
geotextile and a thin (0.2-mm-thick) latex membrane to avoid
drainage and provide only shear reinforcement.
Fig. 12 shows the dynamic total head isochrones for Models 3
and 4 during the Kobe motion. As expected, the use of draining
granular columns (Model 3) slightly reduced the net excess pore
pressures generated during shaking (for example, it reduced the
peak ru values to slightly below 1.0 in the liquefiable layer) and
increased the rate of dissipation after shaking in comparision with
the nondraining columns in Model 4 (with latex). Fig. 13 shows the
faster rate of the pore pressure dissipation in Model 3 during all
shaking events. The faster drainage in Model 3 limited the
settlements both at the top and toe of the slope by limiting both
Fig. 10. (Color) Time frequency (Stockwell spectra) of transverse
accelerations at the bottom, middle, and top of liquefiable layer and
container base in sloped sites with granular columns of Ar ¼ 0, 10,
and 20%, during the Kobe motion.
Fig. 11. (Color) Cumulative vertical and horizontal displacements re-
corded in sloped sites with granular columns of Ar ¼ 0, 10, and 20%.
© ASCE 04018065-10 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
sedimentation and shear deformations. Similar effects were ob-
served in the comparison of lateral displacements shown in Fig. 13,
both at the top and toe of the slope. The nondraining granular col-
umns with latex extended the duration of large excess pore pres-
sures, which notably amplified lateral displacements in the slope
(for example, an increase from 25 to 50 cm at the slope toe during
Kobe).
Fig. 14 shows a comparison of the cumulative vertical and hori-
zontal displacements, as well as rough estimates of the slope angle,
in Models 3 and 4 during the three consecutive motions (in a man-
ner similar to that shown in Fig. 11). A reduction of approximately
20% was observed in the total surface settlements after two motions
due to the draining effect of granular columns. Similarly, a reduc-
tion of approximately 74% was observed in the lateral toe
displacements after three consecutive motions due to enhanced
drainage, even though Model 3 had a slightly steeper slope follow-
ing the first shake. The vertical and lateral deformations in Model 4
with the nondraining columns were quite similar to those in Model
2 without any treatment, which indicated a negligible influence
from shear reinforcement on slope performance in terms of defor-
mations for Ar ¼ 10%. The photographs taken during model
Fig. 12. (Color) Dynamic total head isochrones recorded in sloped
sites with granular columns of Ar ¼ 10%, draining (without latex,
Model 3), and nondraining (with latex, Model 4).
Fig. 13. (Color) Base acceleration, excess pore pressure ratio, vertical and horizontal displacement time histories in sloped sites with granular
columns of Ar ¼ 10%, draining (Model 3), and nondraining (Model 4).
© ASCE 04018065-11 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
excavation (Fig. 15) and the deformation of initially vertical col-
ored sand columns confirmed the significantly larger shear defor-
mations in loose Ottawa sand in Model 4 with nondraining
granular columns in comparison with Model 3, and similar to
Model 2. The photographs also show minimum lateral deforma-
tions in Model 1, which was consistent with the recordings
shown previously. These observations point to the critical impor-
tance and benefits of maintaining the drainage capabilities of
granular columns (for example, reducing the possibility of clog-
ging in the field after subsequent events) and of not relying solely
on the shear reinforcement mechanism of mitigation for reducing
slope deformations.
Fig. 16 shows the time-frequency Stockwell spectra of trans-
verse accelerations at the container base and various depths within
the loose layer of Ottawa sand during the Kobe motion in Models 2,
3, and 4. The same color scale is used at all depths to facilitate this
comparison. For all models, a significant deamplification of accel-
erations occurred from the base toward the surface due to soil soft-
ening. The higher shear stiffness provided by the nondraining
granular columns in Model 4 reduced the accelerations over a range
of frequencies. In this case, the high-frequency accelerations were
reduced in Model 4 in comparison with the untreated Model 2,
which led to a notable reduction in surface PGA (for example,
during Kobe, 0.31 g in the untreated Model 2 and 0.19 g in Model
4 with nondraining columns). The trends were consistent in other
motions. But the enhanced drainage (even by a slight amount in
Model 3 with Ar ¼ 10%) was shown to nullify this reduction
(PGA ¼ 0.29 g) or even cause notable amplifications of
accelerations in the surrounding soil (for example, as observed pre-
viously for Model 1 with Ar ¼ 20%, or near the granular column in
Model 0). The influence of granular columns on accelerations is
therefore expected to depend strongly on the column’s drainage
Fig. 14. (Color) Cumulative vertical and horizontal displacements in
sloped sites with granular columns of Ar ¼ 10%, draining (Model 3),
and nondraining (Model 4).
Fig. 15. (Color) Deformation of colored sand rows and columns after
the test during model excavation in sloped sites: draining granular
columns of Ar ¼ 20% (Model 1), no drains Ar ¼ 0 (Model 2), draining
granular columns of Ar ¼ 10% (Model 3), and nondraining granular
columns of Ar ¼ 10% (Model 4).
Fig. 16. (Color) Time frequency response (Stockwell spectra) of
transverse accelerations at different depths in sloped sites with granular
columns of Ar ¼ 10%, draining (Model 3) and nondraining (Model 4),
and Ar ¼ 0% (Model 2) during the Kobe motion.
© ASCE 04018065-12 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
ability and Ar as well as the soil properties, preshake slope geom-
etry, and motion characteristics.
Conclusions
A series of highly instrumented dynamic centrifuge model tests
were performed to evaluate the effectiveness of granular columns
as a liquefaction countermeasure and its properties on the perfor-
mance of level and gently sloping sites. The primary conclusions
drawn from the experimental results are the following:
• Drains in isolation or at the corner of a group cannot be relied
upon for preventing liquefaction if a large influx of fluid from
the far-field is possible. The conventional design charts do not
account for such cases. However, even drains without adjacent
drains can be somewhat effective in speeding up the dissipation
of excess pore pressures in their vicinity after shaking, and they
tend to reduce net volumetric settlements.
• The use of granular columns in gentle, layered, liquefiable
slopes was found to be effective in reducing the lateral and ver-
tical deformations. The drain’s area replacement ratio (Ar) was
experimentally shown to be a critical parameter in reducing both
settlements and lateral spreading in the slope. In this study, the
best performance was achieved when liquefaction was pre-
vented. The data presented here will be useful for the validation
of future numerical models used for optimizing design pro-
cedures.
• The influence of granular columns on the seismic demand and
accelerations experienced by the surrounding soil was shown to
depend on their Ar, drainage capability, and the frequency of
interest. Generally, shear reinforcement provided by the col-
umns tended to deamplify the accelerations over a range of fre-
quencies and reduced the PGA. However, enhanced drainage
could simultaneously increase the rate of excess pore pressure
dissipation and recovery of shear stiffness (and hence, reduction
in damping) in its surrounding soil, amplifying accelerations,
particularly at higher frequencies. The possible increase in
accelerations (and in particular, the PGA) at the surface of a
treated, draining site must be considered in evaluating liquefac-
tion triggering, consequences, and response of the overlying
structures.
• The possible reduction in accelerations and seismic demand due
to the shear reinforcement provided by granular columns cannot
be entirely relied upon for reducing the likelihood of liquefac-
tion triggering nor its consequences in terms of deformation. For
the particular Ar investigated (10%) here, the beneficial influ-
ence of granular columns on slope deformations was shown
to be due almost entirely to their ability to enhance drainage,
and not to shear reinforcement. However, this may not be the
case for higher Ar values or for cases in which notable ground
densification is expected during column installation. Neverthe-
less, the experimental results demonstrate that the influence of
enhanced drainage is vital. It is critical for engineers to avoid
clogging in these columns by using appropriate geotextile
filters.
• For soils with a high silt content, the hydraulic conductivity, and
thereby the zone of influence, is reduced. Therefore, the benefits
of drainage will be less visible (Adalier et al. 2003). This may
necessitate the use of higher area replacement ratios or possibly
different mitigation strategies.
Despite the inherent simplifications and uncertainties in centri-
fuge modeling, these experimental results and discussion are aimed
to guide the future field, physical, and numerical model studies to-
ward more reliable and performance-based liquefaction mitigation
design methodologies. Additional tests considering a wider range
of soil, slope, drain, and ground motion properties are needed with
parallel nonlinear numerical simulations in order to clarify the seis-
mic response of liquefiable deposits treated by granular columns
and the relative importance of reinforcement, drainage, and ground
densification in different kinds of soils; improve the existing design
procedures; and provide recommendations for practitioners.
Acknowledgments
The authors would like to acknowledge the support of the depart-
ment of Civil, Environmental, and Architectural Engineering at the
University of Colorado Boulder and the assistance of Mr. Balaji
Paramasivam, Mohamed Elmansouri, and Simon Petit in the execu-
tion of the centrifuge experiments presented in this paper.
References
Adalier, K. 1996. “Mitigation of earthquake induced liquefaction hazards.”
Ph.D. thesis, Dept. of Civil Engineering, Rensselaer Polytechnic
Institute, 659.
Adalier, K., A. Elgamal, J. Meneses, and I. J. Baez. 2003. “Stone columns
as liquefaction counter-measure in non-plastic silty soils.” J. Soil Dyn.
Earthquake Eng. 23 (7): 571–584. https://doi.org/10.1016/S0267-7261
(03)00070-8.
Asgari, A., M. Oliaei, and M. Bagheri. 2013. “Numerical simulation of
improvement of a liquefiable soil layer using granular column and
pile-pinning techniques.” Soil Dyn. Earthquake Eng. 51: 77–96.
https://doi.org/10.1016/j.soildyn.2013.04.006.
Baez, J. I. 1995. “A design model for the reduction of soil liquefaction by
vibro-stone columns.” Ph.D. thesis, Dept. of Civil and Environmental
Engineering, Univ. of Southern California.
Baez, J. I. 1996. “Seismic ground improvement for bridge sites.” In Proc.,
4th CALTRANS Seismic Research Workshop. Sacramento, CA: Califor-
nia Dept. of Transportation, Engineering Service Center.
Baez, J. I., and G. R. Martin. 1993. “Advances in the design of vibro sys-
tems for the improvement of liquefaction resistance.” In Proc., Symp. on
Ground Improvement. Vancouver, BC, Canada: Vancouver Geotechni-
cal Society.
Bartlett, S. F., and T. L. Youd. 1992. Empirical analysis of horizontal
ground displacement generated by liquefaction-induced lateral spread.
Technical Rep. No. NCEER-92-0021. Buffalo, NY: State Univ. of
New York.
Boulanger, R., I. Idriss, D. Stewart, Y. Hashash, and B. Schmidt. 1998.
“Drainage capacity of stone columns or gravel drains for mitigating
liquefaction.” In Vol. 1 of Proc., Geotechnical Earthquake Engineering
and Soil Dynamics III, 678–690. Reston, VA: ASCE.
Dashti, S., J. D. Bray, J. M. Pestana, M. R. Riemer, and D. Wilson. 2010.
“Centrifuge testing to evaluate and mitigate liquefaction-induced build-
ing settlement mechanisms.” J. Geotech. Geoenviron. Eng. 136 (7):
918–929. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000306.
Elgamal, A., J. Lu, and D. Forcellini. 2009. “Mitigation of liquefaction-
induced lateral deformation in a sloping stratum: Three-dimensional
numerical simulation.” J. Geotech. Geoenviron. Eng. 135 (11):
1672–1682. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000137.
Howell, R., E. M. Rathje, R. Kamai, and R. Boulanger. 2012. “Centrifuge
modeling of prefabricated vertical drains for liquefaction remediation.”
J. Geotech. Geoenviron. Eng. 138 (3): 262–271. https://doi.org/10.1061
/(ASCE)GT.1943-5606.0000604.
INA (International Navigation Association). 2001. Seismic design guide-
lines for port structures. Tokyo: A.A. Balkema.
ISSMGE (International Society of Soil Mechanics and Geotechnical
Engineering). 2001. Case histories of post-liquefaction remediation.
Rep. Technical Committee for Earthquake Geotechnical Engineering
TC4. Tokyo: Japan Japanese Geotechnical Society.
Ketchum, S. A. 1989. “Development of an earthquake motion simulator
for centrifuge testing and the dynamic response of a model sand
© ASCE 04018065-13 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
embankment.” Ph.D. dissertation, Dept. of Civil, Environmental, and
Architectural Engineering, Univ. of Colorado Boulder.
Koelling, M., and S. Dickenson. 1998. “Ground improvement case histories
for liquefaction mitigation of port and near-shore structures.” In Vol. 1 of
Proc., Geotechnical Earthquake Engineering and Soil Dynamics III,
614–626. Reston, VA: ASCE.
Kramer, S. L., S. S. Sideras, and M. W. Greenfield. 2016. “The timing of
liquefaction and its utility in liquefaction hazard evaluation.” Soil Dyn.
Earthquake Eng. 91: 133–146. https://doi.org/10.1016/j.soildyn.2016
.07.025.
Mitchell, J. K. 1986. “Material behavior and ground improvement.”
In Vol. 2 of Proc., Int. Conf. on Deep Foundations. 1.1–1.17. Beijing.
Mitchell, J. K. 1988. “Densification and improvement of hydraulic fills.”
In Proc., Conf. on Hydraulic Fill Structures, 608–633. Reston, VA:
ASCE.
Mitchell, J. K., C. D. P. Baxter, and T. C. Munson. 1995. “Performance
of improved ground during earthquakes.” In Proc., Soil Improvement
for Earthquake Hazard Mitigation, 1–36. Reston, VA: ASCE.
Ramirez, J. C., M. Badanagki, M. Rahimi, M. A. ElGhoraiby,
M. T. Manzari, S. Dashti, A. Barrero, M. Taiebat, K. Ziotopoulou,
and A. Liel. 2017. “Seismic performance of a layered liquefiable site:
Validation of numerical simulations using centrifuge modeling.”
In Proc., 2017 GeoFrontiers. Reston, VA: ASCE.
Rayamajhi, D., S. A. Ashford, R. W. Boulanger, and A. Elgamal. 2016a.
“Dense granular columns in liquefiable ground. I: Shear reinforcement
and cyclic stress ratio reduction.” J. Geotech. Geoenviron. Eng. 142 (7):
4016023. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001474.
Rayamajhi, D., S. A. Ashford, R. W. Boulanger, and A. Elgamal. 2016b.
“Dense granular columns in liquefiable ground. II: Effects on deforma-
tions.” J. Geotech. Geoenviron. Eng. 142 (7): 4016024. https://doi.org
/10.1061/(ASCE)GT.1943-5606.0001475.
Seed, H. B. 1987. “Design problems in soil liquefaction.” J. Geotech. Eng.
113 (8): 827–845. https://doi.org/10.1061/(ASCE)0733-9410(1987)
113:8(827).
Seed, H. B., and J. R. Booker. 1977. “Stabilization of potentially liquefiable
sand deposits using gravel drains.” J. Geotech. Eng. Div. 103 (7):
757–768.
Seed, H. B., and I. M. Idriss. 1970. Soil moduli and damping factors
for dynamic response analyses. Technical Rep. No. EERC 70-10.
Berkeley, CA: Earthquake Engineering Research Center, Univ. of
California.
Seed, H. B., R. T. Wong, I. M. Idriss, and K. Tokimatsu. 1986. “Moduli and
damping factors for dynamic analyses of cohesionless soil.” J. Geotech.
Eng. 112 (11): 1016–1032. https://doi.org/10.1061/(ASCE)0733-9410
(1986)112:11(1016).
Stewart, D. P., Y. R. Chen, and B. L. Kutter. 1998. “Experience with the use
of methylcellulose as a viscous pore fluid in centrifuge models.”
Geotech. Test. J. 21 (4): 365–369. https://doi.org/10.1520/GTJ11376J.
Stockwell, R. G., L. Mansinha, and R. P. Lowe. 1996. “Localization of the
complex spectrum: The S transform.” IEEE Trans. Signal Process.
44 (4): 998–1001. https://doi.org/10.1109/78.492555.
Stringer, M., and S. Madabhushi. 2009. “Novel computer-controlled
saturation of dynamic centrifuge models using high viscosity fluids.”
Geotech. Test. J. 32 (6): 559–564. https://doi.org/10.1520/GTJ102435.
Tan, T. S., and R. F. Scott. 1985. “Centrifuge scaling considerations for
fluid-particle systems.” Geotechnique 35 (4): 461–470. https://doi
.org/10.1680/geot.1985.35.4.461.
Taylor, R. N. 1995. Geotechnical centrifuge technology. 1st ed. New York:
Blackie Academic and Professional.
Tokimatsu, K., and A. Asaka. 1998. “Effects of liquefaction-induced
ground displacements on pile performance in the 1995 Hyogoken-
Nambu earthquake.” Soils Found. 38: 163–177. https://doi.org/10
.3208/sandf.38.Special_163.
Walker, A. J., and H. E. Stewart. 1989. “Cyclic undrained behavior of
nonplastic and low plasticity silts.” Technical Rep. No. NCEER-89-
0035. Buffalo, NY: National Center for Earthquake Engineering
Research.
© ASCE 04018065-14 J. Geotech. Geoenviron. Eng.
J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065
Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.

