HOSTED BY Available online at www.sciencedirect.

com

ScienceDirect
Soils and Foundations xxx (2018) xxx–xxx
www.elsevier.com/locate/sandf

Investigating the effect of geogrid on stabilization of high

railway embankments
Morteza Esmaeili ⇑, Behnood Naderi, Hossein Kalantar Neyestanaki, Alireza Khodaverdian
School of Railway Engineering, Iran University of Science and Technology, Tehran, Iran

Received 18 February 2017; received in revised form 20 October 2017; accepted 2 December 2017

Abstract

This paper investigates the effect of geogrid on controlling the stability and settlement of high railway embankments using laboratory
testing and finite element modeling. To do this, five series of embankments with 50 cm height were constructed, at a scale of 1:20 and then
were uniformly loaded on the crest in a loading chamber in dimensions of 240
235
220 cm. In this regard, the embankments of the
first series were constructed without geogrid reinforcing layers. Following to preliminary numerical simulations for determining the
appropriate level of geogrid layers installation, the second to fifth series of embankments were constructed. These embankments were
reinforced with one to four layers of geogrid respectively and finally, the results of their load in terms of settlements were compared.
In these studies, the reinforced embankments with a single geogrid layer had 7.14% raise in bearing capacity and 11.24% reduction in
settlement respectively, in comparison with the unreinforced embankment. The obtained results for the third to fifth series of embank-
ments were respectively in order of (19, 36.14), (26.3, 52.8) and (28.9, 53.42)%. In the next stage, by modeling the embankments in the
PLAXIS 2D software, the results were validated by the values obtained through laboratory models. In continuation of the study, real
embankments with heights of 5, 10, 15, and 20 m were simulated and placed under LM71 loading pattern (Eurocode, 2003). In this
respect, the impact of important effective parameters such as number of geogrid layer, soil characteristics, embankment dimensions,
interface coefficient between soil and geogrid and tensile strength of geogrid on bearing capacity and settlement have been studied.
The numerical results like the experimental ones, confirmed the increase in bearing capacity and settlement diminishing with definite
increase in the geogrid layers, so that more geogrid layers do not affect these parameters. With respect to improving the soil character-
istics and reducing the height of embankments, the FEM models showed decreasing effect of geogrid tensile strength on embankment
crest settlement. On the other side, the value of geogrid-soil interface coefficient has minor effect on both settlement and sliding safety
factor.
Ó 2018 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society.

Keywords: Settlement reduction; Embankment stability control; High railway embankments; Geogrid layer; Laboratory model; Finite element simulation

1. Introduction embarked on shortening the routes length and travel time.

This has made them to pass the railways through areas
Concurrent with the development of railway transporta- with many difficulties, and consequently, construction of
tion network during the recent years, the designers have technical infrastructures such as high embankments or
bridges has become inevitable. From a geotechnical point
of view, controlling the stability and settlement of high
Peer review under responsibility of The Japanese Geotechnical Society. embankments under operational loads or in special loading
⇑ Corresponding author.
E-mail addresses: m_esmaeili@iust.ac.ir (M. Esmaeili), behnoodnx@
conditions, e.g. earthquake, is always a challenging issue
gmail.com (B. Naderi), hk.neyestanaki@gmail.com (H.K. Neyestanaki), for railway tracks designers. A review on the technical lit-
ali.khodaverdian@aut.ac.ir (A. Khodaverdian). erature suggests various methods such as application of

https://doi.org/10.1016/j.sandf.2018.02.005
0038-0806/Ó 2018 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
2 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

Nomenclature

u distance of topmost layer from the loading bed CC curvature coefficient (dimensionless)
(m) q uniform loading on the embankment crest
4
Qvk

B load width (m) ¼ ð3aþ2b kN
Þ
B m2
h distance between geogrid layers (m) Qvk concentrated load of train (kN)
b width of geogrid layers (m) a&b geometrical parameter equal to 1.6 and 0.8
N number of geogrid layers respectively (m)
D loading distance from edge of slope (m) a Impact factor = a ¼ 1 þ 5:21 VR (dimensionless)
b slope angle from the horizon (°) V train speed (km/h)
CU uniformity coefficient of Soil (dimensionless) R diameter of train wheel (mm)

pile, micropile, deep soil mixing, berm construction on the the development of the numerical model of the embank-
sides of the embankment, injection, tieback installation, ments in PLAXIS 2D software, the results were then vali-
etc., for stabilization of the embankments, enhancement dated by the laboratory results. In the next stage, using
of bearing capacity and reduction of settlement. Some of validated numerical models, the behavior of 5, 10, 15,
these methods have been mentioned in the UIC Code and 20 m reinforced embankments by the geogrid under
719R (1994) under ‘‘Earthworks and Track-bed Layers actual railway load was studied. In the final stage, a series
for Railway Lines”. Among the research papers on the of sensitivity analyses on the effective parameters such as
mechanical stabilization of railway embankments, the the number of geogrid layer, soil characteristics, embank-
experimental and numerical study of micropiles to rein- ment dimensions, interface coefficient between soil and
force high railway embankments by Esmaeili et al. (2013) geogrid and tensile strength of geogrid were performed
and railway embankments stabilization by tied back-to- and their effect on the results of numerical models were
back system by Esmaeili and Arbabi (2015) are of particu- investigated.
lar merit.
During the recent years, with the development of
2. Literature survey
geosynthetics, they have been abundantly used in road
and railway projects. The majority of the projects deal with
In all research work, which have been carried out in the
the placement of reinforcing geosynthetics layers in super-
field of slope stabilization of slope with geogrid, various
structure layers of roads and railways. In this matter, and
factors play important role. Amongst, the normalized dis-
particularly regarding the reinforcement of railway sub-
tance (u/B) of topmost layer from the loading bed, normal-
structures, the published technical reports by Coleman
ized distance (h/B) between the layers, normalized width
(1990), Webster (1991), Helstrom et al. (2007), Penman
(b/B) of the layers and number (N) of the geogrid layers,
and Priest (2009), Lee et al. (2012), and Parsons et al.
normalized loading distance (D/B) from the edge of the
(2012) are noteworthy.
slope as well as the slope angle (b) from the horizon can
The main objective of this research is recognizing the
be pointed out. It should be noted that in the above cases,
performance mechanism of reinforcing layers of geogrid
B is the width of the loading area. The mentioned param-
materials in high railway embankments. In this regard,
eters are depicted in Fig. 1.
the focus of study is sliding control in embankment body
Yoo (2001) allocated his research to the investigation of
and decreasing the crest settlement as main effecting factors
bearing capacity of a strip foundation on a geogrid-
in railway embankment serviceability. According to the
reinforced slope. In his research an extensive range of con-
requirements of the UIC Code 719R, poor graded sand
material (QS2) for subgrade and well graded sand (QS3)
for embankment were adopted in five series of embank-
ments, each of which had 50 cm height and were in scale
of 1:20 with side slopes of 1:1 in loading chamber of 240

