You are on page 1of 22

Geotextiles and Geomembranes 19 (2001) 235–256

Bearing capacity of strip footings supported on


geocell-reinforced sand
Sujit Kumar Dash, N.R. Krishnaswamy, K. Rajagopal*
Geotechnical Engineering Division, Department of Civil Engineering, Indian Institute of Technology Madras,
Chennai 600 036, India
Received 7 April 2000; received in revised form 8 December 2000; accepted 21 January 2001

Abstract

This paper presents the results from laboratory-model tests on a strip footing supported by
a sand bed reinforced with a geocell mattress. The parameters varied in the testing program
include pattern of geocell formation, pocket size, height and width of geocell mattress, the
depth to the top of geocell mattress, tensile stiffness of the geogrids used to fabricate geocell
mattress and the relative density of the sand. With the provision of geocell reinforcement,
failure was not observed even at a settlement equal to 50% of the footing width and a load as
high as 8 times the ultimate bearing capacity of the unreinforced sand. Based on the model test
results, the depth of placement and the dimensions of the geocell layer for mobilising
maximum bearing capacity improvement have been determined. In addition to the tensile
strength of reinforcement, the aperture size and orientation of ribs of the geogrid used to
fabricate geocell mattress must be taken into account while evaluating its contribution to the
improvement in the performance. # 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Bearing capacity; Geocell reinforcement; Model test; Strip footing; Reinforced sand

1. Introduction

Over the last two decades, the beneficial effects of using planar reinforcement to
increase the bearing capacity of sand have been clearly demonstrated by several
investigators (Binquet and Lee, 1975, Akinmusuru and Akinbolade, 1981, Fragaszy
and Lawton, 1984, Guido et al., 1986, Khing et al., 1993, etc.). The more recent
advancement of reinforced soil is to provide three-dimensional confinement to the

*Corresponding author. Tel.: 91-44-4458298; fax: 91-44-235-0509.


E-mail address: gopal@civil.iitm.ernet.in (K. Rajagopal).

0266-1144/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 6 - 1 1 4 4 ( 0 1 ) 0 0 0 0 6 - 1
236 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

soil by using geocells. The geocell foundation mattress consists of a series of


interlocking cells, constructed from polymer geogrids, which contains and confines
the soil within its pockets. It intercepts the potential failure planes because of its
rigidity and forces them deeper into the foundation soil, thereby increasing the
bearing capacity of the soil.
Several investigations have reported the beneficial use of geocells. Rea and
Mitchell (1978) and Mitchell et al. (1979) conducted a series of model plate load tests
on circular footings supported over sand-filled square shaped paper grid cells to
identify different modes of failure and arrive at optimum dimensions of the cell.
Schimizu and Inui (1990) carried out load tests on single six-sided cell of geotextile
wall buried in the subsurface of the soft ground. Krishnaswamy et al. (2000) have
conducted load tests on geocell supported model embankments over soft clay
foundation. Cowland and Wong (1993) reported case studies on geocell mattress
supported road embankment. Bush et al. (1990) have proposed a methodology to
calculate the increase in bearing capacity of the soft soil due to the presence of
geocell mattress on the top of it.
The present study investigates the reinforcing efficiency of the geocell mattress
within a homogeneous sand bed supporting a strip footing. The experimental study
has involved performing more than 60 laboratory model tests by varying different
parameters as detailed in the paper in a later section.

2. Laboratory model tests

The model tests were conducted in a steel tank with a length of 1200 mm, width of
332 mm and a height of 700 mm. The two long sides of the tank were made of 15 mm
thick perspex sheet and were braced laterally on the outer surface with mild steel
angles to avoid yielding during the tests. The model foundation used was made of
steel and measured, 330 mm length  100 mm width  25 mm thickness. The base of
the model footing was made rough by cementing a thin layer of sand to it with epoxy
glue. The footing was centered in the tank, with the length of the footing parallel to
the width of the tank. In order to create plane strain conditions within the test
arrangement, the length of the footing was made almost equal to the width of the
tank. On each side of the tank, a 1 mm wide gap was given to prevent contact
between the footing and the side walls. The side wall friction effects on the model test
results were reduced by coating the inside of the perspex walls with petroleum jelly.
The two ends of the footing plate were polished to have smooth surface and also
coated with petroleum jelly to minimise the end friction effects.
The soil used in the present investigation was a dry river sand with coefficient of
uniformity (Cu) of 2.318, coefficient of curvature (Cc) of 1.03 and effective size of
particle (D10) of 0.22 mm. The soil can be classified as poorly graded sand with letter
symbol SP according to the Unified Soil Classification System. The maximum and
minimum dry unit weights of the sand are found to be 17.41 and 14.30 kN/m3. The
model footing tests were performed at relative densities of 30, 40, 50, 60 and 70%.
These relative densities were achieved in the test tank using a sand raining technique
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 237

