Nat Hazards (2012) 64:797–820

DOI 10.1007/s11069-012-0272-8

ORIGINAL PAPER

Experimental and numerical investigation of slope

stabilization by stone columns

Mustafa Vekli • Mustafa Aytekin • S. Banu İkizler • Ümit Çalik

Received: 27 April 2011 / Accepted: 23 June 2012 / Published online: 14 July 2012
Ó Springer Science+Business Media B.V. 2012

Abstract In this study, an investigation has been performed on a small-scaled laboratory

model and its numerical model by the code of PLAXIS to see the effect of stone columns
(SCs) placed vertically in a soft soil slope in terms of slope stability, bearing capacity, and
settlements. Also, several hypothetical cases have been examined by the code. Effect of
s/D ratios (distance between the vertical axes of SCs/diameter of SCs) was also investi-
gated on slope stability, ultimate bearing capacity, and settlement of a footing rested on top
of the slope on the laboratory model. Firstly, ultimate bearing capacity and settlement
properties of soil were determined for unreinforced soil that is no SCs were considered.
Then, some values of soil were determined after the installation of stone columns with
various ratios of s/D. The ratios of s/D were 2, 3, 3.5, and 4. The tests carried out on the
laboratory model were simulated and numerically analyzed in two dimensions under plain-
strain conditions by Mohr–Coulomb model. In the analyses, PLAXIS computer code,
which is based on finite elements method, has been employed. Then, a parametric
investigation was carried out to see the effect of SCs on the stability of the slope. In the
parametric investigation, several hypothetical cases that were one layer of soil and two
layers of soil with the presence of water in the reservoir side of the slopes were examined.
The analyses in the investigation were performed by the PLAXIS code for various slope

M. Vekli (&)
Department of Civil Engineering, Engineering and Architecture Faculty,
Bozok University, Yozgat, Turkey
e-mail: mustafavekli@gmail.com

M. Aytekin
Department of Civil Engineering and Architecture, University of Bahrain,
Isa Town, Kingdom of Bahrain
e-mail: maytekin1@gmail.com

S. Banu İkizler
Department of Civil Engineering, Blacksea Technical University, Trabzon, Turkey
e-mail: banuh73@gmail.com

Ü. Çalik
10th Region Directorate of Turkish Highways, Trabzon, Turkey
e-mail: infoumitcalik@hotmail.com

123
798 Nat Hazards (2012) 64:797–820

angles b, ratios of c/(cH), and ratios of s/D. From the test results of the laboratory model,
and the results obtained from the numerical analyses, it was observed that the bearing
capacity of the footing constructed on the top of the slope in soft soil was increased;
settlements were decreased after the improvement with SCs. From the analyses performed,
it was found that the SCs increased the stability of slope 1.18- to 1.62-fold as a relative
effect of different parameters.

Keywords Finite element analyses
Laboratory model
Slope stability

Soil improvement
Stone column

1 Introduction

In last decades, rapid increment of population rate, economic, and social developments has
forced to use lands that were unsuitable for construction earlier, so that new engineering
solutions must be generated to use these types of lands. Improvement of soils in these lands
became an important matter for civil engineers and especially for geotechnical engineers.
In the years of 1950 s, soil improvement has become an important application area in
geotechnical engineering and many numbers of improvement techniques were developed.
Stone column technique that is one of the soil improvement techniques in which some
additive materials are added into the soil can be defined as a good improvement technique
to improve bearing capacity, settlement, reduction on liquefaction risk, and slope stability
in soft and loose soils that do not have the proper quality for the planned super structure.
Stone column technique has been used commonly for 30 years in all over the world. It has
been proved by the in situ researches that they have good effect on the improvement of
soils in many field applications (Osmanoglu 1999; Connor and Gorski 2000; Kumar 2001;
Nalçakan 2004; Heitz et al. 2005). Munfakh 1984; Han and Ye 2002 pointed out that the
SCs have the rate of consolidation increased. In addition, they mentioned that the pressures
on SCs were increased during the consolidation while the pressures were reduced on
surrounding soil. Furthermore, they showed that the pressure concentration factor was
become stable at the end of consolidation.
Use of SCs in soft soils has emerged as a major ground improvement technique for the
last three decades. Stone columns are used to increase the bearing capacity of soft soils
(Madhav and Vitkar; 1978; Murugesan and Rajagopal 2006; Ambily and Gandhi 2007), to
reduce settlements (Lee and Pande 1998; Nalçakan 2004; Tan and Khine 2005; Patel and
Shroff 2005), to decrease the risk of liquefaction (Baez 1995; Adalier et al. 2003; Han and
Ye 2002), and to increase stabilities of natural slopes and embankments (Christoulas et al.
1997; Connor and Gorski 2000; Kirsch and Sondermann 2003).
Many researchers have developed theoretical solutions to estimate bearing capacity and
settlement behavior of soils improved by SCs (Greenwood 1970; Hughes et al. 1975;
Madhav and Vitkar 1978; Aboshi et al. 1979; Greenwood and Kirsch 1984). On the other
hand, experimental studies have shown that the magnitude of improvement on bearing
capacity and settlements by SCs (Ambily and Gandhi 2007; Guetif et al. 2007; Black et al.
2007). However, there is not much intention about improvement on the bearing capacity
and settlement characteristics of the footings rest on soft soil near a slope. Investigations
about the slope stability have been focused on the improvement of engineering properties
of natural soils lying under man-made slopes, especially under filled areas. Thus, the main
objective of these investigations was to improve the stability of these kinds of slopes.