235
220 cm in conjunction with 60 cm of substructure,
all uniformly loaded on crest. All reinforced laboratory
models were made based on the results of the preliminary
numerical modeling to gain optimal level for placing the
geogrid layers. To that end, first series of embankments
were constructed without geogrid reinforcement and the
series second to fifth were reinforced with one to four geo-
grid layers each. Having finished the laboratory tests, with
Fig. 1. One way slope reinforced with geogrid layers.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 3

ditions, including unreinforced models, was evaluated by struct slopes with better stability. Second, for increasing
varying the parameters such as geogrid length, number of the running speed or the axle load of vehicles on the exist-
geogrid layers, vertical spacing and depth to topmost layer ing railway tracks, the embankment reinforcement will be
of geogrid. Subsequently many Finite Element Models essential and hence, the geogrid layers can be used in
(FEM) were developed to assess prototype slopes as well embankments for controlling the stability and reducing
as the constructed models in laboratory. Results revealed the settlement.
that failure wedge tends to become wider and deeper than As it seen, all of the referred works have been deal with
that for the unreinforced slope. one way slopes stabilization which their stability under the
In another study El Sawwaf (2007) studied the potential vertical surcharge is the main goal of the mentioned
assistances of reinforcing a replaced layer of sand con- researches while in the case of railway embankment as
structed adjacent an earth slope crest. Several parameters two way slope both stability and settlement control have
including the depth of replaced sand layer and the location major role in railway serviceability. In the frame work of
of foundation relative to the slope crest were studied. Par- practical projects many works can be referred which the
ticular emphasis was paid on the reinforcement configura- application of geogrid for railway embankment stabiliza-
tions including number of layers, spacing, layer length and tion have been reported but no definite research can be
depth to ground surface. Subsequently, many finite element found in the literature with the mentioned focus. There-
models were developed in order to investigate the behavior fore, focusing on embankment material properties, a series
of slopes in prototype scale. According to results, the opti- of laboratory tests and the numerical simulations in this
mum amount for (u/b) and (h/b) parameters were deter- research were carried to show the efficiency of geogrid lay-
mined 0.6 and 0.5 respectively. ers in enhancing the bearing capacity and limiting the crest
Alamshahi and Hataf (2009) assessed the effect of a new settlement of railway embankment.
type of geogrid inclusion on the bearing capacity of a rigid
strip footing constructed on a sand slope. An extensive ser- 3. Laboratory tests
ies of boundary conditions, including unreinforced cases,
were tested by varying parameters such as geogrid type, 3.1. Scaling law
number of geogrid layers, vertical spacing and depth to
topmost layer of geogrid. In the following, many finite ele- In this research, a 10 m-height embankment with a crest
ment models were developed in order to investigate behav- width of 4.8 m and sides slope of 1:1.5 was selected as an
ior of embankments in prototype scale. Regarding to the embankment built according to the requirements of Ira-
obtained results, the optimum embedment depth and verti- nian Railway standards (Periodical 288, 2004). Thereafter,
cal spacing of the reinforcement layer, which resulted in the on the basis of the method proposed by Wood (2004), the
maximum bearing capacity of the geogrid-reinforced slope, scaling law was applied on the aforementioned embank-
was about 0.75 times the width of the foundation. In addi- ment with a scale of 1:20 for the laboratory tests program.
tion, the optimum number of geogrid reinforcements was Considering the height of the loading chamber as well as
two. the 10 cm course of hydraulic jack, the embankment was
Choudhary et al. (2010) assessed bearing capacity of a constructed with a maximum height of 50 cm over a 60
strip foundation resting on the top of a geogrid reinforced cm thickness foundation soil (subgrade) at the chamber.
fly ash slope using laboratory model tests. In this research Given the limited space (2.35
2.4 m) inside the available
resemble to previous studies some effective parameters loading chamber to prevent the impact of chamber walls on
including unreinforced cases were conducted by varying the embankment bearing capacity and to avoid probable
parameters such as location and depth of embedment of uplift on the sides of the laboratory model, 50 cm distance
single geogrid layer, number of geogrid layers, location of from the wall on either side was provided. Considering
footing relative to the slope crest, slope angles and width these restrictions and assuming the crest width of 24 cm,
of footing were assessed. the side slope of 1:1 was selected for the laboratory model.
El Sawwaf and Nazir (2012) presented a laboratory It should be mentioned that the selected slope, in compar-
study of the effect of geosynthetic reinforcement on the ison with the normal slope of 1:1.5, imposes more critical
cumulative settlement of repeatedly loaded rectangular sliding stability condition on the embankment under the
model footings placed on geogrid reinforced sand. The vertical load. Furthermore, it better shows the impact of
studied parameters included the initial monotonic load applying geogrid in embankment stabilization. Table 1 pre-
levels, the number of load cycles, and the relative density sents the strength and geometrical parameters of the actual
of sand along with geosynthetic parameters including size embankment and the laboratory model. At the same time,
and number of layers. the laboratory model has been specifically shown in Fig. 2.
With respect to the studies already carried out, using It should be noted, the embankment model adopted in
geogrid in railway embankments can be justified in two laboratory study is reduced to a certain scale while the soil
ways. First, during the construction of railway embank- and geogrids were the same in the model and the prototype
ments, in the case where the borrow area with good quality analysis. Therefore, model the soil, may not play the same
materials is scarce, geogrids can be used in order to con- role as in the prototype and it might lead to some influence

