See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/251194389

Settlement evaluation of soft clay reinforced with stone columns using the
equivalent secant modulus

Article  in  Arabian Journal of Geosciences · April 2012

DOI: 10.1007/s12517-010-0145-y

CITATIONS READS

28 1,671

2 authors:

Ali Zahmatkesh A. J. Choobbasti

Islamic Azad University Babol Noshirvani University of Technology
5 PUBLICATIONS   150 CITATIONS    98 PUBLICATIONS   1,786 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Master of science View project

Laboratory and Numerical Analysis of Piled Rafts and Offshore Foundations Considering Interaction and Bearing Mechanism View project

All content following this page was uploaded by Ali Zahmatkesh on 13 June 2014.

The user has requested enhancement of the downloaded file.

Arab J Geosci
DOI 10.1007/s12517-010-0145-y

ORIGINAL PAPER

Settlement evaluation of soft clay reinforced with stone

columns using the equivalent secant modulus
A. Zahmatkesh & Asskar J. Choobbasti

Received: 5 January 2010 / Accepted: 19 March 2010

# Saudi Society for Geosciences 2010

Abstract A method is proposed to evaluate settlement of zone of influence have to be improved in order to make
soft clay reinforced with stone columns. Finite element them suitable to support the given loads. Ground improve-
analyses were carried out using 15-noded triangular ment for the purpose of foundation construction essentially
elements with PLAXIS. A drained analysis was carried out means increasing the shear strength of the soil and reducing
using Mohr–Coulomb’s criterion for soft clay, stones, and the compressibility to a desired extent. A number of
sand. The stress due to column installation has been techniques have been developed in the last 50 years. Stone
considered in the analysis. At the interface between the columns are extensively used to improve bearing capacity
stone column and soft clay, interface elements have been of poor ground and reduce settlements of structures built on
used. The settlement ratio (SR) of the soil has been them. The techniques were first employed in Europe in the
estimated using the equivalent secant modulus. The results 1830s and have been used there extensively since the late
are compared with those available in the literature, and the 1950s (Ambily and Grandhi 2007).
advantages of the numerical analysis were highlighted. The mechanics of ground improvement depend largely
Based on the results of this analysis, the SR decrease with on the type of soil. Cohesive soils can be improved using
compaction surrounding soft soil, but decrease of SR is stone columns. Stone columns may also be used in sand
mainly due to a stiffer column material in soft clay. deposits but have particular application in soft, inorganic,
cohesive soils. The construction of stone columns is
Keywords Settlement ratio . Stone columns . generally carried out using either a replacement or a
Finite element method . Area replacement percent . displacement method. In the replacement or wet method,
Equivalent secant modulus native soil is replaced by stone columns in a regular pattern
where the holes are constructed using a vibratory probe
accompanied by a water jet. In the displacement or dry
Introduction method, native soil is displaced laterally by a vibratory
probe using compressed air. When the probe is reached to
Man does not have any control on the process of soil favorite depth, the stones are added and are compacted by
formation. The existing soil on a given site may not be the vibrating probe. The displacement method is appropri-
suitable for supporting the desired facilities such as ate where ground water level is low and in situ soil is firm
buildings, bridges, dams, and so on because safe bearing (Lee and Pande 1998). The installation of stone column is
capacity of a soil may not be adequate to support the given accompanied by vibration and horizontal displacement of
loads. In such cases, the properties of the soil within the soil. For considering horizontal displacement of soil during
the installation of stone column, many researchers consid-
ered coefficient of lateral earth pressure (K*) bigger than the
A. Zahmatkesh : A. J. Choobbasti (*) coefficient of lateral earth pressure at rest (K0) (Priebe
Department of Civil Engineering,
1976; Elshazly et al. 2006; Elkasabgy 2005; Pitt et al.
Babol University of Technology,
P.O. Box 484, Babol, Iran 2003). Elshazly et al. (2007) presented an interesting
e-mail: [email protected] relation between the inter-column spacing and K* in
 Arab J Geosci