123
Nat Hazards (2012) 64:797–820 799

Fig. 1 Grain size distribution for soft clay and stones

Table 1 Properties of soft clay and stones

Material Water Modulus Poisson Cohesion, / w cdry cbulk
content, of elasticity, ratio, l cu (kPa) (°) (°) (kN/m3) (kN/m3)
w (%) E (kPa)

Soft clay 80 3,000 0.33 5 4 0 9.8 17.64

Stone – 75,000 0.3 0.5 41 9 13.8 15.9

Fig. 2 Model footing

In the present study, a small-scaled experimental model in the laboratory has been
developed to investigate the bearing capacity and settlement behavior of a footing rested on
a soft clay slope. Then, the effect of (s/D) ratios of distances (s) between SCs axes and SCs
diameters (D) on the bearing capacity and settlement behavior of footing was investigated
by the code of PLAXIS. At the end of the experiments, bearing capacity improvement
factors (BCIFs) and settlement reduction ratios (SRRs) on footing have been estimated.
Experimental and theoretical investigations supporting experimental results are very lim-
ited on the usage of SCs even if the SCs have been used in practice in several locations.

123
800 Nat Hazards (2012) 64:797–820

Fig. 3 Experimental setup

Fig. 4 Installed stone columns seen from the top

Therefore, in this study, an experimental and numerical investigation have been performed
on a soft clay slope that was improved by SCs to see the effect of them on the stability of
slope as well as on the bearing capacity and settlement of shallow footings rested on top of
the slope. The results of the experimental model and the results of the numerical model in
which Plaxis (Plaxis, BV 2006) code was employed have been compared. On the other
hand, a parametric study in which several hypothetical cases were considered has been
performed to see the effect of SCs on the stability of slopes for various soil conditions and
different slope angles.

2 Experimental program

2.1 Properties of materials

The basic materials used in this study were soft clay, stones, and model footing that have
been described as follows.

123
Nat Hazards (2012) 64:797–820 801

Fig. 5 Different applications of the SCs for different ratios of s/D

Fig. 6 The boundary conditions and generated mesh for soft clay slope unreinforced soil SCs and with SCs
for s/D = 3.5

2.1.1 Soft clay

The soft clay used in the experiments was classified as CH by the plasticity chart. It was
excavated from Fatih campus of Karadeniz Technical University. The soft clay was sieved
through 4.75-mm sieve to remove the coarser fraction. Its grain size distribution curve is
shown in Fig. 1. The other properties are the liquid limit = 94 %, plastic limit = 38 %,

123
802 Nat Hazards (2012) 64:797–820

Fig. 7 Slope geometry of one-layered clay case

Fig. 8 The geometry of two-layered soil

Fig. 9 Presence of water in the reservoir side of the slope

optimum moisture content = 27 %, max dry unit weight = 16.10 kN/m3, and specific
gravity = 2.78. The soft clay was thoroughly mixed by a mixer and placed into the model
box with slope angle. Elastic parameters (elasticity modulus and Poisson’s ratio) of the soft
clay were determined in several series of undrained triaxial shear tests at the moisture
content used in the tests. Other properties of the soft clay are given in Table 1.

2.1.2 Stones

Grain size distribution of aggregate (stones) used in the formation of SCs is given in Fig. 1.
The cmax and cmin of the aggregate are 15.9 and 13.8 kN/m3, respectively. The angle of
internal friction has been determined using direct shear tests. The stones were compacted

123
Nat Hazards (2012) 64:797–820 803

Fig. 10 Variations of q with S/B for model slope with different values of s/D and no SCs

Fig. 11 Variations of q with S/B for FEM model footing with different values of s/D and no SCs

Table 2 Ultimate bearing change by s/D

No stone columns s/D = 2 s/D = 3 s/D = 3.5 s/D = 4

q (kPa) experimental 18.96 32.56 28.56 27.72 23.76

q (kPa) FEA 20.48 33.99 29.6 28.3 23.98

to a unit weight of 15 kN/m3, which could be carried out while constructing the SCs for the
experiments and sheared at a constant rate of 1 mm/min under normal pressure of 75, 100,
250 kPa. Modulus of elasticity of the stones is the constrain modulus obtained by loading