More Related Content

What's hot

IRJET - Influence of Submergence on the Behaviour of Shallow Footing
IRJET - Influence of Submergence on the Behaviour of Shallow FootingIRJET - Influence of Submergence on the Behaviour of Shallow Footing
IRJET - Influence of Submergence on the Behaviour of Shallow FootingIRJET Journal
 
Soil mechanics note
Soil mechanics noteSoil mechanics note
Soil mechanics noteSHAMJITH KM
 
Consolidation settlement with sand drains – analytical and numerical approaches
Consolidation settlement with sand drains – analytical and numerical approachesConsolidation settlement with sand drains – analytical and numerical approaches
Consolidation settlement with sand drains – analytical and numerical approachesUmed Paliwal
 
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATION
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATIONEFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATION
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATIONLakshmi Narayanan
 
The role of deformation bands in modifying reservoir performance in porous sa...
The role of deformation bands in modifying reservoir performance in porous sa...The role of deformation bands in modifying reservoir performance in porous sa...
The role of deformation bands in modifying reservoir performance in porous sa...TomAdamson7
 
Time-lapse Stress Effects in Seismic Data
Time-lapse Stress Effects in Seismic DataTime-lapse Stress Effects in Seismic Data
Time-lapse Stress Effects in Seismic DataAli Osman Öncel
 
Swelling correlations
Swelling correlationsSwelling correlations
Swelling correlationsAli Rehman
 
Goetech. engg. Ch# 03 settlement analysis signed
Goetech. engg. Ch# 03 settlement analysis signedGoetech. engg. Ch# 03 settlement analysis signed
Goetech. engg. Ch# 03 settlement analysis signedIrfan Malik
 
6 site investigation
6 site investigation6 site investigation
6 site investigationRaghav Gadgil
 
Juscélia testing soil encasing materials for measuring hydraulic conductivit...
Juscélia  testing soil encasing materials for measuring hydraulic conductivit...Juscélia  testing soil encasing materials for measuring hydraulic conductivit...
Juscélia testing soil encasing materials for measuring hydraulic conductivit...Juscélia Ferreira
 
Settlement and bearing capacity of foundations with different
Settlement and bearing capacity of foundations with differentSettlement and bearing capacity of foundations with different
Settlement and bearing capacity of foundations with differentAlexander Decker
 
Effect of Soil Flexibility on Analysis and Design of Building
Effect of Soil Flexibility on Analysis and Design of BuildingEffect of Soil Flexibility on Analysis and Design of Building
Effect of Soil Flexibility on Analysis and Design of BuildingIJERA Editor
 
Rock Mass Classification
Rock Mass ClassificationRock Mass Classification
Rock Mass ClassificationAndi Anriansyah
 
Geotechnical investigation for road work
Geotechnical investigation for road workGeotechnical investigation for road work
Geotechnical investigation for road workRakesh Meena
 
ING IABSE December 2015 Paper 8
ING IABSE December 2015 Paper 8ING IABSE December 2015 Paper 8
ING IABSE December 2015 Paper 8Anusha Nandavaram
 

What's hot (20)

IRJET - Influence of Submergence on the Behaviour of Shallow Footing
IRJET - Influence of Submergence on the Behaviour of Shallow FootingIRJET - Influence of Submergence on the Behaviour of Shallow Footing
IRJET - Influence of Submergence on the Behaviour of Shallow Footing
 
Soil mechanics note
Soil mechanics noteSoil mechanics note
Soil mechanics note
 
Consolidation settlement with sand drains – analytical and numerical approaches
Consolidation settlement with sand drains – analytical and numerical approachesConsolidation settlement with sand drains – analytical and numerical approaches
Consolidation settlement with sand drains – analytical and numerical approaches
 
K013166871
K013166871K013166871
K013166871
 
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATION
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATIONEFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATION
EFFECT OF GROUTING ON STABILITY OF SLOPE AND UNSUPPORTED STEEP EXCAVATION
 
The role of deformation bands in modifying reservoir performance in porous sa...
The role of deformation bands in modifying reservoir performance in porous sa...The role of deformation bands in modifying reservoir performance in porous sa...
The role of deformation bands in modifying reservoir performance in porous sa...
 
Time-lapse Stress Effects in Seismic Data
Time-lapse Stress Effects in Seismic DataTime-lapse Stress Effects in Seismic Data
Time-lapse Stress Effects in Seismic Data
 
Swelling correlations
Swelling correlationsSwelling correlations
Swelling correlations
 
Goetech. engg. Ch# 03 settlement analysis signed
Goetech. engg. Ch# 03 settlement analysis signedGoetech. engg. Ch# 03 settlement analysis signed
Goetech. engg. Ch# 03 settlement analysis signed
 
L1303077783
L1303077783L1303077783
L1303077783
 
F04653441
F04653441F04653441
F04653441
 
6 site investigation
6 site investigation6 site investigation
6 site investigation
 
Juscélia testing soil encasing materials for measuring hydraulic conductivit...
Juscélia  testing soil encasing materials for measuring hydraulic conductivit...Juscélia  testing soil encasing materials for measuring hydraulic conductivit...
Juscélia testing soil encasing materials for measuring hydraulic conductivit...
 