Table 1
Properties of geogrids

Descriptions

Parameter BX NP-1 NP-2

Ultimate tensile strength (kN/m) 20 4.5 7.5


Failure strain (%) 25 10 55
Initial modulus (kN/m) 183 75 95
Secant modulus at 5% strain (kN/m) 160 70 70
Secant modulus at 10% strain (kN/m) 125 45 50
Aperture size (mm) 35  35 50  50 87
Aperture opening shape Square Square Diamond

by changing the height of fall. These heights of fall to achieve different relative
densities were determined a priori by performing a series of trials. The relative
densities achieved during the tests were monitored by collecting samples in small
aluminium cans of known volume placed at different locations in the test tank. The
difference in densities measured at various locations was found to be less than 1%.
The friction angle of the sand at three relative densities 30, 50 and 70%, as
determined from standard triaxial compression tests are 39.28, 418 and 42.28,
respectively.
The geocells were formed using three different types of geogrids; one of these is a
biaxial grid (BX) made of oriented polymer while the other two were made of non-
oriented polymers, referred to as NP-1 and NP-2 grids. The properties of the
geogrids were determined from standard wide width tension tests (American Society
for Testing and Materials, 1986) and are listed in Table 1. The load–strain behaviour
of these geogrids is presented in Fig. 1. The geocell mattress was prepared by cutting
the geogrids to required length and height from full rolls and placing them in
transverse and diagonal directions with bodkin joints (plastic strips) inserted at the
connections (Bush et al., 1990). The two different patterns (chevron and diamond)
used to form geocell mattress are illustrated in Fig. 2.
The geocell mattress was introduced at the required depth below the footing and
was filled with sand by the raining technique. The density of the soil placed within
the geocell mattress was also monitored by collecting soil samples from this layer as
explained earlier. The reduction in soil density in this layer due to the obstruction
from geocell pockets was observed to be less than 1% in all the test cases. As the
geocells were made of geogrids having more than 80% open area, the free flow of
sand during the sand raining might not have been much affected leading to this
marginal reduction in placement density.
The geometry of the test configurations considered in this investigation is shown in
Fig. 3. The pocket size (d ) of the geocells is taken as the diameter of an equivalent
circular area of the geocell pocket opening. Eight series of tests were conducted by
varying patterns of formation of geocell mattress, pocket size of geocells (d ), height
of geocell layer (h), the width of the geocell mattress (b), depth to the top of geocell
238 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Fig. 1. Load–strain behaviour of geogrids used in the test program.

layer below the footing (u), relative density of soil and type of reinforcement used to
form the geocell. The details of the various tests are given in Table 2.
The footing was loaded by a hand operated hydraulic jack supported against a
reaction frame. The hydraulic jack was connected to the footing through a pre-
calibrated proving ring to measure the loads applied on the footing. A ball bearing
was positioned between the proving ring and the footing to ensure that no
extraneous moment was applied to the footing. The load was applied in small
increments. Each load increment was maintained constant until the footing
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 239

Fig. 2. Patterns used for the formation of geocells.

Fig. 3. Geometric parameters of geocell-reinforced foundation bed.

settlement has stabilised. Settlements on the footing were measured through two dial
gauges located on either side of the footing. The deformations (heave/settlement) of
the soil surface on either side of the footing were also measured by dial gauges. In the
absence of a clear-cut failure, the loading was applied until a footing settlement of
around 50 mm is reached.
Many of the above tests were repeated once again to verify the consistency of the
test data. Apart from the footing load-settlement and surface deformations, the
measured parameters during the tests include strain in the reinforcement, pressure at
the base of the geocell mattress and the pattern of movement of soil below the geocell
layer. This paper presents only some of the above results.

3. Results and discussion

Typical pressure-settlement responses that illustrate the effect of the width of


geocell layer are shown in Fig. 4. It could be observed that the unreinforced sand had
a clearly defined ultimate (failure) pressure, which is much lower than that with
geocell reinforcement. In the case of geocell-reinforced tests, there was no
pronounced peak pressure. At a settlement in the range of 15–20% of footing
240 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Table 2
Details of model test series

Test series Details

A Tests on unreinforced sand with relative densities (ID) of 30, 40, 50, 60, 70%

B Variable parameter: Patterns of geocell formation: chevron, diamond


Constant parameters: d/B =1.5, h/B=0.8, b/B=12, u/B=0.1, geocell made of BX
grid, ID=70%

C Variable parameter: d/B=1.2, 1.5, 2.7


Constant parameters Chevron pattern, h/B=0.8, b/B=12, u/B=0.1, geocell made of
BX grid, ID=70%

D Variable parameter: h/B=0.8, 1.6, 2.0, 2.75, 3.14


Constant parameters: Chevron pattern, d/B=1.2, b/B=12, u/B=0.1, geocell made of
BX grid, ID =70%

