Pullout Behavior of Geotextiles: Numerical Prediction

Sieira, A. C. C. F. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 11, ( Part -4) November 2016, pp.15-18
www.ijera.com 15 | P a g e
Pullout Behavior of Geotextiles: Numerical Prediction
Sieira, A. C. C. F.*
*(Structures and Foundations Department, Rio de Janeiro State University, Brasil)
ABSTRACT
The mechanism of soil-geosynthetic interaction is usually complex and depends on the nature of the
reinforcement, as well as on the characteristics of the surrounding soil. The strength parameters of the interface
are the key for the design of reinforced soil slopes. Usually, these parameters are defined from laboratory
pullout tests. The absence of test results implies on conservative assumptions and higher costs. The possibility
of using computer programs for analyzing the load transfer mechanism arises as an attractive design tool. This
paper presents the numerical simulation of pullout tests, conducted in large equipments. The numerical
predictions of the load and displacement distribution along the geosynthetic length were compared to
instrumented test results, available in the literature. The analyses revealed to be satisfactory and consistent with
the experimental results. Thus, it becomes possible to reduce the uncertainties in the design of the anchorage
length for the reinforcement by previously performing studies with computer programs that simulates stress x
strain behavior of geotechnical problems.
Keywords - Numerical prediction, Geotextiles, Pullout tests
I. INTRODUCTION
The design of reinforced soil structures
requires knowledge of the resistance parameters at
the soil-reinforcement interface, usually obtained
from field or laboratory pullout tests.
Field pullout tests reproduce more
adequately the actual conditions. However, these
tests are costly and require a complex infrastructure.
Laboratory tests, in turn, are easier to interpret since
both envi-ronmental and boundary conditions are
fully con-trolled. Nevertheless, they usually require
large equipments to overcome scale effects.
Numerical or analytical simulation appears
as an attractive approach that combines low cost and
speed, allowing the evaluation of different soil-
geosynthetics and boundary conditions. Once the
model is validated, parameters may be assigned for
geotechnical design purposes. Moreover, the simula-
tions are an efficient tool for understanding the
stress-strain behavior of geosynthetics.
Many analytical [1-3] and numerical [4-5]
studies have been conducted in order to better
understand the stress-strain behavior of
geosynthetics.
This paper presents the results of numerical
simulations of pullout tests carried out in the
laboratory in large equipment [6]. The experimental
program consisted of a series of pullout tests with 2
types of geotextiles, embedded in sandy soil. The
analyses were accomplished with a commercially
available finite element program (PLAXIS).
II. LABORATORY TESTS
The pullout tests were carried out at
CEDEX Geotechnical Laboratory, in Spain. Two
types of geotextiles were used and will be identified
in the present paper as G1 e G2. Both are nowoven
needle-punched geotextiles and their main physical
and me-chanical properties are presented in Table 1.
The tests were carried out with silty sand
with 10% moisture content and 80% relative density.
Characterization tests indicated specific gravity of
solids (G) equal to 2.71.
The shear strength parameters were
obtained from drained triaxial tests carried out with
large di-mension samples (22.9 cm in diameter),
under different confining stress levels. The tests
provided friction angle (’) and cohesion (c’) equal
to 37o
and 16 kPa, respectively.
Pullout tests were carried out in large
dimension equipment (1.0 m x 1.0 m x 0.6 m), under
different confining stress levels equal to 12.5; 25 and
50kPa. The geotextile specimens were instrumented
with strain gages, installed at 5 different points along
the geotextile length. Test set up is shown Figure 1.
Table 1. Geotextile properties [6]
Property Geotextile
G1
Geotextile
G2
Longitudinal
resistance (kN/m)
50 200
Transverse resistance
(kN/m)
14 14
Elongation at failure
(%)
12 13
Thickness (mm) 2.3 2.9
RESEARCH ARTICLE OPEN ACCESS

