Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

EXPERIMENTAL AND THEORETICAL EVALUATION OF STONE

COLUMN IN SOFT CLAY
A.P. Ambily
Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, INDIA
Email: ambilychn @rediffmail.com

S.R. Gandhi
Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, INDIA
Email: [email protected]

ABSTRACT: In recent years stone columns have been increasingly used for improvement of soft soils to increase the
load bearing capacity and to reduce the settlements. This ground improvement technique has been successfully applied
for the foundations of structures like oil storage tanks, earthen embankments, raft foundations etc where large settlement
is possible. Despite the wide spread use of stone columns, present design methods are largely empirical. No codes have
been developed. In the present work experimental studies are carried out to evaluate the behaviour of stone column by
varying spacing, shear strength of soft clay, moisture content etc. The results obtained are analysed using the finite
element package PLAXIS. The paper describes literature review on the subject, details of experimental works carried
out, numerical analysis using finite element package, comparison of results and discussion.

1 INTRODUCTION 2 LITERATURE REVIEW

India has large coastline exceeding 6000kms. In view of Several researchers have worked on theoretical,
the developments on coastal areas in the recent past, large experimental and field study on behavior of stone
number of ports and industries are being built. In addition, columns. However very little information is available on
the availability of land for the development of design procedure that can be used for a given situation.
commercial, housing, industrial and transportation, Semi empirical design approach based on the allowable
infrastructure etc. are scarce particularly in urban areas. stress on stone columns and the undrained shear strength
This necessitated the use of land, which has weak strata, of clay have been proposed by Greenwood (1970),
wherein the geotechnical engineers are challenged by Hughes and Withers (1974), Saha (1992,1999). Semi-
presence of different problematic soils with varied empirical design approach based on pressuremeter theory
engineering characteristics. Many of these areas are was proposed by Hughes et al (1975). Cavity expansion
covered with thick soft marine clay deposit, with very low approach proposed by Vesic (1972) have been used by
shear strength and high compressibility. Ranjan and Rao(1986) and Datye and Nagaraju(1981)
Out of several techniques available for improving the In experimental approach, Hughes and Withers
weak strata, stone columns have been used to a large (1974) carried out series of model tests in normally
extend for several applications. The design of stone consolidated clay. The test results indicated that ultimate
column is still empirical, based on past experience and capacity of stone column was governed primarily by the
needs field trials before execution. No well-defined maximum radial reaction of the soil against the bulging
guidelines or codes are available. and the extend of vertical movement in the stone column
To study the stiffness and deformation behaviour of the was limited to about 4 times the diameter. Shankar and
improved ground for various column spacing, Shroff (1997) conducted experimental studies to study the
experimental study has been carried out in a cylindrical effect of pattern of installation of stone columns and
tank representing one stone column and surrounding soft showed that triangular pattern seems to be optimum and
clay equivalent to the effective area of a column (unit rational. Mitra and Chatopadhyay (1999) studied the
cell). Ten tests have been carried out for three different effect of different factors influencing the capacity of stone
spacing and moisture content. The experimental results column improved ground from the available literature and
are compared with finite element analysis using PLAXIS showed that in the case of columns failing by bulging the
and found to be comparable. critical length is about 3 to 5 times the stone column

Department of Civil Engineering, IIT Bombay, Mumbai, INDIA 201

Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

diameter. Mitchell and Huber (1985) compared the field Load

performance of stone columns with the predictions by
finite element analysis and reported that the agreement 30mm
1
was generally good. It was concluded further that
settlement predictions using other simpler methods also 2
gave values, which agreed reasonably with the measured
values. Saha et al (2000) studied the load response behavior
of stone columns in soft soil environment by using a finite
element software package (ANSYS). The parameters 3
studied are the variation of lateral deformation, lateral
pressure and vertical stress with depth for various intensity
of loading. The results of computer aided numerical 4 500 mm
solutions are presented in the form of non-dimensional
5
quantities. Madhav (2000) presented an overview of
recent contributions for the analysis and design of stone 450mm
columns. Different equations available in the literature for
finding bearing capacity and settlement of stone column
improved ground have also being given.

3 EXPERIMENTAL WORK 210 to 420 mm φ

The experimental study is carried out with following

objectives:
( i ) To estimate the load carrying capacity of a stone 1. Loading plate 4. Stone column(100mmφ)
column. 2. Sand pad 5. Cylindrical test tank
( ii ) To estimate the stiffness of the improved soil for 3. Soft clay
different shear strength of surrounding clay and for
different spacing. Fig. 1. Typical Test arrangement (entire area loaded)
Accordingly, two types of tests are carried out in test
tank where a stone column of 100mm diameter is
constructed at the center of a cylindrical tank, which is Cylindrical tank of height 500mm is used as model tank.
filled with soft clay of required consistency. Tank Diameter of the tank is taken as the diameter of the area of
diameter is chosen to represent a required spacing zone of influence around each column. Stone column
between the columns. To estimate the load carrying diameter used for the test is 100mm. To study the effect of
capacity of the column, column area alone is loaded. spacing the diameter of the tank was varied from 210mm
Whereas the entire area represented by the column is to 420mm. Assuming triangular pattern and spacing to
loaded to estimate the stiffness of the improved ground. diameter ratio (s/d) as 2, 3 and 4 the model tank diameters
The test program is given in Table1. used are 210, 315 and 420mm. In the tank clay is placed
for a height of 450mm in which the stone column is
Table 1: Test program installed at the center.

