Você está na página 1de 24

Computers and Geotechnics, Vol. 18, No. 4, pp.

261 290, 1996


Copyright 7; 1996 Elsevier Science Ltd
Printed in Great Britain. All rights reserved
0266-352X(95)00034-8 0266-352X/96/%1 5.00 + 0.00
ELSEVIER

Deformation Analysis of Soft Ground Reinforced by


Columnar Inclusions

M. Alamgir,” N. Miura,a H. B. Poorooshasbh & M. R. Madhav’

ODepartment of Civil Engineering, Saga University, 1 Honjo, Saga 840, Japan


*Department of Civil Engineering, Concordia University, Montreal, Canada
‘Department of Civil Engineering, Indian Institute of Technology, Kanpur, India

(Received 5 December 1994; revised version received 12 April 1995; accepted 10 July 1995)

ABSTRACT

A simple theoretical approach to predict the deformation behaviour of soft


ground reinforced by columnar inclusions such as stone columns/granular piles,
sand compaction piles, lime or cement columns, etc., is presented in this paper.
The analysis is performed based on the deformation properties of the column
material and the surrounding soil. The interaction shear stresses between the
column and the surrounding soil are considered to account for the stress transfer
between the column and the soil. The solution is obtained by imposing compati-
bility between the displacements of the column and the soilfor each element of the
column-soil system. Numerical evaluations are made for a range of parameters
to illustrate the influence of various parameters on the predictions. The pro-
posed method is verified with finite element analysis and a reasonable agree-
ment is obtained between the predictions. Copyright 0 1996 Elsevier Science Ltd.

INTRODUCTION

Amongst the various techniques for improving in situ ground conditions,


columnar inclusions are considered as one of the most versatile and cost
effective ground improvement techniques. The columnar inclusions can be of
the form such as stone columns/granular piles, sand compaction piles, lime
or cement columns, etc., which are stiffer and stronger than the ambient soil.
These ground improvement techniques have been and are being used in
many difficult foundation sites throughout the world, to increase the bearing
capacity, reduce settlement, improve slope stability and the resistance to
liquefaction. In order to evaluate the behaviour of the ground improved by
261
268 hf. Alamgir et al

columnar inclusions, several empirical and theoretical approaches are avail-


able considering different failure modes and load sharing mechanisms,
homogenization techniques, etc. [l-9]. Almost all of the existing approaches
are based on the behaviour of composite ground encountered for a particular
type of columnar inclusion. In the existing design approaches, the load
sharing between the two components of the composite ground, i.e. column
and the surrounding soil, is generally calculated on the assumption that
horizontal sections in the ground remain horizontal in the course of settle-
ment (“equal strain” theory) [lo]. Since, the deformation of the two compo-
nents is different for flexible loading [ 11, 121, the approaches based on the
assumption of “equal strain” are not able to predict the actual behaviour of
the uniformly loaded columnar reinforced ground. However, Balaam et al.
[ 131 presented a finite element solution to determine the settlement of gran-
ular column reinforced ground subjected to uniform vertical pressure
imposed by a flexible raft foundation or embankment. Irrespective of their
finite element analysis, there is still a need for a simple, analytical method to
be developed for obtaining the solution to this problem by practicing engi-
neers. This is, in fact, the aim of the present paper.
Different techniques of installation have been conceived for various types
of columnar inclusions in a wide variety of soils ranging from loose sandy
soils to soft compressible soils depending on their technical ability, efficiency
and local conditions. The physical and mechanical properties of the column
materials are also different from each other. This behaviour leads the column
to behave in a different manner in the post-installation phase and during
loading. However, the ground reinforced by columnar inclusions is a com-
posite ground consisting of a relatively stiffer and stronger column and softer
surrounding soil. From the standpoint of foundation analysis, different types
of columnar inclusions can be categorized as a single foundation type which
allows, with some idealization, a unified approach to evaluate the deforma-
tion behaviour of any soft ground reinforced by columnar inclusions.
This paper presents a simple theoretical approach to evaluate the beha-
viour of uniformly loaded ground reinforced by columnar inclusions. This
study takes into account the phenomenon of different deformation (“free
strain”) of the two components of the composite ground, i.e. column and the
surrounding soil, and the interaction between them to determine the load
sharing and the resulting settlements of the reinforced ground. The column-
soil system is divided into a number of uniformly loaded elements and the
solutions are obtained by imposing compatibility between the displacements
of the column and the soil for each element of the column-soil system. The
displacements of the column and the soil elements are obtained considering
the elastic deformations of column and soil. It should be noted that yielding
may occur in the surrounding soil and even in the column, which leads to
Deformation behaviour of soft ground 269

redistribution of stresses and hence the load sharing between the components
of the system. For simplicity, in the proposed approach only the elastic
solution is considered. Moreover, Balaam et al. [ 131 suggested that since the
settlement values obtained from elastic and elasto-plastic analysis differs
very little, the relative effects of various changes in a column-soil unit could
be determined satisfactorily by means of elastic analysis, thus avoiding the
lengthy calculations involved in elasto-plastic analysis.
A parametric analysis is conducted to examine the influence of various
design parameters and to investigate the distribution of shear stresses with
depth and radial distance, to obtain the resultant settlements of the system
and the nature of sharing of the loads between the column and the sur-
rounding soil. It is found that the shear stresses are maximum at the top and
at the column-soil interface and they decrease with depth and radial dis-
tance. At the surface, the stresses on the column and the soil are the same.
They increase in the column and decrease in the soil with depth. The stress
on the soil is small adjacent to the column and increases with radial distance
towards the outside boundary of the unit cell. The settlement profile shows
the deformation response of the column-reinforced soft ground as expected
in the field for the case of uniform loading. The proposed method is compared
by the finite analysis for a typical problem. Good agreement is found between
the results obtained by the proposed method and the finite element analysis.