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
4 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

Table 1 as 16.9 kN/m3 according to the ASTM D4253. In this

Geometrical parameters and dimensions of the actual embankment and study, for compacting the 5 cm layers of subgrade soil, a
the laboratory model.
12 kg laboratory roller passed 9 times for achieving the
Parameter Real embankment/ Laboratory 70% compaction ratio. The compaction ratio was con-
Unit (SI) model
trolled by doing the in place density measurement using
Embankment length 48.00 m 2.40 sand cone according to the ASTM D1556-82. On the other
Embankment height 10.00 m 0.50
Slope length 18.00 m 0.70
hand, in order to attain the shear strength parameters of
Embankment crest width 4.80 m 0.24 the subgrade soil, many direct shear tests in drained condi-
Subgrade Depth 12.00 m 0.60 tion were performed on remolded samples according to the
Depth of modified subgrade 2.00 m 0.10 ASTM D3080. In this regard, the subgrade soil was poured
Width of the bed sides 12.00 m 0.56 in direct shear test sample box and it was compacted to
Tensile strength of geogrid 140 kN/m 7
reach the 70% of the maximum laboratory dry density.
Consequently, the desired direct shear tests were carried
out. Since, all reported soil mechanical parameters have
been obtained from low loading rate direct shear tests, no
definite control have been possible to be made on pore
water pressure. Therefore, only the low loading rate of
the sample grantee the achievement of drained parameters
of the soils. As all samples have been in semidried condi-
tion so it can be claimed that the obtained shear strength
parameters are allocated to drained condition. Moreover,
the plate loading test (PLT) has also been performed in
loading chamber according to the ASTM D1194-94 for
determining the soil elasticity modulus. Table 2 shows the
summary of subgrade soil mechanical and physical properties.

3.3. Properties of embankment material

Fig. 2. Embankment with 10 m height in 1:20 scale at the loading
chamber (all dimensions in m). According to the ASTM D422 standard, the well graded
sand (SW) with uniformity coefficient of 8.33 and curvature
on the experimental results, which is named scale effects coefficient of 1.00 was adopted as the embankment mate-
(Vesic, 1973). These differences happen predominantly rial. This soil in known as QS3 based on the UIC Code
because of the differences in level of stress between the 719R definitions. Using the standard proctor test according
reduced scale model and the field tests. On the word, it to the ASTM D698, the maximum dry density of this soil
should be emphasized that in present study the self- has been obtained equal to 2.05 g/cm2 in the optimum
weight stress level has not been scaled in present work. moisture content of 11.1%. In order to make the laboratory
Similar to other 1 g experiments, the focus of current model, the mentioned soil was compacted with optimal
research is understating the failure mechanism of embank- moisture in 5 cm layers by using a 36 kg laboratory roller.
ment including the geogrid layers and so the stress values Overall, by passing the roller for 15 times over each layer,
are not so reliable. Moreover, it was preferred to use a pro- the maximum compaction ratio of 84% has achieved
totype geogrid of lowest allowable stress (standard one) according to the conducted in place density measurement
instead of using some kind of hand-made geotextile in compliance with the ASTM D1556 proposed method.
(anonymous one) for resemblance. Consequently, due to Similar to the subgrade soil, many direct shear tests
scale effects and the use of prototype geogrid in the model (ASTM D3080) have been performed on remolded samples
tests, the results of embankment model may be influenced. of the embankment soil to assess its shear strength param-
eters in drained condition. In this regard, the embankment
3.2. Properties of subgrade material soil was poured in the same condition of construction
phase in a metal box and consequently many undisturbed
The standard sand 06 (sand with maximum grain size of samples were taken for direct shear tests. Since, all reported
6 mm) poorly graded (SP), with uniformity coefficient (CU) soil mechanical parameters have been obtained from low
of 1.45 and curvature coefficient (CC) of 1.08 according to loading rate direct shear tests, no definite control have been
the ASTM D422 has been used as the subgrade material. possible to be made on pore water pressure. Therefore,
This soil is known as QS2 classification according to the only the low loading rate of the sample grantee the achieve-
UIC Code 719R. In the laboratory, the minimum specific ment of drained parameters of the soils. As all samples
weight of this sand (in its loosest state) was attained as have been in semidried condition so it can be claimed that
14.3 kN/m3 according to the ASTM D4254 and its maxi- the obtained shear strength parameters are allocated to
mum specific weight (in its densest state) was calculated drained condition. In addition, the PLT (ASTM D1194-

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 5

Table 2
Embankment and subgrade soils properties respect to the maximum unit weight.
Soil type Friction Cohesion Elasticity modulus Wet soil unit Lab compaction
angle u (°) C (kN/m2) (%) (kN/m2) weight ct (kN/m3) ratio (%)
Embankment (SW) 36 10 6000 17.2 84
Subgrade (SP) 38 1.8 14,900 15.9 70
SW: well-graded sand, SP: poorly-graded sand.