Experimental set-up comprises of a cylindrical tank

s/d w Su Loading condition filled with soft clay and with a stone column of 100mm
(%) (kPa) Entire Column diameter at its center. A sand layer of 30 mm thick is
area alone placed at top as a blanket. Vertical load is applied either
2 25 30 over the entire tank area or over diameter equal to that of
2 30 12 stone column. Sand layer of 30mm height is placed on the
2 35 6.5 clay bed around the column in the case of column alone
3 30 12 loaded. The load was applied through a proving ring at a
4 30 12 constant strain rate of 1.2mm/min.

3.1 Experimental Set - up 3.2 Properties of Material Used

Typical test arrangement is shown in fig 1. Two basic materials are used for this study: Clay and
Stones

202 Department of Civil Engineering, IIT Bombay, Mumbai, INDIA

Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

(a) Properties of clay weight of the gravel placed was weighed and the dry
density of material as placed considering diameter of
The clay used is natural clay from IIT Madras campus. 100mm is 16 kN/m3. A sand layer of 30mm height is
Surface clay was excavated after removal of vegetation, placed over the prepared stone column and clay bed.
air-dried and pulverized. Properties of clay used are:

Type of Clay : CH 3.5 Test Procedure

Specific Gravity : 2.492
Liquid Limit (%) : 52 After preparing the stone column, the load deformation
Plastic Limit (%) : 21 behavior of the column is studied by loading it in a triaxial
Plasticity Index (%) : 31 loading frame at a strain rate of 1.2mm/minute. To load
Clay Content (%) : 25 the stone column area alone a loading plate of 98mm
Silt Content (%) : 45 diameter is placed exactly at the center of the stone
Max. Dry Density (kN/m3) : 16.63 column and the load is applied till failure. To load the
Optimum Moisture Content (%) : 19.25 entire tank area a loading plate of diameter 5mm less than
the inside diameter of the test tank is placed over the sand
(b) Properties of Stones blanket and the load is applied over the plate. Load is
observed for equal intervals of settlements up to an
Crushed stones (aggregates) of size below 10mm have average deformation of 14mm.
been used to form stone column. The finer fraction
passing through 2mm was removed by wet sieving and 4 FINITE ELEMENT ANALYSIS
used after drying. Typical properties of aggregate for
stone column are: Predictions of the load test results by a finite element
analysis were compared with the model test results to
Angle of Internal Friction : 420 evaluate the ability of the method to model the actual
behaviour of stone column effectively. The finite element
Size : 2 – 10 mm analysis is done using a package – PLAXIS, developed for
Uniformity Coefficient :2 the analysis of soil and rock. Validation of the package is
done and reported below.
Coefficient of curvature : 1.125
Modulus of Elasticity (kPa) : 48000 4.1 Validation of finite element package used

3.3 Preparation of Soft Clay Bed Finite element package (PLAXIS) is validated by
analyzing the load test result published based on the work
Oil is applied in the tank wall to reduce any friction done by Narasimha rao et al (1992). The test tank is 650
between clay and tank wall. Required quantity of clay is mm diameter and height of clay bed prepared is 350mm.
mixed with different moisture content of 25%, 30% and A stone column of diameter 25mm and height 225mm
35%. Thoroughly mixed paste is filled in the tank in was made at the center of the clay bed and loaded with a
layers of 50mm each giving uniform compaction to plate of diameter equal to two times diameter of stone
achieve a uniform dry density of 15.4kN/m3, 14kN/m3 and column. Properties of clay and stones are given below
13kN/m3 respectively. Care was taken to ensure that no
significant air voids are formed in the test bed. At the (a) Properties of clay:
center of clay bed prepared, vane Shear test is carried out Modulus of elasticity : 2000kPa
to measure the shear strength of the clay bed. The results Poissons ratio : 0.45
are shown in table 1. Shear strength : 20kPa
3.4 Construction of Stone Column (b) Properties of stones
The column is constructed by replacement method. A Modulus of elasticity : 40000kPa
100mm outer diameter thin open-ended seamless pipe is Poissonns ratio : 0.3
pushed into the clay at its center up to bottom of tank. The Angle of internal friction : 380
clay within the pipe is scooped out using a helical auger.
Gravels of size 2 – 10mm are charged into the hole in Figure 2 compares the results obtained from model test
layers of 50mm each giving uniform compaction to each and finite element analysis. Load settlement curve, and
layer. Casing pipe is raised in stages ensuring minimum ultimate load obtained from finite element analysis
5mm penetration below the gravel placed. The total matches well with experimental result.