STATEMENT OF THE PROBLEM

Consider a general case in which a soft ground reinforced by a group of


columnar inclusions is subjected to a uniform loading over the area. The
columns rest on an undeformable bearing layer and the uniform load is
applied on the ground through a flexible platform. The solutions are
obtained by using the “unit cell” concept, which consists of the column and
the surrounding soil within a column zone of influence. It is recognized that
the behaviour of the unit cell adequately represents the behaviour of this type
of column-reinforced ground [ 14- 171. A typical column-reinforced ground
and the column-soil unit are shown in Fig. 1. The column is considered to be
a cylinder, of length H, and diameter d,( = 2a). The unit cell of diameter
d,( = 2b), is loaded with a uniform load a,,. Balaam and Booker [7] relate the
diameter d,, with the spacing of the columns s,, as

d, = cgsc (1)

where cs is a geometry dependent constant equal to 1.05, 1.13 and 1.29 for
triangular, square and hexagonal arrangements, respectively. The soil is
270 M. Alumgir et al

Fig. 1. The foundation system to be analysed: (a) plan and (b) elevation

assumed to behave as a linearly deformable homogeneous material defined


by a constant modulus of deformation Es, and a constant Poisson’s ratio v,,
that are not influenced by the presence of columns and remain constant
throughout the loading process. The column material is also assumed to
behave as a linearly deformable homogeneous material defined by a modulus
of deformation EC and a constant Poisson’s ratio uC. Since it is assumed that
the column-soil interface remains elastic and no slip occurs, the displace-
ments of column and soil at the interface are equal. From symmetry of load
and geometry the shear stresses at the outside boundary of the unit cell must
be equal to zero. According to this condition the uniform load applied over
the unit cell will remain within the unit cell. In this analysis, only the vertical
deformations of the column-soil system are considered, the radial deforma-
tions are considered to be negligible.

ANALYSIS

The expressions to describe the behaviour of column and soil under loading
and their formulations are obtained by the following steps.
Deformation behaviour of soff ground 271

(9 discretization of the column and the surrounding soil into a num-


ber of elements,
(ii) assumption of a deformation mode for the composite ground and
postulation of an equation to express the assumed deformation
pattern,
(iii) development of equations to obtain the interaction shear stresses,
stresses on the column and the soil by stress transfer mechanism
and satisfying equilibrium of vertical forces,
(iv) development of equations to obtain the displacements of column
and soil by applying elastic deformation characteristics of the col-
umn and the soil,
(v) imposing compatibility of displacements for the column and the
soil elements,
(vi) obtaining solutions for the unknown parameters used in the pos-
tulation, i.e. determining the displacement field and thence the
interaction shear stresses, stresses on column and soil and the set-
tlement of the composite ground.

Displacement pattern and shear stresses

The key to the analysis presented here is to assume a deformed shape of the
column-soil system. The displacement of the column is assumed to remain
the same over its area. The displacement of the surrounding soil increases
from the column-soil interface towards the outside boundary of the unit cell.
The assumed deformed shape of the column-soil system and the co-ordinate
axes are shown in Fig. 2. This type of deformation mode has been selected
for the sake of simplicity and for the reason that this deformation mode has
already been used in the analysis of two layer systems [18]. In fact, the

Fig. 2. Assumed mode of deformation


272 M. Alamgir et al

present deformation mode is nothing but an extension of the one proposed in


[18]. It has also been used in the field of reinforced earth to analyse the
influence of cavities appearing underneath a reinforced layer [19] and to
analyse the geosynthetic-reinforced soil system [20]. The experimental evi-
dence reported by Haung and Tatsuoka [21] also supports the rationality of
the assumed deformation mode. The mode of deformation of the column-
reinforced ground considered here, is expressed by the following equation

u’,; = u’,, + a,, [Y/LIZ


- e’c(rla-‘)] for adr<h (2)

where a and b are the radii of column and unit cell, respectively, r is the
radial distance measured from the centre of column, w,., is the displacement
of the soil element at a depth z and at a radial distance Y, w,., is the
displacement of the column element at a depth z, oCZ and /3, are the dis-
placement parameters.
Since the vertical displacement of the column-soil system varies both with
depth and radial distance, the mobilized shear strain and shear stresses also
vary in both directions. Differentiating eqn (2) with respect to Y and by sub-
stituting the value of shear modulus G, = &/2( 1 + II,), the expressions for the
shear strain yrZ, and the shear stress r,.‘, are derived as

Yrz =
~
,&
= 5 [l _ pee/w-1)]

E&z [1 _ /j,iW”-‘)I.
rrT,;
= 2a( 1 + vs)

Due to the symmetry of load and geometry, the shear stress at the outside
boundary of the unit cell is zero, i.e. at Y = b, r,, = 0. Applying this condition
to eqn (4) and noting that CQ, is not equal to zero, leads to the equation

Pc+(“-‘) _ 1 = 0,
(5)
Equation (5) reveals that ,& depends only on the spacing ratio, n = b/a of the
columns. The variation of ,& with II is plotted in Fig. 3. The value of p,
decreases from a value of one to about 0.11 for II = 20.