vibro-installation technique (Fig. 1).This relation was columns. Poorooshasb and Meyerhof (1996) proposed the
inferred from analyses for load settlement records of performance ratio, which is defined as the ratio of the
various field load tests, performed for stone column settlement of the improved ground to that of the unim-
arrangements with different inter-column spacing values. proved ground under identical surcharges. They considered
A well-documented case history, involving three-column linear elastic behavior for stone column. Balaam et al.
patterns (group of columns with three different spacing) (1978) proposed a finite element approach for soft clay
along with their relevant field and laboratory test results, treated with granular piles and reported the effect of
was utilized for this study. Moreover, a well-tested-coupled stiffness of granular pile on load deformation behavior.
finite element model was employed in the analysis. The Mitchell and Huber (1985) compared the field performance
analysis is inversely posed to determine the soil initial of stone columns by an axisymmetric finite element model
stresses, based on the recorded settlements and the post- with groups of columns surrounding the central column
installation material properties. replaced by a ring of stone material having equivalent
Many researchers have developed theoretical solutions thickness. Ambily and Grandhi (2007) conducted experi-
for estimating bearing capacity and settlement of founda- mental and numerical analysis on singles and groups of
tions improved by stone columns (Greenwood 1970; stone columns. They presented improvement factor without
Hughes et al. 1976; Aboshi et al. 1979). Priebe (1976) considering stress due to installation of stone columns.
proposed a method to estimate the settlement of foundation Field observations showed that stone columns could also
resting on the infinite grid of stone columns based on unit accelerate the rate of consolidation of soft clays (Han and
cell concept. In this concept, the soil around a stone column Ye 1992). Han and Ye (2001) developed a simplified and
for area represented by a single column, depending on closed form solution for estimating the rate of consolidation
column spacing, is considered for the analysis. As all the of the stone column reinforced foundations accounting for
columns are simultaneously loaded, it is assumed that the stone column soil modular ratio. It is also reported in
lateral deformations in soil at the boundary of unit cell are the paper that during the process of consolidation the stress
zero. The settlement improvement factor is derived as a on stone column increases with time, whereas the stress on
function of area ratio and angle of internal friction of soil decreases. At the end of consolidation, a steady stress
column material. The calculation of the improvement factor concentration ratio is approached.
was done by considering the stone columns material to be This paper is presented in the following sequences. First,
incompressible and column is founded a rigid layer (end- the simulation of stone column in soft soil in plain strain is
bearing). Priebe (1995) considered the effect of compress- introduced. Next, settlement improvement factor is calcu-
ibility of the column material and the overburden. He lated using the equivalent Young’s modulus. Finally, results
developed design charts to calculate the settlement of single are compared with existing theories.
and strip footing reinforced by a limited number of stone

Finite element analysis

3
Numerical modeling was performed using the PLAXIS V8
Lower bound program. PLAXIS is used for the analysis of deformation
2.5 Most problale value and stability in geotechnical engineering. The improved soil
Upper bound
is modeled with 15 nodes triangular finite elements. In the
2 area of reinforced ground, because stresses and displace-
ments are higher in this area, the considered medium mesh
*
size was refined.
1.5
k

Geometric modeling
1
Each column acts within a cylindrical cell with a
radius of influence denoted by Re (Fig. 2). Balaam and
0.5
Booker (1981) related the radius of influence to the actual
column spacing by the relation Re = c·S, where S is the
0 actual spacing (from center to center of the columns) and
1.2 1.4 1.6 1.8 2 2.2 c is a constant having values of 0.525 and 0.564 for
colums spacing(m)
triangular and square patterns, respectively. For most
Fig. 1 Variation of K* with columns spacing (Elshazly et al. 2007) practical cases, the diameter of influence may be assumed
Arab J Geosci