Department of Civil Engineering, IIT Bombay, Mumbai, INDIA 203

Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

Load (N)
Center line of stone column
0 100 200 300 400 500
0

10
Settlement (mm)

15 Stone
colum
20 model test n Tank wall (radial
deformation) = 0
plaxis
25
450mm
30

35

Fig. 2. Validation of PLAXIS Soft clay

4.2 Analysis of Stone Column

A non – linear, plastic, axisymmetric analysis is done..

Material properties used for the analysis are given below.
(a) Properties of Clay 50mm dia Fixed boundary
For the three different consistency of clay used in the
experiments, the compressibility modulus (Ec) was
obtained from the consolidation test and Poisson's ratio is Fig. 3. Finite element mesh for test model (s/d = 3)
assumed based on Bowels (1988) as given below;

w Ec Poisson's su Center line of Center line of

γdry γwet
(%) (kPa) ratio (kPa) (kN/m ) 3
(kN/m ) 3 stone column stone column

25 4500 0.45 30 15.4 19.25

30 2000 0.45 12 14 18.2
Sand
35 1200 0.45 6.5 13 17.8
layer

(b) Properties of Stones

The compressibility modulus and angle of internal friction
have been found by laboratory experiments and Poisson's Soft
clay
ratio is assumed based on Bowles (1988) as given below;

Modulus of compacted stones : 48000kPa

Poisson's ratio : 0.3
Angle of internal friction : 420 Stone
column
The basic axisymmetric finite element mesh and
boundary conditions used to represent the individual load
tests are given in figure 3. 15-noded triangular elements
are used for meshing. Along the periphery radial
deformation is restricted but settlement is allowed. Along
the bottom of the tank both radial deformation and Deformed Mesh

settlement are restricted.

An equal settlement analysis is done when column alone
loaded. Analysis for entire area loaded is done by (a) Column alone loaded (b) Entire area loaded
applying uniform load over the area. In case of column
alone loaded failure is by bulging of the column at a depth Fig. 4. Deformed mesh

204 Department of Civil Engineering, IIT Bombay, Mumbai, INDIA

Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

of 0.5 to 1 times diameter of the column. Figure4(a)

Load intensity (kN/sq.m)
shows a typical deformed mesh at the time of failure when 0 100 200 300 400
column alone loaded. When entire area loaded the 0
analysis is done for a load capacity corresponding to that
of model test. Figure 4(b) shows a typical deformed mesh 5

Settlement(mm)
when entire area loaded. No bulging of the column is seen
10 su = 12kPa
as in the case of model test.
15

5 RESULTS AND DISCUSSIONS 20 Model test

FEM s/d = 2
25 s/d = 4
5.1 Column Area Alone Loaded s/d = 3

Fig.5 compares the load settlement curves from model test

and finite element analysis for different shear strength Fig. 6. Load – settlement curves for different s/d ratios
with same s/d ratio of 2. As seen from the figure failure is
clear and found to be due to bulging of the column. The 5.2 Entire Area Loaded
ultimate load capacities obtained from the model test and
finite element analysis are given in table2 Figure 7 compares load settlement curves from model test
Load intensity (kN/sq.m)
and finite element analysis for different shear strength and
same s/d ratio of 2. In model test because of the confining
0 200 400 600 800
effect from the tank wall and since the full area is loaded,
0 failure has not taken place. Curves from the finite element
5 s/d =2 analysis also follow the same pattern. The stiffness of the
improved ground can be obtained by the back calculation
Settlement(mm)

10
from the curves, which can be adopted for settlement
15 analysis. Table 3 compares the stiffness values obtained
20 from model test and finite element analysis.
6.5kPa
12kPa Load intensity (kN/sq.m)
25
0 50 100 150 200 250
30 Model teat 0
35 FEM 30kPa
2 Model test
FEM
Settlement(mm)

4
Fig. 5. Load – settlement curves at different shear
strength 6
s/d = 2
8
Figure 6 compares the load settlement curve from model
test and finite element analysis for different spacing to 10
12kPa
diameter ratio with same shear strength of 12kPa. The 30kPa
12 6.5kPa
curves obtained from the finite element analysis compares
well with that from model test. Table 2 compares the
ultimate load capacities obtained from model test and Fig. 7. Load – settlement curves for different shear
finite element analysis strength