Column displacements

The uniform load applied on the soft ground reinforced by columnar inclu-
sions is shared by the column and the surrounding soil on the basis of their
stiffness, geometry and the interfacial shear stress. Due to the surface load-
ing, the stress transfers from the soil to the relatively stiffer and stronger
Deformation behaviour of soft ground 213

I I I I I I I
0 5 10 1.5 20
n( =b/a)

Fig. 3. Variation of p, with the spacing of the columns.

column through the interaction shear stresses at the column-soil interface.


To obtain the solutions, the column is divided into L number of elements
having equal thickness, AH( = H/L). The discretization of the column and
the stresses acting on a typical element are shown in Fig. 4. Equilibrium of
vertical forces on a typical element, see Fig. 4(b), leads to

da,, 4
i.e. -
dz = zraz.

I AH=WL

Column
zlj
Element j

%j+l
Element L

(4 @I
Fig. 4. (a) Discretization of the column and (b) stresses acting on the jth element of the column.
274 M. Alamgir et al

Combining of eqn (4) for Y= a, with eqn (7) results in the equation

do,;
__ ZzE
Es(l - Pc)~cz
dz a2(l + VS)

Integrating eqn (8) numerically, the normal stress acting on the jth element
of column is obtained as

(T,j+, = oc, + (AHla)(l- PC)ESW j

41 + 4)

where 0c.j and gcj+ 1 are the normal stresses acting at the top and bottom of
the jth element of column, respectively, j varies from 1 to L and Qcj is the
displacement parameter.
In calculating the deformation of the column elements, only the axial
deformation of the column is considered. The vertical deformation of each
element of column is related to the stress acting at the mid section of that
element and deformation modulus of the column material. Deformation of
the jth element of the column, Wc:jis derived from eqn (9) as

w _ AHo + ww2v - Pcbs%, (10)


CJ ‘,J
EC X(1 + 4

Soil displacements

Due to transfer of stresses from soil to column, the stresses on the soil
decrease along the depth on the basis of the mobilized shear stresses acting at
the vertical planes. The shear stress is maximum at the columnsoil interface
and decreases with radial distance and becomes zero at the outside boundary
of the unit cell. Therefore, the normal stress acting on the surrounding soil
varies with vertical and radial distances. To obtain the stress field, the soil
surrounding the column is divided into L number of elements the same as the
column but subdivided into N number of annular elements having a radial
width equal to Ar (= (b-a)/JV). The discretization of the soil and the stresses
acting on the elements are shown in Fig. 5, where i varies from 1 to N. From
boundary conditions, the shear stresses at the outside boundary of the unit
cell is zero, i.e. at r = 6, ?-bz= 0. From the equilibrium of vertical forces acting
on the Nth element of the soil shown in Fig. 5(b), one gets

(wz + AcwJ[nb2-r(b - AT)‘] - asN+b2 - r(b - Ar)2]


(11)
+ rNz2r(b - Ar)Az = 0
Deformation behaviour of soft ground 215

jL$yy$$Lpq$
%ij*l %Nj+l
(b) E

Fig. 5. (a) Discretization of the surrounding soil and (b) stresses acting on the soil elements.

d%NZ (n - AR)TN~
i.e. -
dz
=
- aAR(n - AR/2)
(12)

where TNZ is the shear stress acting at a radial distance r/a=n-AR and at a
depth z, AR = Arla. Combining eqn (4) for r/a = n-AR with eqn (12), leads to
the equation

dcsNz (n - AR)[l - @cePC(“-AR-l)]ESacz


- dz = - 2a*AR(n - AR/2)( 1 + vS) ’
(13)

Integrating eqn (13) numerically, the normal stress acting at the (N&h ele-
ment is obtained as

(AH/a)(n - AR)[l - ,&ePc(n-AR-l)]ESo+


flsNj+l= ac.Nj- (14)
2aAR(n - AR/2)( 1 + vS)

where aSNj and are the normal stresses acting


a,Nj+ 1 at the top and bottom of
the (Nj)th element, respectively.
The vertical deformation of each supporting soil element surrounding the
column is obtained from the uniform vertical stress acting at the mid section
of the soil element and the deformation modulus of the soil. Compression of
276 M. Alamgir et al

the (N,j)th element of soil (soil element adjacent to the boundary of unit
cell), n’svj is derived from eqn (14) and expressed as

AH (AH/cz)~(~ - AR)[l - PceL’C(n-AR-‘)],c~


~~'s,Vj
= E, uTs.Uj- (15)
4AR(n - AR/2)(1 + v,) .

From the equilibrium of vertical forces acting on an soil element (Y,z) shown
in Fig. 5(b), in the absence of body forces, the following differential equation
is obtained where the stress system is symmetric about the z-axis

(16)

Equation (16) is reduced into the following equation to calculate the normal
stresses acting at the (i,j)th element of soil as

(17)

where gsi,l and o,~;+ tjtl are the uniform stresses acting at the top and
bottom of the (ij)th element of soil, respectively, yiti and yi+ I>;are the shear
strains for the (ij)th and (i+ 1j)th elements of soil, respectively. The shear
strains are obtained by the following equations, in which the displacements,
r+‘rz,i.e. naii, are predicted from eqn (2) by substituting the values of unknown
parameters as

&vi+ 1,./+ I + u’i+ 1..j _ wi.j +2kb’i.j+I llAY,


Yi,j = ( (18)
2

Wi+2,j+l + Wi+2,j w'i+l.j +2’vi+‘~j+~)lAr,


Yi+l.j = ( (19)
2

Displacement compatibility

The displacement compatibility between the column and the soil element is
considered here as an important criterion for the development of the pro-
posed approach. For no slip conditions, the compatibility of column and
soils along the full depth are considered. The displacement of thejth element
of column, W, and (N&h element of soil, is related as the following, WsNj

obtained from’eqn (2) and by substituting for r/a=n-AR/2

WsNj = Wcj f Qcj[n - AR/2 - &(n-AR/2-‘)], (20)


