(\ Proceedings: Second International Conference on Recent Advances In Geotechnical Earthquake Engineering and Soil Dynamics,

~ March 11-15, 1991 St. Louis, Missouri, Paper No. 1.28

Behavior of Geotextile Reinforced Soil under Cyclic Loading

G. Venkatappa Rao G. P. Singh
Professor In Civil Engineering, Indian Institute of Technology, New Formerly Graduate Student, Department of Civil Engineering, Indian
Delhi, India Institute of Technology, New Delhi, India

K. K. Gupta
Assistant Professor in Civil Engineering, Indian Institute of Technol-
ogy, New Delhi, India

SYNOPSIS The vast potential of geotextiles for separation(reinforcement in

pavements and rail roads is yet to be fully exploited in India, despite the initial
successful trials. This is partly due to the lack of sufficient test data on the behaviour
under repetitive loading of indigeneous geotextiles. This paper reports typical triaxial
test results on reinforced and unreinforced specimens before and after cyclic loading. The
reinforcement used is a needle punched polypropylene geotextile manufactured in India. The
results reported include strength before and after cyclic loading (cyclic loading under both
drained and undrained conditions), modulus of resilience and plastic strain. Limited test
results on simulated rail road (pavement) specimens clearly reveal the absence of
intermixing and improvement in strength and hence possible rutting.

INTRODUCTION subgrade materials to intrude and vice

versa.
Construction of unpaved roads is done by
placing the most readily available base( In order to understand the behaviour of
subbase materials over the subgrade. The roads and rail roads it is desirable to
movement of traffic causes intermixing of predict the irrecoverable deformation of the
base and subgrade, resulting in rutting, soil. In the present study an attempt has
particularly in soft soils. This been made to study the behaviour of soil
necessitates placement of additional base reinforced with geotextile when subjected to
material. On the other hand, placement of a cyclic loading in drained and undrained
geotextile between the base and subgrade conditions as well as simulated conditions
reduces the thickness requirement of base of road pavement and rail road. The detailed
course, and also reduces the maintenance experimental programme and the results are
during the life of the road. This improved presented.
performance of the road is due to the
various functions performed by geotextile
viz., separation, reinforcement, filtration LITERATURE REVIEW
and drainage. The reinforcement function may
be due to restraint effect, membrane effect, Friedli and Anderson (1982) studied the
and local reinforcement. effect of placing woven fabric between a
railway base course and a fine grained
A road subgrade is subjected to repeated subgrade by performing cyclic triaxial
wheel loads through the pavement layers tests. Saxena and Chiu (1982) have studied
resulting in permanent deformation. In a rail-road simulation to evaluate the fabric
rail road, the wheel load stresses are performance in a rail road system. In order
transmitted through the ballast on a point to investigate the stress-strain behaviour
to point contact. Loading in this case is of of the soil-fabric-ballast system, triaxial
extended repetitive duration and the tests with measurement of pore water
magnitude of loads is also higher. pressure on l5x30 em composite samples of
Intermixing of sub-base material into saturated soil-ballast with and without
ballast reduces the frictional strength of fabric were performed. A comparison of the
ballast significantly. In the ballast maximum deviator stress of the soil-fabric-
reinforcing function, the geotextile helps ballast samples with the soil ballast sample
to keep the individual ballast particles in clearly indicated that the strength of the
place and in their same relative positions. former is 2.7 times higher unde2 a confining
Therefore, one characteristic that is pressure 68.9 kPa (0.703 kgjcm). Based on
required is high surface friction that can the_dynamic kg tes~s they concluded that the
be performed by a geotextile or interlocking reslllent ehavlour and deformation
with ballast that can only be performed by a characteristics of soil-fabric-ballast
geogrid. Also a high tensile strength, system were improved. However, Friedli and
particularly in the cross-track direction is Anderson (1982) showed that resilient
required. Another characteristic of modulus decreases with number of cycles.
importance with regard to ballast
reinforcement is the ability to elongate Saxena and Wang (1982) found that the
around localized stresses. Otherwise, the geotextile makes the rail-ballast-soil
ballast will rupture the fabric allowing system to behave closer to an elastic