Legend: SG = strain-gage
Figure 1. Pullout test instrumentation [6]
III. NUMERICAL MODEL
The geometry, boundary conditions, and
load systems, adopted in the numerical analyses, are
summarized in Figure 2. The test box is 0.60 m high
and 1.0 m long. The box walls were represented by
rigid plates. The geotextile length was equal to
0.90m. The confining stress was simulated by an
uniformly distributed load (A), acting on the soil
surface. The horizontal load was simulated by a
point load (B) at the front end of the geotextile
specimen. Horizontal movements of the vertical
walls were inhibited and no displacement was
allowed at the base of the test box. A 6mm vertical
spacing between the two halves of the front wall was
imposed to allow free move-ment of geotextile.
The geotextile was simulated by an element
presenting tensile resistance only. The axial stiffness
(EA) of geotextiles G1 and G2 were defined from
tensile tests results. Steel properties were used to
compute the stiffness of box walls, consider-ing a
wall thickness equal to 2.0 mm.
Figure 2. Numerical model [5]
Soil-geotextile behavior was simulated by
pre-scribing an interface element above and below
the interface contact. This element minimizes the
shear strength parameters of the soil adjacent to the
rein-forcement, by applying a correction factor
(Rinter), according to the following equations:
0,1int 
c
c
R a
er
(1)
0,1
tan
tan
int 


erR (2)
In the present study, interface elements
were pre-scribed not only at the soil-geotextile
contacts but al-so at the front wall contact to
simulate internal lubrication. Correction factors
(Rinter) were defined according to experimental data
provided by Espinoza [6]. At the contact between
soil sample and front wall, Rinter was defined
according to reference [7] indicating a light
lubrication.
Hardening Soil constitutive model was
selected to reproduce the stress-strain behavior of
the compacted sandy soil sample. This model allows
the variation of the Young modulus with stress level.
Table 2 summarizes all parameters and constitutive
models adopted in the numerical analyses.
IV. RESULTS DISCUSSION
Figure 3 compares the experimental results
with numerical prediction of G1 pullout test, under
25 kPa of confinement. The results (Figure 3a) show
a good agreement at the rear end of the geotextile.
According to reference [6], the strain-gage located
200 mm from the front edge of the box was damaged
during the test and the SG3 (400 mm from the front
end) presented malfunction.
Table 1. Constitutive models and material
parameters
Material Constitutive model Parameters
Silty Sand
soil
Hardening Soil
c’ = 16 kPa



ref
E50 


Geotextile Linear elastic


Plates Linear elastic



Soil-Geotextile G1 interface 
Soil-Geotextile G2 interface 
Soil-Frontal wall interface 
Note: c’ = cohesion; ’ = friction angle;  = specific
weight; E50
ref =
deformability modulus corresponding
to 50% failure; m = HS model exponent parameter; 
= Poisson coefficient; EA = axial stiffness; e =
thickness.

(a) Predicted and measured strains
(b) Predicted and computed horizontal loads
Figure 3. Geotextile G1 pullout response at failure,
confining stress = 25 kPa
Based on the load and strain curve provided
by the tensile tests, the axial loads were predicted
and compared to the experimental data. The results,
shown in Figure 3b, followed a similar trend ob-
served in Figure 3a.
The results of the geotextile G2 test, under
50kPa of confinement, are presented in Figure 4.
The predicted deformation is slightly lower
(maximum difference of 1%) than the experimental
data (Figure 4a). On the other hand, the axial loads
showed a better agreement, particularly at the rear
end of the reinforcement (Figure 4b).
(a) Predicted and measured strains
(b) Predicted and computed horizontal loads
Figure 4. Geotextile G2 pullout response at failure,
confining stress = 50 kPa
It is worthwhile to mention that the
geotextile is simulated by a linear and elastic
element. Figure 5 compares the curve obtained in
tensile tests with GA and the curve adopted by
Plaxis. It is expected that for higher strain levels the
program provides higher values of tensile load. The
same behavior is expected in the simulations with
the geotextile G2. Additionally, the linear and elastic
constitutive mod-el does not incorporate an ultimate
load.
0
10
20
30
40
50
60
70
80
0 2 4 6 8 10 12 14
Strain (%)
Tensileload(N/mm)
Plaxis - linear elastic model
Manufacturing
Figure 5. Difference among the constitutive models
Geotextiles G1 and G2 show a significant
difference of axial stiffness, which definitely
influences the comparison of the pullout responses.
Figure 6 presents experimental and numerical
results, at failure, for the specimens under 50 kPa of
confinement. It is interesting to note that
experimental data with geotextile G1 shows higher
strains at the front end of the reinforcement and a
sharp reduction along the geotextile length; at the
rear end the strain is quite negligible. This behavior
is typical of low stiffness geotextiles. Due to the
linearity of F vs. curve and, hence, the constant
relationship between F and  the numerical results