Deformation behaviour qf soft ground 277

Hence, evaluating wCifrom eqn (10) and WsNj from eqn (15) and substituting
in eqn (20) Cycjis obtained as

(21)

in which,
A, = @w42u- AL% (24
2&(1 + -4

A2 = (AH/Lz)~(~ - AR)[l - ,&ePc(“-ARP’)]


4AR(n - AR/2)( 1 + v,)
(23)

(24)
Due to the “free strain” consideration, at the surface, the stresses on the
column and soil are the same as the applied stress, i.e. oCl = osl = uo. Apply-
ing this condition, eqn (21) can be solved for all the elements, i.e. j= 1 to L.
Once the values of ,& and acj are known, the displacements, the shear stres-
ses, the stresses on the column and soil for all elements are calculated from
eqns (2), (4), (9) and (17), respectively. Since the stresses on the soil sur-
rounding the column vary with depth and radial distance, the stress con-
centration ratio ~1, (ratio of stress on column to stress on soil), varies with
depth and radial distance as
acj
F&j = _ (25)
usij

The overall settlement at the surface of the column-reinforced ground is


evaluated by adding the displacement of all elements Cj= 1 to L). Thus the
general equation for the settlement (Sri) of the column-treated ground is
obtained as
j=L

Sli = C Wij. (26)


j=l

The settlement of the untreated ground (So) is estimated as

So=gH (27)
s

and a settlement ratio, psi is defined as

psi = 2. (28)
278 M. Alamgir et al.

RESULTS AND DISCUSSION

Results have been obtained for a range of parameters to illustrate the influ-
ence of various parameters on the proposed approach in predicting the
behaviour of the composite ground improved by the columnar inclusions. It
is considered that the cylindrical columns are installed in a group, resting on
an undeformable bearing strata (bedrock, for example) and uniformly loaded
through a flexible platform at the top. The values of the parameters con-
sidered are H/a = 20, rr,/E, = 0.10, EJE, = 5-100, v, = 0.25-0.45 and n = 2210.
The unit cell is divided into 20 equal elements and the surrounding soil is
also subdivided into 20 annular elements. Twenty elements provide a high
degree of accuracy although only 10 elements have been found to give results
of satisfactory accuracy in solutions for an incompressible pile by Poulos and
Davis [22], for a compressible pile by Mattes and Polous [23] and for a
granular column by Alamgir et al. [24]. In the following sections, solutions
are presented and discussed for the variations of shear stresses, stresses on
the column and soil and the resultant settlement of the composite ground.

Distribution of shear stresses

The distribution of shear stresses, r/a 0, along the depth of column, Z/H,
for various column spacings, n, and modular ratios, EC/Es, are presented in
Fig. 6(a) and (b), respectively. The predictions are made for so/Es = 0.10,
v,=O.35, n= 2.0 to 5.0 and EC/Es= 5 to 100. From these figures, it can be
seen that for all cases, the shear stresses are high near the top of the column
and decrease gradually with depth to the bottom of the column. Figure 6(a)

Normalized Shear Stress, T / CTo Normalized Shear Stress, ir / 0 o


0.0 0.1 0.2 0.3 0.0 0.1 0.2
I

v ,=0.3.5

lil---
H/a=20
5

(a)

Fig. 6. Distribution of shear stresses along the depth of the column for different (a) spacings
and (b) modular ratios.
Deformation behaviour of soft ground 219

shows that for the closer spacing of a column, the shear stress is very high in
the top region (0.0 < Z/H< 0.15) but decreases sharply beyond a certain
depth (Z/H > 0.30). At the top, r/u0 varies from 0.13 to 0.39 for n = 5.0 to 2.0
but at the bottom r/a,, varies from 0.00 to 0.02 for n = 2.0 to 5.0. From these
results, it can be seen that there is a marked decrease of shear stress along the
column-soil interface especially for close spacing. The shearing stress is
developed only at the upper one-third of the column and the remaining two-
thirds is almost shear stress free. This may be explained as, the soil in the
upper portion arches between the columns and thus prevents transmission of
forces to the lower portions of column. The arching of soil increases with the
closer spacing of columns, which is, of course, expected. Figure 6(b) reveals
that the interaction shear stress increases with the increase of modular ratio.
The normalized shear stress -r/a0 varies from 0.18 to 0.23 at the top, and
0.002 to 0.012 at the bottom of the column for EC/Es= 5 to 100. The expla-
nation of this behaviour is very simple. Since the stiffness of the column
increases, the difference in the deformation between the column and the
adjacent soil increases which leads to the increase of magnitude of mobilized
shear stress. The results from Fig. 6(b) also indicate that the influence of
modular ratio on the rate of increase of shear stress is more significant for
lower values of modular ratio, E,/E,= 5 to 20, than for higher values of
modular ratio, EC/Es= 50 to 100.

Stresses on column and soil

The sharing of applied stress between the two components of the composite
ground, i.e. the column and the surrounding soil, is shown in the Figs 7 and 8,
respectively, for various parameters. The variations of stresses on the column

Normalized Stress on Column, 0 d 0 O Normalized Stress on Column, LT,j c~ o


1 2 3 1 2 3 4

Fig. 7. Variation of stress on the column with depth for (a) spacings and (b) modular ratios.
280 M. Alamgir et al.