S The area replacement percent is considered between

10% and 30%. For values less than 10%, no significant
improvement in the ground properties is achieved (Hu et al.
1977); whereas, there would be installation difficulties for
the area replacement percent more than 30%. Stone
S
columns usually are extended to bedrock or a hard layer,
Re but occasionally floating columns are also installed. In this
investigation, it is supposed that stone columns are
extended to a hard layer. In most practical cases, a soil
Re=0.564 S layer is placed at the top soft clay reinforced with stone
columns, so a sand layer of 20-cm thick was placed at the
Fig. 2 Plan of stone column a triangular pattern and b square pattern top of model such layer. The analysis was carried out on
stone column with the diameter of 0.8, 1.0, and 1.2 m and
equal to the actual column spacing. The analyses were depth of 10 m. Because of symmetry, only half of the
carried out assuming columns were arranged in a square geometry is modeled.
pattern.
In this investigation, it was assumed that the raft is rigid, Material modeling
and both the stone column and soft clay undergo the same
amount of settlement. There are no interface elements Appropriate choices of material properties are necessary in
placed between the soil and the footing, so any slippage order to have an accurate simulation of reinforcement
between footing and soil occurs within the soil. This is system in the numerical modeling. The properties of soft
realistic because concrete footings poured against the clay, stone column, and sand can be found in the literature
ground form a very rough interface. Fixed supports were (Ambily and Grandhi 2007; Guetif et al. 2007). Plain strain
considered at the bottom of geometry and roller supports analyses were carried out considering elastoplastic behavior
were on the vertical boundaries. At the interface between for soft clay, sand, and stones. A drained behavior is
the stone column and soft clay, interface elements have assumed for all the materials. In this investigation, the
been used. This can be explained by the fact that the constitutive law of Mohr–Coulomb was used for the stone
deformation of the column is mainly by general failure and columns, sand, and soft soil. The input parameters of
which produces significant shear between clay and stone Mohr–Coulomb model (E, μ, φ , ψ, c , and γ) are given in
column (Etezad 2006). Table 1. Before the columns are installed, the horizontal
The arrangement of the test columns is generally 3D. For component of stress in the ground is given by the equation
modeling column in plain strain, the use of equivalent strip k0γz, where z is the depth below ground surface and k0 is
is necessary. The idealization formula for the equivalent the coefficient of the rest earth pressure for the soft clay.
strip is given in Fig. 3. The area replacement percent (ρ), The coefficient of lateral earth pressure at rest value was
total area of stone columns over original area of unrein- estimated from the Jacky’s formula:
forced soil, assuming columns were arranged in a square
k0 ¼ 1  sin f ð2Þ
pattern was considered as:
Installation of the columns increases this stress to a
d2 higher value. Horizontal component of stress due to
r¼  100 ð1Þ
ð1:13S Þ2 installation of the columns, the coefficient of the rest earth