123
804 Nat Hazards (2012) 64:797–820

Table 3 Experimental and FEA settlements’ change by s/D

No stone columns s/D = 2 s/D = 3 s/D = 3.5 s/D = 4

Settlement (mm) experimental 40.48 14.99 18.17 20.98 26.86

Settlement (mm) FEM 43.78 15.62 17.17 19.92 27.82
Bearing Capacitiy Increment Ratio

1,80
Plaxis

Experimental
1,60

1,40

1,20

1,00
1 1,5 2 2,5 3 3,5 4 4,5
s/D

Fig. 12 s/D-bearing capacity increment ratio

0,8
Settlement Reduction Ratio

0,6

0,4

0,2
Plaxis
Experimental
0
1 1,5 2 2,5 3 3,5 4 4,5
s/D

Fig. 13 s/D-settlement reduction ratio

to the stones in a cylindrical mold of 152.4 mm in diameter and 177.8 mm in height at an

initial unit weight of 15 kN/m3. Other properties of the stones are also shown in Table 1.

2.1.3 Model footing

A model strip rigid footing made of steel with spherical cavity at its top center to
accommodate loading ring was used. The footing had 200 mm width, 150 mm length, and
20 mm thickness. The footing is shown in Fig. 2.

123
Nat Hazards (2012) 64:797–820 805

2,500
c/( H)= 0.1
c/( H)= 0.05
2,000 c/( H)= 0.025
Safety Factor c/( H)= 0.01

1,500

1,000

0,500

0,000
15 17 19 21 23 25 27 29 31 33 35
Slope Angle

Fig. 14 Relationship between safety factor and slope angle for s/D = 2

2,500
c/( H)= 0.1
c/( H)= 0.05
2,000 c/( H)= 0.025
c/( H)= 0.01
Safety Factor

1,500

1,000

0,500

0,000
15 17 19 21 23 25 27 29 31 33 35
Slope Angle

Fig. 15 Relationship between safety factor and slope angle for s/D = 3

2,0
c/( H)= 0.1
c/( H)= 0.05
c/( H)= 0.025
1,5
c/( H)= 0.01
Safety Factor

1,0

0,5

0,0
15 17 19 21 23 25 27 29 31 33 35
Slope Angle

Fig. 16 Relationship between safety factor and slope angle for s/D = 4

123
806 Nat Hazards (2012) 64:797–820

Table 4 Values of INSF for one-layered clay with different slope angles and s/D ratios
Slope angle INSF

s/D = 2 s/D = 3 s/D = 4

c/(cH) c/(cH) c/(cH)

0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01

18.4 1.32 1.39 1.51 1.53 1.20 1.30 1.38 1.45 1.16 1.19 1.26 1.34
26.6 1.34 1.50 1.54 1.68 1.21 1.33 1.49 1.53 1.14 1.26 1.27 1.37
33.7 1.40 1.50 1.58 1.82 1.21 1.33 1.39 1.66 1.20 1.18 1.22 1.41

3,0

2,5
Safety Factor

2,0

1,5

1,0
c/( H)= 0.1
c/( H)= 0.05
0,5 c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 17 The relationship between c2/c1 and the safety factors for s/D = 2 with the slope of (3:1)

2,5

2,0
Safety Factor

1,5

1,0
c/( H)= 0.1
0,5 c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 18 The relationship between c2/c1 and the safety factors for s/D = 3 with the slope of (3:1)

2.2 Experimental setup

In the experimental setup, the best in situ condition in the laboratory was established, so
that the problem could be seen closely, and the parameters that affect the slope behavior
could be controlled. In the experimental model, it was aimed to determine the total and

123
Nat Hazards (2012) 64:797–820 807

2,5

2,0
Safety Factor
1,5

1,0
c/( H)= 0.1
0,5 c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 19 The relationship between c2/c1 and the safety factors for s/D = 4 with the slope of (3:1)

2,0

1,5
Safety Factor

1,0

c/( H)= 0.1

0,5 c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2/c1

Fig. 20 The relationship between c2/c1 and the safety factors for s/D = 2 with the slope of (2:1)

2,0

1,5
Safety Factor

1,0

c/( H)= 0.1

0,5 c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 21 The relationship between c2/c1 and the safety factors for s/D = 3 with the slope of (2:1)

123
808 Nat Hazards (2012) 64:797–820

2,0

Safety Factor 1,5

1,0

c/( H)= 0.1

0,5
c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 22 The relationship between c2/c1 and the safety factors for s/D = 4 with the slope of (2:1)

1,5
Safety Factor

1,0

0,5
c/( H)= 0.1
c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 23 The relationship between c2/c1 and the safety factors for s/D = 2 with the slope of (3:2)

1,5

1,0
Safety Factor

0,5 c/( H)= 0.1

c/( H)= 0.05
c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 24 The relationship between c2/c1 and the safety factors for s/D = 3 with the slope of (3:2)

123
Nat Hazards (2012) 64:797–820 809

1,2

1,0

Safety Factor 0,8

0,6

0,4
c/( H)= 0.1
c/( H)= 0.05
0,2 c/( H)= 0.025
c/( H)= 0.01
0,0
0 0,5 1 1,5 2 2,5
c2 /c1