Table 3 The analyses on the embankment stability have been con-

Technical properties of bi-oriented geogrids used in reinforcement of
embankment.
ducted to find the failure load in the optimal state of the u,
h and N parameters. As shown in Fig. 3.a, the maximum
Specifications Standard Type (value)
load capacity of the reinforced soil with one layer of geogrid
Mesh type Rectangular apertures in terms of different u/B values has been studied. This has
Standard color Black
Polymer type Polypropylene
been also conducted in the cases of embankments with 2,
Aperture size mesh 41 mm 3, and 4 layers of geogrid by varying the u/B and h/B for
Mass per unit area Iso 9864 250 g/m2 the first layer location and varying the other geogrid layers
Strength at 2% strain Iso 10,319 7 kN/m levels respect to the first layer. Finally, the optimal level of
Strength at 5% strain Iso 10,319 14 kN/m all geogrid layers has been obtained. With respect to the
Peak tensile strength Iso 10,319 20 kN/m
Fig. 3.a to 3.d, it is observed that the placement of layers
at u/B = 0.3 and h/B = 0.4 has led to the best performance
94) has been conducted on the embankment soil in the of the geogrid layers in the embankment. The trend of place-
chamber for measuring the soil elasticity modulus. A sum- ment of geogrid layers in the embankment was stopped at N
mary of the tests results is presented in Table 2. = 4. This increasing trend has continued to the point that the
impact of number of layers on the bearing capacity increase
3.4. Properties of geogrid and the settlement decrease proved to be minor.
The construction process of laboratory models in the
In construction of the laboratory models, bi-oriented loading chamber has been performed based on the
geogrid (Table 3) has been utilized. This type of geogrid obtained results from the numerical analyses. The trans-
was selected based on the works carried out on reinforced parent front wall of the chamber facilitates the evaluation
slopes by various researchers (http://www.tenax.net/ of the failure condition in the loading process. During
pdf_geo_e/TENAX_LBO_220_SAMP_e.pdf). Moreover, the construction of laboratory models, the soil has been
the geometry and the dimensions of laboratory models as compacted in 5 cm layers until the geogrid layer optimum
well as the load and the quality of applied soil were impor- location reached. Thereafter, the earthwork was stopped
tant factors in this matter. and the geogrid layer was positioned at the desired loca-
tion. Afterwards, the geogrid layer was covered with
3.5. Experimental tests design and procedure another layer of compacted soil. To make sure of proper
access to the anticipated density in the embankment con-
In order to construct the embankments in laboratory struction, the in situ density test was conducted according
scale, the loading chamber located in advanced substructure to the ASTM D1556 standard. Although, the moisture loss
laboratory, School of Railway Engineering, Iran University after embankment models preparation before loading tests
of Science and Technology was utilized. In this study, the may be occurred, but it should be stated that this condition
embankments were named in the following abbreviated for is the same for all tested embankment models so in a com-
ease in presenting the results: EM0 embankment denotes parative manner the effect of moisture loss of soil on the
an embankment without geogrid layer and EM1G to outer side of the model can be neglected.
EM4G embankments infer the embankments with one to
four geogrid layers respectively. Before construction of lab- 3.6. Model instrumentation
oratory models, the numerical models of the embankments
were created using the PLAXIS V.8.2 software to determine For monitoring the settlement of embankment crest, a
the optimal level of the geogrid layers to the crest of the linear variable displacement transducer (LVDT) with the
embankment (u), distance between the layers (h), and the capacity of 100 mm and two another ones with the capacity
number of layers (N). It should be mentioned here in prelim- of 25 mm were utilized for controlling the uplift of
inary numerical studies, the B parameter is whole width of embankment sides, as illustrated in Fig. 2. These instru-
embankment crest, which will be loaded during the experi- ments were connected to the TMR-211 data logger device,
mental tests in the next section. In these numerical analyses, manufactured by the Tokyo Sokki Kenkyujo company
the soils material parameters and the geogrid specifications (http://www.tml.jp/e/product/transducers/catalog_pdf/
have been selected according to Tables 2 and 3, respectively. CDP.pdf). It should be noted that for reading the load on

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
6 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

Fig. 3. Results of preliminary numerical analyses regarding the optimal level for positioning the geogrid layers.

the hydraulic jack, an analog gauge was attached to the EM2G embankment endured the vertical load of 10 tons
jack itself. or the vertical stress of 181.1 kN/m2 before the failure. In
addition, the recorded crest settlement under this load was
3.7. Loading the laboratory models about 21.8 mm. For the embankments EM3G and EM4G,
the failure loads were respectively 10.6 tons and 10.8 tons
Upon construction of the laboratory models, the load- corresponding to the vertical stresses of 192 kN/m2 and
ing was applied by a 30-ton hydraulic jack in a step by step 196 kN/m2. Similarly, the crest settlement under these loads
manner. The jack was positioned between the upper sup- were respectively 16.1 mm and 15.9 mm. Fig. 5 depicts the
port of loading chamber and the reinforced beam IPE220 load-settlement diagrams of each embankment. Moreover,
placed on the embankment crest for exerting uniform load. Table. 4 demonstrates the percentage of increase in bearing
The force thus imposed on the model in each step was capacity and settlement decrease upon adding geogrid layers
equal to 2.5 kN. The exerted force has been recorded by to the lab models. Noting to Table 4, by adding the geogrid
the gauge during the jack loading and the results of the layers from one to four layers, the load capacity has
data logger device in the form of load-settlement diagrams increased by 7, 19, 26, and 29% respectively and the corre-
have been drawn for the laboratory tests. It should be sponding crest settlement has reduced by 11, 36, 53, and 53.5%.
noted that loading on the model were continued until the Focusing on overall slope stability equation of the
failure occurrence. In this respect, the failure appeared nor- model in limit equilibrium condition and defining the geo-
mally with the reduction in load against the increase in set- grid effect on total elasticity modulus of reinforced soil in
tlement of the embankment crest. one side and in the other side by consideration of scaling
relations between the laboratory model and prototype ones
4. Results of laboratory tests according to Table 1, it can be expected that both of failure
load and settlement of prototype model are at least n times
4.1. Load-settlement results of the relevant values in laboratory model. This finding is
Table 4
With regard to the laboratory tests, EM0 embankment
Results of failure load and settlement in laboratory models.
failed under the vertical load of 8.4 tons corresponding to
Embankment Failure load (kN/m2) Settlement (mm)
the uniform vertical stress of 152.17 kN/m2. The embank-
ment settlement at the failure load was equal to 34.16 mm. Failure Increase Settlement at Percent
load (%) failure load decrease (%)
In the EM1G embankment, the analog gauge showed a
load of 9 tons corresponding to the value of 163.04 kN/m2 EM 0 152.17 0 34.14 0
EM 1G 163.04 7.14 30.3 11.24
as the failure stress for the embankment. Furthermore, the
EM 2G 181.1 19 21.8 36.14
crest settlement under this load was recorded as 30.3 mm EM 3G 192 26.3 16.1 52.8
by the LVDT. EM 4G 196 28.9 15.9 53.42

