Geotextiles and Geomembranes 44 (2016) 358e365

Contents lists available at ScienceDirect

Geotextiles and Geomembranes

journal homepage: www.elsevier.com/locate/geotexmem

Technical note

Geosynthetic-reinforced soil structures with concave facing profile

Farshid Vahedifard a, *, Shahriar Shahrokhabadi a, Dov Leshchinsky b, c
a
Department of Civil and Environmental Engineering, Mississippi State University, Mississippi State, MS 39762, USA
b
Dept. of Civil and Environmental Engineering, University of Delaware, Newark, DE 19716, USA
c
ADAMA Engineering, Inc., P.O. Box 90217, Portland, OR 97290, USA

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents a new method to determine the optimal profile of facing elements in geosynthetic-
Received 3 September 2015 reinforced soil structures. Flexibility of some facing systems and advances in construction technology
Received in revised form allow construction of reinforced soil structures with a non-planar cross section. In this study, the facing
1 January 2016
profile of a concave geosynthetic-reinforced soil structure (referred to as CGRSS) is idealized by a circular
Accepted 14 January 2016
Available online 8 February 2016
arc defined by a single variable, the Mid-Chord Offset (MCO). For a given setback and elevation change,
the optimal facing profile is determined by seeking the MCO which, for a given margin of safety, yields
the least tensile load in the reinforcement layers. The proposed procedure for finding the optimal facing
Keywords:
Geosynthetics
profile is incorporated into a limit equilibrium-based log spiral formulation to determine the required
Limit equilibrium methods tensile strength of the reinforcement. Results are presented in a set of charts showing the required
Reinforced soils unfactored tensile strength, MCO, and mode of failure for various friction angles, batter angles, and
Optimization seismic coefficients. It is shown that CGRSSs can decrease the required tensile strength of the rein-
Concave profile forcement by up to 30% under static and pseudo-static conditions. This observation justifies employing
Seismicity concave facing profiles in practice.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction construction of GRSS with multi-tiers or curved profile in longitu-

Geosynthetic-reinforced soil structures (GRSSs), including
geosynthetic-mechanically stabilized (MSE) walls and
geosynthetic-reinforced slopes (GRS), are widely used for stabiliz-
ing steep slopes and walls. Well-established design methodologies,
relative ease of construction, and satisfactory record of perfor-
mance under normal and extreme loading conditions have estab-
lished GRSS as an economically and technically viable choice for
both public and private sectors. Main components of a GRSS include
the reinforced soil, geosynthetic reinforcement layers, and a facing
system. Various facing systems are available for GRSSs including
geosynthetic wrap-around facing, masonry block units, gabions,
welded wire mesh, vegetation, etc (e.g., FHWA, 2009). Facing sys-
tems provide protection against backfill sloughing and erosion, and
in some cases, serve as drainage paths. Since facing elements are
the only visible component of a GRSS, the choice of facing system
will also influence aesthetics of the structure. Flexibility in the
arrangement of a majority of facing systems has allowed
* Corresponding author.
E-mail addresses: farshid@cee.msstate.edu (F. Vahedifard), ss2804@msstate.edu
(S. Shahrokhabadi), dov@udel.edu (D. Leshchinsky).
dinal section. Using bi-linear and multi-tiered configurations can
reduce tensile loads in reinforcement layers, especially for taller
GRSSs. These configurations can enhance the aesthetics of the built
structure and also can lead to a more economical design (e.g.,
Leshchinsky and Han, 2004; Ruan et al., 2015).
Several analytical and numerical studies have been conducted to
optimize the number, arrangement, and strength of reinforcement
layers in GRSSs (e.g., Basha and Babu, 2012; Leshchinsky, 2014; Xie
and Leshchinsky, 2015). However, similar to other engineered
slopes and earthen structures, GRSSs are commonly designed and
built only with a planar facing profile in cross section. Inspired by
some natural concave slopes, recent studies have shown that un-
reinforced slopes with a concave profile in cross section offer higher
stability (e.g., Sokolovskiĭ, 1960; Utili and Nova, 2007; Jeldes et al.,
2013) and better erosion resistance (e.g., Rieke-Zapp and Nearing,
2005; Schor and Gray, 2007; Jeldes et al., 2014) when compared
with the equivalent planar slopes. Few attempts have been made to
analyze concaves slopes and to quantify the contribution of such
concave profiles to stability and erosion resistance of earthen slopes
(e.g., Sokolovskiĭ, 1960; Utili and Nova, 2007; Jeldes et al., 2013).
Sokolovskiĭ (1960) used the slip-line field theory and showed that
the slope surface at the limit equilibrium (LE) state has a concave

http://dx.doi.org/10.1016/j.geotexmem.2016.01.004
0266-1144/© 2016 Elsevier Ltd. All rights reserved.
 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365 359

