Vous êtes sur la page 1sur 3

Bouassida, M., de Buhan, P. & Dormieux, L. (1996). Geotechnique 46, No.

3, 570572

DISCUSSION

Bearing capacity of a foundation resting on a soil improved by a


group of columns
M . B O UA S S I DA , P. D E B U H A N & L . D O R M I E U X ( 1 9 9 5 ) . G e o t e ch n i q u e 4 5 , N o . 1 , 2 5 3 4 .

D. D. Nagpure & M. R. Madhav, IIT Kanpur,


India
The authors have presented an interesting paper
dealing with a general lower-bound estimate for
the bearing capacity of a foundation resting on a
soil reinforced with a group of columns (stone
columns and lime columns). The authors have
derived analytical expressions for the bearing
capacity of such a composite soil by using
equivalent soil parameters (C eq and eq ), regardless of the shape of the foundation and the
geometrical arrangement of the columns. The
purpose of this discussion is to make a few notes
that seem important in appreciating the signicance of the paper.
The authors have illustrated a procedure to
estimate a lower-bound solution of a bearing
capacity by using a simple static approach
assuming a piece-wise constant stress eld. This
procedure is exactly similar to the approach
suggested by Enoki et al., (1991) who have further
considered the effect of stress anisotropy. In order
to improve this static approach, the authors have
considered a three-zone axisymmetric stress eld
and derived the expressions for the bearing
capacity of reinforced ground. Based upon the
improved static approach, the expressions for the
equivalent cohesion C eq and equivalent frictional
angle eq are given in equations (29) and (30),
depending upon the value of the coefcient of
passive earth pressure, Kp . In these equations, the
expressions for the equivalent cohesion C eq are
different, whereas those for the equivalent frictional angle eq are identical. Fig. 4 shows the
variation of the ratio of equivalent cohesion to the
cohesion of the unreinforced soil, C eq /C, with the
friction angle of the column material , assuming
that column material is cohesionless (C c 0). As
expected, with increasing area replacement ratio
(surface ratio ), the ratio C eq /C decreases for a
given value of friction angle of column material .
However it is interesting to note that for a given
value of the area ratio , the ratio C eq /C decreases
with increasing values of upto 27308. For
. 308, this ratio is observed to increase rather
than decrease with . We feel that this is due to

the attempt made to improve the static approach by


using a three-zone axisymmetric stress eld and
the subsequent assumption made regarding the
stress vectors.
Further, for the case of the ground improved
by stone columns, Kp > 2 (since . 1958), the
equivalent cohesion C eq is determined from equation (30). This involves the use of a function, g(),
given by equation (15). A closer look will reveal
that for any value of . 1958 and 10, the
ratio of cohesions C eq /C does not reduce to zero.
Hence the expression for C eq in equation (30) does
not satisfy the end condition for the case of fully
reinforced ground.
Using the simple static approach (i.e. equation
(29) only), the equivalent cohesion C eq is
determined, and the variation of the ratio of
cohesions C eq /C with the area ratio for different
values of friction angle of column material is
shown in Fig. 10.
From Fig. 10 it can be seen that for all values
of considered, the cohesion ratio C eq /C
continuously decreases with increasing values of
area ratio . Further, for the end condition 10;
the ratio of the cohesions C eq /C decreases to zero,
indicating that this approach satises the end
condition for the case of fully reinforced ground.
Hence this simple, static approach is applicable for
the entire range of the area ratio, as well as for the
range of angle of shearing resistance of the column
material.
We agree with the authors' view that increasing

C eq/C

1.0

= 30

0.8

= 35

0.6

= 40

Cc = 0

= 45

0.4
0.2
0.0
0.0 0.1

0.2 0.3

0.4

0.5

0.6

0.7

0.8

Area ratio,

Fig. 10. Variation of C eq /C with area ratio

570

0.9

1.0

DISCUSSION

the bearing capacity of the reinforced soil with the


stone columns is not the prime objective of this
type of the ground improvement technique. For a
friction angle of column material 408 and for
the area ratio 0307, the bearing capacity
enhancement factor R is found to be in the range
139191 by using equation (43a) and 154226
from equation (43b).
Barksdale & Bachus (1983) have suggested a
method to determine the bearing capacity of a
reinforced soft soil with a group of stone columns
similar to a single, isolated stone column. In this
method, ultimate bearing capacity of a composite
soil is given by
qult N c .cu

(44)

where Nc is a bearing capacity factor and cu is the


undrained shear strength of surrounding soil. They
have suggested the range of factor Nc to be 1822,
depending upon the compressibility of in situ soil.
Earlier, Mitchell (1981) recommended a value of
Nc of 25 for vibro-replacement stone columns.
Based on a number of case histories, Datye &
Madhav (1988) found that the factor Nc ranges
between 31 and 60 for the rammed stone columns.
Of course the equipment, construction technique
and type of in situ soil have a signicant inuence
on the factor Nc . Taking Nc 514 for the
unreinforced soil, the enhancement factor R is
observed to be in the range 3501167. Hence one
can conclude that both the authors' approaches
predict very low improvement in the bearing
capacity of such soils.
In conclusion it seems that the improved
approach using a three-zone axisymmetric stress
eld does not predict the acceptable values of
the equivalent cohesion Ceq , and both of these
approaches predict very low improvement in the
bearing capacity for the case of reinforced soil with
stone columns. Hence it may be preferable to adopt
the simple, static approach for the determination of
the equivalent cohesion of such a composite soil.
Further there is a need to develop or to modify the
present approach to determine the bearing capacity
of a soil reinforced with a group of columns so that
the predicted bearing capacity enhancement factors
will be in the range of the actual observed values.