Settlement and bearing capacity of foundations with different
Settlement and bearing capacity of foundations with differentSettlement and bearing capacity of foundations with different
Settlement and bearing capacity of foundations with different
 
Effect of Soil Flexibility on Analysis and Design of Building
Effect of Soil Flexibility on Analysis and Design of BuildingEffect of Soil Flexibility on Analysis and Design of Building
Effect of Soil Flexibility on Analysis and Design of Building
 
Rock Mass Classification
Rock Mass ClassificationRock Mass Classification
Rock Mass Classification
 
Geotechnical investigation for road work
Geotechnical investigation for road workGeotechnical investigation for road work
Geotechnical investigation for road work
 
ARMA_08-354
ARMA_08-354ARMA_08-354
ARMA_08-354
 
Iz3615681571
Iz3615681571Iz3615681571
Iz3615681571
 
ING IABSE December 2015 Paper 8
ING IABSE December 2015 Paper 8ING IABSE December 2015 Paper 8
ING IABSE December 2015 Paper 8
 

Similar to Influence of Dense Granular Columns on the Performance of Level and Gently Sloping Liquefiable Sites

The effect of soil improvement on foundation super structure design
The effect of soil improvement on foundation  super structure designThe effect of soil improvement on foundation  super structure design
The effect of soil improvement on foundation super structure designIAEME Publication
 
Performance of square footing resting on laterally confined sand
Performance of square footing resting on laterally confined sandPerformance of square footing resting on laterally confined sand
Performance of square footing resting on laterally confined sandeSAT Publishing House
 
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...Renee Heldman
 
Bearing capacity of sand overlying clay strip footing - paxis 3 d
Bearing capacity of sand overlying clay   strip footing - paxis 3 dBearing capacity of sand overlying clay   strip footing - paxis 3 d
Bearing capacity of sand overlying clay strip footing - paxis 3 dGonzalo Nureña Zegarra
 
"A Review of Field Performance of Stone Columns in Soft Soils"
"A Review of Field Performance of Stone Columns in Soft Soils""A Review of Field Performance of Stone Columns in Soft Soils"
"A Review of Field Performance of Stone Columns in Soft Soils"Remedy Geotechnics Ltd
 
International Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentInternational Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentIJERD Editor
 
11CCEE_23Jul2015_Final
11CCEE_23Jul2015_Final11CCEE_23Jul2015_Final
11CCEE_23Jul2015_FinalUpul Atukorala
 
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...Experimental Study of Bearing Capacity in Single and Group Stone Columns With...
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...Samirsinh Parmar
 
Evaluating 2D numerical simulations of granular columns in level and gently s...
Evaluating 2D numerical simulations of granular columns in level and gently s...Evaluating 2D numerical simulations of granular columns in level and gently s...
Evaluating 2D numerical simulations of granular columns in level and gently s...Mahir Badanagki, Ph.D.
 
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...IAEME Publication
 
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay""Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"Remedy Geotechnics Ltd
 
Triaxial test on soil important insights -formatted paper
Triaxial test on soil   important insights -formatted paperTriaxial test on soil   important insights -formatted paper
Triaxial test on soil important insights -formatted paperSamirsinh Parmar
 
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...IAEME Publication
 
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...IAEME Publication
 
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...IAEME Publication
 
Hyperbolic constitutive model for tropical residual soils
Hyperbolic constitutive model for tropical residual soilsHyperbolic constitutive model for tropical residual soils
Hyperbolic constitutive model for tropical residual soilsIAEME Publication
 
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...Alexander Decker
 

Similar to Influence of Dense Granular Columns on the Performance of Level and Gently Sloping Liquefiable Sites (20)

Ijciet 06 10_008
Ijciet 06 10_008Ijciet 06 10_008
Ijciet 06 10_008
 
The effect of soil improvement on foundation super structure design
The effect of soil improvement on foundation  super structure designThe effect of soil improvement on foundation  super structure design
The effect of soil improvement on foundation super structure design
 
Performance of square footing resting on laterally confined sand
Performance of square footing resting on laterally confined sandPerformance of square footing resting on laterally confined sand
Performance of square footing resting on laterally confined sand
 
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...
Understanding Permeability of Hydraulic Fracture Networks A Sandbox Analog Mo...
 
Bearing capacity of sand overlying clay strip footing - paxis 3 d
Bearing capacity of sand overlying clay   strip footing - paxis 3 dBearing capacity of sand overlying clay   strip footing - paxis 3 d
Bearing capacity of sand overlying clay strip footing - paxis 3 d
 
"A Review of Field Performance of Stone Columns in Soft Soils"
"A Review of Field Performance of Stone Columns in Soft Soils""A Review of Field Performance of Stone Columns in Soft Soils"
"A Review of Field Performance of Stone Columns in Soft Soils"
 
International Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentInternational Journal of Engineering Research and Development
International Journal of Engineering Research and Development
 
11CCEE_23Jul2015_Final
11CCEE_23Jul2015_Final11CCEE_23Jul2015_Final
11CCEE_23Jul2015_Final
 
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...Experimental Study of Bearing Capacity in Single and Group Stone Columns With...
Experimental Study of Bearing Capacity in Single and Group Stone Columns With...
 
Evaluating 2D numerical simulations of granular columns in level and gently s...
Evaluating 2D numerical simulations of granular columns in level and gently s...Evaluating 2D numerical simulations of granular columns in level and gently s...
Evaluating 2D numerical simulations of granular columns in level and gently s...
 
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...
LABORATORY MODEL TESTS TO EFFECT OF DENSITY TO FILL MATERIAL ON THE PERFORMAN...
 
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay""Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"
"Observed Installation Effects of Vibro Replacement Stone Columns in Soft Clay"
 
Triaxial test on soil important insights -formatted paper
Triaxial test on soil   important insights -formatted paperTriaxial test on soil   important insights -formatted paper
Triaxial test on soil important insights -formatted paper
 
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
 
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
A NUMERICAL STUDY ON INTERFERENCE EFFECTS OF CLOSELY SPACED STRIP FOOTINGS ON...
 
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...
MATHEMATICAL MODEL TO MONITOR STIFF CLAY COMPRESSION INDEX IN WET LAND AREA O...
 
cut-off wall
cut-off wallcut-off wall
cut-off wall
 
Hyperbolic constitutive model for tropical residual soils
Hyperbolic constitutive model for tropical residual soilsHyperbolic constitutive model for tropical residual soils
Hyperbolic constitutive model for tropical residual soils
 
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...
Baghdad subgrade resilient modulus and liquefaction evaluation for pavement d...
 
20120130406011
2012013040601120120130406011
20120130406011
 

Recently uploaded

HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVHARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVRajaP95
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )Tsuyoshi Horigome
 
microprocessor 8085 and its interfacing
microprocessor 8085  and its interfacingmicroprocessor 8085  and its interfacing
microprocessor 8085 and its interfacingjaychoudhary37
 
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝soniya singh
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...Soham Mondal
 
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ
 
power system scada applications and uses
power system scada applications and usespower system scada applications and uses
power system scada applications and usesDevarapalliHaritha
 
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130Suhani Kapoor
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSCAESB
 
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130Suhani Kapoor
 
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfAsst.prof M.Gokilavani
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidNikhilNagaraju
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineeringmalavadedarshan25
 
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2RajaP95
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxwendy cai
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSKurinjimalarL3
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCall Girls in Nagpur High Profile
 
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxArtificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxbritheesh05
 

Recently uploaded (20)

HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IVHARMONY IN THE NATURE AND EXISTENCE - Unit-IV
HARMONY IN THE NATURE AND EXISTENCE - Unit-IV
 
SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
 
microprocessor 8085 and its interfacing
microprocessor 8085  and its interfacingmicroprocessor 8085  and its interfacing
microprocessor 8085 and its interfacing
 
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
Model Call Girl in Narela Delhi reach out to us at 🔝8264348440🔝
 
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
OSVC_Meta-Data based Simulation Automation to overcome Verification Challenge...
 
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
VICTOR MAESTRE RAMIREZ - Planetary Defender on NASA's Double Asteroid Redirec...
 
power system scada applications and uses
power system scada applications and usespower system scada applications and uses
power system scada applications and uses
 
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
VIP Call Girls Service Hitech City Hyderabad Call +91-8250192130
 
GDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentationGDSC ASEB Gen AI study jams presentation
GDSC ASEB Gen AI study jams presentation
 
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130
VIP Call Girls Service Kondapur Hyderabad Call +91-8250192130
 
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
 
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfid
 
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineering
 
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2HARMONY IN THE HUMAN BEING - Unit-II UHV-2
HARMONY IN THE HUMAN BEING - Unit-II UHV-2
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
 
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service NashikCollege Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
College Call Girls Nashik Nehal 7001305949 Independent Escort Service Nashik
 
Artificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptxArtificial-Intelligence-in-Electronics (K).pptx
Artificial-Intelligence-in-Electronics (K).pptx
 

Influence of Dense Granular Columns on the Performance of Level and Gently Sloping Liquefiable Sites