Notation R log spiral radius

Sv vertical spacing between reinforcement layers
A1 log spiral constant T resultant of all reinforcement forces
c design cohesion Tmax-i maximum tensile force in the ith reinforcement
D height of the line of action of the resultant force from W weight of failure mass
toe x, y coordinates of any point along the curved facing profile
Di height of the line of action of the ith reinforcement xcc, ycc coordinates of the center of the curved facing profile
force from toe xF, yF coordinates of any point along the log spiral failure
FS factor of safety surface
H slope/wall height xCG, yCG coordinates of the center of gravity of the failure mass
I arc angle xCL, yCL coordinates of the pole of log spiral
kh, kv horizontal and vertical seismic coefficients Ye height where the slip surface exits the facing profile
LC long cord a backslope angle
MCO mid-chord offset g unit weight
MCOmax maximum MCO forming a vertical tangent at the crest DA soil volume reduction per unit length of the slope
MCOopt MCO representing the optimum concavity (i.e., b1, b2 angle of rotation to the points where the log spiral exits
yielding the minimum T) and enters the slope
MRh moment due to Rh z angle between toe and circular arc of the planar facing
Mq moment induced to the uniform surcharge q profile
n number of reinforcement layers j tan(f)
q uniform surcharge f design internal angle of friction
Rc circular arc radius u batter (¼90  Average slope angle)
Rh horizontal resistance of the facing at the bottom of the
slope

profile. Jeldes et al. (2013) used the Sokolovskiĭ solution for a yields the least tensile load in the reinforcement. An optimization
weightless medium and presented an approximate solution rep- problem is defined based on the consideration that the locations of
resenting the optimal concave profile of slope. Jeldes et al. (2014) toe and crest are prescribed by the project.
used the latter approximate solution along with an erosion model Fig. 1 shows the notation and details of the idealized circular arc
and showed that concave slopes yield 15e40% less sediment than geometry which is used to idealize the concave facing profile. As
alternative planar slopes with the same factor of safety. Utili and shown in Fig. 1b, the curvature of the arc is controlled via MCO,
Nova (2007) used two log spirals: one spiral to represent the which is a commonly used term in surveying to define circular arcs.
concave slope surface and the other for the slip surface, in the The single parameter MCO is the distance between Points 4 and 5 in
context of the upper bound limit analysis (LA) method. All the Fig. 1b.
aforementioned methods only investigated concave profiles for It is desired to obtain the coordinates of any point along a
unreinforced slopes under static loading. concave profile as a function of MCO, the batter, u, and the slope/
Flexibility of some facing systems (e.g., wrapped-faced, masonry wall height, H. Note that u is defined as the orientation of a straight
block units) and advances in construction technology allow us to line connecting the toe (Point 1 in Fig. 1a) to the crest (Point 3 in
take advantage of the higher stability of concave profiles in rein- Fig. 1a) relative to the vertical axis. For the given CGRSS geometry,
forced soil structures and to build GRSS with concave facing pro- the Long-Chord (LC) is known, which is defined as the straight line
files. In this study, the facing profile of a concave GRSS (referred to between Points 1 and 3 (i.e., the length of the equivalent planar
as CGRSS) is idealized by a circular arc defined by a single variable, profile). According to Fig. 2b, MCO and LC can be defined as:
the Mid-Chord Offset (MCO). The effects of using concave facing
profile under static and pseudo-static conditions are studied by


I
implementing the proposed concave profile into a LE-based log MCO ¼ Rc 1  cos (1)
2
spiral formulation. Theoretically, this study shows the impact of
face geometry on the stability and required tensile strength of


reinforcement in GRSSs. To the best of the authors' knowledge, this I
impact has not been investigated for reinforced earthen structures LC ¼ 2Rc sin (2)
2
yet. Practically, the proposed procedure for employing concave
profiles can potentially lead to more economical designs by where Rc and I are the arc radius and the arc angle, respectively, and
decreasing the required tensile strength of reinforcement and/or can be determined as:
backfill volume. The proposed approach is developed by employing
a conventional design procedure (i.e., LE), suggesting that it can be
H2   MCO
readily implemented in practice. Rc ¼ 1 þ tan2 u þ (3)
8MCO 2