Fig. 25 The relationship between c2/c1 and the safety factors for s/D = 4 with the slope of (3:2)

Table 5 The INSF values estimated for two-layered clay for different slope angles, c2/c1, and s/D rates
c2/c1 b = 18.4(3:1) b = 26.6(2:1) b = 33.7(3:2)

c/(cH) c/(cH) c/(cH)

0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01

s/D = 2
0.5 1.37 1.41 1.52 1.51 1.40 1.60 1.63 1.66 1.36 1.44 1.55 1.59
1 1.30 1.37 1.48 1.49 1.33 1.49 1.55 1.65 1.37 1.45 1.55 1.62
2 1.33 1.43 1.52 1.54 1.20 1.47 1.56 1.72 1.36 1.50 1.54 1.66

s/D = 3
0.5 1.21 1.32 1.44 1.44 1.27 1.41 1.42 1.52 1.23 1.32 1.41 1.49
1 1.18 1.28 1.38 1.43 1.22 1.32 1.37 1.55 1.23 1.34 1.35 1.51
2 1.23 1.34 1.39 1.47 1.17 1.32 1.40 1.53 1.30 1.37 1.36 1.55

s/D = 4
0.5 1.17 1.20 1.28 1.38 1.15 1.31 1.32 1.44 1.19 1.23 1.31 1.42
1 1.15 1.18 1.24 1.36 1.14 1.26 1.28 1.42 1.20 1.25 1.29 1.46
2 1.21 1.26 1.29 1.38 1.12 1.24 1.31 1.41 1.20 1.24 1.28 1.50

local deformations of the footing rested on the soft clay slope that was built in the labo-
ratory and to establish a relationship between vertical load and deformation under the
footing on soft clay slope after the SCs were installed.
The test box, having dimensions of 1,000 9 200 mm in plan and 800 mm in depth, was
manufactured from fiberglass with the front wall made of 20 mm and the other walls (back,
right, and left walls) made of 25 mm thickness plexiglas and supported by two horizontal
steel beams as seen in Fig. 3. The fiberglass side of the box (front wall) allowed the visual
inspection of the sample during the construction of soft clay slope and the observation of
slope deformations during the tests.

123
810 Nat Hazards (2012) 64:797–820

Fig. 26 Relationships between

safety factor, slope angle, and
fullness percentages for s/D = 2
and c/(cH) = 0.1

Fig. 27 Relationships between

safety factor, slope angle, and
fullness percentage for s/D = 2
and c/(cH) = 0.05

The test box was built rigid enough to maintain plane strain conditions by minimizing
out-of-plane displacement. In order to achieve this purpose, the box was taken in frames
that were made of six L-shaped steel profiles. During the test, lateral displacement was
measured by gages if there was any displacement. No lateral displacement was observed as
expected. The inner surface of the test box walls was made of fiberglass to minimize the
friction between the soil and inside surface of the walls. The loading system was consisted
of a hydraulic jack and calibrated load ring. The capacities of the jack and the load ring
were 50 and 4.5 kN, respectively. The load ring was made of steel with a diameter of
200 mm. The load was applied to the center of the footing through a proving ring at a
constant displacement rate of 0.5 mm/min. The load was observed for equal intervals of
settlements until the failure occurred. The dial gauges were placed on opposite sides across
the center of the footing (Fig. 3).

123
Nat Hazards (2012) 64:797–820 811

Fig. 28 Relationships between

safety factor, slope angle, and
fullness percentages for s/D = 2
and c/(cH) = 0.025

Fig. 29 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 2 and c/(cH) = 0.01

2.2.1 Preparation of clay soil

Water was added to the air-dried clay to have a water content of 80 % and mixed by an
electric mixer to get a homogeny distribution of water in the clay. Then, the mixture was
allowed to 24-h curing period to become homogeneous in terms of water content distri-
bution. The consistency limit and liquidity index of the clay were determined in the
laboratory as 0.25 and 0.75, respectively. After 24 h, moisture content of clay was mea-
sured to make sure it was 80 %. Then, the slope was constructed in the test box by three
units in vertical and two units in the horizontal direction (3:2). The slope was constructed
layer by layer with a thickness of 50 mm under the control of a constant unit weight. Equal
masses of clay were used for each layer. In addition, samples were taken from each layer
after the establishment of layers. The tests were started after it had been seen that the

123
812 Nat Hazards (2012) 64:797–820

Fig. 30 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 3 and c/(cH) = 0.1

Fig. 31 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 3 and c/(cH) = 0.05

density of the samples was close enough. In order to build the slope in correct angle, plates
with a width of 20 mm and a thickness of 2 mm were employed. The built-up procedure
was repeated until 700 mm slope high the capacity of the test box was reached. The
loading system was consisted of a hydraulic jack and calibrated load ring. The deforma-
tions occurred under the model footing during the test were measured by two dial gauges
with a sensitivity of 0.001 mm placed two sides of the model footing. The tests were
stopped when deformations were increasing but the load stayed in a constant. Each test was
performed twice, and similar results were obtained.