  • 1. Influence of Dense Granular Columns on the Performance of Level and Gently Sloping Liquefiable Sites Mahir Badanagki, S.M.ASCE1 ; Shideh Dashti, M.ASCE2 ; and Peter Kirkwood3 Abstract: Dense granular columns are often used as a liquefaction mitigation measure to (1) enhance drainage; (2) provide shear reinforce- ment; and (3) densify and increase lateral stresses in the surrounding soil during installation. However, the independent influence and con- tribution of these mitigation mechanisms on the excess pore pressures, accelerations (or shear stresses), and lateral and vertical deformations are not sufficiently understood to facilitate a reliable design. This paper presents the results of a series of dynamic centrifuge tests to fun- damentally evaluate the influence of dense granular columns on the seismic performance of level and gently sloped sites, including a lique- fiable layer of clean sand. Specific consideration was given to the relative importance of enhanced drainage and shear reinforcement. Granular columns with greater area replacement ratios (Ar), for example Ar greater than about 20%, were shown to be highly effective in reducing the seismic settlement and lateral deformations in gentle slopes, owing primarily to the expedited dissipation of excess pore water pressures. The influence of granular columns on accelerations (and therefore, the shear stress demand) in the surrounding soil depended on the column’s Ar and drainage capacity. Increasing Ar from 0 to 10% was shown to reduce the accelerations across a range of frequencies in the surrounding soil due to the shear reinforcement effect alone. However, enhanced drainage simultaneously increased the rate of excess pore pressure dissipation, helping the surrounding soil regain more quickly its shear strength and stiffness. At short drainage distances or higher Ar values (for example, 20%), this could notably amplify the acceleration and shear stress demand on soil, particularly at greater frequencies that influence PGA. The experimental insight presented in this paper aims to improve our understanding of the mechanics of liquefaction and lateral spreading mitigation with granular columns, and it may be used to validate the numerical models used in their design. DOI: 10.1061/(ASCE)GT.1943-5606.0001937. © 2018 American Society of Civil Engineers. Author keywords: Soil liquefaction; Granular columns; Drains; Centrifuge modeling; Lateral spreading; Site performance. Background and Introduction Past earthquakes have provided many examples of damage to important geotechnical structures such as slopes, retaining walls, and embankments caused by soil liquefaction. For example, exten- sive lateral slope deformations were reported by Seed (1987) and Tokimatsu and Asaka (1998) during the earthquakes in Niigata, Japan, in 1964; Loma Prieta, California, in 1989; and Kobe, Japan, in 1995. Mitigation techniques are often warranted to prevent ex- cessive lateral deformations in slopes founded on liquefiable depos- its and the subsequent damage to infrastructure. A reliable and performance-based design of liquefaction remediation techniques for slopes requires a clear understanding of the influence of various mitigation mechanisms that control performance. Bartlett and Youd (1992) characterized liquefaction-induced lat- eral spreading in mild (0.3–5%) slopes underlain by loose, saturated granular soils. Despite the low angle of these slopes, the lateral deformations produced during earthquake loading could still exceed several meters, leading to extensive damage. Many ground-improvement techniques such as densification, reinforce- ment, and methods that enhance drainage can be used to reduce the risk of liquefaction and its associated ground deformations. In particular, the installation of dense granular columns made of gravel or stone is an attractive mitigation method for slopes and embankments (Seed and Booker 1977). Depending on its installa- tion procedure, this method is believed to mitigate the soil lique- faction hazard and its consequences through a combination of densification, increased lateral earth pressures, shear reinforcement, and enhanced drainage (Baez 1995; INA 2001; Rayamajhi et al. 2016a). Previous case histories have generally demonstrated a success- ful performance of different types of compacted granular columns in loose, saturated cohesionless soils (e.g., Mitchell et al. 1995; Mitchell 1986, 1988; Adalier 1996; Baez 1996; Boulanger et al. 1998; Koelling and Dickenson 1998; ISSMGE 2001). However, although valuable insights can and must be drawn from case his- tories, the influence and relative importance of each mechanism of mitigation on the performance of the site and slope cannot be re- liably evaluated from case histories alone in a systematic manner, as is necessary for a reliable, performance-based mitigation design. Seed and Booker (1977) introduced an analytical method for the design of drains based on radial consolidation or excess pore pres- sure (Δu) dissipation that is commonly used in practice. They rec- ommended the use of gravel or stone columns with a hydraulic conductivity at least two orders of magnitude greater than that of the surrounding soil, to avoid significant Δu generation in the drains. This study had a number of limitations, including (1) the assumption of an infinite granular column permeability and ignor- ing the well resistance or clogging potential; (2) the assumption of a 1 Graduate Research Assistant, Dept. of Civil, Environmental, and Architectural Engineering, Univ. of Colorado Boulder, Boulder, CO 80309. Email: mahir.badanagki@colorado.edu 2 Associate Professor, Dept. of Civil, Environmental and Architectural Engineering, Univ. of Colorado Boulder, Boulder, CO 80309 (corresponding author). Email: shideh.dashti@colorado.edu 3 Research Associate, Dept. of Civil, Environmental, and Architectural Engineering, Univ. of Colorado Boulder, Boulder, CO 80309. Email: peter .kirkwood@colorado.edu Note. This manuscript was submitted on May 19, 2017; approved on March 29, 2018; published online on July 6, 2018. Discussion period open until December 6, 2018; separate discussions must be submitted for indi- vidual papers. This paper is part of the Journal of Geotechnical and Geoenvironmental Engineering, © ASCE, ISSN 1090-0241. © ASCE 04018065-1 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 2. level ground condition (no slope); and (3) not considering shear reinforcement in the gravel or stone columns. In practice, however, engineers sometimes rely on the shear reinforcement provided by granular columns and use design procedures (based on shear strain compatibility) to reduce the cyclic stress ratios induced in the sur- rounding soil (Rayamajhi et al. 2016a). Rayamajhi et al. (2016a) numerically evaluated the contribution of the shear reinforcement provided by granular columns in isola- tion, with and without the generation and redistribution of excess pore pressures. Their nonlinear, coupled, three-dimensional (3D) finite-element analyses indicated that granular columns experience shear strain deformations that are not compatible with the surround- ing soil. As a result, the reduction in the cyclic stress ratios induced in the treated soil was notably less than that predicted by the conventional design approaches that assume strain compatibility. They subsequently provided a modified design approach to esti- mate the reduction in cyclic stress ratios provided by dense granular columns. Elgamal et al. (2009) and Asgari et al. (2013) performed a numerical parametric study with a unit cell and showed that granu- lar columns could help reduce lateral displacements in treated soil due to a combination of drainage and shear reinforcement. Rayamajhi et al. (2016b) numerically showed that granular col- umns can be effective in reducing lateral spreading through shear reinforcement, even if liquefaction or generation of large excess pore pressures is not prevented. They also numerically identified the surface pressure, length and diameter of the drains, area replace- ment ratios, liquefiable soil depth, hydraulic conductivity, and slope angle as some of the key parameters influencing the lateral displacement. However, the results presented in these numerical studies were not validated with rigorous physical model studies or case history observations. Although limited in quantity and scope, a number of physical model studies (for example, 1g shake table and centrifuge tests) have been conducted to evaluate the influence of different types of drains on site response and lateral spreading. Due to the diffi- culties of drain installation in flight, the centrifuge experiments fo- cused primarily on the reinforcement and drainage mechanisms of mitigation, as opposed to installation-induced densification or an accurate representation of the increase in lateral earth pressures in the surrounding soil, which may not be reliable in all cases, even in the field. As an example, Adalier et al. (2003) performed cen- trifuge experiments with stone columns in a uniform and level de- posit of saturated, loose silt with an overlying model of a rigid footing. This study showed that stone columns can be effective in reducing the foundation settlement by reducing shear strains and generation of excess pore pressures in the underlying soil. They did not, however, evaluate the influence of granular columns and their properties on the performance of sites with more realistic layering and geometry and under different earthquake motions. Centrifuge tests were more recently conducted by Howell et al. (2012) to evaluate the performance of liquefiable slopes mitigated with prefabricated vertical drains (PVDs). This method did not in- troduce notable shear reinforcement and relied primarily on the drainage mechanism of mitigation. The study showed that PVDs could be effective in expediting the dissipation of excess pore pres- sures and reducing the resulting lateral slope deformations, depend- ing on the characteristics of the earthquake motion. The combined or independent influence of shear reinforcement and enhanced drainage in slopes has not been experimentally evaluated. It is not clear, for example, whether the added shear reinforcement pro- vided by stiffer drains (e.g., dense granular columns in comparison with PVDs) would reduce the seismic demand or the deformations in slopes. Dashti et al. (2010) performed a series of centrifuge experiments to identify the dominant mechanisms of settlement on layered liquefiable soil deposits. The study classified the primary volumet- ric settlement mechanisms active in the free-field as (1) sedimenta- tion or solidification (εp-SED), which occurs after significant strength loss, as settling particles from the uppermost part of the liquefiable sand accumulate at the bottom to form a solidified zone. This zone increases in thickness with time and concurrently con- solidates under its own weight; (2) reconsolidation (εp-CON), which occurs as the excess pore pressures dissipate and soil’s effective stress regains its initial value; and (3) partial drainage and loss of water during cyclic loading (εp-DR) that occur according to the 3D transient hydraulic gradients. Volumetric deformations caused by partial drainage can be notable, as they occur while the hydraulic gradients are kept at their peak during shaking. Drainage-enhancing granular columns, such as those investigated in this study, locally reduce the duration, or the amplitude, of the elevated excess pore pressures. This, along with the altered shear stress and strain profiles, likely reduces the contribution of sedi- mentation to the volumetric strains. However, the increased rate of drainage likely increases the volumetric strains due to partial drainage. The net effect of granular columns on the competing mechanisms leading to volumetric strains is uncertain and requires investigation. This paper presents the results of three centrifuge experiments that systematically evaluated the influence of dense granular col- umns and their various mitigation mechanisms and properties on the seismic performance of level and gently sloping sites underlain by a layered liquefiable deposit consisting mostly of clean sand. The data obtained from these experiments facilitated (1) the mecha- nistic evaluation of the influence of granular columns and their dif- ferent properties (for example, the area replacement ratio, shear reinforcement, and enhanced drainage) on the site and slope per- formance in terms of the acceleration, excess pore pressure, settle- ment, and lateral spread; and (2) the provision of data for the calibration and validation of advanced numerical models in the future. Following validation, such models can be more reliably used in the design or development of design guidelines for the mitigation of lateral spreading hazards using dense granular columns. Experimental Procedure Centrifuge Testing Plan A series of three centrifuge experiments comprising five separate models were designed and conducted at the University of Colorado (CU) Boulder’s 400g-t (5.5-m-radius) centrifuge facility. In these