2. Formulation of concave facing profile


2MCO cos u
I ¼ 4 tan1 (4)
H
The concave facing profile of a CGRSS for given height is
formulated using a circular arc defined by a single parameter, MCO. The center of the circular arc (xcc, ycc) is located on the bisector of
The optimal facing profile is determined by seeking the MCO that LC and can be defined as:
360 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365

(a) The y-coordinate of any point on the circular arc can be deter-
Pole of log spiral mined using the equation of a circle defined as:
(xCL, yCL)
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
R2 = A
1 exp(-ψ 2 q y ¼ ycc  R2c  ðx  xcc Þ2 (8)
Ψ = tan( 2)
2
)
3 where Rc, xcc, and ycc can be calculated using Eqs. (3) and (7),
Tmax-7 respectively. For a given u and H, the optimal facing profile can be
R1 =

1
Sv found by varying MCO. The lower limit of MCO is zero, representing
Tmax-6
a planar profile. The upper limit of MCO in the optimization search
A 1 ex

kvW
CG Tmax-5 needs to be selected by taking into account the physical admissi-
p(-ψ

khW bility of the final profile. For steep slopes, the most stable facing
H W Tmax-4 T profile found in the optimization search tends to generate a nega-
1)

tive slope at the crest, which represents an overhanging cliff (e.g.,

Tmax-3 Utili and Nova, 2007). However, a negative slope at the crest is
y
D difficult to construct while implying the potential development of
Tmax-2 D3 tension cracks and therefore, is not considered as permissible in the
1 D2
Tmax-1 optimization process. To exclude negative slopes at the crest, a
Rh D1
(0, 0) x physical constraint is defined to limit the maximum feasible con-
cavity to the case with a vertical tangent at the crest. Geometrically,
it means that minimum ycc is equal to H and is located on the
(b)
bisector of LC. Subsequently, the upper and lower bounds of MCO
Center of circular arc used in the optimization process can be defined as follows:
(xcc, ycc) 3
Rc 0  MCO
I I sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2

 2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
H 1 1
 1þ  1 cot u þ H tan u  ðH  2Þ2 ðcsc uÞ2
LC

2 2 2
MCO I
2 (9)

H 5
4 3. Implementation in limit equilibrium formulation
R
c

6
I The proposed procedure for finding the optimal facing profile is
y incorporated into the LE log spiral reinforced slope stability
formulation. As shown in Fig. 1a, the problem is formulated by
ζ considering rotational failure of a CGRSS along a log spiral slip
x
1 surface. LE formulation for the body defined by the log spiral slip
surface satisfies all static equilibrium conditions without resorting
Fig. 1. Notation and geometry used in the formulation: a) Rotational failure of a CGRSS along to statical assumptions. The log spiral validity has been demon-
a log spiral slip surface; b) Detail of circular arc used in optimizing the facing profile.
strated for GRSSs through several experimental and numerical in-
vestigations (e.g., Leshchinsky et al., 2009, 2014). It has been
successfully used in commercial slope stability programs (e.g.,
xcc ¼ Rc cos z (5a) Leshchinsky, 1997), as well as in complex problems such as seismic
analysis of reinforced soil structures (e.g., Vahedifard et al., 2013).
ycc ¼ Rc sin z (5b) The moment limit LE equation is derived for the reinforced soil.
At a limit state, it is assumed that the soil strength is fully mobilized
where z is the angle of rotation between the toe and the center of along the failure surface (i.e., an active wedge is formed), and the
the circular arc with respect to the horizontal (see Fig. 1b): stability of the system then hinges on the reinforcement. Further, it
is assumed that the maximum tensile force for each layer i, Tmax-i,
I acts horizontally and occurs at its intersection with the critical
z¼uþ (6)
2 failure surface. The moment equilibrium equation around the pole
of the log spiral (xCL, yCL) can be written as (Fig. 1a):
Substituting Eqs. (3), (4) and (6) into Eq. (5) will yield the co-
Z
ordinates of the circular arc as function of H, MCO, and u:
Wð1  kv ÞðxCG  xCL Þ þ kh WðyCL  yCG Þ  c ðxF  xCL Þdy