1913
foundation. On the basis of laboratory model
tests of such a system they concluded that
the reinforcing effect of the geotextile
causes reduced deflection and strains
within the whole system.
Despite the availability of the above quoted
and other research findings conclusively
demonstrating the advantages of geotextiles
as separator and reinforcement and also
their use throughout the world, it remains a
fact that the vast potential is yet to be
fully utilised in India. This is not
withstanding the fact that nearly 6,000 km
of rail line has been identified for track
bed stabilization and that field trials have
been very encouraging (Azeem 1988, Yog et
al., 1989). This has been attributed to the
fact that there is lack of sufficient test
data on the behaviour under repetitive
loading of geotextile reinforcedjsepara ted
granular; clayey soils using Indian made
geotextiles. with this in view a preliminary
investigation is taken up to demonstrate the
mechanical behaviour of geotextile
reinforced soil system under cyclic loading
using a needle punched geotextile made in
India.
Saturated sand was deposited into the
membrane lined split sample former in 3 or 5
layers (depending on whether 2 or 4
geotextile discs were used) placed over the
pedestal of a triaxial cell. Each layer was
compacted to achieve required density by
vibration. After that the geotextile disc
was placed in position. The next layer was
then placed in the same manner.
Specimens thus prepared were consolidated
under the desired cell pressures for 30
minutes by which time consolidation is found
to be complete. The triaxial cell was
assembled in a 10 ton Instron universal
testing machine. For cyclic loading, the
load was varied from zero to the required
level and then back to zero. The term cyclic
stress level is defined as the ratio of the
cyclic shear stress to the shear stress at
failure in a corresponding static test. For
applying the cyclic loads, the cross head
speed was kept at 0.33 mmjsec. Thus,
depending upon the stiffness of the specimen
the frequency of load cycling was found to
be around 0.25 to 0.75 Hz. The load was
recorded on one pen autographic recorder.
The permanent deformation of the specimen
was recorded through a dial gauge.

EXPERIMENTAL WORK All sand specimens were sheared at the end

of cyclic-loading at a deformation rate of
Two types of soils viz., fine grained sand 0.0167 mmjsec. The pavement simulated
(Dmax = 0.6 mm, D in= 0.04 mm, c = 1.62, specimens were tested at a rate of 0.0083
Cc=l.13) and silt (W. sand= 15, % s~lt = 70, mmjsec.
Clay = 15, L.L. (%) = 32, P.L. (%) = 22)
were used in the investigation. The
geotextil~ used is needle punched type, its RESULTS AND OBSERVATIONS
characterlstics are given in Table 1.
Reinforced and unreinforced specimens were Static and Cyclic Drained Tests on Sand
shea~ed both before and after applying
cycl1c stress at different confining The stress-strain-volu me change curves
pressures, for different number of cycles. obtained in drained triaxial tests on
In the first series drained and undrained reinforced and unreinforced sand specimens
triaxial compression tests on 38 mm diameter are presented in Fig.1. In general one can
and 76 mm high specimens were conducted on observe that the strength of sand increases
unreinforced and reinforced specimens. on reinforcement. This is in conformity with
that of Broms (1977) and the earlier work
carried out at I.I.T. Delhi (Venkatappa Rao,
TABLE 1: Properties of Geotextile used et al. 1987, 1989). It is observed that the
st~ength ratio, defined as strength of the
Mass per unit area (gjm 2 ) 275 re1nforced, specimen to that of the
Thickness (mm) 2.86 unreinforced, is higher for 4 disc
Tensile strength (kg/m) reinforcement (3.95 and 2.24 respectively
for 200 mm wide strip for a =100 and 400 kPa) when compared with
3
X-machine direction 1440 correspond1ng values for 2 disc
Machine direction 1403 reinforcement (1.87 and 1.71 respectively).
10% secant modulus The variation of volumetric strain with
Machine direction 110 axial strain indicates a shift towards
X machine direction 76 compression side with increase in
reinforcement and confining pressure.