E Variable parameter: b/B=1, 2, 4, 6, 8, 10, 12


Constant parameters: Chevron pattern, d/B=1.2, h/B=2.75, u/B=0.1, geocell made
of BX grid, ID=70%

F Variable parameter: u/B=0.0, 0.1, 0.25, 0.5, 0.75, 1.0, 1.5


Constant parameters: Chevron pattern, d/B=1.2, h/B=2.75, b/B=8, geocell made of
BX grid, ID=70%

G Variable parameter: Geocell materials: BX, NP-1, NP-2 grids


Constant parameters : Chevron pattern, d/B=1.5, h/B=1.2, b/B=8, u/B=0.1,
ID=70%

H Variable parameter: ID=30, 40, 50, 60, 70%


Constant parameters: Chevron pattern, d/B=1.2, h/B=1.6, b/B = 8, u/B=0.1,
geocell made of BX grid.

width, there is a slight reduction in the slope of the pressure-settlement response after
which the slope had remained constant. Even when the data was plotted in log–log
scale, the break point where the slope has changed significantly was not clearly
apparent. A similar trend was observed in most other test data. Typical pressure-
settlement responses with different relative heights (h/B) are shown in Fig. 5. It could
be observed that when the h/B ratio is 0.80, the ultimate pressure is clearly defined
and for other h/B ratios, the response is almost linear even at settlements equal to
about 50% of the footing width. Even for the case of h/B=0.8, the ultimate pressure
has increased by about 5 times the ultimate capacity of the unreinforced sand.
In the responses shown in the earlier two figures, it is interesting to note that, the
load-settlement response is linear up to a settlement of about 5% of footing width in
the unreinforced sand while this limit has gone up to about 20% when reinforced
with geocell layer. Even at a settlement equal to about 50% of the footing width,
clear signs of failure were not evident in the case of geocell-reinforced foundations.
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 241

Fig. 4. Variation of bearing pressure with settlement for different widths of geocell mattress}test series E.

At this stage, the load on the footing was almost 8 times higher than the ultimate
capacity of the unreinforced sand.
In comparison to this response, the systems reinforced with planar reinforcement
fail at settlements equal to about 20% of footing width, with clearly defined peak
loads in the range of about 4–4.5 times the ultimate capacity of unreinforced soil,
Omar et al. (1993) and Khing et al. (1993). This is because the planar reinforcement
242 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Fig. 5. Variation of bearing pressure with settlement for different heights of geocell mattress}test
series D.

deflects downwards when the soil undergoes shear failure (Yetimoglu et al., 1994),
initiating a catastrophic failure while much of its tensile strength remains
unmobilised. In the case of geocell reinforcement, however, the total reinforcing
system, being an interconnected cage, derives tremendous anchorage from both
sides of the loaded area due to the frictional and passive resistance developed at the
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 243

soil–geocell interfaces. Further, because of shear and bending rigidity, the geocell
layer supports the footing even after the shear failure of the sand inside its pockets
below the footing. This action is similar to that of a beam supporting a column load.
The support for this beam action is derived from the anchoring mechanism
developed at both the sides. Hence catastrophic failures are not observed in the case
of geocell reinforcement. This was confirmed from the post test observation that, the
horizontal and vertical ribs of the geogrids making up the geocell wall showed high
strains and buckling type deformations below the footing. This type of deformations
was observed to extend over the full height of the geocell indicating that the entire
geocell mattress acted as a single entity.
The improvement due to the provision of geocell reinforcement is represented
using a non-dimensional factor, called bearing capacity improvement factor (If).
This is defined as the ratio of footing pressure (q) with geocell-reinforced soil at a
given settlement to the pressure on unreinforced soil (q0) at the same settlement.
When this ratio is calculated at settlements beyond the ultimate capacity of
unreinforced soil, the ultimate capacity of the footing (qult) is used instead of q0.
The footing settlement ‘‘s’’ and the surface deformation (settlement/heave) ‘‘d’’ were
also expressed in non-dimensional form in terms of the footing width as s/B (%) and
d/B (%).
The variation of the bearing capacity improvement factor (If) with footing
settlement for different cases are shown in Table 3 and Figs. 6, 8 and 9. The surface
settlement/heave on either side of the loaded area for different groups of tests are
presented in Table 4 and Figs. 7 and 10, where the solid and dotted lines indicate the
measurements at points 250 mm (2.5B) to the left and right of the footing center line.
The footings on the unreinforced sand had undergone severe rotation at failure (at a
settlement of about 10% of footing width) typical of general shear failure of footings
(Note: the rotation of footings is indicated by large difference between the solid and
dotted lines). On the other hand, the footings supported on the geocell layer did not
show rotations until much larger settlements. The data on the footing rotation could
also be used to decide whether the footing has reached its ultimate capacity or not.
For example, the footings that did not show clear failure in the pressure-settlement
response have also not undergone significant rotation. There was significant heaving
in the case of unreinforced soils while footings with geocell reinforcement have
shown much lesser heave. In fact, the provision of geocell reinforcement has led to
uniform surface settlement. This is because the footing load is distributed over a
much larger area due to the rigidity of geocell layer leading to lower pressures
transferred to the soil.