H2   Z X
n
xcc ¼ 1 þ tan2 u c ðyCL  yF Þdx þ Mq  MRh  Tmaxi ðyCL  Di Þ ¼ 0
8MCO


  i¼1
MCO 2MCO cos u
þ cos 2 tan1 u (7a) (10)
2 H
where W is the weight of the failure mass, c is the design cohesion,

H2   kh is the horizontal seismic coefficient, kv is the vertical seismic
ycc ¼ 1 þ tan2 u
8MCO coefficient (upward positive), Mq is the moment induced to the


  uniform surcharge q, MRh is the moment due to Rh, where Rh is the
MCO 2MCO cos u
þ sin 2 tan1 u (7b) horizontal resistance of the facing at the bottom of the slope, n is
2 H
the number of reinforcement layers, and Di is the height of the line
 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365 361

of action of the ith reinforcement force from toe. As shown in

Fig. 1a, ðxCG ; yCG Þ are the coordinates of the center of gravity of the
failure mass, and ðxF ; yF Þ are the coordinates of any point along the
slip surface. Current GRSS design guidelines (e.g., FHWA, 2009)
commonly recommend to ignore cohesion. While it is demon-
strated that including cohesion can have a substantial impact on
the design of GRSSs, it can be an unreliable long-term design
parameter (e.g., Vahedifard et al., 2015). Further, it is common in
design to set Rh to zero. While toe resistance is shown to counter-
balance a portion of the lateral thrust of the backfill (e.g.,
Leshchinsky et al., 2010; Leshchinsky and Vahedifard, 2012; Ehrlich
and Mirmoradi, 2013), its contribution is commonly ignored in
design (e.g., FHWA, 2009) considering potential excavation or
scouring in front of the toe. Along with several other factors, toe
resistance is referred to as “implicit safety margin” or “redundancy”
which are not accounted for in limit state design methods
(Leshchinsky and Tatsuoka, 2013).
An assumed log spiral failure surface can be fully defined using
ðxCG ; yCG Þ and log spiral constant, A1. The log spiral radius, RL, at any
point is defined as RL ¼ A1 ejb , where j ¼ tan f (where f is the
design internal angle of friction) and b is the angle of rotation be-
tween the point along the slip surface and the pole of log spiral
measured from the vertical as shown in Fig. 1a. The explicit terms of
the moment LE equation for the log spiral mechanism can be found
in the literature (e.g., Vahedifard et al., 2012). Concavity of the
facing profile affects W, and any increase in MCO will change the
associated terms in the moment equilibrium equation. Using the
moment equilibrium equation, one can find individual Tmax-i values
by using, for example, the top-down procedure (e.g., Leshchinsky
et al., 2014). Alternatively, a single equivalent force, T, represent-
ing the resultant of all Tmax-i values can be replaced in the formu-
lation as:

X
n
T¼ Tmaxi (11)
i¼1

The resultant force, T, acts at the height D above toe. In this

approach, one needs to assume a distribution function (e.g., linear,
triangular or trapezoidal) and then determine D based on the
selected distribution function. For example, D ¼ H/3 or H/2 for
triangular and uniform distribution functions, respectively
(Leshchinsky et al., 2010).
For an assumed distribution function, T is calculated through a
maximization process yielding the most critical active wedge. For
given input parameters, the solution procedure starts with finding T
for a planar profile (i.e., MCO ¼ 0). Subsequently, MCO is incre-
mentally increased and the T value corresponding to each MCO is
recorded. Reduction in the soil volume per unit length of CGRSS
(DA) for each MCO can be calculated by finding the area confined
between the tested concave profile and the planar profile. Another
parameter of interest is the height above the toe where the critical
failure surface emerges (Ye), referred to herein as the exit point. For
homogenous GRSSs with planar profiles, the critical failure surface
passes through the toe (referred to as toe failure), except for cases
with low f and high u (i.e., mild slopes) values where the surface
emerges away from the toe, forming deep-seated failures
(Leshchinsky, 1997). Such deep-seated failures are not considered
in this work (i.e., the foundation soil is considered competent).
Consequently, when MCO ¼ 0, Ye ¼ 0. However, in CGRSSs the slope
surface curvature may render critical failure surfaces emerging
above the toe (i.e., Ye > 0) and such possibility needs to be
considered. The latter failure is referred to as face failure. For each
Fig. 2. Results for static loading: a) Normalized resultant (T/gH2); b) Exit point (Ye); c)
MCO, in order to capture the most critical failure surface (i.e.,
Soil volume reduction per unit length (DA/H2).
largest T), a given circular profile must first be divided into several
362 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365
equally-spaced distinct points along its height. Further, each point
along the facing profile is treated as a potential exit point for the
failure surface. Subsequently, the local resultant (Ti) corresponding
to each potential Ye is determined and recorded. The largest Ti and
its corresponding Ye are taken as the global T and Ye for that MCO.
The same procedure is repeated for each MCO that is examined in
the optimization search.
After obtaining T following the aforementioned solution pro-
cedure, the following equation can be used to distribute T amongst
the individual layers to determine Tmax-i values using (Leshchinsky
et al., 2010):
DT
Tmaxi ¼ T Pn maxi (12)
i¼1 DTmaxi
where DTmaxi is a nondimensional distribution parameter. This
parameter relates the force in reinforcement layer i to the
maximum force that is mobilized among all layers. It can be
calculated using the relationship: DTmaxi ¼ Tmaxi =maxðTmaxi Þ.
DTmaxi is determined based on the selected distribution function
(Leshchinsky et al., 2010). In the case of a face failure, the calculated
T should be distributed amongst the reinforcement layers which are
above Ye.
It should be noted that the proposed approach for finding the
optimal facing profile can be used along with rigorous LE methods
for inhomogeneous (e.g., Spencer, Morgenstern-Price) or other
advanced numerical methods to deal with more complex problems.
Since the log spiral formulation does not require any static as-
sumptions, the current formulation creates a benchmark for such
rigorous methods that involve static assumptions.
4. Results and discussions
The LE-based method is used to generate a few representative
charts in this section. The charts are presented in the form of T
versus MCO for different angles of friction and batters under static
and seismic loading conditions. Due to space limitation, the results
are shown only for a uniform distribution function (D ¼ H/2), which
is commonly used in LE-based methods. Further, unless otherwise
noted, Rh, kv, a and q are set to zero. It is noted that the following
charts show results for some representative cases. However, one
can use the proposed formulation and its associated solution pro-
cedure to generate results for any other set of input parameters.
4.1. Parametric study results for static loading
The optimum configuration for a CGRSS profile can be obtained
based on the concavity concept. The input parameter for the opti-
mization algorithm is MCO/MCOmax ð0  MCO=MCOmax  1Þ,
where MCOmax represents the maximum concavity defined as the
upper bound in Eq. (9). The results in Fig. 2 demonstrate the impact
of f and MCO/MCOmax (termed relative MCO) on T, Ye, and reduc-
tion in soil volume per unit length (DA) for different u values. The
results are presented in normalized form.
Fig. 2(a) shows the relationship between normalized T and
relative MCO at f ¼ 30 and 40 . For lower batters (u ¼ 10 and
20 ), it can be seen that as the relative MCO increases, T decreases.
This means that a minimum T is obtained when the relative MCO
approaches unity. However, in the case of higher batters (u ¼ 30
and 40 ), based on f, a minimum value for T is obtained at a
relative MCO less than one, which indicates that the mode of
failure changes from toe to face failure. Moreover, as expected,
lower batters corresponding to lower values of f lead to an in-
crease in T.
Fig. 2(b) illustrates the change in the mode of failure (i.e., toe to