tests, the primary goal was to systematically evaluate the influence and relative importance of various granular column parameters on the performance of a level site and a gentle slope underlain by a layered liquefiable deposit during one-dimensional horizontal earthquake loading. Fig. 1 shows the detailed plan and elevation- view drawings of the model tests and their instrumentation layouts. The dimensions are presented at both prototype and model scales following the accepted scaling relations (Tan and Scott 1985). For all tests, as shown in Fig. 1, a dense layer (8 m thick in the prototype scale) of Ottawa sand F65 was dry pluviated to attain a relative density (Dr) of approximately 90% at the bottom of a flexible-shear-beam (FSB) container constructed of aluminum and rubber at CU. Subsequently, a loose layer of Ottawa sand (8 m thick) with a Dr ≈ 40% was pluviated as the liquefiable material. This layer was subsequently overlaid by a 0.5-m-thick layer of © ASCE 04018065-2 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 3. Silica silt (Sil-Co-Sil 120) to restrict the rapid excess pore pressure dissipation vertically and to represent a case with void redistribu- tion, followed by a 1.5-m-thick layer of coarse Monterrey sand 0/30 with Dr ≈ 90%, to create a nonliquefiable, thin crust. For all the treated models (Models 0, 1, 3, and 4), the granular columns with diameter = 1.75 m (in prototype scale) were placed vertically at the bottom of the container prior to sand pluviation to avoid localized densification during their installation. This was done to isolate the influence of drains from ground densification and to keep the den- sity of the surrounding soil controlled and uniform within each layer. Test 1 (Model 0) examined the behavior of a single draining granular column in level ground, before adding the complications associated with multiple drains and a slope. Vertical arrays of ac- celerometers (Accs), pore pressure transducers (PPTs), and linear variable differential transformers (LVDTs) were placed at three dif- ferent radial distances from the single drain to track the wave propa- gation, pore water pressure generation and dissipation, and volumetric deformations throughout the soil profiles, as shown in Fig. 1. In Tests 2 and 3, all soil models were constructed of two symmetric slopes on the two sides of the container separated by an open channel, as shown in Fig. 1. In Test 2, one side of the slope (Model 1) was treated with a grid of 1.75-m-diameter dense granu- lar columns encased in geotextile filters (to avoid clogging) sepa- rated by 3.5 m (from center to center) with an Ar of 20%. The Ar is ratio of the area of granular columns to the total treatment area (Baez and Martin 1993). The other side of the slope (Model 2) in Test 2 was left untreated to evaluate the effectiveness of the granular columns as a mitigation technique. In Test 3, both sides of the slope (Models 3 and 4) were treated with 1.75-m-diameter dense granular columns separated by 5 m (from center to center) with Ar ¼ 10%. In Model 3, the granular columns were encased in geotextile filters, similarly to Model 1, whereas in Model 4 they were encased in geotextile and a thin (0.2-mm-thick) latex mem- brane to evaluate the influence of stiffness alone without drainage. The groundwater table was at the soil surface in all tests (the highest elevation in the cases with slope), and all model specimens were spun to a centrifugal acceleration of 70 g prior to the application of earthquake motions. Properties of Sand, Silt, and Gravel The response of Ottawa sand F65 at various relative densities was evaluated under monotonic and cyclic, drained and undrained tri- axial tests prior to the centrifuge experiments presented here (de- tailed by Ramirez et al. 2017). The engineering properties of this sand were measured by the authors as follows: maximum void ratio ðemaxÞ ¼ 0.81; minimum void ratio ðeminÞ ¼ 0.53; coefficient of uniformity ðCuÞ ¼ 1.56; specific gravity ðGsÞ ¼ 2.65; and hy- draulic conductivity ðkÞ ¼ 1.19 × 10−2 to 1.41 × 10−2 cm=s at 1g with water (which were the same at 70g with viscous fluid 70 times more viscous than water) for the relative densities (Dr) of 90 and 40%, respectively. The critical friction angle (ϕcs) of this sand was estimated at approximately 31° (Ramirez et al. 2017). Coarse Monterey 0/30 sand was selected for the thin, high- permeability, dense surface layer. The properties of Monterey sand, obtained from previous studies, were emax ¼ 0.84; emin ¼ 0.54; Cu ¼ 1.3; k ¼ 5.29 × 10−2 cm=s; and Gs ¼ 2.64 (Dashti et al. 2010). Silica silt (Sil-Co-Sil 120) was selected as the thin, low- permeability cap above the liquefiable layer. The properties of silica silt, also obtained from previous studies (Walker and Stewart 1989), were Cu ¼ 7.3; k ¼ 3 × 10−5 cm=s; and Gs ¼ 2.65; and the critical friction angle (ϕcs) was approximately 25°. In this study, the material 428mm 428mm [30m] [30m] Treated side (Model 1) Untreated side (Model 2) 3° 60° Silica silt Monterey sand (Dr=90%) Ottawa sand Dr=90% Ottawa sand Dr=40% Gravel column (Diam.=25mm) [1.75m] 150 mm Model 10.5 m [Prototype] 0 75 5.30 Acc. (A) PPT (P) LVDT (D) Latex Shaking table Without latex (Model 3) With latex (Model 4) Model 0 Test #1 Test #2 Test #3 177mm144mm158mm [12.4m][10m][11m] 956mm [26.3m] 376mm [66.9m] R 35 [2.5m ] R121 [8.5m] [17.4m] R249 Shaking table Porous stone Shaking table 228mm 328mm [16m] [23m] 21mm 7mm [1.5m] [0.5m] [8m] 115mm [8m] 115mm Shaking Shaking 3° Fig. 1. (Color) Three centrifuge experiments with their instrumentation layout. © ASCE 04018065-3 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 4. representing the granular columns was made of relatively uniform, clean, medium gravel with Cu ¼ 1.54. It had the critical and peak friction angles (ϕcs and ϕpeak) of 35° and 40°, respectively, at a dry unit weight of 17 kN=m3 , obtained from isotropically consolidated drained triaxial tests conducted by the authors. The hydraulic con- ductivity of these columns was estimated at k ≈ 2.9 cm=s from con- stant head permeability tests (roughly 200 times more permeable than Ottawa sand F65). Fig. 2 shows the grain-size distribution curves for all four materials used in this study. Model Construction The Ottawa and Monterey sand layers were prepared using the air pluviation reconstitution method with the automated pluviator at CU [Fig. 3(a)], to improve the repeatability among tests and the uniformity within each model. The silt layer was dry pluviated from a constant height using a #10 sieve to avoid lumps and distribute silt across a controlled area. It was then gently compacted using a static surcharge pressure of 5 kPa to achieve a dry unit weight of approx- imately 16 kN=m3 with a final thickness of approximately 0.5 m in the prototype scale. To prepare the dense granular columns, geotextile filters were formed to the desired column’s shape and diameter, and they were subsequently plugged and glued at the bottom. Then the gravelly soil was poured inside and tightly compacted in the geotextile col- umn in three layers (previously calibrated to achieve the desired dry unit weight of approximately 17 kN=m3 ) and subsequently glued from the top by adhesive tape, to prevent sand from entering the columns during pluviation. For Model 4, the granular columns were encased first in geotextile filters and then in 0.2-mm-thick latex membranes. At their base, the membrane latex was sealed with silicone sealant and end-cap plates. This sealing process pre- vented any water in the soil from moving into the constructed granular columns in Model 4 from the bottom or sides. The granu- lar columns in Models 0, 1, and 3 were capable of drainage and were not encased with a latex membrane. The shear modulus ratio (Gr), ratio of the small-strain shear modulus (Gmax) of granular col- umns to that of the surrounding loose Ottawa sand, was approxi- mately 3. The Gmax value was calculated for sand and gravel on the basis of the empirical procedures detailed by Seed and Idriss (1970) and Seed et al. (1986), respectively. The construction of dense granular columns inside geotextile may not be practical in loose saturated sandy deposits, based on the existing construction techniques. Nevertheless, in this experi- mental study, the geotextile was used to prevent clogging and preserve the drainage capability of granular columns during sub- sequent motions. This design also provided an upper-bound con- dition for the shear stiffness and drainage ability of these columns, enabling an easier evaluation of the underlying mechanisms. Fig. 2. (Color) Particle size distribution curves for all soil types used in this study. Fig. 3. (Color) Photographs taken (a) during air pluviation of Ottawa Sand; (b) after model completion of Test 2; (c) showing horizontal LVDT holders and setup designed for the centrifuge experiments; and (d) after model completion of Test 3. © ASCE 04018065-4 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 5. To create accurate 3° slopes that were consistent among the models, the sand surface was vacuumed rather than scraped (to minimize soil disturbance) with the aid of guide rails clamped along the length of the container. After the model completion in Tests 2 and 3, the top surface of Monterey sand was sprayed with a sugar solution (5% by mass), lightly heated, and left overnight to dry. This treatment kept the slopes intact during the model transport across the lab floor to the centrifuge, but cementation effects were removed when the sugar dissolved during saturation. For all the treated models (Models 0, 1, 3, and 4), the granular columns were placed vertically at the bottom of the container at predetermined positions in a grid prior to the pluviation of sand. Rows and vertical columns of colored sand (Ottawa) were also placed at various locations throughout the model, to enable meas- urement and evaluation of various modes of deformation during model excavation. Figs. 3(b and d) show photographs of the fully assembled models. Figs. 3(b and c) also show the setup for meas- uring vertical and horizontal soil movements. The vertical LVDTs measured the settlement of bearing plates placed on the soil surface. The vertical LVDT rods were free to slide horizontally across the surface of the bearing plate, while the plate displaced with the soil owing to the nails placed around its perimeter. The nails had a diameter of 0.9 mm and height of 9 mm in the model scale, and their effect on the soil response was assumed to be negligible. The horizontal LVDTs recorded lateral displacements along an ini- tially vertical slot in an aluminum plate. This permitted the vertical settlement of soil without influencing horizontal displacement mea- surements [Fig. 3(c)]. The bearing plate for horizontal LVDTs was embedded in Monterey sand to resist the overturning moment aris- ing from the inertia of LVDT rods. To minimize this moment, the LVDT rod was initially positioned near the base of the slot. After dry preparation, the models were saturated with a hydrox- ypropyl methylcellulose solution of 64 cSt viscosity, prepared per Stewart et al. (1998) and measured before use. The viscosity was 70 times greater than that of water, to satisfy both the diffusive and the dynamic scaling laws (Taylor 1995). A computer-controlled sat- uration system was designed and implemented to improve the qual- ity and rate of saturation, similar to that proposed by Stringer and Madabhushi (2009). Initially, the soil model was flushed with CO2 from the bottom of the container for about 1 h, after which it was kept under constant vacuum. The fluid tank was placed on a scale, and its vacuum level was subsequently controlled automatically to maintain a safe and constant flow rate below that required for flow- induced liquefaction (in this case, 19 g=minute just prior to reach- ing the silt layer and then 2 g=min, based on Stringer and Madabhushi 2009). Complete saturation of each model required approximately 96 h. Properties of Input Motions Once they were under 70 g of centrifugal acceleration, a series of one-dimensional horizontal earthquake motions were applied to the base of each model using a servo-controlled hydraulic shake table available at CU (Ketchum 1989). These motions were scaled versions of the horizontal component of recorded earthquake motions, selected to cover a range of characteristics in terms of am- plitude, frequency content, and duration, thus enabling evaluation of the impact of these properties on system performance. Table 1 provides a summary of the sequence and properties of the first three earthquake motions applied at the base of the model container. Fig. 4 shows the acceleration response spectra (5% damped) and the Arias Intensity time histories of the base motions achieved in the centrifuge (the mean and their range of variability among the three tests). Ample time was allowed between each phase of shak- ing to allow and monitor full excess pore pressure dissipation. The 1992 Mw 7.3 Landers Earthquake recording at the Joshua Tree sta- tion was selected because of its longer duration and slower rate of energy buildup