2.3 Stone column installation

All the stone columns had a diameter of 100 mm in the tests in which the stone columns
were used. The stone used in the construction of the SCs had a grain size distribution curve

123
Nat Hazards (2012) 64:797–820 813

Fig. 32 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 3 and c/(cH) = 0.025

Fig. 33 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 2 and c/(cH) = 0.01

as seen in Fig. 1. The stone was washed and made it ready to build in the SCs. The reason
that the stones were washed out was to avoid clay size materials could be mixed into the
SCs. The SCs were installed by a replacement method. An open-ended steel pipe with a
wall thickness of 2.2 mm was used for installation of the SCs. Slight grease was applied on
both inner and outer surface of the pipe to facilitate its easy penetration and withdrawal
without any significant disturbance to the surrounding soil. Pipe with an inter diameter of
100 mm was penetrated to soft soil slope in a vertical direction in their places that had been
determined earlier. In order to estimate the critical shear surface, the STABL5 (1978) code
was employed for the experimental model and the critical shear surface was assumed the
surface that have the minimum factor of safety. The clay within the pipe was scooped out
using a helical auger of 90 mm diameter. Then, the stones washed earlier were charged
into the penetrated pipe layer by layer with a relative density of 61 % and a unit weight of
15 kN/m3. Each layer was compacted equally, and the pipe was taken up slowly until the
SCs were constructed up to the surface of the slope. The completed SCs are seen in Fig. 4.

123
814 Nat Hazards (2012) 64:797–820

Fig. 34 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 4 and c/(cH) = 0.1

The SCs’ lengths were determined of 100 mm below the critical failure surface. The
distance (s) between the vertical axes of the SCs and the diameter (D) of the SCs were
considered while selecting the number of the SCs installed to the slope. The tests were
performed for the s/D ratios of 2, 3, 3.5, and 4. The estimated critical failure surface, the
lengths and placement of SCs are shown for different s/D ratios in Fig. 5.

3 Finite element analyses (FEA)

In the computer code PLAXIS, a two-dimensional geometrical model of the test box was
modeled by considering the geometrical shape of the sloped clay placed in it without
reinforced soil and with SCs. The two-dimensional geometrical model had a width of
1,000 mm and a total high of 700 mm as seen in Fig. 6. In the modeling of the soft clay, 15
nodal triangle elements were employed.
Fig. 35 Relationships between safety factor,
slope angle, and fullness percentages for
s/D = 4 and c/(cH) = 0.05

123
Nat Hazards (2012) 64:797–820 815

Fig. 36 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 4 and c/(cH) = 0.025

In the analyses, the behavior of clay and the stone used to build SCs was introduced to
the code as the Mohr–Coulomb (MC) soil model. In the code, modulus of elasticity and
Poisson’s ratio estimated by the undrained-unconsolidated (UU) triaxial tests were used.
The coefficient of earth pressure at-rest (K0) was estimated by the formulae 1 - sin/,
which is used by PLAXIS. The estimated parameters of MC are given in Table 1. The
safety factor of the slope unreinforced soil and no load was found to be 3.4 using by
PLAXIS’ ‘‘reduction strength method.’’

4 Parametrical study

The most important factors on the analysis of slope stability are slope geometry, soil
properties, and the loads acting on the slope (Hammouri et al. 2008). Once the experi-
mental and numerical works had been ended up with results that were very close to each
other, a parametrical study was performed to investigate the effect of the SCs on the
stability of slopes. In the parametrical study, it was aimed to investigate the variation on
the factor of safety against sliding. For this objective, three hypothetical cases were con-
sidered. Firstly, one-layered slope, then two-layered slope in clay, finally a clay slope
behind that water existed were analyzed with SCs and unreinforced soil. In the hypothetical
cases, the high of slope and slope angle are H and b, respectively. The shear strength
parameters of the soil were taken as c and /, unit weight of soil was taken as c.
In addition to this objective, the ratios of the factor of safety for slopes stabilized by the
SCs to the factor of safety for slopes without SCs were employed. This ratio is called as
‘‘Improvement Number of Safety Factor, INSF’’. An INSF value has been estimated for
each hypothetical case.