face failure) as the relative MCO increases for various values of f
and u. As explained previously, when Ye is zero, the slip surface
passes through the toe and when Ye is greater than zero, the slip
surface emerges through the face. Out of the eight cases shown in
Fig. 2(a), only four cases (u ¼ 30 and 40 for each f) are shown in
Fig. 2(b). For the other four cases, Ye ¼ 0 for entire range of MCO/
MCOmax. That is, for u ¼ 10 and 20 , regardless of the value of the
relative MCO, the toe failure always dominates the failure type.
However, for the case of u ¼ 30 and 40 , the occurrence of the face
failure is contingent upon the magnitude of the relative MCO.
Specifically, as the value of the relative MCO increases, the proba-
bility of the face failure increases. For u ¼ 40 and f ¼ 40 , it can be
seen that the face failure covers the widest range of relative MCOs.
On the other hand, for u ¼ 30 and f ¼ 30 , the toe failure covers
the widest range of relative MCOs.
Fig. 2(c) shows a linear trend between DA and the relative MCO.
Since DA solely depends on MCO and u, the trend for different
values of f is similar for a given u. As illustrated for milder batters
(i.e., u ¼ 30 and 40 ), the magnitude of DA is larger than that
obtained for steeper batters (i.e., u ¼ 10 and 20 ). Additionally, as
expected, the maximum value of DA increase as u increases.
Fig. 3 depicts the performance of the proposed concave facing
profile with respect to a conventional GRSS with a planar facing
profile. Fig. 3 associates the relationship between the ratio of T from
concave and planar configurations (Tconcave/Tplanar) and different
values of the relative MCO. As can be seen, the maximum reduction
of about 30% in the magnitude of Tconcave/Tplanar occurs at u ¼ 40
and f ¼ 40 . Further, a non-linear trend is observed in Fig. 3 for the
cases of mild batters. Comparing Figs. 2(b) and 3 shows that for
mild batters the optimum concavity, MCOopt (i.e., the concavity
yielding the minimum T) coincides with the transition from toe to
face failure. However, for steep batters, Tconcave/Tplanar linearly de-
creases up to the point where the relative MCO is unity (i.e.,
MCOopt ¼ MCOmax).
It has been demonstrated that by increasing both u and f, the
toe failure dominates as the mode of failure for concave facing
profiles. For mild batters face failure occurs with large concavity.
For such cases, when the face failure governs, the trend for the ratio
of Tconcave/Tplanar is upward. This means that an increase in the
relative MCO corresponds to an increase in Tconcave. On the other
hand, for lower values of u, the trend for the ratio of Tconcave/Tplanar
is linearly decreasing as the magnitude of the relative MCO
increases.
4.2. Effect of seismicity
Figs. 4 and 5 show the TeMCO relationship for different kh
values for two u values of 10 and 40 , respectively. These batters
are selected to represent a small batter (i.e., a reinforced wall) and a
large batter (i.e., a reinforced slope).
Fig. 4 shows the effect of f and the relative MCO on T, the ratio of
Tconcave/Tplanar, and DA/H2 at various levels of seismicity (i.e.,
kh ¼ 0.1, 0.3, and 0.5), for u ¼ 10 . In Fig. 4(a), as expected, it is
apparent that a higher kh and/or lower u yields an increase in T for a
given MCO. Generally, the trends for all three values of kh are lin-
early decreasing. Fig. 4(b) shows the relationship between the ratio
of Tconcave/Tplanar versus the relative MCO. As illustrated, using a
concave profile for a CGRSS with u ¼ 10 can decrease T between 1
and 9%. It can be concluded that an increase and decrease in kh and
f, respectively, yields a higher ratio of Tconcave/Tplanar. As demon-
strated in Fig. 4(c), DA increases linearly with an increases in the
relative MCO, regardless of the value of kh. It should be noted that
for u ¼ 10 , Ye was always zero (i.e., toe failure) through the entire
range of MCO for all the cases which were examined.
 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365
Fig. 3. Reduction in the resultant force using concave facing with respect to planar
configuration.
Fig. 5(a) depicts the values of normalized T with respect to
different values of the relative MCO for u ¼ 40 . It can be noted that
the trend in Fig. 5(a) is similar to Fig. 4(a), however there is a slight
non-linearity as f increases. The performance of concave configu-
ration with respect to planar profile is depicted in Fig. 5(b). As can
be seen, the trend is non-linear as kh and f decrease and increase,
respectively. Furthermore, a maximum reduction of about 30% in T
occurs when kh ¼ 0.1 and f ¼ 40 . On the other hand, Tconcave is
slightly lower than Tplanar when kh ¼ 0.5 and f ¼ 30 Fig. 5(c) as-
sociates values of DA/H2 versus the relative MCO. As it can be seen,
there is a linearly increasing trend and the trend is independent of
the level of seismicity.
Fig. 6 depicts the changes in Ye with respect to the relative MCO
for u ¼ 40 at different levels of kh. Out of six cases shown in
Fig. 5(a), only four curves are shown in Fig. 6. For the two other
cases not shown in Fig. 6, Ye ¼ 0 for the entire range of MCO and
Fig. 4. Results for seismic analysis for u ¼ 10 : a) Normalized resultant force (T/gH2); b) Reduction in resultant force; c) Soil volume reduction per unit length (DA/H2).
363
various seismicity levels which were examined. It means that for
u ¼ 40 , Ye depends on the level of kh and f.
In summary, the concave configuration, in relation to the planar