Normalized Stress on Soil, (7 $ (5 o


1
0.4 0.6 0.8 1.0 0
b
B
b

w
2
.Z 0.6

Fig. 8. Variation of stress on the soil (a) with depth at different radial distances and (b) with
radial distance at different depths.

with depth are represented in the crC/uOversus Z/H diagrams in Fig. 7(a) and
(b), and the variations of stresses on the soil with depth and radial distance
are shown in the ~~~~~ versus Z/H and (T,/u, versus v/a diagrams in Fig. 8(a)
and (b), respectively. All these figures show similar trends in that the stresses
on the column and soil are the same and equal to the applied stress at the
surface (Z/H= 0) of the column-reinforced ground, but the stress increases in
the column and decreases in soils with depth upto the bottom of the column-
soil system. Figure 7(a) shows that in the top region of the column (0.0 <Z/
H < 0.15), the stress on the column (TJcJ~, decreases with an increase of spa-
cing n, but in the lower part of column the stress on the column increases
with an increase of IZ. It is also observed from Fig. 7(b) that gC.cr, increases
with an increase of EC/Es. This figure also reveals that the rate of increase
with EC/Es is significant for lower values than that for higher values of
modular ratio. The values of CJ~C/Q~ at Z/H= 1.O are found to vary from 2.22
to 3.52 for n = 2.0 to 5.0 in Fig.7(a) and 2.60 to 4.09 for EC/Es= 5 to 100 in
Fig.7(b). The applied stress shared by the soil surrounding the column varies
with depth and radial distance. To show such variations the results are
plotted in Fig. 8(a) and (b) for the values of a,/Es =O.lO, EC/Es= 10,
V, = 0.35, y1= 3.5 and H/a= 20. In Fig. 8(a), the variation of stress on soil a,/
crO, with depth Z/H, at radial distances v/a= 1.06 to 3.44 are plotted. It is
observed that the stress on the soil increases with depth up to the bottom of
the unit cell and the rate of decrease of stress on the soil depends on the
radial distance from the column-soil interface. The rate of decrease of stress
on the soil with depth is minimum near the outside boundary of the unit cell
and maximum near the column-soil interface. In Fig. 8(a), the values of crS/uO
Deformation behaviour of soft ground 281

at Z/H= 1.0 are found to vary from 0.41 to 0.98 for r/a= 1.06 to 3.44. The
variation of normalized stresses on the soil (a,/~~) with radial distance (r/a)
at depths of Z/H= 0.0 to 1.O are shown in Fig. 8(b). All the lines in this fig-
ure show a similar trend in that the stress on the soil is minimum adjacent to
the column-soil interface and it increases almost linearly with radial distance
up to the outside boundary of the unit cell. In this figure, the values of gs/gO
at r/a= 1.06 are found to vary from 1.O to 0.41 for Z/H= 0.0 to 1.O. The
predictions presented in this section reveal that while the stress on the col-
umn increases, the stress on the soil decreases with depth. For a particular
depth the stress on the soil is minimum adjacent to the column-soil interface
and increases towards the outside boundary of the unit cell. Since a uniform
flexible load is acting over the surface of the column-reinforced ground and
the column and the surrounding soil have different stiffnesses, the softer soil
media transfers the loads into the adjacent stiffer column. It can be seen from
the predictions that the spacing of the column and the relative stiffness of the
column and soil have a noticeable influence on the magnitude of stress
transfer because both of these parameters influence the mobilization of shear
stresses and the distribution of load sharing.

Stress concentration ratio

The variation of the stress concentration ratio n,, the ratio of the stress in the
column gC, to the stress in the soil gs, with depth Z/H, at radial distances
r/a = 1.06 to 3.44, is plotted in Fig. 9(a) for the values of so/Es = 0.10, EC/
Es = 10, Y, =0.35, y1= 3.5 and H/a= 20. It is found that n, increases with
depth up to the bottom of the unit cell and at a particular depth it decreases
with radial distance. The variations of stress concentration ratio with radial

Stress Concentration Ratio, nc

(b)

Fig. 9. Variation of the stress concentration ratio (a) with depth at different radial distances
and (b) with radial distance at different depths.
282 M. Alamgir et al.

distance, v/a, at the depth Z/H=O.O to 1.0 are shown in Fig, 9(b). It is
observed that for any depth, ~2,decreases with r/a and the rate of decrease of
n, with depth is minimum near the outside boundary of the unit cell and
maximum adjacent to the column. The influence of spacing and modular
ratio on the variation of stress concentration ratio with depth are shown in
Fig. 10. It is observed that the value of n, decreases with the increase of y1
while it increases with the increase of EC/Es. This figure also reveals that the
rate of increase of n, with depth is maximum for closer spacings and higher
modular ratios while it is minimum for higher spacings and lower modular
ratios. The results presented in Figs 9 and 10 reveal that the magnitude of
stress concentration ratio varies noticeably with depth and radial distance
depending on the spacing of columns and the relative stiffness of the column
and the surrounding soil. The values of stress concentration ratios reported
by Vautrain [25], Barksdale and Bachus [14], Ichimoto and Suematsu [26]
and Goughnour [27] for different types of column-reinforced composite
ground measured in field and laboratory studies, vary from 2 to 6 with the
area replacement ratio, i.e. spacing. From laboratory studies on sand col-
umns [28] and full-scale load tests on stone columns [29], a higher value of
stress concentration ratio around 9 is obtained. From the results obtained by
the proposed method (Figs 9 and lo), it can be seen that the value of n, is
unity at the surface and after that it increases non-linearly with depth up to
the bottom of column. At a particular depth, the value of ~1, is maximum
adjacent to the column-soil interface and decreases with radial distance and
becomes minimum at the outside boundary of the unit cell. For a particular
case [Fig. 10(a)], having parameters such as co/Es =O.lO, EC/Es= 10,
v, =0.35, H/a=20, n = 3.5, the value of ~1, adjacent to the columnsoil