4.1 One-layered clay

The safety factors of one-layered homogeny and drained clay slope that has a geometry as
seen in Fig. 7 were estimated by PLAXIS for the ratios c/(cH) of 0.1, 0.05, 0.025, 0.01 and
slope angles b of 18.4° (3:1), 26.6° (2:1), 33.7° (3:2) unreinforced soil. New safety factors

123
816 Nat Hazards (2012) 64:797–820

Fig. 37 Relationships between safety factor,

slope angle, and fullness percentages for
s/D = 4 and c/(cH) = 0.01

Table 6 Values of INSF for different fullness percentages

Fullness b = 18.4(3:1) b = 26.6(2:1) b = 33.7(3:2)
percentage (%)
c/(cH) c/(cH) c/(cH)

0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01 0.1 0.05 0.025 0.01

s/D = 2
100 1.13 1.21 1.35 1.41 1.11 1.31 1.37 1.52 1.20 1.32 1.44 1.54
80 1.18 1.26 1.36 1.39 1.17 1.36 1.43 1.53 1.26 1.35 1.40 1.58
60 1.22 1.30 1.40 1.39 1.22 1.40 1.46 1.52 1.31 1.35 1.42 1.49
40 1.25 1.33 1.41 1.38 1.25 1.44 1.52 1.55 1.33 1.35 1.41 1.48
20 1.27 1.33 1.44 1.37 1.28 1.49 1.54 1.49 1.33 1.35 1.40 1.45
0 1.27 1.30 1.39 1.37 1.28 1.48 1.54 1.47 1.33 1.33 1.39 1.45
s/D = 3
100 1.07 1.17 1.30 1.35 1.06 1.20 1.26 1.41 1.12 1.23 1.33 1.50
80 1.11 1.21 1.31 1.34 1.12 1.24 1.31 1.46 1.17 1.26 1.28 1.53
60 1.14 1.24 1.35 1.36 1.17 1.27 1.33 1.46 1.19 1.24 1.32 1.41
40 1.16 1.26 1.38 1.33 1.18 1.32 1.35 1.50 1.18 1.28 1.34 1.41
20 1.17 1.26 1.36 1.32 1.19 1.37 1.36 1.45 1.16 1.29 1.35 1.32
0 1.17 1.25 1.37 1.33 1.20 1.33 1.35 1.45 1.16 1.25 1.29 1.32
s/D = 4
100 1.06 1.11 1.19 1.31 1.04 1.17 1.20 1.31 1.10 1.18 1.20 1.28
80 1.08 1.15 1.22 1.32 1.07 1.19 1.24 1.38 1.14 1.22 1.20 1.24
60 1.10 1.17 1.25 1.34 1.09 1.21 1.25 1.38 1.16 1.21 1.26 1.29
40 1.12 1.18 1.26 1.33 1.10 1.22 1.28 1.44 1.18 1.23 1.24 1.30
20 1.12 1.18 1.26 1.32 1.11 1.25 1.28 1.43 1.16 1.23 1.23 1.28
0 1.13 1.16 1.26 1.33 1.12 1.24 1.28 1.43 1.16 1.22 1.22 1.29

123
Nat Hazards (2012) 64:797–820 817

were estimated in the same manner for same ratios of c/(cH) and slope angles b but there
were SCs with the s/D ratios of 2, 3, and 4. In all the hypothetical cases, internal friction
angle (/c) of clay was assumed as 10°, internal friction angle (/s) of stone used for the SCs
was assumed as 40°, and the diameter of the SCs was taken as 1,000 mm, the s/D ratios
were 2, 3, and 4.

4.2 Two-layered clay

The slope geometry for two-layered clay is seen in Fig. 8. The ratios of c/(cH) in one-
layered clay case were used in this case as the ratios of c1/(cH). In this case, bottom end of
the SCs assumed to be inserted the bottom layer that has higher bearing capacity than the
bearing capacity of top layer and vice versa to investigate the effect on the safety factor
against sliding. Analyses were performed for the ratios c2/c1 of 0.5, 1, and 2 in order to see
the variation.

4.3 Presence of water in the reservoir side of the slope

In this case, the slope angles were again taken as same as the previous cases b = 18.4°
(3:1), b = 26.6° (2:1), and b = 33.4° (3:2). The ratio of high of the water behind the slope
(Hw) to the height of slope was taken as percentages and defined as fullness percentage
(Fig. 9). The analyses were performed for the fullness percentages of 0, 20, 40, 80, and
100 % for the ratios of c/(cH) given in one-layered case. The safety factors of slopes
against sliding were estimated unreinforced soil first. Then, they were estimated in case of
the SCs with s/D ratios given in one-layered case.

5 Results and discussions

In the present study, results of an investigation performed to see the effect of the stone
columns inserted in clay slopes with different distances of s/D ratios are given. Further-
more, the improvements on the bearing capacity and settlement properties of a footing
rested on top of the slope are shown by the results of five small-scaled laboratory tests.
An additional numerical study on the effect of reinforcing soft clay by the SCs on the
behavior of a footing was carried out using the finite element model and compared with
experimental results. Once the experimental and numerical works had been ended up with
results that were very close to each other, a parametrical study was performed to inves-
tigate the effect of the SCs on the stability of slopes.