REFERENCES
Barksdale, R. D. & Bachus, R. C. (1983). Design and
construction of stone columns. Report No. FHWA/
RD-83/026, National Information Service, Springeld, Virginia, USA.
Datye, K. R. & Madhav, M. R. (1988) Case histories of
foundations with stone columns. Proc. Int. Conf. on
Case Histories in Geotech. Engng St. Louis, Paper
No. 537, 112.

571

Enoki, M., Yagi, N. & Yatabe, R. (1991). Evaluation of


bearing capacity analysis method of improved ground.
Geo-Coast, Yokohama, 305310.
Mitchell, J. K. (1981). Soil ImprovementState-of-theArt Report. Proc. Xth Int. Conf. on Soil Mech. &
Foundn. Engng, Stockholm, 4, 509565.

Authors' reply
The contribution to the discussion on our paper
made by Nagpure and Madhav and their related
comments give us the opportunity to clarify some
points about the practical signicance of some
results. The paper is divided into two separate
parts: a rst part where the analysis is focused on
the composite cell model, and a second part
devoted to the derivation of lower-bound estimates
for the bearing capacity of a rigid footing resting
on a soil reinforced by columns.
As regards the problem of the composite cell
subjected to a triaxial loading, equation (9) represents a rst lower-bound estimate deduced from a
static approach making use of a piecewise homogeneous stress eld. Improved lower-bound estimates are then calculated by considering more
elaborate axisymmetric stress elds. They prove to
be the exact values (equations (27)), since they
coincide with upper-bound estimates obtained by
the kinematic approach of yield design (or limit
analysis). Finally, the notions of equivalent cohesion and friction angle for the composite soil as
such are introduced (equations (29) and (30)), and
several charts giving the variations of these global
strength parameters are presented.
The authors of the discussion observe that,
according to equation (30) which is valid for
Kp . 2 (i.e. . 1958), C eq does not reduce to
zero when the volume fraction of reinforcement
is taken equal to one with C c 0 (reinforcement
by stone columns made of a purely frictional
material), which is obviously in contradiction with
what would be expected in such a case. Actually,
it should be noted that the maximum possible
value of allowed within the framework of the
composite cell model is obtained when the circular
reinforcing column is becoming tangent to the cell's
boundary, that is referring to Fig. 3, when a b or
/4 0785. As a consequence, any greater
value of would correspond to overlapping
columns, which falls beyond the range of validity
of the proposed model. Thus, equation (30) is no
longer applicable, and as rightly pointed out in the
contribution, should be replaced by equation (29),
which still represents a lower-bound estimate, since
the piecewise constant stress eld upon which it is
based does not involve any particular geometry of
the column cross-section.
Concerning the evaluation of the bearing
capacity of a foundation resting upon a purely

572

DISCUSSION

cohesive soil (undrained saturated clay), reinforced


by a network of vertical columns placed beneath
the footing, the authors of the contribution misuse
equations (43) which give an estimate of the
bearing capacity enhancement factor R. Indeed,
for the selected value of 408, Kp . 2, so that
only equation (43b) and not equation (43a) is to
be considered as a better lower-bound value of R.
Nevertheless, they notice that such a theoretical
value seems to underestimate signicantly those
usually adopted for the design of stone columns'
foundations.
In order to meet such an objection, we refer to
the particular situation of one single trench of
granular material located beneath a strip footing.
This case has been thoroughly investigated by
Bouassida & Hadhri (1995) under the plane strain
assumption, by implementing both the static and
kinematic methods of yield design. Their analysis
clearly shows that, leaving aside the effects of
gravity forces and water pore pressure, the bearing
capacity factor as dened by equation (44) lies
approximately between 13 (lower-bound value) and
15 (upper-bound value) for a trench of same width
as the footing lled up with a material of friction
angle 408. Consequently the corresponding

value of R is no more than 3, that is among the


smallest values quoted by Nagpure and Madhav,
even if, as expected, it remains higher than the
lower-bound values (1522) derived in our paper
for the general case.
Nonetheless, we do agree with the idea that, in
the bearing capacity problem, it is still necessary
to improve further the static approach by exploring
more complex stress elds than those considered in
the paper. Owing to the strong heterogeneity of the
foundation soil, this is likely to be a quite difcult,
if not impossible, task. Following the suggestion
already outlined in the paper, a homogenization
method should preferably be used. This requires
as a preliminary, to determine the macroscopic
strength properties of the composite soil regarded
as a homogeneous but obviously anisotropic continuum. Such a promising approach is currently in
progress.

REFERENCES
Bouassida, M. & Hadhri, T. (1995). Extreme load of soils
reinforced by columns: the case of an isolated column
Soils Fdns, 35, No. 1, 2136.

Vous aimerez peut-être aussi