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 7

justifiable due to assuming the same material properties 5. Numerical analysis of geogrid reinforced railway
(soil and geogrid) for two models in present study with embankment
ignorable soil cohesion value, considering the linear rela-
tion between elasticity modulus of soil and the applied In this research, the PLAXIS V.8.2 finite element soft-
stress, ignoring the geogrid thickness. ware has been used for numerical analysis of the embank-
ments. PLAXIS has high capability in processing and
4.2. Failure mechanism of laboratory models analyzing the geotechnical plane strain problems associated
with slopes considering the dry and saturated soil condi-
As it was expected, during the data record of installed tions. In this study, with respect to the length of 2.4 m
LVDTs on both embankment sides, the displacement and the width of 0.24 m of the crest as well as the plate
(uplift) of the base on the sides of the embankment was form of geogrid layers, the plane strain conditions for
at zero level and so no uplift was observed. This shows suf- two-dimensional modeling of the problem has been consid-
ficient stability of the foundation soil against the failure ered for numerical model. In the modeling by PLAXIS, the
and the lack of extending the sliding surfaces to the base. ‘‘plastic” phase and the ‘‘stage construction” options have
Fig. 4 displays the sliding pattern on the upper part of been used for determining the stress-strain variation in the
the lab embankments. In EM0 embankment, i.e. the unre- model. Moreover, for determination of the embankment
inforced embankment, the sliding surface covers the one- safety factor, the ‘‘phi-c reduction” method has been uti-
third upper part of the embankment and some fractures lized (Bringkgreve and Vermeer, 1998).
are observed in the lower part of the embankment. Upon
positioning the geogrid layers in the embankment, the slid- 5.1. Geometry, mesh generation, and boundary conditions
ing surface appears in the upper part of the reinforced area
and at the interface between the reinforcing layers and the The geometry and dimensions of the model, as shown in
soil. The geogrid layers prevent the extension of sliding sur- Fig. 2, were considered in the numerical modeling. With
face to be deepen into the soil. In the embankments rein- respect to the conditions of the embankments in the labo-
forced with three and four layers of geogrid, the sliding ratory and for imposing the boundary conditions, the
surface appears only in the upper part of the first layer. ‘‘standard fixities” option was selected in the software,
i.e. the base of the model has been fixed in both vertical
and horizontal directions and the side walls, due to possi-
bility of vertical deformation under the loading are free
in vertical direction and they are fixed in horizontal direc-
tion. Meshing of the model was conducted in fine quality in
the form of 15-node triangular elements with 12 integral
points for soil and 5-node element with 5 strain integral
points for geogrid. Moreover, the geogrids were modeled
using geogrid element with a tensile strength considered
in the software for this purpose. Also fully bonded condi-
tion was considered in order to model interface between
geogrid and soil.

5.2. Material properties and constitute law

During the lab model simulation, the materials proper-

ties for the embankment, modified subgrade (which has
the same embankment material) and subgrade and geogrid
have been taken from Tables 2 and 3. It should be noted
that due to the slow speed of construction and the vertical
loading, the drained condition was adopted for both the
subgrade and embankment soil materials. Furthermore,
with respect to the shear strength tests carried out on the
embankment and subgrade soils and noting to the nature
of embankment stability problem, elasto-plastic material
model of Mohr-Coulomb for both embankment and sub-
grade soils. In addition, with respect to the results of inter-
nal force of geogrid layers during the lab model tests, linear
Fig. 4. Failure mechanism of laboratory embankments: a. EM0. b. elastic material model of Hooke has been assigned to geo-
EM1G. c. EM2G. d. EM3G. e. EM4G. grid material.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
8 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

applicability of the geogrid in enhancing the bearing capac-

ity and reducing the settlement is emphasized. To that end,
the models of embankments in heights of 5, 10, 15 and 20
m with side slopes of 1:1 and 1:1.5 and the crest width of
4.8 m are developed. In the beginning, each embankment
was modeled without reinforcement and its minimum
safety factor of sliding was evaluated under the loading
pattern LM71 (EN 1991-2, 2003). In continuation, the
embankments were reinforced by geogrid using the derived
pattern for placement of geogrid layers in the laboratory
Fig. 5. Comparison of load-settlement diagrams in laboratory models to be under the same load for achievement of the
embankments. minimum safety factor of 1.5. This safety factor has been
referred by various guidelines such as US Army Corp
EM 1110-2-1902 (2003) and UIC 719R (1994). In all men-
5.3. Numerical model validation
tioned embankments, the sensitivity analyses has been car-
ried out on the important embankment design parameters
In this section, the results of load-settlement attained
such as soil characteristics, the geogrid tensile strength
from the numerical models are compared with the labora-
and interface parameter between soil and geogrid in order
tory results in section 4 and the validity of the numerical
to attain the impact of each on the settlement. I should be
models is investigated. To that end, the names of the
noted that in all numerical analyses, the geogrid locations
PLAXIS models are adopted corresponding to those of
in embankment for first layer Bu and other layers Bh have been
laboratory tests. In this regard, PM0 refers to an embank-
ment without geogrid and PM1G to PM4G refer to the considered as 0.3 and 0.4 respectively where B is effective
embankments with one to four geogrid layers, respectively. width of loading area on embankment crest equal to 3.2
According to Fig. 6, which shows the load-settlement considering the sleeper length 2.6, ballast depth 0.5 and
diagrams for both numerical and laboratory models, the stress distribution of 1:1 beneath the sleeper in ballast.
results based on the Mohr-Coulomb constitute law, has
illustrated a good fitting with the laboratory results. Table 5 6.1. Materials properties
depicts the difference in bearing capacity between the
numerical and laboratory models in terms of some definite In the modeling of real railway embankments, the same
settlements. It should be noted that the failure condition in properties of embankment and subgrade materials of the
numerical models has been presumed as stopping the soft- laboratory models have been utilized. In all initial sensitiv-
ware analysis at a specific applied load; while in the labora- ity analyses, the soil properties of laboratory model were
tory models, it has been supposed as a constant load value, utilized. However, in the order to investigate the effects
which causes the incremental settlement of the soil. With of embankment soil characteristics on performance of rail-
respect to these results, the difference in numerical and lab- way embankments, 5 various soil types have been adopted
oratory models of the unreinforced embankments stands at according to Table 6. These parameters practically cover
16%. Upon placement of the geogrid layers in the embank- the range of poor to high strength embankment soils which
ment and reinforcing the numerical and laboratory models, here are named from ST1 to ST5.
the Mohr-Coulomb constitute law has proven to be enough
consistency to the laboratory results. This trend of 6.2. Loading model
decreasing difference is also observed in reinforced models
with two geogrid layers. However, upon adding the third For loading the models, the longitudinal loading pattern
and fourth geogrid layers and increasing the stiffness of LM71 presented in the Euro code standard of EN 1991-2
the upper part of embankment, the amount of this differ- has been utilized. For uniform loading on the embankment
ence increases. Moreover, the slope of load-settlement crest, the following equation is used.
diagram in the numerical model of unreinforced embank-  
4
Qvk kN
ment (Fig. 6a) reduces in comparison with the laboratory q¼ ð1Þ
model of reinforced models with three and four layers ð3a þ 2bÞ
B m2
(Fig. 6g and i).
where Qvk is the concentrated load of 250 kN, according to
Fig. 7, a and b are the geometrical parameters which are
6. Parametric study equal to 1.6 and 0.8 m respectively. Moreover, B denotes
the load width is equal to a railway sleeper length of 2.6
After calibration of numerical models by the results of m. Considering the sleeper length 2.6 and ballast depth
laboratory tests and selecting the Mohr-Coulomb consti- 0.5 m, stress distribution angle of 1:1 beneath of sleeper
tute law for embankment and subgrade soils, the full- in ballast layer a 115 kN/m2 uniform load was applied over
scale modeling of railway embankments for assessing the effective loading width of 3.2 over the embankment crest.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 9