yield a more uniform strain reduction along the
geotextile length.
Accordingly, both numerical and
experimental results with geotextile G 2 present
lower strains due to its higher stiffness. Moreover, it
is observed a good agreement between both
approaches.
Figure 7 compares the pullout loads
computed at the strain gage (SG2), located 20 cm
from the front wall with the tensile resistance
provided by the manufacturer. The geotextile G2,
which is stiffer and more resistant, is less mobilized
than geotextile G2. In engineering practice, the
design with geotextile G2 would be conservative
and, hence, economically inappropriate.
Figure 6. Predicted and measured strains at failure,
under 50 kPa of confinement
Figure 7. Geotextiles G1 and G2: Pullout loads
under 50 kPa of confinement
V. CONCLUSION
This paper presented the results of
numerical simulations of pullout tests with
geotextiles. The numerical analysis was performed
with PLAXIS software. The analyses were
compared to experimental results of instrumented
pullout tests, available in the literature.
The numerical results are in agreement with
the experimental tests response and, therefore, imply
in considering the numerical analysis as a powerful
tool to simulate pullout tests on geotextiles. Besides
they may be used as a design tool to help engineers
to reduce the uncertainties in defining the anchorage
length of geotextiles.
The geotextile G2, which is stiffer and
more resistant, is less mobilized than geotextile G1.
In engineering practice, the design with geotextile
G2 would be conservative and, hence, economically
inappropriate.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the
financial support for this work provided by the
Brazilian Science Foundation’s CAPES, CNPq and
FAPERJ.
REFERENCES
[1] Beech, J. F., (1987). Importance of stress-
strain relationships in reinforced soil
system designs. In: Geosynthetics 87
Conference. Vol.1, pp. 133-144.
[2] Ochiai, H.; Hayaschic, S.; Ogisako, E.;
Sakau, A. (1988). Analysis of polymer grid:
reinforced soil retaining wall. International
Conference on Numerical Methods in
Geomechanics, 6. Innsbruck. Balkema,
Rotterdam. Vol. 2, pp. 1449-1454.
[3] Sieira, A.C.C.F, Gerscovich, D.M.S and
Sayão, A.S.F.J. (2009). Displacement and
Load Transfer Mechanisms of Geogrids
under Pullout Condition. Geotextiles and
Geomembranes. 27, pp. 241-253.
[4] Sobhi, S. and Wu, J.T.H. 1996. An
interface pullout formula for extensible
sheet reinforcement. Geosynthetics
International, Vol. 3, n. 5, pp. 565-582.
[5] Ferreira, L. H. T. 2009. Analytical and
numerical models for simulation of pullout
tests with geotextiles. (in Portuguese). MSc
thesis. State University of Rio de Janeiro.
134p.
[6] Espinoza, M.E.D. (2000). Study of the
stress-strain behavior of geosynthetics
under pullout condition with respect to the
design of reinforced soil structures. PhD
thesis. Polytechnic University of Madrid,
350p.
[7] Palmeira, E.M. and Dias, A.C 2007.
Experimental and numerical study of the
behavior of geogrids on large pullout tests
(in Portuguese). REGEO. V. 1, pp. 1-8.

Pullout Behavior of Geotextiles: Numerical Prediction

More Related Content

What's hot (16)

Viewers also liked (20)

Similar to Pullout Behavior of Geotextiles: Numerical Prediction (20)

Recently uploaded (20)

Pullout Behavior of Geotextiles: Numerical Prediction