Stress Concentration Ratio, nc Stress Concentration Ratio, n,


1 4 8 12
0.0
g &=O.l
v ,=0.35
n=3.5
H/a=20
q

Fig. 10. Variation of the stress concentration ratio with depth for (a) spacings and (b) modular
ratios.
Deformation behaviour of soft ground 283

interface (r/a = 1.06) is found to vary from unity at the surface to 7.8 1 at the
bottom of the column. It is also found from Fig. 9(a), that at the bottom of
the unit cell, the value of n, varies from 7.81 adjacent to column-soil inter-
face (r/a= 1.06) to 3.29 at the outside boundary of the unit cell (v/a= 3.44).
The stress concentration ratio is a very important concept which accounts
for much of the beneficial effect of improving marginal ground reinforced by
columnar inclusions. These results bring into focus that, the use of a single
value of stress concentration ratio for the design of column-reinforced
ground will not be rational and may lead to considerable errors in predictions.

Settlement of the ground

The influence of column spacing on the settlement of the column and the
surrounding soil is shown in Fig. 11 for the values of parameters a,/
E,=O.lO, E,/E,= 10, v,=O.35 and H/a= 20. The settlements of the sur-
rounding soil are predicted at radial distances, r/a = (n + 19)/20, (n + 1)/2 and
II. From Fig. 1 l(a) it is observed that, for very close spacing (1 < n < 1.25)
the settlements are almost the same all over the area of composite ground
which occurs because of the high arching effect of the soil between the columns
due to close spacing. The differences between the column settlements and
those of the surrounding soil increase with the increase of spacing. The dif-
ferences in the settlements between the column and the surrounding soil
increase with radial distance. It is also observed that the settlement over the
column area increases with the increase of n up to a certain value, beyond
which it decreases with further increase of n while the settlement of the
surrounding soils increases gradually with IZ. For very large spacing, the

s 0.04
8
E
%
3 0.06
z
4
& 0.08

z I (4
I I
2 4 6 8 10 0 2 4
n(=b/a) r/a
Fig. 11. Settlement of the column reinforced ground (a) with spacing at different radial
distances and (b) with radial distance for different spacing.
284 M. Alamgir et al.

settlement of the column approaches (a,/E,)H and the settlement of soil at


the outside boundary of the unit cell approaches (a,/E,)H. These results
indicate that for higher spacing, IZ> 10, the column and the soil (at r/a = n)
settle independently according to their own properties irrespective of the
interaction between the column and the surrounding soil. This aspect of the
behaviour of the ground reinforced by columnar inclusions can also be ver-
ified by the findings of Balaam et al. [13] for a particular type of columnar
inclusion. They reported that the stone columns with spacing more than five
times the diameters (n > 10) could be assumed to act individually. The results
plotted in Fig. 1 l(a) also reveal that it is possible to minimize the differences
in the settlements between the column and the surrounding soils by selecting
a suitable value of column spacing. The settlement (S,/H) profiles of the
composite ground with radial distance (r/a) for the values of the spacing
n = 2.0, 3.0, 4.0 and 5.0 are shown in Fig. 1 l(b). The settlement over the
column is minimum, increases with radial distance and becomes maximum at
r/a = b/a = ~1, i.e. at the outside boundary of the unit cell. There is a marked
decrease in the magnitude of ground settlement for the close spacing of the
column as indicated by Fig. 1 l(b). This is, of course, expected.
The reduction of settlement of the in situ ground is measured here by the
settlement ratio 0, (the ratio of the settlement of treated ground to the set-
tlement of untreated ground). The variations of 0, with spacing and modular
ratios are plotted in Fig. 12(a) and (b). The settlement of the treated ground
over the area of the column (O<v/a,< 1) and the soil at v/a= ~1, is considered
for the prediction of S, and is plotted in Fig. 12(a) and (b), respectively. In
Fig. 12(b), it is observed that the value of ,8, decreases with the increase of
EC/Es and for a particular value of EC/Es, the value of ,LI,increases with the

I I I , I , I ,

1 I I I I I I
2 4 6 8 10 2 4 6 8 10
n(=b/a) n(=b/a)
Fig. 12. Influence of modular ratio on the variation of settlement ratio with the spacing of the
columns at (a) r/a= 1 and (b) r/a = n.
Deformation behaviour of soft ground 285

0.6

r/a
Fig. 13. Influence of (a) modular ratio and (b) the Poisson’s ratio of the soil on the variation
of settlement ratio with radial distance.

increase of n. But in Fig. 12(a), it is found that for the case of the column
area only (0 < r/ad l), the value of ,& increases with n up to a certain value
beyond which it decreases with further increase of n. The variation of p, with
r/a for different EC/Es and V, is plotted in Fig. 13(a) and (b), respectively. The
figures show that ps decreases with the increase of EC/Es but increases with
the increase of v,. These results also show that for a particular value of EC/Es
and v,, p, increases with r/u. From these figures, it is revealed that the rela-
tive stiffness of the column and soil has a significant effect on the reduction
of settlement of reinforced ground but the Poisson’s ratio of soil has little
influence on it.