relative to the other records. The 1995 Mw 6.9 Kobe Earthquake recording at the Takatori station had a more rapid buildup of energy and a shorter duration, but it was scaled to have a lower amplitude in comparison with the Landers event (for exam- ple, in terms of the PGA). Finally, the 1994 Mw 6.7 Northridge Earthquake recorded at the Newhall-WPC station was a near-fault record with a significant velocity pulse. The three selected motions also differed in their frequency content, as shown in Fig. 4, for the achieved or measured base motions. Centrifuge Experimental Results During all tests, the data were recorded at a sampling rate of 3,500 samples per second. In the following sections, all the test results are presented and discussed in prototype units. The responses of Model 0 (level ground with a single drain), Model 1 (slope with 36 draining granular columns and Ar ¼ 20%), Model 2 (slope without granular columns), Model 3 (slope with 25 granular col- umns and Ar ¼ 10%), Model 4 (slope with 25 granular columns and Ar ¼ 10%, encased with latex membrane) were analyzed and compared on the basis of the recorded accelerations, pore water pressures, and vertical and horizontal displacements, in order to provide insight into the effects of granular columns and their vari- ous properties on site and slope performance. Due to space limi- tations, only selected, representative results are presented. Table 1. Ground motion properties recorded at the base of the container during Test 3 Motion number Ground motion identifier Event Station PGA (g) Significant duration, D5–95 (s) Mean period, Tm (s) Arias intensity, Ia (m=s) 1 Kobe 1995 Kobe Takatori 0.38 11.9 0.86 1.92 2 Joshua 1992 Landers Joshua Tree 0.58 27.6 0.66 7.54 3 Northridge 1994 Northridge Newhall-WPC 0.75 16.6 0.94 4.94 Fig. 4. (Color) Acceleration response spectra (5% damped) and Arias Intensity time histories of the base motions recorded during the three tests. The solid lines show the average of the three tests, and the highlighted regions show the variation among tests. © ASCE 04018065-5 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 6. In this paper, toe refers to the location of the instrument array closer to the slope toe but still within the slope, whereas top refers to the location of the instrument array closer to the slope top and container boundaries. Influence of a Single Granular Column on Site Performance The primary objectives of Test 1 (Model 0) were (1) to evaluate site performance in the far field and in the vicinity of a single granu- lar column in a level, layered soil profile in terms of accelerations, pore pressures, and settlements; and (2) to provide a data set for the calibration and validation of numerical models in the future, before adding the complexities of a slope and multiple drains. Furthermore, drain systems are often designed for a unit drain cell surrounded by other drains on all sides. The performance of a drain outside the unit cell framework (for example, drains at the edge of a group or in isolation) has not been adequately studied experimentally. Fig. 5 shows a summary of the time histories of excess pore pressure, vertical displacement, and horizontal acceleration recorded at three different radial distances from the single granular column at different depths in Model 0 during the Kobe motion. Liquefaction, defined as an excess pore pressure ratio (ru ¼ Δu=σ0 vo) of 1.0, was observed quickly in the loose sand layer. Large excess pore pressures were also generated in the lower dense layer of Ottawa sand, causing liquefaction during the first motion but at a slightly slower pace. The accelerations generally showed large, high-frequency spikes due to dilation within dense Ottawa sand. These were attenuated in loose sand. The pore pressure recordings showed that a single granular col- umn was unable to prevent liquefaction even at a radial distance of 2.5 m, a result that was not surprising given that the drain spacing was effectively infinite. However, the rate of excess pore pressure dissipation was increased after shaking ceased, leading to a faster rate of dissipation near the granular column, particularly at the lower depths. The net dissipation of excess pore pressures was slower at higher elevations because of the upward flow from lower elevations. The influence of a single granular column considered in this study (in terms of its effects on pore pressures) appeared to fall within a radius of 2.5–8.5 m. Further, the ru values greater than 1.0 at the top of the liquefiable layer may indicate the possible forma- tion of a water film below the silt interface and/or the slight settle- ment of PPTs with respect to the surrounding soil at that location. The slight movement of those PPTs with respect to colored sand was confirmed during the excavation. The dynamic total head isochrones were used to show the di- rection and magnitude of transient hydraulic gradients at different times and the resulting flow tendencies around the drain. Fig. 6 shows a comparison of the dynamic total head isochrones obtained from the PPT recordings at three different radial distances from the granular column center during the Kobe and Joshua motions. At different radial distances, the isochrones indicated that liquefac- tion (ru ¼ 1.0) was achieved quickly within the loose layer of Ottawa sand during all motions. The hydraulic gradients were generally formed upward from the dense layer toward the surface. Fig. 5. (Color) Acceleration, excess pore pressure, and vertical displacement time histories during the first motion (Kobe) at three radial distances from the single granular column in Test 1. © ASCE 04018065-6 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 7. The presence of a drain led to a radially inward and upward flow pattern in its vicinity, which slightly reduced the duration of large excess pore pressures near the drain, particularly in deeper layers, in comparison with the far-field (for example, a radius of 17.4 m in this experiment) that experienced only vertically upward flow. More rapid drainage, particularly for the deeper liquefiable layers, would be expected to reduce the duration of large lateral spreads in the slopes and benefit their performance; this is explored in the next section. The acceleration time histories showed spikes at large strains after significant pore pressure generation and softening of the soil. The amplitude of the acceleration spikes increased as the shaking intensity and soil density increased in the subsequent events (not shown, for brevity), which may be attributed to increased soil di- lation and restiffening (Dashti et al. 2010). The relative density of the loose sand layer increased from approximately 40% prior to Kobe to about 60% prior to the Northridge motion on the basis of far-field LVDT recordings, assuming uniform settlement across the far-field. Fig. 7 shows the time-frequency Stockwell spectra (Stockwell et al. 1996; Kramer et al. 2016) of the transverse accelerations at the container base as well as the bottom, middle, and top of the lique- fiable layer during the Kobe motion. These plots enable visualiza- tion of the spectral changes in acceleration over time, which is more appropriate and insightful than a Fourier transform in a highly non- linear and nonstationary system. In this figure, the color scale is constant across all plots to facilitate comparison, and the peak ground acceleration (PGA) measured during the response is shown in the top right corner of each plot. Significant deamplification of accelerations occurred from the base toward the surface due to ex- cessive softening, particularly in loose Ottawa sand. Within the loose Ottawa sand layer, the acceleration’s frequency content varied with distance from the granular column. Fig. 5 shows that the excess pore pressures developed in this layer were relatively unaffected by the column for the duration of strong shak- ing. Therefore, enhanced drainage was likely not primarily respon- sible for this change in acceleration’s frequency content. It is proposed that the shear stiffness increment due to the presence of a gravel column shifted the site’s modal frequencies as a function of the radius (R). Owing to the increased shear stiffness, the lower frequency shear waves (less than approximately 1 Hz) experienced deamplification in soil closer to the column in comparison with the far-field. The frequency bandwidth of the deamplification may have been more extensive in the absence of enhanced drainage. However, the high-frequency accelerations (around 4–10 Hz), were amplified near the granular column in comparison with the far-field due to a slightly faster dissipation of excess pore pressures, even during shaking, and hence the instances of recovery in the soil shear stiffness (and the reduction in damping). This hypothesis is tested in subsequent sections in which the effect of drainage is separated from that of reinforcement. When used on a gently sloping site (explored in the next sec- tions), the increased shear stiffness close to the granular column may help limit the lateral spreading. However, designers should Fig. 6. (Color) Dynamic total head isochrones at three different radial distances from the granular column during the Kobe and Joshua motions. Fig. 7. (Color) Time frequency (Stockwell spectra) of transverse accelerations at the bottom, middle, and top of liquefiable layer and container base at three different radial distances from the granular column during the Kobe motion. (The PGA value is noted at the top right corner of each figure.) © ASCE 04018065-7 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 8. be aware of a consequential increase in PGA. The PGA of the sur- face motion (controlled by higher-frequency vibrations) and hence the cyclic stress ratio (commonly used in liquefaction-triggering analyses as the measure of seismic demand) were increased by ap- proximately threefold when the radial distance (R) to the drain was decreased from 17.4 m (PGA ¼ 0.21 g) to 2.5 m (PGA ¼ 0.61 g), during the Kobe motion. The adverse influence of the granular col- umns on surface PGA and the cyclic demand must be considered in design. The readings from vertical LVDTs provided a measure of set- tlement (positive readings) at the top and bottom of the loose sand layer during different shaking events at three different radial dis- tances from the drain center, as shown in Fig. 5, during the Kobe motion. Generally, significant settlements occurred both in the dense and loose layers of Ottawa sand during shaking due to sed- imentation and partial drainage (εp-SED and εp-DR), followed by a smaller contribution from post-shaking consolidation (εp-CON). As the distance to the drain decreased, settlements within both loose and dense layers of Ottawa sand generally reduced. Even though the drain tended to amplify volumetric strains due directly to partial drainage (εp-DR), it was shown to reduce the net volumetric settle- ments at the surface by limiting the duration of large excess pore pressures and therefore the extent of sedimentation (εp-SED). The patterns were similar during other ground motions. Influence of Granular Columns and Area Replacement Ratio on Slope Performance In this section, the effectiveness of granular columns is assessed in reducing the extent of excess pore pressure generation, accelera- tions, and lateral spreading in a gently sloping site. The influence and relative importance of the area replacement ratio (Ar) of granu- lar columns was assessed experimentally. Models 2, 3, and 1 con- tained a gentle slope treated with granular columns that had Ar ¼ 0% (no drains); Ar ¼ 10%; and Ar ¼ 20%, respectively. Fig. 8 shows a comparison of the dynamic total head isochrones for the Kobe motion, which were obtained from the PPT recordings for Models 1 through 3 at the depths shown in Fig. 1. These isochrones show that liquefaction was achieved quickly within the loose layer of Ottawa sand in Models 2 and 3 with the Ar of 0 and 10%, respectively. Large excess pore pressures were mea- sured in the dense layer of Ottawa sand, but liquefaction was typ- ically not reached in this layer in the presence of a gentle slope (even for the untreated case of Ar ¼ 0%). Importantly, the greater Ar of 20% in Model 1 successfully reduced the extent and duration of large excess pore pressures in all layers in comparison with the other two models. Model 3 (Ar ¼ 10%) also appeared to slightly increase the rate of excess pore pressure dissipation in comparison with the untreated Model 2 (Ar ¼ 0%), but not as well as Model 1. These patterns were consistent during different motions. Fig. 9 shows a comparison of the excess pore pressure ratio (ru) time histories at various depths together with the vertical and hori- zontal displacements recorded at the top and toe of the slope during the three shaking events in the three models. The container base acceleration time histories are also provided for comparison. The horizontal displacement is considered positive in the fall line or downslope direction for all models. The rate of excess pore pres- sure buildup at all depths was generally