Fig. 3 Idealization of stone

columns in plane strain
 Arab J Geosci

Table 1 Parameters used in the

numerical analysis E (Kpa) v φ (˚) ψ (˚) c (Kpa) γ (kN/m2) Rinter

Soft clay 4,000 0.35 21 0 5 17 0.7

Stone column 55,000 0.3 43 10 0 19 0.9
Sand 20,000 0.3 30 4 0 16 -

pressure increased. In PLAXIS, the material properties of of the soft soil surrounding the stone columns during the
interface are related to the soil properties and are entered in installation of stone column. The third factor is the vertical
the same data sets as the soil properties. For interaction drainage provided by stone columns (Guetif et al. 2007).
between stone column and soft clay, the interface is weaker Therefore, the insertion of stone columns into weak soils is
than the soil layer, which means that interface strength not just a replacement operation and stone column can
(Rinter) should be less than 1. Suitable value for interface change both the material properties and the state of stresses
strength (Rinter) between stone column and soft clay was in the treated soil mass. In this analysis, the effect of
found in the literature (Brinkgreve and Vermeer 1998). stiffness of column material and the densification of the
surrounding soft soil during the installation of stone column
Analysis were considered.
Figure 4 shows a model of group of stone columns in
The initial vertical stress due to gravity load has been finite element analysis when entire area is loaded. A
considered in the analysis. The stress caused by column uniform vertical displacement (ε=2%) was prescribed to
installation depends on the method of construction and type the model. The average settlement (Se) can be calculated by
of soil. In this investigation for considering the stress due to the following equation (Christian and Carrier 1978):
column installation, initial horizontal stress (K0) is in- qB
Se ¼ m 0  m 1  ð3Þ
creased. Groundwater was supposed to be more than 10 m E
below the ground surface. Hence, there was no need to Where q is the applied footing load, E is elastic Modulus
enter groundwater condition. of the soil, and μ0 and μ1 values depend on the depth of the
In this analysis, the improvement of the stiffness footing and the thickness between the footing base and hard
(reduction of settlement) of the treated ground was strata, respectively. Assuming the whole soil medium to be
evaluated. Improvement of a soft soil by stone columns is homogeneous, the equivalent secant modulus values (Eeq)
due to three factors. The first factor is inclusion of a stiffer have been calculated as
column material (such as crushed stones, gravel, and so
s
on...) in the soft soil. The second factor is the densification Eeq ¼ ð4Þ
"

Fig. 4 Simulation of reinforce-

ment system in the numerical
analysis a initial model and b
deformed mesh Assumed strips of
Soft soil
stone columns 10m

30m
Arab J Geosci

1
where
S
"¼ ð5Þ 0.9 K0
L 1.25K0
where, σ is the average applied stress, ε is the average 0.8
1.50K0
1.75K0
strain, S is the settlement of the footing, and L is the
thickness of the clay bed (=10 m). Figure 5 shows typical 0.7
axial stress versus settlement behavior for improved ground

SR
based on finite element analysis at different area replace-
0.6
ment percentage. The vertical stress versus settlement
relation is almost linear. The equivalent Young’s modulus
0.5
of the composite ground can be obtained from average
slope of the plot. Settlement ratio (SR), settlement of the
composite ground divided by settlement of ground without 0.4

stone column at the same stress level, was calculated. Using

Eq. 3, SR can be expressed as 0.3
0 5 10 15 20 25 30
E0 Area replacement percent
SR ¼ ð6Þ
Eeq Fig. 6 Effect of soil compaction on SR
where, E0 is Young’s modulus of ground without stone
column. SR for each area replacement percent. Use of k0 means that
The installation of stone column is accompanied by the installation of stone column does not have effect on the
vibration and horizontal displacement of soil. The lateral surrounding soft soil. Figure 6 shows that stone column
expansion generates large strains approximately 45% in soft improves the ground mainly due to the higher stiffness of
clay next to the column. The surrounding soft soil is the columns compared to the soil. In other words, stone
compressed and coefficient of lateral earth pressure in- column decreases largely settlement even if the insertion of
crease. The value of coefficient of lateral earth pressure stone columns into weak soils be considered just a
after the installation of stone column depends largely on the replacement operation without any effect on surrounding
type of soil, spacing of stone columns, and installation soft soil.
method of stone column. Figure 6 shows the value of SR An increase of 25% in k0 resulted in a decrease of about
for four different values of K* =(k0, 1.25k0, 1.50k0, 1.75k0). 4% of SR. Therefore, decrease of SR is not only due to
The average of SR of various diameters was calculated as stiffer column material soft clay but SR decreases with
compaction of surrounding soft soil. With increase of K*,
value of SR is decreased. The value of coefficient of lateral
Stress(KPa)
earth pressure after the installation of stone column depends
0 50 100 150 200 250 300
0 largely on spacing of stone columns. Elshazly et al. (Mitchell
and Huber 1985) presented the interesting relation between
Unimproved the spacing of stone columns and K* in vibro-installation
ρ=10 % technique. For considering horizontal component of stress
5
ρ=15 % due to installation of the columns, Fig. 1 was used.
ρ=20 % The diameter of the finished stone column depends on
Settlement(cm)