COMPARISON OF RESULTS BETWEEN THE PROPOSED METHOD


AND THE FINITE ELEMENT ANALYSIS

The results for a typical problem using the proposed method are compared
with those obtained from the finite element analysis. The problem selected is
that of a uniformly loaded typical columnar reinforced ground having
H/dc=20, n =4.0, a,/E,=O.lO, EC/Es= 10, v,=O.20 and ~,=0.40. For the
finite element analysis, the program CRISP [30] is used. The mesh covering
the solution region is shown in Fig. 14. Because of symmetry only one half of
the problem requires modelling and this is done using 152 eight-noded linear-
strain quadrilateral elements. Each element has nine integration points at
which the stresses and strains are calculated. The load is applied as an uni-
form pressure at the top. The boundary conditions for the mesh are the same
as those considered in the proposed method, that is, the outer boundary is
286 M. Alamgir et al.

-
~OIUITI Soil
H
:,

Fig. 14. Finite element mesh used for the analysis.

restrained in the horizontal direction and is assumed to be smooth, i.e. free


to move in the vertical direction. The base of the layer is assumed to be rigid
and smooth and hence is restrained in the vertical direction but free to move
in the horizontal direction.
The settlement profile of the composite ground, the distribution of shear
stress and the load sharing between the components are shown in Figs 15-
17, respectively. The settlement profile of the composite ground along the
radial distance is shown in Fig. 15. It can be seen that the proposed approach
underpredicts the settlement compared to the finite element method from the
centre of column up to the radial distance r/a = 2.5, while after that it over-
predicts up to the outer boundary (r/a= n) of the unit cell. The differences
in the magnitudes of settlements are 5.25% over the column and 5.88% at
the outer boundary of the unit cell. From this figure, it can also be seen that
the predicted deformation patterns of the reinforced ground obtained by the
proposed method and the finite element analysis are quite similar.
The distribution of the shear stress along the depth of column is shown in
Fig. 16. It is observed that there is good agreement between the predictions
of the proposed method and those of the finite element analysis except that
Deformation behaviour of soft ground 287

Normalized Radial Distance, r/a

0 1 2 3 4
o- , I I 1 I I I I I
o-o The Proposed Method
_ - Finite Element Analysis
5

I I I 1 1 I I I I I

Fig. 15. Settlement profile of the composite ground.

Normalized Shear Stress, 5 / CJo

0.0

Z/H

0-O The Proposed Method


1.0 + Finite Element Analysis _
I I I I I I I I

Fig. 16. Distribution of the shear stress along the depth of the column.

the proposed approach underpredicts slightly at the top. At the top, the
value of T/U~ is found to be 0.28 by the proposed method while it is 0.35 by
the finite element analysis. The variation of the stress concentration ratio
along the depth at a radial distance r/a= 2.25 is shown in Fig. 17. In both
predictions, the stress concentration ratio increases rapidly with depth and
after a certain depth it becomes almost constant. At the top portion (Z/H
<0.20), the predictions by the proposed method and the finite element
288 M. Alamgir et al.

Stress Concentration Ratio, II,


1.0 3.0 5.0

Cl a,,/E,=O.lO EJE.=lO
1 1

Fig. 17. Variation of the stress concentration ratio along the depth of the column

analysis is almost the same while at the lower part (Z/H> 0.20) the pro-
posed method overpredicts somewhat when compared to the results from the
finite element analysis. At the bottom the predicted value of n, is 5.37 by the
proposed method, while it is 4.90 by the finite element analysis.
These results justify the acceptability of the proposed method and the use
of the particular displacement function in the proposed analysis for obtain-
ing a solution in a simple manner.

CONCLUSIONS

A simple theoretical approach to predict the deformation behaviour of uni-


formly loaded soft ground reinforced by columnar inclusions is presented.
The approach incorporates the free strain condition, the distribution of shear
stresses and the load sharing between column and soil. A simple deformation
mode of the columnsoil system is assumed and the compatibility of the
displacements between the column and the soil is considered. Results are
presented for a uniformly loaded, column-reinforced ground for various
cases by changing the values of spacing of the column, modular ratio and
Poisson’s ratio of the soil. The predictions show that the effects of spacing
and modular ratios on the distribution of the shear stresses, the load sharing
between the column and the soil and the settlements of the ground are
significant. However, the Poisson’s ratio of the soil has little influence.
Deformation behaviour of soft ground 289

In order to illustrate the accuracy and usefulness of the proposed method,


the results from the proposed method and the finite element analysis have
been compared. It is found that the predicted settlement profiles, the shear
stress distributions and load sharings are very similar in magnitude and pat-
tern for the two techniques. This demonstrates the merits of the proposed
method, which is indeed simple in concept and requires very minimal com-
putational effort.