decreased and the rate of post- shaking dissipation increased when the Ar increased from 0 to 10% and to 20%, as expected. Similar to Model 0, the fastest rate of dissipation was observed at lower elevations, due to an upward flow tendency, which is also shown in Fig. 8. The time-frequency response (Stockwell spectra) of transverse accelerations recorded in the three models during the Kobe motion is shown in Fig. 10. The figure shows that the amplitude of soil surface accelerations generally tended to decrease at lower frequen- cies over an extended period of time, as the Ar increased from 0 to 10% (the left two plots on the top row of the figure). On the other hand, the acceleration amplitudes increased considerably when Ar was increased from 10 to 20% (the right two columns in the figure), particularly at frequencies greater than 0.7 Hz during and after strong shaking. This led to a notable increase in PGA in the top half of the liquefiable layer. This response may be explained by the counteracting effects of shear reinforcement and enhanced drainage provided by granular columns. For both treated Models 3 (Ar ¼ 10%) and 1 (Ar ¼ 20%), the added shear stiffness of the col- umns alone was expected to reduce the amplitude of accelerations and shear stresses in the surrounding liquefiable soil, particularly at lower frequencies. In the experiments, an Ar of 10% provided shear reinforcement, resulting in a slight reduction of accelerations in comparison with the unmitigated slope (Ar ¼ 0%) at frequencies less than around 1 Hz. This model also enhanced the drainage rate, but not sufficiently to noticeably amplify accelerations at higher frequencies. At the accelerometer locations (midway between the columns), the accelerations were in this case only marginally am- plified or preserved at higher frequencies. For the greater Ar of 20%, the drainage rates were increased sufficiently to prevent soil liquefaction in most cases. Therefore, the average soil stiffness was increased and the damping decreased in comparison with Models 2 (Ar ¼ 0%) and Model 3 (Ar ¼ 10%). This impacted the propagation and amplification of accelerations, particularly above the dense Ottawa sand layer. The result was more intense accelerations measured near the soil surface (the top of loose Ottawa sand) across a range of frequencies, concentrated around the strain-compatible fundamental frequency of the treated slope (approximately 0.7–0.8 Hz as determined from the Stockwell spectra) due to the combined effect of reinforcement and drainage. These observations are in line with those in Model 0 and were consistent during various motions, despite their varia- tions in intensity, frequency content, and duration. These effects on the amplitude and frequency content of the motions near granular Fig. 8. (Color) Dynamic total head isochrones recorded in sloped sites with granular column treatment of Ar ¼ 0, 10, and 20%, during the Kobe motion. © ASCE 04018065-8 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 9. columns must be considered in evaluating liquefaction triggering and consequences as well as the design of overlying structures. Fig. 9 shows the surface settlements at the top and bottom of the slope for each of the three models. These settlements were domi- nated primarily by volumetric strains and to a smaller degree by deviatoric or shear strains, due to the slope’s small angle. For ex- ample, the settlements recorded at the top of the slope in Model 2 were quite similar to those of a similar profile with level ground in Model 0. The settlements recorded at the top of the slopes generally decreased with an increasing Ar due to the reduction in the extent and duration of large excess pore pressures and hence the reduction in volumetric strains caused by sedimentation and deviatoric strains caused by the existing static and dynamic shear stresses. These fac- tors were more significant than the increase in volumetric strains due to partial drainage. At the toe of the slope, the settlements were more strongly affected by the deviatoric strains and rotational movements from the slope above, at times causing bulging or heave in the central channel and close to the channel. Larger deviatoric displacements downslope, for example, in Model 2 (Ar ¼ 0%), led to smaller net settlements at the toe of the slope in comparison with the top, or with other models, during the first motion (Kobe) due to the bulging effect induced by rotational movements toward the cen- tral channel. The settlements generally reduced during the sub- sequent motions due to densification in the prior motions as well as the cumulative reduction in slope’s angle. Lateral displacements were recorded in the slopes with the LVDT setup shown in Fig. 3(c) and the locations shown in Fig. 1. It is noteworthy that these measurements were sensitive to the rotation of the slotted aluminum plate and in many cases malfunc- tioned. However, in comparison with the plates placed perpendicular to the slope in previous experimental studies, the horizontal LVDT measurement system had a far smaller impact on the response of the slope and the development of lateral spread- ing. The results from horizontal LVDTs that were subject to Fig. 9. (Color) Base acceleration, excess pore pressure ratio, vertical and horizontal displacement time histories recorded in sloped sites with granular column treatment of Ar ¼ 0, 10, and 20%, during the Kobe, Joshua, and Northridge motions. © ASCE 04018065-9 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 10. excessive rotation are not shown in Fig. 9. The successful record- ings as well as measurements obtained during model excavation highlight the effectiveness of the granular columns, particularly with Ar ¼ 20%, in reducing permanent lateral spreading, which is a critical measure of slope performance. This was due partly to the added shear reinforcement and partly to the faster dissipation of excess pore pressures. The contribution of each mechanism is subsequently evaluated for one case. Lateral displacements were generally reduced as the Ar in- creased. These lateral deformations, governed by deviatoric strains under the existing static and dynamic shear stresses, were in all cases greatest at the toe (shown in the second row in Fig. 9) of the slope in comparison with the top (shown in the first row in Fig. 9), and could be unacceptable even with Ar ¼ 20% (that is, on the order of 15 cm or more during the first motion). The per- manent lateral displacements reduced significantly after the first motion (Kobe) in the treated slopes (both Ar ¼ 10% and Ar ¼ 20%), whereas the large lateral deformations continued in the untreated model (Ar ¼ 0%). However, the slope angle was re- duced significantly in the untreated Model 2 after the first two mo- tions (from 3° to about 0.6° during the Kobe and to about 0° during the Joshua, calculated roughly with the two vertical LVDT recordings on the soil surface), which indicated negligible static shear stresses driving the permanent lateral deformations be- yond this point. Following the first two motions, the lateral dis- placement recordings in the untreated Model 2 were therefore less reliable and were strongly affected by the rotation of the aluminum plate (hence, not presented during the Northridge). Nevertheless, the recordings indicated excessive lateral spread potential in an untreated slope. Fig. 11 shows the cumulative settlements (greatest at the top of the slope), lateral displacements (greatest at the toe of the slope), and slope angles (calculated roughly from the two vertical LVDT recordings on the surface of the Monterey sand), as recorded during the application of three consecutive motions, when available. The results generally support the conclusion that using granular col- umns reduces the volumetric settlements and deviatoric lateral displacements in gentle slopes. The magnitude of this reduction was shown to depend strongly on Ar, the characteristics of shaking, the initial preshaking soil properties, and the slope geometry. The use of granular columns was highly effective in reducing the mag- nitude of lateral spreading. The vertical settlements were also re- duced, but additional mitigation measures may be required to limit the deformations for structures or other sensitive infrastructure. Relative Influence of Shear Stiffness and Drainage on Slope Performance The influence and relative importance of mitigation arising from the mechanisms of shear reinforcement and enhanced drainage in granular columns were evaluated experimentally by comparing the responses of Models 3 and 4. Both models contained a gentle slope treated with granular columns at Ar ¼ 10%. In Model 3, the granular columns were encased only in geotextile to avoid clog- ging, combining the drainage and shear reinforcement mechanisms. In Model 4, the dense granular columns were encased in both a geotextile and a thin (0.2-mm-thick) latex membrane to avoid drainage and provide only shear reinforcement. Fig. 12 shows the dynamic total head isochrones for Models 3 and 4 during the Kobe motion. As expected, the use of draining granular columns (Model 3) slightly reduced the net excess pore pressures generated during shaking (for example, it reduced the peak ru values to slightly below 1.0 in the liquefiable layer) and increased the rate of dissipation after shaking in comparision with the nondraining columns in Model 4 (with latex). Fig. 13 shows the faster rate of the pore pressure dissipation in Model 3 during all shaking events. The faster drainage in Model 3 limited the settlements both at the top and toe of the slope by limiting both Fig. 10. (Color) Time frequency (Stockwell spectra) of transverse accelerations at the bottom, middle, and top of liquefiable layer and container base in sloped sites with granular columns of Ar ¼ 0, 10, and 20%, during the Kobe motion. Fig. 11. (Color) Cumulative vertical and horizontal displacements re- corded in sloped sites with granular columns of Ar ¼ 0, 10, and 20%. © ASCE 04018065-10 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 11. sedimentation and shear deformations. Similar effects were ob- served in the comparison of lateral displacements shown in Fig. 13, both at the top and toe of the slope. The nondraining granular col- umns with latex extended the duration of large excess pore pres- sures, which notably amplified lateral displacements in the slope (for example, an increase from 25 to 50 cm at the slope toe during Kobe). Fig. 14 shows a comparison of the cumulative vertical and hori- zontal displacements, as well as rough estimates of the slope angle, in Models 3 and 4 during the three consecutive motions (in a man- ner similar to that shown in Fig. 11). A reduction of approximately 20% was observed in the total surface settlements after two motions due to the draining effect of granular columns. Similarly, a reduc- tion of approximately 74% was observed in the lateral toe displacements after three consecutive motions due to enhanced drainage, even though Model 3 had a slightly steeper slope follow- ing the first shake. The vertical and lateral deformations in Model 4 with the nondraining columns were quite similar to those in Model 2 without any treatment, which indicated a negligible influence from shear reinforcement on slope performance in terms of defor- mations for Ar ¼ 10%. The photographs taken during model Fig. 12. (Color) Dynamic total head isochrones recorded in sloped sites with granular columns of Ar ¼ 10%, draining (without latex, Model 3), and nondraining (with latex, Model 4). Fig. 13. (Color) Base acceleration, excess pore pressure ratio, vertical and horizontal displacement time histories in sloped sites with granular columns of Ar ¼ 10%, draining (Model 3), and nondraining (Model 4). © ASCE 04018065-11 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 12. excavation (Fig. 15) and the deformation of initially vertical col- ored sand columns confirmed the significantly larger shear defor- mations in loose Ottawa sand in Model 4 with nondraining granular columns in comparison with Model 3, and similar to Model 2. The photographs also show minimum lateral deforma- tions in Model 1, which was consistent with the recordings shown previously. These observations point to the critical impor- tance and benefits of maintaining the drainage capabilities of granular columns (for example, reducing the possibility of clog- ging in the field after subsequent events) and of not relying solely on the shear reinforcement mechanism of mitigation for reducing slope deformations. Fig. 16 shows the time-frequency Stockwell spectra of trans- verse accelerations at the container base and various depths within the loose layer of Ottawa sand during the Kobe motion in Models 2, 3, and 4. The same color scale is used at all depths to facilitate this comparison. For all models, a significant deamplification of accel- erations occurred from the base toward the surface due to soil soft- ening. The higher shear stiffness provided by the nondraining granular columns in Model 4 reduced the accelerations over a range of frequencies. In this case, the