Fig. 6. Comparison of load-settlement diagrams in numerical and laboratory models in the case of unreinforced embankment (a and b) and embankments
with one (c and d), two (e and f), three layers (g and h) and four layers of geogrid reinforcements (I and j).

In railway engineering, the vehicle loads have a dynamic dynamic interaction, the standard loading was multiplied
nature. However, from geotechnical perspective and for by an impact factor. In this research, the following equa-
slope stability control the dynamic load can be substituted tion introduced by AREMA (2006) was used to assess
by quasi-static load through multiplying the static load to the impact factor.
dynamic impact factor. In present study for the quasi-
V
static loading, to include the effects of the wheel-rail a ¼ 1 þ 5:21 ð2Þ
R

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
10 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

Table 5 Table 7
Difference in failure load and the associated displacement between the Number of required geogrid layers to meet the safety factor of 1.5 for
numerical and laboratory models. various loads and embankments.
Embankment Settlement at Failure Difference Embankment Number of required geogrid Number of required
failure stress (mm) stress (kN/m2) (%) geometry layers to meet the safety geogrid layers to
factor of 1.5 (under the achieve no more
EM 0 15.99 133.3 16
PM 0 111 vertical stress of 115 kN/m2) bearing capacity
EM 1G 19.42 154.95 9.6 5 m – slope 1:1.5 1 2
PM 1G 140 5 m – slope 1:1 1 1
EM 2G 20.8 166.93 7.7 10 m – slope 1:1.5 1 2
PM 2G 181 10 m – slope 1:1 2 3
EM 3G 16.89 171.42 9.7 15 m – slope 1:1.5 2 2
PM 3G 190 15 m – slope 1:1 4 6
EM 4G 16.56 170.9 17.8 20 m – slope 1:1.5 2 3
PM 4G 208 20 m – slope 1:1 7 8

Table 6 embankments need reinforcement to meet the safety factor

Soil characteristics have been employed in parametric study. of 1.5 under the vertical load of 115 kN/m2. This is in brief
Soil Dry density Elasticity Poisson’s Friction Cohesion c in Table 7. On the other side, Fig. 8and 9show the maxi-
type cd (kN/m3) modulus E ratio angle u (°) (kN/m2) mum bearing capacity of embankments with heights of 5,
(kN/m2)
10, 15, and 20 m with side slope of 1:1 and 1:1.5 respec-
ST1 17 2.0e4 0.45 25 20 tively required to meet the safety factor of 1.5 for various
ST2 18 4.0e4 0.4 28 23
number of geogrid layers. As evidenced in Fig. 8(a)–(d)
ST3 19 6.0e4 0.35 32 27
ST4 20 8.0e4 0.3 35 30 and 9(a)–(d), the least number of geogrid layers is the min-
ST5 21 1.0e5 0.3 38 33 imum that the reinforced embankment meets the safety fac-
tor of 1.5 under the applied vertical load of 115 kN/m2.
Furthermore, the increase in bearing capacity and decrease
In the above equation, a is impact factor, V is the train in settlement diminish with the increase in the geogrid lay-
speed in terms of km/h, and R is the diameter of wheel in ers, so that more geogrid layers do not affect these
terms of millimeter. In this study, assuming a velocity of parameters.
160 km/h and a wheel diameter of 1 m for train, the uni-
formly distributed load of 115 kN/m2 has been calculated.
6.3.2. Effect of the soil characteristics
In order to assess the effect of soil characteristics on the
6.3. Results and discussion behavior of railway embankments, according to Table 6,
sensitivity analyses have been conducted on 5 various soil
6.3.1. Effect of the number of geogrid layers types. As evidenced in Table 8, the number of required geo-
In this research, the minimum safety factor of 1.5 should grid layers to meet the safety factor of 1.5 in various
be met as a main embankment stability criterion. All embankment heights and side slope of 1:1 have been illus-
numerical analyses have been conducted for various num- trated for various soil types. According to this table, ST5
bers of geogrid layers in the presence of their associated does not require any reinforcements in order to meet safety
failure loads as well as the vertical load of 115 kN/m2 men- factor of 1.5 in all conditions. In the other soil types, the
tioned in regulations. In this section, all analyses have been number of geogrid layers and associated settlement have
conducted on soil parameters which used for constructing been mentioned respectively. As can be seen, by decreasing
embankment in laboratory. the height of embankments and improving the soil charac-
Various analyses on the embankments with different teristics, required geogrid layers and associated settlement
heights and side slope of 1:1 and 1:1.5 suggest that these have similarly decreased. In ST1, by increasing the

Fig. 7. Loading model LM71.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 11

Fig. 8. Diagram of load-settlement of embankments with side slope of 1:1 to achieve minimum safety factor of 1.5 in terms of the number of various
geogrid layers. a. Embankment height of 5 m. b. embankment height of 10 m. c. embankment height of 15 m. d. embankment height of 20 m.