ρ= 25 %
the strength and consistency of the soil, the energy of
10 ρ= 30 %
compaction, and diameter of probe in replacement or
displacement method. In the softer soil, the diameter of
the column is increased because compaction of the
15 aggregate pushes the stone into the surrounding soil.
Figure 7 shows effect of diameter of column on SR. In
general, the SR increases with increase in the diameter of
column. With increase of area replacement percent, the
20
effect of diameter of column on SR increases. This can be
explained by the fact that in small diameters, column
Fig. 5 Stress settlement behavior under loading for different area spacing decreases and consequently, surrounding soft soil is
replacement percent (finite element results) compacted properly.
 Arab J Geosci

0.9
ment, whereas in the present work both elastic and plastic
displacement is considered. Therefore the present work
0.8 ρ= 10 %
predicts higher settlement compared to Poorooshasb's
method.
0.7 Priebe (1976) performed the calculation of the basic
improvement factor (SR) by considering the stone columns
ρ= 20 %
0.6 material to be incompressible. Therefore, any settlement
of the load area results in a bulging of the column which
SR

0.5 remains constant all over its length. However, column

material is compressible, and failure of stone column is
ρ= 30 %
in the form of shear failure (Etezad 2006). Therefore,
0.4
calculated settlement in the present work will be more
than Priebe’s method. Besides, for uncompressible mate-
0.3
rial, in the case where the area replacement percent
increases up to 100%, the SR approaches to zero. But in
0.2 the present work, with increase of area replacement
0.8 1 1.2
Diameter of stone column(m) percent up to 100%, the SR decreases to 0.09 (Young’s
modulus of soft soil divided by Young’s modulus of stone
Fig. 7 Effect of diameter of column on SR column). The actual SR does not achieve to zero.
Therefore, the values of SR obtained from present work
In order to judge the results of the numerical analysis, a are close to actual SR.
comparison is made with the results of standard analytical
design methods. Figure 8 shows SR obtained from the Example
present work for different area ratio. The following
relationship is obtained between SR and replacement area: Using the results of the present analysis, settlement of a
circular foundation with the diameter of 10 m is calculated.
SR ¼ 2:78r2  2:86r þ 1 ð7Þ This foundation is placed on clay layer of 10-m thick and
imposes vertical stress of 100 kPa. The average properties
Results were compared with the existing theories. The of soft clay are: cohesion=5 kPa, angle of internal friction=
present work predicts an upper SR compared to Poorooshasb 21° and modulus of elasticity=4,000 kPa. The stones used
et al. (Poorooshasb and Meyerhof 1996) and Priebe (1976; for columns have modulus of elasticity=55,000 kPa and
1995). Poorooshasb et. al considered only elastic displace- angle of internal friction=43°. Soil is improved using
100 cm diameter stone column with spacing of 200 cm
center to center. The settlement of untreated ground using
1 Eq. 3 is 75 mm. Using Eq. 7 and for ρ=0.2, SR is
Priebe(1976) calculated as 0.54. So, settlement of foundation after
0.9 Priebe(1995) treatment ¼ 0:54  75 ¼ 40:5 mm.
Poorooshasb and Meyerhof(1996)
0.8 PLAXIS

Conclusions
0.7

A series of numerical analysis has been carried out to

SR

0.6
evaluate settlement of soil reinforced by a group of stone
columns. The clay layer was assumed to be uniform. The
0.5
analyses employed an elastic, perfectly plastic constitutive
0.4
model following the Mohr–Coulomb failure criterion.
Based on the results of this numerical study, the following
0.3 conclusions can be drawn:

0.2
0 5 10 15 20 25 30
1. The load settlement behavior of model with an entire
Area replacement percent
area loaded is almost linear, and it is possible to find
Fig. 8 Comparison of SR with existing theories the stiffness of improved ground.
Arab J Geosci