5.1 Comparison of experimental tests and finite elements analysis

The footing settlement (S) is also expressed in terms of the footing width (B) as the ratio
(S/B, in percent). Typical variations of base pressure (q) with settlement ratios (S/B) for the
laboratory model and FEA model footing for different values of s/D and unreinforced soil
are shown in Figs. 10 and 11.
Both Figs. 10 and 11 confirm the significant increase on the bearing capacity of the
footing with the lower rates of s/D. The ultimate bearing capacities for both the experi-
mental and FEA model footing are determined from the load–displacement curves as the
pronounced peaks, after which the footing collapses and the load decreases. Ultimate

123
818 Nat Hazards (2012) 64:797–820

bearing capacity change by s/D is given Table 2. Settlement changes by no SCs and with
SCs for different s/D ratios are given Table 3.
The percentage of bearing capacity increment is named as the bearing capacity
increment ratio comparing to bearing capacity of footing for unreinforced soil and with
SCs for the different s/D rates. The relationship between bearing capacity increment ratio
and s/D for the comparison of results from experimental work and FEA is seen in Fig. 12.
The ratios of the settlement decrement is the ratio of the difference of settlements
between the settlement of footing resting on the slope unreinforced soil and the settlement
of the footing resting on the slopes with SCs. The ratio is named as settlement reduction
ratio. The variation of the settlement reduction ratio with s/D for the comparison of results
from experimental work and FEA is seen in Fig. 13.
Since the experiments showed a remarkable increase in water content of the stone
column material, stone columns showed a kind of trench feature. Because of this feature
can be identified numerical modeling with Plaxis in full minor differences compared with
experimental and numerical results have occurred.

5.2 Parametric study

For different s/D ratios, Figs. 14, 15, and 16 show the variation of a safety factor estimated
by PLAXIS with slope angles for one-layered clay with different c/(cH) ratios. The
Improvement Number of Safety Factor, INSF, is given in Table 4.
The relationship between c2/c1 and the safety factors estimated by PLAXIS for the two-
layered clay for different s/D ratios seen in Figs. 17, 18, 19, 20, 21, 22, 23, 24, and 25.
The values of INSF that were estimated by numerical analyses for two-layered clay are
given in Table 5.
The relationship between slope angles and safety factors estimated by PLAXIS for the
case of the presence of water in the reservoir side of the slope with different fullness
percentages and s/D ratios are given in Figs. 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36,
and 37.
In the case of the presence of water in the reservoir side of the slope, estimated INSF
values by numerical analyses are given in Table 6.

6 Conclusions

In this study, firstly, an experimental investigation has been performed on a small-scaled

experimental model in the laboratory. In the model, a slope was built using soft clay. The
SCs were inserted into the slope with different ratios of s/D to see its effect on the safety
factor against sliding, on the bearing capacity, and settlement behavior of a footing rested
on top of the slope. For a parametrical study, three hypothetical cases were considered. The
hypothetical cases of the one-layered clay, the two-layered clay, and the water existed
behind the slope were investigated for the objective of understanding effect of the SCs on
slope stability. Following conclusions were reached according to the results of the
experimental and numerical studies performed.
1. SCs have an increment on bearing capacity of sloped soft clay about 1.25–1.71 times
of the bearing capacity of sloped soft clay with no reinforced soil. Very limited
increment on the bearing capacity of sloped soft clay was observed on SCs that the
ratio of s/D higher than 3.5.

123
Nat Hazards (2012) 64:797–820 819

2. As a result of the treatment of soft clay by SCs, settlement of footing is decreased

about ratios between 0.37 and 0.66. An increment on the ratio of s/D makes bearing
capacity to be lowered and settlement decrement ratios to be increased.
3. The experimental and FEA works performed in this study give a compatibility of
92–99 % on the bearing capacity and settlement values.
4. The results of PLAXIS analyses showed that the factor of safety decreased while the
slope angle increased for all of the s/D ratios in the case of one-layered clay. This
decrement is linear for the ratios c/(cH) of 0.025 and 0.01. In the ratios of 0.1 and
0.05, the decrement on the safety factor is linear and even more when the slope angle
is more than 26.60° (2:1).
5. In the case of one-layered clay, the values of INSF are found as 1.51 for s/D = 2,
1.38 for s/D = 3, and 1.25 for s/D = 4. When the slope angle is increased and the
c/(cH) ratio is decreased, the value INSF is increased in all of the s/D ratios.
6. The result of the PLAXIS analyses showed that the safety factor for two-layered clay
is increased when the ratio of c2/c1 is increased for all the s/D and c1/(cH) ratios.
7. The average value of INSF is increased when c1/cH ratio is decreased in the case of
two-layered clay for all the slope angles.
8. The value of INSF is not affected much more by the order of layers such as the lower
end of the SCs is inserted the bottom layer that has a higher bearing capacity than the
bearing capacity of top layer, or vice versa for different ratios of s/D. Thus, it is
concluded that whether the length of the SC is under the level of critical shear
surface, the SCs would serve good enough in depended from the bottom layer is firm
or weak.
9. Due to the performed analyses, the safety factor is decreased when the slope angle is
increased in all ratios of s/D and all fullness percentages in the case of presence of
water in the reservoir side of the slope. On the other hand, the safety factor is
increased when the fullness percentage is increased for all the c/(cH) ratios.
10. The analyses have shown that INSF is increased while fullness percentage is down to
40 %, and getting a maximum value of INSF at the 40 % fullness in the case of
presence of water in the reservoir side of the slope. The value of INSF is decreased
very little or not affected for fullness percentages of less than 40 %.
11. Due to the present study, INSFs and the safety factors of slopes stabilized by SCs are
presented in graphical and tabular forms. It is the authors’ belief that the graphic and
tables would be helpful to engineers in practice during the pre-design of slopes in a
shorter time.