Table 2: Comparison of ultimate load capacity Table 3: Comparison of stiffness

s/d su (kPa) Ultimate load capacity (kPa) s/d su (kPa) Stiffness (kPa)
Model test FEM Model test FEM
2 30 780 725 2 30 16070 16200
2 12 365 345 2 12 9640 9000
2 6.5 195 185 2 6.5 8020 8090
3 12 325 305 3 12 7230 7740
4 12 255 225 4 12 3870 3620

Department of Civil Engineering, IIT Bombay, Mumbai, INDIA 205

Ambily & Gandhi, Evaluation of Stone Column in Soft Clay ICGGE-2004

Figure 8 compares the load settlement curves from model REFERENCES

test and finite element analysis for different s/d ratios and
same shear strength of 12kPa. Table 3 compares the Bowles (1988). Foundation Analysis and Design, 4th
stiffness obtained from model test and finite element edition, McGraw – Hill International Editions,New
analysis. Delhi.
Datye, K.R. and Nagaraju,S.S. (1981). Design Approach
and Field Control for Stone Columns, Proc. Tenth Int.
Load intensity(kN/sq.m)
Conf. On SMFE., Stockholm, Vol. 3, 637 – 640.
Greenwood, D.A. (1970). Mechanical Improvement of
0 20 40 60 80 100 120 140
Soils Below Ground Surfaces, Proc. Ground
0
Engineering Conf., Institution of Civil Engineers,
2 Model test London, 11-22.
FEM Hughes, J.M.O. and Withers, N.J. (1974). Reinforcing of
Settlement(mm)

4 Soft Cohesive Soils with Stone Columns, Ground

Engineering, Vol.7, No.3, 42-49.
6
Hughes, J.M.O., Withers, N.J. and Greenwood, D.A.
su =12kPa s/d = 2 (1975). A Field Trial of Reinforcing Effect of Stone
8
Column in Soil , Geotechnique, Vol.25, No.1, 32-44.
10 Madhav, M.R.(2000).Granular Piles - Recent Contributions.
s/d = 3 A short term course on Ground Improvement and Deep
12
s/d = 4 foundations held at IIT Madras, Dec 2000, MRM1 -
MRM38.
Mitchel, J.K. and Huber, T.R. (1985). Performance of a
Stone Column Foundation. Journal of Geotechnical
Fig. 8. Load – settlement curves for different s/d ratios Engineering, Vol. 111, No. 2, ASCE.
Mitra, S. and Chathpadhyay, B.C. (1999). Stone Columns
and Design Limitations. Proc. of Indian Geotechnical
Conference held at Culcutta , 201 – 205.
6 SUMMARY AND CONCLUSIONS Narasimharao, S., Madhiyan, M. and Prasad, Y. V. S. N.
(1992). Influence of Bearing Area on the Behaviour of
The present work compares the load settlement behaviour, Stone Columns, Proc. of Indian Geotechnical
ultimate axial capacity of stone column and stiffness of Conference held at Culcutta, 235 – 237.
the improved ground from experimental work with finite Ranjan, G. and Rao, B.G. (1986). Granular Piles for Ground
element analysis. Following conclusions are drawn based Improvement, Proc. of Int. Conf. On Deep Foundations,
on the study. Beijing, Vol.1
Saha, S. (1992). Design Approach and Performance of
1. When column area alone is loaded, the failure is by Stone Columns. Proc. of Indian Geotechnical
bulging of the column with maximum bulging at 0.5 to 1 Conference held at Culcutta, 195 – 198.
times the column diameter below the top. Saha, S. and Saha, S. (1999). Optional Design of Ground
2. When the column and surrounding area represented by Improvement for Large Oil storage Tanks. Proc. of
the column is loaded with confinement of tank wall, the Indian Geotechnical Conference held at Culcutta.,
bulging failure cannot take place. Vol.1, 163 – 166.
3. The load settlement behaviour when entire area is Saha, S., Saha, S. and Roy, A. (2000). Analysis of Stone
loaded is almost linear and it is possible to arrive at the Column in Soft Ground, Indian Geotechnical
stiffness of the improved ground. The stiffness obtained Conference 2000 held at Bombay, 297 – 300.
from model test compares well with that obtained from Sankar, K. and Shroff, A.V. (1997). Experimental Study
the finite element anaysis. on Floated Stone Column in Soft Kaolinite Clay. Proc.
4. Compared to the load settlement for s/d of 2 and 3, s/d of Indian Geotechnical Conference held at Vadodara.,
of 4 is not having any significant improvement. 265-268.
5. The load settlement behaviour and the ultimate axial Vesic, A.S. (1972). Expansion of Cavities in Infinite Soil
capacities obtained from model test compares well with Mass, J. SM and FE Div., ASCE, Vol 98, SM3, 265 –
that of finite element analysis. 290.

206 Department of Civil Engineering, IIT Bombay, Mumbai, INDIA