Fig. 9. Diagram of load-settlement of embankments with side slope of 1:1.5 to achieve minimum safety factor of 1.5 in terms of the number of various
geogrid layers. a. Embankment height of 5 m. b. embankment height of 10 m. c. embankment height of 15 m. d. embankment height of 20 m.

embankment height from 5 m to 10, 15 and 20 m, settle- to 10, 15 and 20 m, the number of required geogrid layers
ment have been increased 2.3, 4 and 6 times respectively. for achieving to safety factor of 1.5 have been increase 1.5,
Also due to increasing the embankment height from 5 m 3 and 4.5 times respectively.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
12 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

Table 8 Table 9
Number of required geogrid layers to meet the safety factor of 1.5 for Safety factor and settlement in embankment with height of 5, 10, 15 and
various soil types. 20 m with side slope of 1:1, Soil type 2 and fully bonded interface.
Embankment Soil Number of required geogrid Maximum Embankment Number of required geogrid layers Safety Settlement
height type layers to meet the safety factor settlement height (m) to meet safety factor of 1.5 (under factor (mm)
of 1.5 (under the vertical (mm) the vertical stress of 115 kN/m2)
stress of 115 kN/m2) 5 1 1.547 29.6
5m ST1 2 41.2 10 2 1.496 68.5
ST2 2 29 15 4 1.505 147.7
ST3 Not required – 20 7 1.538 219.4
ST4 Not required –
ST5 Not required –
10 m ST1 3 93.7 out on tensile strength of geogrid layers. As the geogrid
ST2 2 73.5 tensile strength has shown minor effect on the embankment
ST3 1 46.7 bearing capacity, in this section only the settlement results
ST4 Not required –
are presented. Fig. 10 illustrated the embankment crest set-
ST5 Not required –
tlement corresponding to the geogrid tensile strength for
15 m ST1 6 163 variety of soil types and embankment heights. It should
ST2 4 141.2
be noted that in all embankments, the number of geogrid
ST3 3 96.1
ST4 1 78.3 layers have been determined based on required layers to
ST5 Not required – achieve safety factor of 1.5. As can be seen in Fig. 10, by
20 ST1 9 250.7
improving the soil characteristics and decreasing the
ST2 7 210.2 embankment height, the effect of geogrid tensile strength
ST3 5 168.5 on crest settlement has decreased. For instance in embank-
ST4 3 134.3 ment with height of 10 m, use of geogrid with tensile
ST5 Not required – strength of 200 kN instead of 10 kN leads to 1.5 times
decrease in settlement of ST1 while in 20 m embankment
this change has decreased the crest settlement 2 times. It
6.3.3. Effect of geogrid tensile strength can be concluded that the effect of geogrid tensile strength
In the following, considering the variety of soil charac- enhancement is more significant in the case of high
teristics, a definite sensitivity analyses has been carried embankments with poor soil quality.

Fig. 10. The effect of geogrid tensile strength on embankment crest settlement in the case of a. 5 m-height, b. 10 m-height, c. 15 m-height and d. 20 m-
height embankment with side slope of 1:1 under the vertical load of 115 kN/m2.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx 13

Fig. 11. The effect of embankment interface coefficient on embankment. a. Safety factor and b. crest settlement, in the case 5, 10, 15 and 20 m-height
embankment with the side slope of 1:1 and geogrid tensile strength of 80 kN/m under the vertical load of 115 kN/m2.

6.3.4. Effect of soil-geogrid interface embankments and select the appropriate elements to vali-
Herein, the effect of interface coefficient between soil date the results of numerical models by the results of labo-
and geogrid layers on the embankment settlement and ratory models. In continuation, of the numerical results
bearing capacity are investigated. In this regard, a series were extended by modeling high embankments in actual
of sensitivity analyses has been carried out on embank- scales and the sensitivity analyses were performed on
ments with height of 5, 10, 15 and 20 m. In these embank- important parameters in design. The results of this research
ments, 5 various interface coefficients of 0.6, 0.7, 0.8, 0.9 can be summarized as the following:
and 1 (fully bond) have been allocated and consequently
the safety factor and crest settlement of embankments have 1. In the 50 cm-height laboratory embankments, using
been evaluated. It should be noted that in all analyses, the three layers of geogrid will result in the maximum bear-
embankment soil type ST2 and the geogrid tensile strength ing capacity of the embankment and the fourth layer
of 80 kN/m have been assigned. Moreover, in order to pre- will have no effect on embankment bearing capacity.
sent this parameter effect on safety factor and settlement in 2. Reinforcement of 5 m-height embankments with the side
sensible manner, all numerical results have been calculated slope of 1:1 by one geogrid layer at a distance of 96 cm
in a comparative condition respect to the model with inter- from the crest of the embankment, as well as reinforce-
face coefficient of 1. In the first stage, the safety factor and ment of 10 m-height embankment with the side slope of
settlement of embankment with fully bonded geogrid have 1:1 by two geogrid layers at a distance of 2.24 m from
been illustrated in Table 9 and consequently the differences the crest was conducted. This configuration can be used
between fully bonded conditions with other cases have for stabilization of the embankment when needed as an
been depicted in Fig. 11. According to this figure, due to option.
reducing the interface coefficient the amount of embank- 3. With regard to the numerical sensitivity analyses,
ment settlement has been increased while the amount of increasing the number of geogrid layers leads to enhance
embankment safety factor has been reduced. For example safety factor and decrease in settlement respectively. On
in embankment with 10 m height by changing the interface the other hand, the increase in bearing capacity and
coefficient from 0.6 to 1, the settlement has increased 9 mm decrease in settlement diminish with the increase in the
but the safety factor has increased 0.083. From another geogrid layers, so that more geogrid layers do not affect
point of view, the interface coefficient value has minor these parameters.
effect on both settlement and safety factor compare to 4. By improving the soil characteristics and reducing the
other aforementioned parameters involved in embankment height of embankments, the effect of geogrid tensile
performance. strength on crest settlement reduction has diminished.
In embankment with height of 10 m, use of geogrid with
7. Conclusion tensile strength of 200 kN instead of 10 kN leads to 1.5
times decrease in settlement for ST1 while in 20 m
This research began with the construction of five series embankment this parameter has decreased 2 times the
of 50 cm-height embankments with the side slope of 1:1 crest settlement.
and placing the geogrid layers in various levels. The bear- 5. By decreasing the soil-geogrid interface coefficient, crest
ing capacity enhancement and settlement reduction due settlement of embankment has been increased while the
to this reinforcement in comparison with the embankment amount of embankment safety factor has been reduced.
without geogrid layers was then studied. Afterwards, the But in general, the interface coefficient value has minor
PLAXIS finite element software was used to model the effect on both settlement and safety factor.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005
14 M. Esmaeili et al. / Soils and Foundations xxx (2018) xxx–xxx