Figu~e 2presents the variat1on of plastic

In the second series of tests, efforts were s~ra1n. Plastic strain generally increases
made to simulate the condition of rail road Wlth number of repetitions and cyclic stress
track by various simplifications. A two level. For instance, at a confining pressure
layer triaxial specimen was used to of 100 kPa with cyclic stress level of 0.25
represent base course and subgrade interface for .two. disc reinforced sand the plastic
using silt for subgrade (lower half of stra1n ls 2.5%, where as it is 6.0% for 4
specimen) and sand for ballast (upper half disc specimen for 500 cycles. This increase
of specimen). The silt was thumb remoulded in plastic strain is due to the increase in
at a relatively high moisture content of the magnitude of cyclic stress. A cyclic
22%. Tests were conducted on both reinforced stress level of 0.25 is equivalent to 190kPa
and unreinforced specimens. for 2 disc reinforced sand while for four
disc reinforcement it is 390 kPa. Thus it is

1914
 14
evident that by providing reinforcement,
repetitive loads inducing higher stress can 13
be allowed which cannot be permitted in
unreinforced sand. For example, a cyclic
stress loading of 390 kPa can be carried out
on sand reinforced with 4 discs. This cyclic
stress can not be applied to unreinforced
sand because of low available strength. It
-'~ ~- 0R ~•o r<Zinforcernoent
2R Two layers of reinforcement
L.R
32

100 500
No. of cycles

0
Fig.2 Variation of plastic strain with
a._ number of drained cycles.
.X.

N is evident that a geotextile is the best

2 alternate when the cyclic loading is higher
X than the bearing capacity of unreinforced
111 sand.
111
...
L. Values of resilient modulus (MR) of sand are
~ presented in Table 2. It is observed that MR
L. increases with number of load repetitions
0 '?0
'f' and confining pressure.
0 ~()
12
...> ...«--,
D ~ TABLE 2: Values of Resilient modulus (MR) I
/.:
/ MPa
/
/ Drained Undrained
/.
/
-- 2RJQ.O kPa
No.
of 03=100 kPa o 3 =400 kPa 03=100 kPa
eye-
les Cyclic Stress Level Cy. Stress
Level

0.25 0.5 0.25 0. 5 0.25 0.5

Unreinforced

100 99.4 148.5 216.8 238.1 49.30 55.1

Axial strain ( "1.) 60.75 69.5
500 106.1 165 234.8 268.3

4 Reinforced with 2 discs

0 10 164.6 172.7
0
100 48.2 65.8 200.0 224. 4 51.30 69.2
c: 500 60.3 90.4 218.2 265.0 61.80 85.2
E "'~2
iii Reinforced with 4 discs
u
L 10 37.1 45.7 121.6 123.8
100 49.7 55.2 131.7 163.0 39.50 71.7
"'
E 500 62.1 74.3 143.7 206.5 52.60 90.2
_2 0..
0 E
> 0
u

-2 Undrained Cyclic Tests

Fig.1 Stress-strain, volume chan::1e curves The results of plastic strain versus number
for reinforced and unreinforced of cycles are shown in Fig.3. The trend is
sand. similar to that of drained cycling. In
undrained cycling the plastic strain

1915
increases much more than that in drained -_..::::::i!o=.5==,5=-oo0.5,100
cycling. It is more than double the value a ----025,500
(L
when the number of repetitions increase from ..X ---0.25,100
10 to 500 in undrained condition whereas in
the drained condition it is small.
0 R -No reinforcement
2 R - Two toyers
11. I. R - Four \ayers
0) - 100 kPa
0
0 0.5, 0.25 Cyc l1c stress level
12 >
0"' 100, 500 No of eye les
0) ::.100 kPa
~ 10
c
t' I. 12 16
v; 8
Axial strain ('/,I
0.5,500
Ill
.2 ~ I.
Q_

c 0..
X
2
iii "' 2
v

-;;;
E 0..
..2EO
0 0 15
> v
-1
0~-------------------L------------~
lO lOO ~0 Fig.5 Stress-strain, curves after drained
No. of eye I es cyclic loading for 2 disc reinforced
Fig.3 Variation of plastic strain with sand.
number of undrained cycles.