3.1. Effect of pattern of geocell formation

The bearing capacity improvement factors (If) at different settlement ratios for the
two patterns of geocell formation (test series-B) are shown in Table 3. It could be
observed that the results with both the patterns are almost the same up to settlement
equal to 10% of footing width. Beyond that settlement, the Chevron pattern has
given slightly higher improvement factors. This is due to the higher rigidity of the
244 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Table 3
Summary of results in terms of bearing capacity improvement factor (If) from test series: B–E

Test Variable Bearing capacity improvement factor (If)


series parameter
(s/B) (s/B) (s/B) (s/B) (s/B) (s/B) (s/B)
3% 5% 10% 15% 20% 30% 40%

B (Pattern)
Chevron 1.40 1.48 1.92 2.77 3.26 4.02 4.57
Diamond 1.42 1.48 1.92 2.67 3.13 3.64 4.18

C (d/B)
1.20 1.50 1.63 2.16 3.04 3.69 4.30 4.98
1.50 1.40 1.48 1.92 2.77 3.26 4.02 4.57
2.70 1.22 1.37 1.69 2.19 2.41 2.82

D (h/B)
0.80 1.50 1.63 2.16 3.04 3.69 4.30 4.98
1.60 2.08 2.02 2.49 3.58 4.35 5.77 7.05
2.00 2.06 2.00 2.64 3.84 4.78 6.33 7.62
2.75 2.08 2.00 2.60 3.80 4.78 6.45 7.93
3.14 2.08 2.06 2.75 3.98 4.88 6.65 8.18

E (b/B)
1 1.25 1.35 1.60 2.28 2.79 3.57 4.21
2 1.40 1.53 1.89 2.64 3.20 4.07 4.89
4 1.53 1.65 2.21 3.24 4.20 5.49 6.55
6 1.68 1.82 2.40 3.45 4.29 5.75 7.09
8 1.68 1.83 2.49 3.61 4.57 6.16 7.61
10 1.68 1.83 2.47 3.68 4.71 6.33 7.82
12 2.00 2.00 2.58 3.80 4.78 6.45 7.93

C and D Aspect ratio (h/d)


0.30 1.22 1.37 1.69 2.19 2.41 2.82
0.53 1.40 1.48 1.92 2.77 3.26 4.02 4.57
0.67 1.50 1.63 2.16 3.04 3.69 4.30 4.98
1.33 2.08 2.02 2.49 3.58 4.35 5.77 7.05
1.67 2.06 2.00 2.64 3.84 4.78 6.33 7.62
2.29 2.08 2.00 2.60 3.80 4.78 6.45 7.93
2.62 2.08 2.06 2.75 3.98 4.88 6.65 8.18

geocell in Chevron pattern resulting from a larger number of joints (Fig. 2) for the
same plan area of geocell. The effect of the formation type is only marginal because
the effective pocket size is the same in both the systems. Until the soil fails in shear at
large deformations, the geocell mattress acts like a composite mat. At this stage, the
confinement effects in both the patterns are the same because of the same size and
same number of pockets in both the patterns. This is evident from the results shown
in Table 4, which shows that the surface deformation is almost the same for both the
cases till a settlement ratio of around 10%. After the soil fails in shear it starts
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 245

Table 4
Summary of results in terms of surface deformations (d/B)% from test series: B–E

Test Variable Surface deformation (d/B)% (settlement is +ve and heave is ve)
series parameter
(s/B) (s/B) (s/B) (s/B) (s/B) (s/B) (s/B)
3% 5% 10% 15% 20% 30% 40%

B (Pattern)
Chevron
Left 0.28 0.30 0.30 0.10 0.80 2.51 4.24
Right 0.23 0.30 0.30 0.20 1.10 3.13 5.58
Diamond
Left 0.20 0.28 0.20 0.32 1.00 2.70 4.78
Right 0.32 0.40 0.40 0.10 0.98 3.10 4.88

C (d/B)
1.2 Left 0.30 0.40 0.45 0.30 0.28 1.54 3.33
Right 0.28 0.40 0.43 0.20 0.42 2.15 4.18
1.5 Left 0.28 0.30 0.30 0.10 0.80 2.51 4.24
Right 0.23 0.30 0.30 0.20 1.10 3.13 5.58
2.7 Left 0.20 0.24 0.01 1.00 2.21 5.10 }
Right 0.15 0.15 0.20 1.10 2.28 4.70 }