2. The SR decrease with compaction surrounding soft Elshazly HA, Hafez D, Mosaad M (2006) Back calculating vibro-
installation stresses in stone columns reinforced grounds. J
soil, but decrease of SR is mainly due to a stiffer
Ground Improve 10(2):47–53
column material in soft clay. Elshazly H, Elkasabgy M, Elleboudy A (2007) Effect of inter-column
3. The SR values depend mainly on column spacing (area spacing on soil stresses due to vibro-installed stone columns:
replacement percent). interesting findings. Geotech Geol Eng (2008) 26:225–236
Etezad M (2006) Geotechnical Performance of Group of Stone
4. In the certain area replacement percent, decrease in
Columns. Ph.D. Thesis, University of Montreal Quebec
diameter of column decreases the value of SR, and this Greenwood DA (1970) Mechanical improvement of soils below
decrease is because of reduction of columns spacing. ground surfaces. Proc., Ground Engineering Conf., Institution
Increase in area replacement percent results in increase of Civil Engineers, London, 11–22
Guetif Z, Bouassida M, Debats JM (2007) Improved soft clay
in the effect of diameter of column on SR. characteristics due to stone column installation. Comput Geotech
34(2007):104–111
Han J, Ye SL (1992) Settlement analysis of buildings on the soft clays
stabilized by stone columns, Proc., Int. Conf. on Soil Improve-
References ment and Pile Found., 118, 446–451
Han J, Ye SL (2001) Simplified method for consolidation rate of stone
column reinforced foundation. J Geotech Geoenviron Eng 127
Aboshi H, Ichimoto E, Harada K, Emoki M (1979) The composer—a (7):597–603
method to improve the characteristics of soft clays by inclusion Hu W, Wood D M, Stewart W (1977) Ground improvement using
of large diameter sand columns. Proc., Int. Conf. on Soil stone column foundation: results of model tests. Proc., Int. Conf.
Reinforcement., E.N.P.C., 1, Paris 211–216 on Ground Improvement Techniques, Macau, 247–256
Ambily AP, Grandhi SR (2007) Behavior of stone columns based on Hughes JMO, Withers NJ, Greenwood DA (1976) A field trial of
experimental and FEM analysis. J Geotech Geoenviron Eng reinforcing effect of stone column in soil. Proc., Ground Treatment
ASCE 133(4):405–415 by Deep Compaction, Institution of Civil Engineers, London, 32–44
Balaam NP, Booker JR (1981) Analysis of rigid rafts supported Lee JS, Pande GN (1998) Analysis of stone column reinforced
by granular piles. Int J Num Anal Methods Geomech 5:379– foundations. Int J Numer Anal Meth Geomech 22:1001–1020
403 Mitchell JK, Huber TR (1985) Performance of a stone column
Balaam NP, Poulos HG, Brown PT (1978) Settlement analysis of soft foundation. J Geotech Engrg 111(2):205–223
clays reinforced with granular piles. Proc., 5th Asian Conf. on Pitt J M, White DJ, Gaul A, Hoevelkamp K (2003) Highway
Soil Engineering, Bangkok, Thailand, 81–92 applications for rammed aggregate piers in Iowa soils, Iowa
Brinkgreve RB, Vermeer PA (1998) Plaxis-finite element code for soil DOT Project TR-443, CTRE Project 00-60, USA.
and rocks analysis. Version 8.2 Rotterdam Brookfield, AA. Poorooshasb HB, Meyerhof GG (1996) Analysis of behavior of stone
Balkema columns and lime columns. Comput Geotech 20(1):47–70
Christian JT, Carrier WD (1978) Janbu, Bjerrum and Kjaernsli’s chart Priebe HJ (1976) Abschactzung des setzungsverhaltns eiens durch
reinterpreted. Canadian Geotechnical Journal 15:123–128 stopfverdichtung verbesserten baugrundees. Die Bautechnik
Elkasabgy MA (2005) Performance of stone columns reinforced 54:160–162 (in german)
grounds. M.Sc. Thesis, Zagazig University, Faculty of Engineering Priebe HJ (1995) The design of vibro replacement. Ground engineering
at Shobra, Cairo 28:31–37

View publication stats