Post-cyclic Behaviour 1.8

Drained Cycling:- Typical stress-strain-
volume change curves obtained in monotonic 0
Q_
triaxial shear tests after cyclic loading ..X

are presented in Figs.4, 5 and 6. A

comparison with the pre-cyclic monotonic "'9
behaviour indicates that reinforced sand X
Ill
Ill
a ~
Q_
..X v;
"'0 ~
0
X
>
Ill
Ill 0"' 0-5,0.25 Cyclic stress level
~ 100 1 500 No. of cycles
"'
~
0.5,0.25 Cyclic stress level 0) = 1.00 kPa
.2
a 100, 500 No. of cycles
> 0) = 100 kPa
0"'
2 I. Ei OL------~----~----~----~16L---~2~0
Axial strain ( '/, 1 Ax1al stra1n ( '/, l
I.
OS, SOO
.S
~
II\ v 0.
v E
~
- 0
E''-'
"'
E
::>
:J
0
- 2
0 0 >
> '-'-1
Fig.6 Stress-strain, curves after drained
Fig.4 Stress-strain curves after drained cyclic loading for 4 disc rainforced
cyclic loading on sand. sand.

1916
exhibits more rigidity after cyclic loading.
The response is linear upto half of the 0
0 .25' 500
(L
maximum deviatoric stress. Overall changes
in stress-strain curves have also been N
.::<.
.--- -=--: 1i .}5:_ iili)_--
__....,.,,_,_,.,., .,.;-- Mon · --
observed. Strength ratios of reinforced sand 9 I.
X
are included in Table 3, which are seen to Vl
increase with number of cycles. Volumetric
strain variation with axial strain is also ~
iii 2
presented in Figs. 4, 5 and 6. All curves
exhibit similar trends i.e., shifting 0
towards expansion side with increase in
number of cycles and with increase in cyclic 0
"'"'
>

stress level. From these curves one may

observe that there is a general improvement
in material properties after cyclic ~ I.
loading. An apparent increase in interface c
frictional angle between sand and geotextile e ci..
X
and densification of sand are possibly iii <ll

responsible for it. u

~

Q;
E
TABLE 3: Strength Ratios for Reinforced Sand ;;J

~ 6 8 10
Drained Undrained
No.
of 03=100 kPa 03=400 kPa 03=100 kPa Fig.7 Stress-strain curves in tests after
eye- undrained cyclic loading on sand.
les Cyclic Stress Level cy. Stress
Level

0.25 0.5 0.25 0.5 0. 25 0.5

0
Unreinforced (L
..X
N
100 1.19 1.18 1. 03 1.12 1.19 1. 04 9
500 1. 25 1. 29 1. 05 1.17 1. 32 1.10 X

Reinforced with 2 discs Ill

~
iii
100 1. 09 1. 20 1. 07 1.17 1.15 1. 03 L

500 1.19 1. 22 1. 08 1.19 1.23 1. 09 0 0.5,0.25 Cyclic stress level

.2 100,500 No. of eye les
>
Reinforced with 4 discs "'
0 2
OJ = 100 k Po
100 1. 07 1.17 1. 06 1. 49 1. 34 1. 01
500 1. 08 1.19 1.13 1. 52 1. 43 1. 05
0
5 10 15
Cy. = Cycles Axial strain ( ·;. l
Undrained Cycling ~ J
c c:i
X
The results of the monotonic triaxial shear
tests after undrained cycling are presented
e "'
iii
in Figs.?, 8 and 9. The strength ratios are
included in Table 3. The strength is found
to increase only at low cyclic stress level
(0.25). At high stress level (0.5) strength
increment is of very small order (1%- 10%).
This is in contrast of the observation
during the drained cycling. In drained
cycling increase in strength is more at Fig.8 Stress-strain curves in tests after
higher cycle stress level than at low cyclic undrained cyclic loading for 2 disc
stress level. However this needs reinforced sand.
verification by more data. At high stress
level failure strain reduced to 3% to 4% The volumetric strain versus axial strain
while at low stress level failure strain is curves are also presented in Figs.7, 8 and
more or less the same after and before 9, which indicate that the reinforced
cyclic loading. When repetitive loading is specimens behave as dense material after
done at higher stress level like 0.6, 0.75 cyclic loading.
the specimen failed at very low number of
cycles. For instance, for 2 disc reinforced A summary of the results of resilient
sand at a cyclic stress level of 0.6, the modulus is presented in Table 2. It is found
failure occurred after 193 cycles, whereas that the resilient modulus is 9enerally
at a cyclic stress level of 0.75 failure higher for reinforced sand.
occurred after only 43 cycles.

1917
 simulated drained triaxial tests with and
without reinforcement at interface. It is
noticed that the strength of simulated
specimen increases from 490 kPa to 540 kPa
after reinforcement, at a confining
pressure, of 100 kPa. This increase is
around 10%. The variation of volumetric
strain with axial strain showed a shift
towards compression side, on reinforcement.