D (h/B)
0.80 Left 0.30 0.40 0.45 0.30 0.28 1.54 3.33
Right 0.28 0.40 0.43 0.20 0.42 2.15 4.18
1.60 Left 0.40 0.55 0.71 0.80 0.78 0.36 0.02
Right 0.42 0.56 0.71 0.72 0.60 0.10 0.40
2.00 Left 0.40 0.53 0.82 1.10 1.10 1.12 1.10
Right 0.38 0.55 0.75 0.90 0.93 0.92 0.90
2.75 Left 0.40 0.62 0.92 1.08 1.10 1.10 1.04
Right 0.38 0.60 0.90 1.08 1.12 1.12 1.00
3.14 Left 0.32 0.50 0.84 0.94 0.99 0.99 1.00
Right 0.38 0.58 0.92 1.10 1.18 1.24 1.22

E (b/B)
1 Left 0.30 0.40 0.53 0.40 0.08 1.20 2.76
Right 0.29 0.46 0.53 0.53 0.30 0.80 2.08
2 Left 0.40 0.56 0.68 0.63 0.28 0.68 1.68
Right 0.37 0.42 0.42 0.38 0.10 1.38 3.12
4 Left 0.40 0.60 0.90 1.02 1.10 1.10 1.06
Right 0.38 0.57 0.83 0.94 0.93 0.69 0.22
6 Left 0.33 0.51 0.72 0.80 0.80 0.64 0.54
Right 0.33 0.50 0.78 0.82 0.82 0.73 0.40
8 Left 0.40 0.60 1.00 1.22 1.30 1.36 1.40
Right 0.40 0.60 0.99 1.18 1.26 1.30 1.30
10 Left 0.37 0.58 1.00 1.26 1.33 1.40 1.40
Right 0.33 0.52 0.78 0.89 0.88 0.82 0.80
12 Left 0.40 0.61 0.92 1.07 1.10 1.10 1.04
Right 0.37 0.60 0.90 1.09 1.12 1.12 1.00
246 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Fig. 6. Variation of improvement factors with settlement for different depths of placement of geocell
mattress}test series F.

moving out leading to heaving of surface and its density reduces (because of volume
expansion). Hence, part of the footing load is directly transferred to the walls of the
geocell. At this stage, the flexural rigidity of the geocell starts playing an important
role in its load carrying capacity. Hence, only at large deformations is there some
difference in the improvement factor (If) between the two patterns.
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 247

Fig. 7. Variation of surface deformation with footing settlement for different depths of placement of
geocell mattress}test series F.

3.2. Effect of pocket size of geocell

The effect of pocket size (d/B) on the bearing capacity improvement factor (If) is
shown in Table 3 (from test series-C). The increase in load carrying capacity with
decrease in pocket size is due to the overall increase in rigidity of the mattress. At the
same time, the confinement offered by cells per unit volume of soil also increases with
248 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Fig. 8. Variation of improvement factors with settlement for geocells made of different types of
geogrids}test series G.

decrease in the pocket size. Both these factors contribute to the overall improvement
of the performance. The effect of the pocket size on the soil behaviour is more clearly
evident from the data presented in Table 4, which shows that the heave of soil is
higher for larger pocket sizes. As the pocket size increases, the confinement reduces
and hence the soil freely moves out of the pockets leading to larger surface heave.
Rajagopal et al. (1999) have also observed similar influence of the pocket size on the
behaviour of geocell-reinforced sands.
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 249

Fig. 9. Variation of improvement factors with settlement for different relative densities of sand}test
series H.

3.3. Effect of height of geocell mattress

The improvement in load carrying capacity with settlement for different heights of
geocell mattress (test series-D) is shown in Table 3 and its effect on the surface
deformations is shown in Table 4. As the height of geocell layer is increased, the
footing load is dispersed over a larger area and the actual pressure transmitted to the
250 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Fig. 10. Variation of surface deformation with footing settlement for different relative densities of
sand}test series H.

soil reduces leading to an improvement in the overall performance. However, beyond


an h/B ratio of 2, the pressure-settlement responses have not shown significant
change. The reason for this is that the geocell walls buckle locally just under the
footing leading to higher settlements. Hence, the bearing capacity improvement
factor (If) has not increased directly in proportion to the height of geocell layer.
From this result, we may say that beyond a certain h/B ratio, the increased flexural
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 251

rigidity of the geocell layer remains unmobilised because of its local buckling below
the footing.
The results are further analysed in terms of surface deformation on both sides of
the footing as presented in Table 4. It is clear that the surface heave decreases with
increase in height of geocell and for h/B 52, the surface undergoes settlements only
instead of heave. This is believed to be due to the fact that, at smaller cell heights the
sand below the footing overcomes the frictional resistance on the vertical cell surface
and moves out through its bottom leading to surface heave. However, with increase
in height, the sand remains confined within the cell and the entire mattress deflects as
a composite mass.