configuration, leads to up to 30% lower T in the presence of seis-
micity. Under analogous conditions where the level of seismicity is
increasing, the dominance of the mode of failure, with respect to
the relative MCO, changes from toe to face failure for milder batters.
5. Illustrative design examples
The application of the proposed LE-based method for design
purposes is illustrated for two representative CGRSSs showing how
a concave facing profile, in comparison with a planar facing profile,
reduces Tmax-i for designing purposes.
The first example, as illustrated in Fig. 7, represents a reinforced
masonry block wall with u ¼ 15 . The face of wall is formed by
20  40 cm blocks, and the reinforcements are placed at every
60 cm (Sv ¼ 60 cm). For this illustrative example, H ¼ 4.8 m,
g ¼ 20 kN/m3, and f ¼ 34 . Based on the location of Ye, the resultant
force T will be uniformly divided among the number of re-
inforcements above Ye to determine Tmax-i. Since it was shown
previously that toe failure always governs for steep batters (e.g.,
u ¼ 15 ), Tmax will be divided equally among all layers of rein-
forcement. In the case of reinforced masonry walls, Rh contributes
to the calculation of resultant force, and in this example, the value
of Rh is taken as being 20% of T. The second illustrative example, as
depicted in Fig. 8, is a wrapped-faced reinforced slope with a batter
u ¼ 45 . The contributory area of each layer is taken as the vertical
height of the wrapped portion, 0.6 m (Sv ¼ 60 cm), and the total
number of reinforcements (i.e., n ¼ 7). For the second example
illustrated in Fig. 8, H ¼ 4.8 m, g ¼ 18 kN/m3, and f ¼ 30 . Based on
the location of Ye, Tmax will be uniformly divided among the
number of reinforcements above Ye. It is important to note that Rh is
not considered for a wrapped-faced reinforced slope. As a result,
the ratio of Rh/T is zero in this example.
For the design examples or any similar design problem, the
proposed formulation can be used in the following manner. For
given friction angle, toe resistance ratio, height and batter, the
feasible range of MCO can be determined using Eq. (9). The lowest
MCO (i.e., zero) generates a planar profile and as MCO increases, the
facing concavity increases. The next step is to use the moment
equilibrium equation (Eq. (10)) and find corresponding T values for
different MCO (e.g., Figs. 2a, 4a and 5a). On these charts, the opti-
mum concavity is depicted by the MCO which leads to the lowest T.
After finding the optimum MCO, the corresponding T is distributed
amongst the reinforcement layers based on the chosen distribution
function (e.g., uniform function as used in this study). The mode of
364 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365
Fig. 5. Results for seismic analysis for u ¼ 40 : a) Normalized resultant force (T/gH2); b) Reduction in resultant force; c) Soil volume reduction per unit length (DA/H2).
Fig. 6. Effect of seismicity on exit point (Ye) for u ¼ 40 .
failure (i.e., toe versus face failure) for each case can be monitored
by checking Ye corresponding to each MCO, as demonstrated in
Figs. 2b and 6 for different cases.
Table 1 shows the results for the abovementioned examples. For
each example, the results are shown for planar facing profile (i.e.,
MCO ¼ 0) and the optimum facing configuration (i.e.,
ω yields the least tensile load in the reinforcement layers. The pro-
γ φ
Fig. 7. Details for CGRSS with masonry block units, u ¼ 15 , MCOopt ¼ 0.326 m
(Example 1).
ω
γ φ
Fig. 8. Details for CGRSS with wrapped facing, u ¼ 45 , MCOopt ¼ 0.527 m (Example 2).
MCO ¼ MCOopt) which leads to the least tensile force in the rein-
forcement. As shown, in the case of reinforced masonry block wall
(i.e., u ¼ 15 ), using CGERS results in a relative difference between
the tensile forces for each layer of reinforcement of approximately
14%. Conversely, CGRSS for the wrapped-faced reinforced slope (i.e.,
u ¼ 45 ) decreases the relative difference in the design tensile
forces by as much as 25% for each layer of reinforcement. As it can
be seen in Table 1 for u ¼ 45 , T is divided among all the layers of
reinforcement. For this case, since Ye is less than the height of first
layer of reinforcement, T is uniformly divided among all layers of
reinforcement.
6. Conclusions
This study proposes a methodology to design geosynthetic-
reinforced soil structures with concave facing profile (referred to
as CGRSS). It is shown that concave slopes offer higher stability,
leading to lower tensile force in reinforcement layers. The proposed
method can be applied in the construction of more efficient
geosynthetic-reinforced soil structures that are simple to construct.
The facing profile of a CGRSS is idealized by a circular arc defined
by a single variable, characterized as the Mid-Chord Offset (MCO).
The optimal facing profile was determined by seeking the MCO that
posed procedure for finding the optimal slope profile was incor-
porated into a limit equilibrium-based formulation considering
rotational failure along a log spiral failure surface. It was shown
that, for a steep CGRSS (i.e., reinforced wall), the resultant force
monotonically decreases as MCO increases. For a CGRSS with a large
batter (i.e., reinforced slope), the change in the forces behaved non-
monotonically and the least tensile load in the reinforcement did
not coincide with the largest concavity.
 F. Vahedifard et al. / Geotextiles and Geomembranes 44 (2016) 358e365 365