Plastic strain:

Cyclic loading of composite specimen has

been carried out under undrained condition.
Plastic strain versus number of cycles is
presented in Fig.11 and values of plastic
strain are tabulated in Table 4. From these
0.25 1 0.50 Cyclic stress level it is clear that the plastic strain reduces
by nearly 70% at a low cyclic stress level
100 1 500 No of eye les
(0.25) due to reinforcement after 1000
0) = 100 kPa cycles. This proves that reinforcement
reduces plastic strain, and hence rutting,
16 20 track settlement thereby decreasing the
maintenance cost.
Axial strain ( •;. l
OR- No reinforcement
L. 1 R- One reinforcement
0.25,0.50 Cyclic stress level emenl
~ e'"'\orc I< po
-
C X
0.
~\\1"\ol.l\ r Q.SO' \R , 270
·-"'
~
iii

"'
E
::JO. u
0.25 OR 122.5 kPa
oE
>O iii
0
0.25. 1R' 135 kPo
'--' n:
Fig.9 Stress-strain curves in tests after
undrained cyclic loading on 4 disc 1000
reinforced sand.

Rail-Road (Pavement) Simulation

Fig.11 Variation of plastic strain vJith
stress-strain curves in static tests number of cycles (undrained) for
simulated specimen.
Figure 10 presents the stress-strain volume
change curves obtained from rail road TABLE 4: Percentage plastic strain during
undrained cyclic loading on
simulated rail road specimen
.X
N
0 Plastic Strain (%)

No. No rft. for One disc rft. for

Vl
Vl of cyclic stress cyclic stress
~ Cy. level of level of
iii
2
6 0.25 0.50 0.25 0.50
0 (122.5 (245 ( 135.0 (270
>
kPa) kPa) kPa) kPa)
0 "'
Axial strain ( •;.) 10 3.80 6.00 2.1 6. 40
100 4.59 9.18 3.0 8.70
1000 5.00 10.40 3.4 9.85
~ 0.1.
6.. rft. Reinforcement; Cy. = Cycle
X
.~
g "' Note: Values vJithin parentheses indicate
Vl cyclic stress.
.\o'
.!=. 0.. But at the higher cyclic stress level (0.5),
"'§ E
oO.I. the plastic strain of reinforced specimen
g'--' does not reduce to that extent. A cyclic
stress level of 0.25 represents 122.5 kPa,
Fig.10 Stress-strain, volume chan~e curves for unreinforced specimen where as it is 135
for simulated specimen. kPa for reinforced specimen.

1918
Resilient modulus:
&
A summary of the results of M with and .:<:
without reinforcement is tabulate~ in Table N

5 for a confining pr~ssure of 100 kPa. It is 9

seen that MR increases by 10% to 26% in X

reinforc3d specimen at the cyclic stress Vl o.2'i '100

Vl
level of 0.25. However, at the high cyclic
~ 6
stress level of 0.5, it decreases by 18% to
Vi
27% in one disc reinforcement. .... Mono;;;;;-
0
TABLE 5: Resilient modulus (MR), for rail 0 I,
road simulated specimen o 3 =100 kPa >
0"' 0.50,0.25 Cyclic stress level
Resilient modulus in MPa
No. 100 J 1000 No of cycles
of No rft. at cycle One disc rft for 0) =100 kPo
cy. stress level of cyclic stress level

0.25 0.50 0.25 0.50 I. 8 12 16 20

1.6 Ax1ol s I rom ( 'I, l 0.5,1000
c: ci.
X
.--··_..- 0.2S, 1000
10 424 666 4 67 541 ...- _.. . . ...... --0.50,100
100 466 718 571 604
0
.... "'
v; ___ ...... --- ..... -~02'i,100
500 518 933 642 734
1000 583 1167 734 856 u / ...- ---~
.--___ .-- / Monolon1c
a:;
rft. Reinforcement; cy. = Cycles E 0.0
~ D.
Resilient modulus is found to increase with 0 E
> 0
u
number of cycles and with cyclic stress
level for both reinforced and unreinforced Fig.l3 Stress-strain, volume change curves
specimens. However, Friedli and Anderson after undrained cyclic loading
(1982) showed that MR decreases with cycling (with one reinforcement).
for K0 condition.
TABLE 6: Strength ratio of rail road
stress-strain curves after repetitive loading simulated specimens after
undrained cycling at o 3 =100 kPa
Drained stress-strain and volumetric change
curves obtained from monotonic drained Strength ratio
triaxial tests after undrained cyclic No.
loading are shown in Figs.12 and 13. The of No rft. at cyclic One disc rft. at
strength ratios after undrained cycling to cy. stress level of cy.stress level of
before cycling are given in Table 6.
0.25 0.50 0.25 0.50
(122.5 (245 (135.0 (270
kPa) kPa) kPa) kPa)