3.4. Effect of width of geocell mattress

It may be seen from the results in Table 3 corresponding to the test series E, that
even with a geocell mattress of width equal to the width of the footing (b/B=1),
significant performance improvement is obtained. As the footing is supported by the
relatively rigid geocell mattress, the footing loads are transferred to the foundation
soil at the base of the geocell. For this test configuration, the geocell base is at a
depth of 285 mm (i.e., 2.85B). Because of the depth effect and the overburden
pressure effects, the surface footing behaves like a footing placed at some depth
below the surface. These factors would have contributed to the improvement in the
overall performance even when the geocell layer does not extend beyond the footing
edges.
As the width of geocell mattress increases, the performance improves up to a b/B
ratio of 4. Beyond that width of the geocell mattress, performance improvement is
negligible. This trend can better be explained with the help of observed surface
deformation data presented in Table 4. The soil surface has heaved up to a geocell
width of b/B=2. At a b/B ratio of 4, there was slight heaving tendency at relatively
large settlements. At higher b/B ratios, the surface has not undergone heaving even
at large settlements. This can be explained from the fact that as the width of the
geocell layer is increased beyond a certain limit, it intercepts all the potential rupture
planes thus preventing the soil heave. From the observed heave patterns in this
investigation, it may be said that the farthest rupture plane would be at a distance of
around 2B from the centre line of the footing on both sides. Similar observations
were made by Selig and Mckee (1961) and Chummar (1972) on the extent of the
rupture planes in a homogeneous sand bed below strip footings. As the geocell layer
of width about 4B completely encapsulates the failure zone of the soil, any increase
in its width would only lead to a marginal improvement due to the secondary factors
such as anchorage at the two ends.

3.5. Effect of depth of placement of geocell mattress

The influence of the depth of placement of geocell layer (defined by u/B ratio) on
the bearing capacity improvement factor (If) determined from the test series-F is
shown in Fig. 6. It is seen that, there is a slight improvement in load carrying
252 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

capacity when u/B was increased from 0 to 0.1 and there after the load carrying
capacity continued to decrease with increase in the depth of placement. The slight
increase in performance improvement with u/B of 0.1 may be due to the effect of soil
cushion spreading the load over larger area of geocell and preventing the geocell wall
from direct contact with the footing that would bring an early buckling. With further
increase in the depth of placement, the soil between the footing and the geocell layer
would squeeze out leading to larger settlements. This is reflected in the reduction of If
for higher u/B ratios. These results suggest that to get maximum benefit, the top of
the geocell mattress should be at a depth of 0.1B from the bottom of the footing.
Up to u/B ratio of 0.25, the footings have not shown evidence of failure even at
large settlements. When u/B was 0.50, the footing had an initial failure at a
settlement of about 0.2B and later started taking higher loads and finally reached its
ultimate load at settlement of about 0.4B. When the u/B ratio was increased beyond
0.5, the footings have reached ultimate pressures at much smaller settlements of
about 0.15B.
The surface settlement data for this test series is shown in Fig. 7. The data clearly
shows that up to a u/B ratio of 0.5, the surface heave is negligible. When the geocell
layer is placed at larger depths, the surface has heaved significantly indicating that
the sand has squeezed out laterally between the footing and the relatively rigid
geocell layer. From these results, it may be said that when geocell layer is placed at
depths greater than 0.5B, its contribution to the improvement is only marginal.
When placed at such depths, its role would be similar to that of a rough, rigid base at
shallow depths below the footings.

3.6. Effect of type of geogrid used to make geocell

The influence of the type of geogrid on the footing performance was studied in the
test series-G and the results are presented in Fig. 8. Both the geocells made of BX
and NP-2 grids have shown the same improvement factor up to a settlement equal to
about 0.2B, though the stiffness of the former is much higher than that of the later.
This is believed to be due to the better confinement offered to sand by NP-2 grids due
to its smaller aperture openings. Although, both NP-1 and NP-2 grids have almost
same stiffness, the performance with NP-1 grids at lower settlements is slightly
inferior because of its higher aperture opening size.
At higher settlements, the performance with geocells made of BX grids is much
better because of its higher stiffness. At this stage, the sand starts moving out of the
geocell pockets and hence the stiffness of the geocell layer has influence on the overall
behaviour.
In the case of geocells made of NP-2 geogrids, a sudden failure was observed at a
settlement of about 20% of the footing width. The exhumed geocells have shown
that the geocell walls have folded under the footing leading to early failure of the
footing in this case. This folding might have happened because the ribs are in
inclined direction (contrary to the other two cases where they are horizontal and
vertical) and are unable to resist vertical compression from the footing in the post
soil shearing stage.
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 253

Based on the findings from this series of tests, it would seem reasonable to argue
that, the tensile strength of the geogrid alone is not a sufficient parameter. The
aperture size and the orientation of the ribs too should be taken into account while
evaluating the performance of a geocell foundation mattress.