References

Aboshi H, Ichimoto E, Harada K, Emoki M (1979) The composer—a method to improve the characteristics
of soft clays by inclusion of large diameter sand columns. In: Proceedings of international conference
on soil reinforcement, ENPC, 1, Paris, pp 211–216
Adalier K, Elgamal A, Meneses J, Baez JI (2003) Stone columns as liquefaction countermeasure in non-
plastic silty soils. Soil Dyn Earthq Eng 23:571–584
Ambily AP, Gandhi SR (2007) Behavior of stone columns based on experimental and FEM analysis.
J Geotech Geoenviron Eng 133(4):405–415
Baez JI (1995) A design model for the reduction of soil liquefaction by vibro-stone columns. PhD Thesis.
University of Southern California

123
820 Nat Hazards (2012) 64:797–820

Black JA, Sivakumar V, Madhav MR, Hamill GA (2007) Reinforced stone columns in weak deposits:
laboratory model study. J Geotech Geoenviron Eng 133(9):1154–1161
Christoulas S, Giannaros C, Tsiambaos G (1997) Stabilization of embankment foundations by using stone
columns. Geotech Geol Eng 15:247–258
Connor SS, Gorski AG (2000) A timely solution for the Nojoqi Grade landslide. Repair US 101 South of
Buellton. In: 51 st Annual highway geology symposium, Seattle, pp 1–11
Greenwood DA (1970) Mechanical improvement of soils below ground surfaces. In: Proceedings of ground
engineering conference, Institution of Civil engineers, London, pp 11–22
Greenwood DA, Kirsch K (1984) Specialist ground treatment by vibratory and dynamic methods. Satae of
the art report. Pilling and ground treatment. Thomas Telford, London, pp 17–45
Guetif Z, Bouassida M, Debats JM (2007) Improved soft clay characteristics due to stone column instal-
lation. Comput Geotech 34:104–111
Hammouri NA, Malkawi AIH, Yamin MMA (2008) Stability analysis of slopes using the finite element
method and limiting equilibrium approach. Bull Eng Geol Environ 8:156
Han J, Ye SL (2002) A theoretical solution for consolidation rates of stone column-reinforced foundation
accounting for smear and well resistance effects. Int J Geomech 2(2):135–151
Heitz C, Kempfert HG, Alexiew D (2005) Embankment project on soft subsoil with grouted stone columns
and geogrids. In: 16th International conference on soil mechanics and geotechnical engineering, Osaka,
pp 1359–1363
Hughes JMO, Withers NJ, Greenwood DA (1975) A field trial of reinforcing effect of a stone column in soil.
Geotecnique 25(1):31–34
Kirsch F, Sondermann W (2003) Field measurements and numerical analysis of the stress distribution below
stone column supported embankments and their stability. In: International workshop on geotechnics of
soft soil theory and practice, Essen, pp 595–600
Kumar S (2001) Reduction liquefaction potential using dynamic compaction and construction of stone
columns. J Geotech Geol Eng 19:169–182
Lee JS, Pande GN (1998) Analysis of stone-column reinforced foundations. Int J Numer Anal Meth Geo-
mech 22:1001–1020
Madhav MR, Vitkar PP (1978) Strip footing on weak clay stabilized with a granular trench or pile. Can
Geotech J 15(4):605–609
Munfakh GA (1984) Soil reinforcement by stone columns-varied case applications. Int Conf In situ Soil
Rock Reinforce, Paris, pp 157–162
Murugesan S, Rajagopal K (2006) Geosynthetic-encased stone columns: numerical evaluation. Geotext
Geomembr 24:349–358
Nalçakan MS (2004) Geotechnical solutions for problematic soils. Eng News Turkey 430:29–31
(in Turkish)
Osmanoğlu U (1999) Improvement of loose and soft soils jet grouting and stone columns. Msc thesis,
Istanbul Technical University, İstanbul (in Turkish)
Patel BR, Shroff AV (2005) Study on composite stone column in Soft Kaolinitic Clay. In: 16th International
conference of soil mechanics and geotechnical engineering, Osaka, pp 1413–1417
Plaxis 8.5 (2006) Finite element code for soil rock analyses. User Manual, Delf
Stabl 5 (1978) Slope stability analyze program, Purdue University
Tan SA, Khine KO (2005) Finite element modeling of stone columns—a case history. In: 16th International
conference of soil mechanics and geotechnical engineering, Osaka, pp 1425–1428

123