References Esmaeili, M., Arbabi, B., 2015. Railway embankments stabilization by

tied back-to-back system. Comput. Geotech. 67, 110–120.
Alamshahi, S., Hataf, N., 2009. Bearing capacity of strip footings on sand Esmaeili, M., Gharouni Nik, M., Khayyer, F., 2013. Static and dynamic
slopes reinforced with geogrid and grid anchors. Geotext. Geomembr. analyses of micropiles to reinforce the high railway embankments on
27, 217–226. loose beds. J. Rehabil. Civ. Eng. 1–2, 80–89.
American Railway Engineering and Maintenance of Way Association European Committee for Standardization, 2003. European Standard (EN
Manual for Railway Engineering, 2006. Economics of Railway 1991–2). Ref. no. EN 1991-2:2003 E.
Engineering and Operations-construction and Maintenance Opera- Helstrom, C.L., Humphrey, D.N., Labbe, J.M., 2007. Performance and
tions, vol. 4, chapter 16, part 10. Effectiveness of a Thin Pavement Section Using Geogrids and
ASTM, 2003. Standard Test Method for Bearing Capacity of Soil for Drainage Geocomposites in a Cold Region. Department of Civil and
Static Load and Spread Footings. D1194. West Conshohocken, PA. Environmental Engineering, University of Maine, Orono, Maine.
ASTM, 2006. Test Method for Direct Shear Test of Soils Under http://www.tml.jp/e/product/transducers/catalog_pdf/CDP.pdf.
Consolidated Drained Conditions. D3080. West Conshohocken, PA. http://www.tenax.net/pdf_geo_e/TENAX_LBO_220_SAMP_e.pdf.
ASTM, 2007a. Standard Test Method for Particle-size Analysis of Soils. International Union of Railways, 1994. Earthworks and Track-bed Layers
D422. West Conshohocken, PA. for Railway Lines. UIC code 714 R, 2nd ed.
ASTM, 2007b. Test Method for Density and Unit Weight of Soil in Place Lee, M.S., Choi, Y.S., Prezzi, M., 2012. Quality Assessment of Geogrids
by Sand-cone Method. D1556. West Conshohocken, PA. Used for Subgrade Treatment. Joint Transportation Research Pro-
ASTM, 2011. Test Methods for Laboratory Compaction Characteristics gram, Indiana Department of Transportation and Purdue University.
of Soil Using Standard Effort (600 kN m/m3). D698. West Con- Management and Planning Organization Office of Deputy for Technical
shohocken, PA. Affairs Technical. Ministry of Road and Transportation Deputy of
ASTM, 2014a. Standard Test Methods for Minimum Index Density and Education Research and Technology, 2004. Railway Geometric
Unit Weight of Soils and Calculation of Relative Density. ASTM Design Code No. 288.
International, D4254-14, West Conshohocken, PA. Parsons, R., Jowkar, M., Han, J., 2012. Performance of Geogrid
ASTM, 2014b. Standard Test Methods for Maximum Index Density and Reinforced Ballast under Dynamic Loading. Mid-America Trans-
Unit Weight of Soils Using a Vibratory Table. ASTM International, portation Center University of Nebraska-Lincoln.
D4253-14, West Conshohocken, PA. Penman, J., Priest, D.J., 2009. The use of geogrids in railroad applications.
Bringkgreve, R.B.J., Vermeer, P.A., 1998. PLAXIS – finite element code Tenser International Corporation. Tenser, Spectra, Mesa and Sierra
for soil and rock analyses. Version 8. Plaxis B.V. Scape are registered trademarks. FP-TCE4C09.
Choudhary, A.K., Jha, J.N., Gill, K.S., 2010. Laboratory investigation of US Army Corps of Engineers, 2003. Engineering and Design, Slope
bearing capacity behavior of strip footing on reinforced fly ash slope. Stability. EM 1110-2-1902.
Geotext. Geomembr. 28, 393–402. Vesic, A., 1973. Analysis of ultimate loads of shallow foundations. J. Soil
Coleman, D.M., 1990. Use of Geogrids in Railroad Track: A Literature Mech. Found. Div.—ASCE 94 (SM3), 661–688.
Review and Synopsis. US Army Corps of Engineering, AD-A221, Webster, S.L., 1991. Geogrid Reinforced Base Course for Flexible
396. Pavements for Light Aircraft: Literature Review and Test Sec-
El Sawwaf, M., 2007. Behavior of strip footing on geogrid reinforced sand tion Design. U.S Department of Transportation, AD-A239 453.
over a soft clay slope. Geotext. Geomembr. 25, 50–60. Wood, D.M., 2004. Geotechnical Modeling. Taylor and Francis.
El Sawwaf, M., Nazir, A.K., 2012. Cyclic settlement behavior of strip Yoo, C., 2001. Laboratory investigation of bearing capacity behavior of
footings resting on reinforced layered sand slope. J. Adv. Res. 3 (4), strip footing on geogrid reinforced sand slope. Geotext. Geomembr.
315–324. 19, 279–298.

Please cite this article in press as: Esmaeili, M. et al., Investigating the effect of geogrid on stabilization of high railway embankments, Soils Found.
(2018), https://doi.org/10.1016/j.sandf.2018.02.005