3.7. Effect of relative density of soil

The influence of the relative density of the soil was investigated in test series-H.
The bearing capacity improvement factors obtained with different relative densities
are shown in Fig. 9. Based on the results of pullout tests on geogrid reinforcements in
sand, Farrag et al. (1993) have reported that, the interfacial frictional resistance
increases with increases in soil density. Therefore, with the increase in relative density
of soil, the frictional resistance between the geocell wall and the sand increases,
thereby increasing the resistance to downward penetration of sand below the geocell
and hence a higher improvement in load carrying capacity. Another factor to
consider is that the soil with higher relative density will dilate more thus mobilising
higher strains in the geocell layer. These two factors will contribute to enhanced
performance with higher relative densities.
The effect of relative density on the surface heaving is shown in Fig. 10. It can be
clearly seen that the soils with higher relative density have undergone more surface
heaving because of volumetric expansion. From the figure, it could be seen that the
surface heaving starts at a settlement equal to about 15–20% of footing width. Only
after this stage, higher strains are mobilised in the geocell layer. This could be the
reason for rapid increase in If values beyond settlement of about 0.15B as compared
to that at lower settlements (Fig. 9). It is interesting to note from this figure that in
the case of soils with 30 and 40% relative density, the rate of increase of If at large
settlements is the same as that at smaller settlements. On the other hand, the soils
with higher relative density had exhibited higher rate of increase of If at higher
settlements as compared to that at smaller settlements. The reason for this is as
explained earlier based on the dilation of the soil and the resulting strains in geocell
layer.

3.8. Effect of aspect ratio

The test results from series-C and series-D have been summarised in Table 3 in
terms of aspect ratio (height to diameter ratio, h/d) of geocell pockets. It is found
that the load carrying capacity of the foundation bed increases with the increase in
aspect ratio up to 1.67, beyond which further improvement was observed to be
marginal. From this result, it may be concluded that the optimum aspect ratio for
strip footings is around 1.67. The same for circular footings was reported to be
around 2.25 by Rea and Mitchell (1978). Krishnaswamy et al. (2000) have reported
that the optimum aspect ratio is unity (1) for geocell supported embankments
constructed on soft clays. The difference in the values of optimum aspect ratio could
be attributed to the differences in the type of loading and the foundation soils.
254 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

4. Summary and conclusions

This paper has presented the results obtained from laboratory model tests on strip
footings supported by homogeneous sand bed reinforced with geocell mattress. The
extrapolation of the results from these model tests to field cases should be done
carefully as the performance of the footings on sand is also dependent on the size of
the footing (Das et al., 1996, Kimura et al., 1985). The different parameters to
consider in such dimensional analyses are discussed by Butterfield (1999) and Fakher
and Jones (1996). In addition, the aspect ratio of the geocells is an important
parameter to be included in such analysis as discussed by Krishnaswamy et al.
(2000).
Based on results obtained from the present investigation, the following
conclusions can be made on the behaviour of strip footings resting on geocell-
reinforced sand beds.

1. The pressure-settlement behaviour of strip footing resting on geocell-


reinforced sand is approximately linear even up to a settlement of about 50%
of the footing width and a load as high as 8 times the ultimate capacity of the
unreinforced one.
2. Very good improvement in the footing performance can be obtained even with
geocell mattress of width equal to the width of the footing, because of the
transfer of footing loads to deeper depths through the geocell layer. The surface
footing in this case behaves like a deeply embedded footing thus improving the
overall performance.
3. The surface heave can be reduced substantially by providing geocell of sufficient
width to restrict the formation of the rupture plane within the foundation soil.
4. Chevron pattern for the formation of geocells is more beneficial than the geocells
in diamond pattern.
5. The performance improvement is significant up to a geocell height equal to 2
times the width of the footing. Beyond that height, the improvement is only
marginal.
6. The optimum width of the geocell layer is around 4 times the footing width at
which stage, the geocell would intercept all the potential rupture planes formed
in the foundation soil.
7. To obtain maximum benefit, the top of geocell mattress should be at a depth of
0.1B from the bottom of the footing.
8. Tensile strength of the grid used to fabricate geocell mattress is not a sufficient
parameter to evaluate the performance of the geocell mattress. Aperture size and
orientation of the ribs in the grid too have an important bearing in load carrying
mechanism of the geocell-reinforced foundation bed.
9. Better improvement in the performance of footing can be obtained by filling the
geocells with denser soils because of dilation induced load transfer from soil to
geocell.
10. The optimum aspect ratio of geocell pockets for supporting strip footings was
found to be around 1.67.
S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256 255

Acknowledgements

The financial support for this work from the Department of Science and
Technology, Government of India, New Delhi under sanction No. III 4(36)/90-ET
and the Netlon India a Division of Parry & Co. Ltd., Vadodara for supplying
different products free of cost under an MOU signed with IIT Madras are gratefully
acknowledged. The authors would like to thank Dr. N. Kumar Pitchumani for his
valuable comments and suggestions.