REFERENCES

1. Greenwood, D. A., Mechanical improvement of soils below ground surface.


Proc. Ground Engineering Conf.. Institution of Civil Engineers, 11-12 June 1970,
pp. 9-20.
2. Baumann, V. & Bauer, G. E. A., The performance of foundation on various
soils stabilized by vibrocompaction method. Can. Geotech. J. 11 (1974) 5099530
3. Hughes, J. M. 0. & Withers, N. J., Reinforcing soft cohesive soil with stone
columns. Ground Engng 7 (1974) 42-49.
4. Priebe, H., Estimating settlements in a gravel column consolidated soil. Die
Bautechnik 53 (1976) 160-162.
5. Madhav, M. R. & Vitkar, R. P., Strip footing on weak clay stabilized with
granular trench or piles. Can. Geotech. J. 15 (1978) 605-609.
6. Goughnour, R. R. & Bayuk, A. A., A field study of long-term settlement of
loads supported by stone columns in soft ground. Proc. Int. Conf Soil Rein-
for.:Reinfor. Earth & Other Techniques, Paris, 1979, Vol. 1, pp. 279-286.
7. Balaam, N. P. & Booker, S. R., Analysis of rigid rafts supported by granular
piles. Ml. J. Numer. Anal. Methods Geomech. 5 (1981) 379-403.
8. Schweiger, H. F & Pande, G. N., Numerical analysis of stone column supported
foundations. Computers Geotech. 2 (1986) 347-372.
9. Canetta, G. & Nova, R., A numerical method for the analysis of ground
improved by columnar inclusions. Computers Geotech. 7 (1989) 99-l 14.
10. Hansbo, S., Foundation Engineering. Elsevier, Amsterdam, 1994.
11. Leung, C. F. & Tan, T. S., Load distribution of soft clay reinforced by sand
columns. Proc. Int. Conf on Soft Soil Engng.: Recent Advances in Soft Soil
Engng, 11-16 November, Guangzhou, China, 1993, pp. 779-784.
12. Ekstrom, J. C., Berntsson, J. A. & Sallfors, G. B., Test fills of clays stabilized
with cement columns. Proc. XIII ICSMFE, New Delhi, 1994, Vol. 3, pp. 1183-l 186.
13. Balaam, N. P., Poulos, H. G. & Brown, P. T., Settlement analysis of soft clays
reinforced with granular piles. Proc. of the 5th Asian Regional Conf., Bangkok,
1977. Vol. 1, pp. 81-92.
14. Barksdale, R. D. & Bachus, R. C., Design and construction of stone columns,
Vol. 1. Report NO FHWA/RD-83/026, National Technical Information Service,
Springfield, VA, 1983.
15. Balaam, N. P. & Booker, J. R., Effect of stone column yield on settlement of
rigid foundations in stabilized clay. Int. J. Numer. Anal. Methods Geomech. 9
(1985) 331-351.
16. Poorooshasb, H. B., Miura, N. & Koumoto, T., Analysis of end bearing gravel
piles. Proc. 9th Asian Regional Conf on SMFE, Bangkok, 1991, Vol 1, pp. 275-278.
290 M. Aiamgir et al

17. Van Impe, W. F. & Madhav, M. R., Analysis and settlement of dilating stone
column reinforced soil. Osterreichische Zng. Arch.-Zschr. 137 (1992) 114-121.
18. Meyerhof, G. G., Ultimate bearing capacity of sand layer overlying clay. Can.
Geotech. J. 3 (1974) 2233229.
19. Giroud, J. P., Bonapart, R., Beech, J. F. & Gross, B. A., Load carrying capacity
of a soil layer supported by a geosynthetic overlying a void. Proc. Int. Geotech.
Symposium: Theory and Practice of Earth Reinforcement (Edited by T. Yama-
nouchi, N. Miura & H. Ochiai). Balkema, Amsterdam, 1988, pp. 185190.
20. Poorooshasb, H. B., Analysis of geosynthetic reinforced soil using a simple
transform function. Computers Geotech. 8 (1989) 2899309.
21. Haung, C. C. & Tatsuoka, F., Prediction of bearing capacity in level sandy
ground reinforced with strip reinforcement. Proc., Int. Geotech. Symp: Theory
and Practice of Earth Reinforcement (Edited by T. Yamanouchi, N. Miura & H.
Ochiai). Balkema, Amsterdam, 1988, pp. 191-197.
22. Poulos, H. G. & Davis, E. H., The settlement behaviour of a single axially loa-
ded incompressible piles and piers. Geotechnique 18 (1968) 35 l-371.
23. Mattes, N. S. & Poulos, H. G., Settlement of a single compressible pile. J. Soil
Mech. Found., ASCE 95 (1969) 189-207.
24. Alamgir, M., Miura, N. & Madhav, M. R., Analysis of granular column rein-
forced ground: I Estimation of interaction shear stresses. Reports Faculty of
Sci. & Engng, Saga University, Saga, Japan, 1993, Vol. 22, pp. 111-l 18.
25. Vautrain, J., Mur en terre armee sur colonnes ballastees. Proc. Znt. Symp. on
Geotech. Aspects of Soft Clays, A.I.T., Bangkok, 1977, pp. 613-628.
26. Ichimoto, E. & Suematsu, N., Deep ground improvement and the practice of
SCP method. Tshuchi-tokiso, JSSMFE 31 (1983) 83-90.
27. Goughnour, R. R., Settlement of vertical load stone columns in soft ground.
Proc. Specialty Session - 8th European Conf on SMFE, Helsinki, 1983 , Vol. 2.
28. Aboshi, H., Ichimoto, E., Enoki, M. & Harada, K., The compozer-A method
to improve characteristics of soft clays by inclusion of large diameter sand
columns. Proc. Int. Conf. Soil Reinforcement: Reinforced Earth and Other
Techniques, Paris, 1979, Vol. 1, pp. 211-216.
29. Bergado, D. T., Haut, S. H. & Kalvade, S., Improvement of soft Bangkok clay
using granular piles in subsiding environment. Proc. 5th Int. Geotech. Seminar
Case Histories in Soft Clay, Singapore, 1987, pp. 219-226.
30. Britto, A. M. & Gunn, M. J., Critical State Soil Mechanics via Finite Elements.
Horwood, Chichester, 1987.

Você também pode gostar