high-frequency accelerations were reduced in Model 4 in comparison with the untreated Model 2, which led to a notable reduction in surface PGA (for example, during Kobe, 0.31 g in the untreated Model 2 and 0.19 g in Model 4 with nondraining columns). The trends were consistent in other motions. But the enhanced drainage (even by a slight amount in Model 3 with Ar ¼ 10%) was shown to nullify this reduction (PGA ¼ 0.29 g) or even cause notable amplifications of accelerations in the surrounding soil (for example, as observed pre- viously for Model 1 with Ar ¼ 20%, or near the granular column in Model 0). The influence of granular columns on accelerations is therefore expected to depend strongly on the column’s drainage Fig. 14. (Color) Cumulative vertical and horizontal displacements in sloped sites with granular columns of Ar ¼ 10%, draining (Model 3), and nondraining (Model 4). Fig. 15. (Color) Deformation of colored sand rows and columns after the test during model excavation in sloped sites: draining granular columns of Ar ¼ 20% (Model 1), no drains Ar ¼ 0 (Model 2), draining granular columns of Ar ¼ 10% (Model 3), and nondraining granular columns of Ar ¼ 10% (Model 4). Fig. 16. (Color) Time frequency response (Stockwell spectra) of transverse accelerations at different depths in sloped sites with granular columns of Ar ¼ 10%, draining (Model 3) and nondraining (Model 4), and Ar ¼ 0% (Model 2) during the Kobe motion. © ASCE 04018065-12 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 13. ability and Ar as well as the soil properties, preshake slope geom- etry, and motion characteristics. Conclusions A series of highly instrumented dynamic centrifuge model tests were performed to evaluate the effectiveness of granular columns as a liquefaction countermeasure and its properties on the perfor- mance of level and gently sloping sites. The primary conclusions drawn from the experimental results are the following: • Drains in isolation or at the corner of a group cannot be relied upon for preventing liquefaction if a large influx of fluid from the far-field is possible. The conventional design charts do not account for such cases. However, even drains without adjacent drains can be somewhat effective in speeding up the dissipation of excess pore pressures in their vicinity after shaking, and they tend to reduce net volumetric settlements. • The use of granular columns in gentle, layered, liquefiable slopes was found to be effective in reducing the lateral and ver- tical deformations. The drain’s area replacement ratio (Ar) was experimentally shown to be a critical parameter in reducing both settlements and lateral spreading in the slope. In this study, the best performance was achieved when liquefaction was pre- vented. The data presented here will be useful for the validation of future numerical models used for optimizing design pro- cedures. • The influence of granular columns on the seismic demand and accelerations experienced by the surrounding soil was shown to depend on their Ar, drainage capability, and the frequency of interest. Generally, shear reinforcement provided by the col- umns tended to deamplify the accelerations over a range of fre- quencies and reduced the PGA. However, enhanced drainage could simultaneously increase the rate of excess pore pressure dissipation and recovery of shear stiffness (and hence, reduction in damping) in its surrounding soil, amplifying accelerations, particularly at higher frequencies. The possible increase in accelerations (and in particular, the PGA) at the surface of a treated, draining site must be considered in evaluating liquefac- tion triggering, consequences, and response of the overlying structures. • The possible reduction in accelerations and seismic demand due to the shear reinforcement provided by granular columns cannot be entirely relied upon for reducing the likelihood of liquefac- tion triggering nor its consequences in terms of deformation. For the particular Ar investigated (10%) here, the beneficial influ- ence of granular columns on slope deformations was shown to be due almost entirely to their ability to enhance drainage, and not to shear reinforcement. However, this may not be the case for higher Ar values or for cases in which notable ground densification is expected during column installation. Neverthe- less, the experimental results demonstrate that the influence of enhanced drainage is vital. It is critical for engineers to avoid clogging in these columns by using appropriate geotextile filters. • For soils with a high silt content, the hydraulic conductivity, and thereby the zone of influence, is reduced. Therefore, the benefits of drainage will be less visible (Adalier et al. 2003). This may necessitate the use of higher area replacement ratios or possibly different mitigation strategies. Despite the inherent simplifications and uncertainties in centri- fuge modeling, these experimental results and discussion are aimed to guide the future field, physical, and numerical model studies to- ward more reliable and performance-based liquefaction mitigation design methodologies. Additional tests considering a wider range of soil, slope, drain, and ground motion properties are needed with parallel nonlinear numerical simulations in order to clarify the seis- mic response of liquefiable deposits treated by granular columns and the relative importance of reinforcement, drainage, and ground densification in different kinds of soils; improve the existing design procedures; and provide recommendations for practitioners. Acknowledgments The authors would like to acknowledge the support of the depart- ment of Civil, Environmental, and Architectural Engineering at the University of Colorado Boulder and the assistance of Mr. Balaji Paramasivam, Mohamed Elmansouri, and Simon Petit in the execu- tion of the centrifuge experiments presented in this paper. References Adalier, K. 1996. “Mitigation of earthquake induced liquefaction hazards.” Ph.D. thesis, Dept. of Civil Engineering, Rensselaer Polytechnic Institute, 659. Adalier, K., A. Elgamal, J. Meneses, and I. J. Baez. 2003. “Stone columns as liquefaction counter-measure in non-plastic silty soils.” J. Soil Dyn. Earthquake Eng. 23 (7): 571–584. https://doi.org/10.1016/S0267-7261 (03)00070-8. Asgari, A., M. Oliaei, and M. Bagheri. 2013. “Numerical simulation of improvement of a liquefiable soil layer using granular column and pile-pinning techniques.” Soil Dyn. Earthquake Eng. 51: 77–96. https://doi.org/10.1016/j.soildyn.2013.04.006. Baez, J. I. 1995. “A design model for the reduction of soil liquefaction by vibro-stone columns.” Ph.D. thesis, Dept. of Civil and Environmental Engineering, Univ. of Southern California. Baez, J. I. 1996. “Seismic ground improvement for bridge sites.” In Proc., 4th CALTRANS Seismic Research Workshop. Sacramento, CA: Califor- nia Dept. of Transportation, Engineering Service Center. Baez, J. I., and G. R. Martin. 1993. “Advances in the design of vibro sys- tems for the improvement of liquefaction resistance.” In Proc., Symp. on Ground Improvement. Vancouver, BC, Canada: Vancouver Geotechni- cal Society. Bartlett, S. F., and T. L. Youd. 1992. Empirical analysis of horizontal ground displacement generated by liquefaction-induced lateral spread. Technical Rep. No. NCEER-92-0021. Buffalo, NY: State Univ. of New York. Boulanger, R., I. Idriss, D. Stewart, Y. Hashash, and B. Schmidt. 1998. “Drainage capacity of stone columns or gravel drains for mitigating liquefaction.” In Vol. 1 of Proc., Geotechnical Earthquake Engineering and Soil Dynamics III, 678–690. Reston, VA: ASCE. Dashti, S., J. D. Bray, J. M. Pestana, M. R. Riemer, and D. Wilson. 2010. “Centrifuge testing to evaluate and mitigate liquefaction-induced build- ing settlement mechanisms.” J. Geotech. Geoenviron. Eng. 136 (7): 918–929. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000306. Elgamal, A., J. Lu, and D. Forcellini. 2009. “Mitigation of liquefaction- induced lateral deformation in a sloping stratum: Three-dimensional numerical simulation.” J. Geotech. Geoenviron. Eng. 135 (11): 1672–1682. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000137. Howell, R., E. M. Rathje, R. Kamai, and R. Boulanger. 2012. “Centrifuge modeling of prefabricated vertical drains for liquefaction remediation.” J. Geotech. Geoenviron. Eng. 138 (3): 262–271. https://doi.org/10.1061 /(ASCE)GT.1943-5606.0000604. INA (International Navigation Association). 2001. Seismic design guide- lines for port structures. Tokyo: A.A. Balkema. ISSMGE (International Society of Soil Mechanics and Geotechnical Engineering). 2001. Case histories of post-liquefaction remediation. Rep. Technical Committee for Earthquake Geotechnical Engineering TC4. Tokyo: Japan Japanese Geotechnical Society. Ketchum, S. A. 1989. “Development of an earthquake motion simulator for centrifuge testing and the dynamic response of a model sand © ASCE 04018065-13 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.
  • 14. embankment.” Ph.D. dissertation, Dept. of Civil, Environmental, and Architectural Engineering, Univ. of Colorado Boulder. Koelling, M., and S. Dickenson. 1998. “Ground improvement case histories for liquefaction mitigation of port and near-shore structures.” In Vol. 1 of Proc., Geotechnical Earthquake Engineering and Soil Dynamics III, 614–626. Reston, VA: ASCE. Kramer, S. L., S. S. Sideras, and M. W. Greenfield. 2016. “The timing of liquefaction and its utility in liquefaction hazard evaluation.” Soil Dyn. Earthquake Eng. 91: 133–146. https://doi.org/10.1016/j.soildyn.2016 .07.025. Mitchell, J. K. 1986. “Material behavior and ground improvement.” In Vol. 2 of Proc., Int. Conf. on Deep Foundations. 1.1–1.17. Beijing. Mitchell, J. K. 1988. “Densification and improvement of hydraulic fills.” In Proc., Conf. on Hydraulic Fill Structures, 608–633. Reston, VA: ASCE. Mitchell, J. K., C. D. P. Baxter, and T. C. Munson. 1995. “Performance of improved ground during earthquakes.” In Proc., Soil Improvement for Earthquake Hazard Mitigation, 1–36. Reston, VA: ASCE. Ramirez, J. C., M. Badanagki, M. Rahimi, M. A. ElGhoraiby, M. T. Manzari, S. Dashti, A. Barrero, M. Taiebat, K. Ziotopoulou, and A. Liel. 2017. “Seismic performance of a layered liquefiable site: Validation of numerical simulations using centrifuge modeling.” In Proc., 2017 GeoFrontiers. Reston, VA: ASCE. Rayamajhi, D., S. A. Ashford, R. W. Boulanger, and A. Elgamal. 2016a. “Dense granular columns in liquefiable ground. I: Shear reinforcement and cyclic stress ratio reduction.” J. Geotech. Geoenviron. Eng. 142 (7): 4016023. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001474. Rayamajhi, D., S. A. Ashford, R. W. Boulanger, and A. Elgamal. 2016b. “Dense granular columns in liquefiable ground. II: Effects on deforma- tions.” J. Geotech. Geoenviron. Eng. 142 (7): 4016024. https://doi.org /10.1061/(ASCE)GT.1943-5606.0001475. Seed, H. B. 1987. “Design problems in soil liquefaction.” J. Geotech. Eng. 113 (8): 827–845. https://doi.org/10.1061/(ASCE)0733-9410(1987) 113:8(827). Seed, H. B., and J. R. Booker. 1977. “Stabilization of potentially liquefiable sand deposits using gravel drains.” J. Geotech. Eng. Div. 103 (7): 757–768. Seed, H. B., and I. M. Idriss. 1970. Soil moduli and damping factors for dynamic response analyses. Technical Rep. No. EERC 70-10. Berkeley, CA: Earthquake Engineering Research Center, Univ. of California. Seed, H. B., R. T. Wong, I. M. Idriss, and K. Tokimatsu. 1986. “Moduli and damping factors for dynamic analyses of cohesionless soil.” J. Geotech. Eng. 112 (11): 1016–1032. https://doi.org/10.1061/(ASCE)0733-9410 (1986)112:11(1016). Stewart, D. P., Y. R. Chen, and B. L. Kutter. 1998. “Experience with the use of methylcellulose as a viscous pore fluid in centrifuge models.” Geotech. Test. J. 21 (4): 365–369. https://doi.org/10.1520/GTJ11376J. Stockwell, R. G., L. Mansinha, and R. P. Lowe. 1996. “Localization of the complex spectrum: The S transform.” IEEE Trans. Signal Process. 44 (4): 998–1001. https://doi.org/10.1109/78.492555. Stringer, M., and S. Madabhushi. 2009. “Novel computer-controlled saturation of dynamic centrifuge models using high viscosity fluids.” Geotech. Test. J. 32 (6): 559–564. https://doi.org/10.1520/GTJ102435. Tan, T. S., and R. F. Scott. 1985. “Centrifuge scaling considerations for fluid-particle systems.” Geotechnique 35 (4): 461–470. https://doi .org/10.1680/geot.1985.35.4.461. Taylor, R. N. 1995. Geotechnical centrifuge technology. 1st ed. New York: Blackie Academic and Professional. Tokimatsu, K., and A. Asaka. 1998. “Effects of liquefaction-induced ground displacements on pile performance in the 1995 Hyogoken- Nambu earthquake.” Soils Found. 38: 163–177. https://doi.org/10 .3208/sandf.38.Special_163. Walker, A. J., and H. E. Stewart. 1989. “Cyclic undrained behavior of nonplastic and low plasticity silts.” Technical Rep. No. NCEER-89- 0035. Buffalo, NY: National Center for Earthquake Engineering Research. © ASCE 04018065-14 J. Geotech. Geoenviron. Eng. J. Geotech. Geoenviron. Eng., 2018, 144(9): 04018065 Downloadedfromascelibrary.orgbyColoradoUniversityatBoulderon07/24/18.CopyrightASCE.Forpersonaluseonly;allrightsreserved.