References

Akinmusuru, J.O., Akinbolade, J.A., 1981. Stability of loaded footings on reinforced soil. Journal of
Geotechnical Engineering Division, ASCE 107 (6), 819–827.
American Society for Testing and Materials, 1986. Standard test method for tensile properties of
geotextiles by wide–width strip method. ASTM D4595.
Binquet, J., Lee, K.L., 1975. Bearing capacity tests on reinforced earth slabs. Journal of Geotechnical
Engineering Division, ASCE 101 (12), 1241–1255.
Bush, D.I., Jenner, C.G., Basset, R.H., 1990. The design and construction of geocell foundation mattress
supporting embankments over soft ground. Geotextiles and Geomembranes 9, 83–98.
Butterfield, R., 1999. Dimensional analysis for geotechnical engineers. Geotechnique, London, UK 49 (3),
357–366.
Chummar, A.V., 1972. Bearing capacity theory from experimental results. Journal of the Soil Mechanics
and Foundations Division, ASCE 98 (12), 1311–1324.
Cowland, J.W., Wong, S.C.K., 1993. Performance of a road embankment on soft clay supported on a
geocell mattress foundation. Geotextiles and Geomembranes 12, 687–705.
Das, B.M., Puri, V.K., Omar, M.T., Evgin, E., 1996. Bearing capacity of strip foundation on geogrid-
reinforced sand-scale effects in model tests. Proceedings of the Sixth International Conference on
Offshore and Polar Engineering, Vol. I. Los Angeles, USA, pp. 527–530.
Fakher, A., Jones, C.J.F.P., 1996. Discussion of bearing capacity of rectangular footings on geogrid
reinforced sand, by Yetimoglu, T., Wu, J.T.H., Saglamer, A., Journal of Geotechnical Engineering,
ASCE 122 (4), 326–327.
Farrag, K., Acar, Y.B., Juran, I., 1993. Pull-out resistance of geogrid reinforcements. Geotextiles and
Geomembranes 12, 133–159.
Fragaszy, R.J., Lawton, E., 1984. Bearing capacity of reinforced sand subgrades. Journal of Geotechnical
Engineering Division, ASCE 110 (10), 1500–1507.
Guido, V.A., Chang, D.K., Sweeney, M.A., 1986. Comparison of geogrid and geotextile reinforced earth
slabs. Canadian Geotechnical Journal 23, 435–440.
Khing, K.H., Das, B.M., Puri, V.K., Cook, E.E., Yen, S.C., 1993. The bearing capacity of a strip
foundation on geogrid reinforced sand. Geotextiles and Geomembranes 12, 351–361.
Kimura, T., Kusakabe, O., Saitoh, K., 1985. Geotechnical model tests of bearing capacity problems in a
centrifuge. Geotechnique, London, UK 35 (1), 33–45.
Krishnaswamy, N.R., Rajagopal, K., Latha, G.M., 2000. Model studies on geocell supported
embankments constructed over soft clay foundation. Geotechnical Testing Journal 23 (1), 45–54.
Mitchell, J.K., Kao, T.C., Kavazanjiam Jr., E., 1979. Analysis of grid cell reinforced pavement bases.
Technical Report No. GL-79–8, US Army Engineers Waterways Experiment Station.
Omar, M.T., Das, B.M., Puri, V.K., Yen, S.C., 1993. Ultimate bearing capacity of shallow foundations on
sand with geogrid reinforcement. Canadian Geotechnical Journal 30, 545–549.
Rajagopal, K., Krishanaswamy, N.R., Latha, G.M., 1999. Behaviour of sand confined with single and
multiple geocells. Geotextiles and Geomembranes 17, 171–184.
256 S.K. Dash et al. / Geotextiles and Geomembranes 19 (2001) 235–256

Rea, C., Mitchell, J.K., 1978. Sand reinforcement using paper grid cells, Reprint 3130, ASCE spring
convention and exhibit. Pittsburgh, PA, pp. 24–28.
Schimizu, M., Inui, T., 1990. Increase in bearing capacity of ground with geotextile wall frame.
Proceedings of the Fourth International Conference on Geotextiles Geomembranes and Related
Products, Vol. 1. Hague, Netherlands. p. 254.
Selig, E.T., Mckee, K.E., 1961. Static and dynamic behaviour of small footings. Journal of the Soil
Mechanics and Foundations Division ASCE 87 (6), 29–47.
Yetimoglu, T., Wu, J.T.H., Saglamer, A., 1994. Bearing capacity of rectangular footings on geogrid-
reinforced sand. Journal of Geotechnical Engineering ASCE 120 (12), 2083–2099.

You might also like