100 1.13 l. 17 l. 3 J 1. 4J
1000 1.19 l. 21 1.G2 1. G4

0.25,0.5 Cyclic stress level rft. Reinforcement; Cy. ~ Cycle

Vl
100, 1000 No. of eye les
Note: Values inside parentheses inclicate
0) = 100 kPa cyclic stress

I, 8 12 16 20 After cycling both reinforced and

Axial stra1n ( '/,)
unreinforced specimens deillonstrate
improvement in the stress-strain b2haviour.
Volumetric strain versus axial strain
;:: 1.2 curves indicate that specimen becomes dense
after cycling and the shift is observed
:: towards expansion side.
~
Vi
Separation
·~
a:; Visual inspection of unreinforced composite
E specimen revealed that migration has
::J
0 occurred from both sides. Silt intruded into
> sand and sand has penetrated into silt. In
reinforced specimen no migration was visible
Fig.12 stress-strain, volume change curves and there was no evidence of deterioration
after undrained cyclic loading of the fabric even after 1000 cycles.
(without reinforcement) .

1919
 Friedli, P. and Anderson, D.G. (1982),
SUMMARY AND CONCLUSIONS "Behaviour of Woven Fabrics under
Simulated Railway Loading". Proc. 2nd
To use geotextiles at the subgradejballast International Conference on Geotextiles,
or subgradejsubbase interface, it is Las Vegas, Vol.2, pp.473-478.
essential to understand the behaviour under
cyclic loading. From the results reported it Saxena, S.K. and Chiu, D. (1982),
is clear that the influence of disc "Evaluation of Fabric Performance in
reinforcement on medium dense fine grained Rail-Road System", Proc. 2nd Int. Conf.
sand is to increase the strength, reduce on Geotextiles, Las Vegas, Vol.2, pp.485-
plastic deformation and increase in general 490.
the modulus of resilience. The limited tests
on simulated behaviour also demonstrated no Saxena, S.K. and Wang, S. (1982), "Mot1.el
intermixing and general improvement in Test of a Rail-Ballast-Fabric-Soil
engineering behaviour. With reinforcement, System", Proc. 2nd Int. Conf. on
a higher cyclic stress level is possible to Geotextiles, Las Vegas, Vol.2, pp.495-
be applied even for poor subgrade 500.
conditions. Thus geotextiles have great
potential for use in pavements and rail Venkatappa Rao, G., Gupta, K.K. and
roads in India. Kachahuwah, R. ( 1987), "Triaxial
Behaviour of Geotextile Reinforced Sand",
Proc. Indian Geotechnical Conference-
REFERENCES 1987, Bangalore, Vol.1, pp.323-328.
Azeem, A., (1988), "Geotextiles for Venktatppa Rao, G., Kate, J.M. and Shamsher,
Improvement of Railway Subgrades- F. (1989), "Drained Triaxial Behaviour of
Potential Evaluation" 3rd National Geotextile Reinforced Sand", Proc.
Convention of Textile Engg. Div. and International Workshop on Geotextiles,
National Seminar on Non-woven and Bangalore, India, Vol.l, pp.55-62.
Geotextiles, Baroda, India, pp.238-250.
Yog, A.K., Balkrishna, and Azeem, A. (1989),
Broms, B.B., (1977), "Triaxial Tests with "Geotextile in Rail Track Foundation".
Fabric Reinforced Soil", C.R.Coll Inst. Proc. Int. Horkshop on Geotextile,
Soils Text. (Paris), Vol.3, pp.129-133. Bangalore, India, Vol.l, pp.97-l02.

1920