Table 1
Design parameters and results for two examples.

u ( ) f ( ) Rh/T MCOmax (m) Facing configuration MCOopt/MCOmax Ye (m) T/gH2 T (kN/m) Tmax-i (kN/m)

15 34 0.2 0.326 Planar 0 0 0.088 40.8 5.8

Concave 1.0 0 0.076 35.1 5.0

45 30 0 1.406 Planar 0 0 0.043 17.7 2.5

Concave 0.375 0.346 0.032 13.4 1.9

Parametric study was performed and the results were presented
in a set of charts to illustrate the relationship between the resultant
force of the reinforcement layers and the MCO, for various batters
under static and seismic loading conditions. The association be-
tween the mode of failure and the MCO was also presented. It was
demonstrated that the required tensile strength of the reinforce-
ment could be reduced by approximately 30% under static and
pseudo-static conditions, if a concave facing profile is used. It was
observed that an increase in the seismic coefficient changed the
location of the exit point, which caused the mode of failure to
transform from face failure to toe failure.
References
Basha, B.M., Babu, G.L.S., 2012. Target reliability-based optimisation for in-
ternal seismic stability of reinforced soil structures. Geotechnique 62 (1),
55e68.
Ehrlich, M., Mirmoradi, S.H., 2013. Evaluation of the effects of facing stiffness
and toe resistance on the behavior of GRS walls. J. Geotext. Geomembr. 40,
28e36.
FHWA, 2009. Mechanically Stabilized Earth Walls and Reinforced Soil Slopes Design
and Construction Guidelines, vol. I. Publication No. FHWA-NHI-10-025, Auth-
ored by Berg, R. R., Christopher, B. R. and Samtani, N.C.
Jeldes, I.A., Vence, N.E., Drumm, E.C., 2013. Approximate solution to the Sokolovskiĭ
concave slope at limiting equilibrium. Int. J. Geomech. 04014049. http://
dx.doi.org/10.1061/(ASCE)GM.1943-5622.0000330.
Jeldes, I., Drumm, E., Yoder, D., 2014. Design of stable concave slopes for reduced
sediment delivery. J. Geotech. Eng. 04014093. http://dx.doi.org/10.1061/(ASCE)
GT.1943-5606.0001211.
Leshchinsky, B., 2014. Limit analysis optimization of design factors for mechanically
stabilized earth wall-supported footings. Transp. Infrastruct. Geotechnol. 1 (2),
111e128.
Leshchinsky, D., 1997. Design Procedure for Geosynthetic Reinforced Steep Slopes.
Geotechnical Laboratory, US Army Corps of Eng., Waterways Experiment Sta-
tion. REMR-GT-23, January 1997, Vicksburg, Mississippi.
Leshchinsky, D., Han, J., 2004. Geosynthetic reinforced multitiered walls. J. Geotech.
Geoenviron. Eng. 130 (12), 1225e1235.
Leshchinsky, D., Kang, B.J., Han, J., Ling, H.I., 2014. Framework for limit state design
of geosynthetic-reinforced walls and slopes. Transp. Infrastruct. Geotechnol. 1
(2), 129e164. Springer.
Leshchinsky, D., Ling, H.I., Wang, J.-P., Rosen, A., Mohri, Y., 2009. Equivalent seismic
coefficient in geocell retention systems. Geotext. Geomembr. 27 (1), 9e18.
Leshchinsky, D., Tatsuoka, F., 2013. Geosynthetic reinforced walls in the public
sector: performance, design, and redundancy. Geosynth. Mag. 31 (3),
12e21.
Leshchinsky, D., Vahedifard, F., 2012. Impact of toe resistance in reinforced ma-
sonry block walls: design dilemma. J. Geotech. Geoenviron. Eng. 138 (2),
236e240.
Leshchinsky, D., Zhu, F., Meehan, C.L., 2010. Required unfactored strength of geo-
synthetic in reinforced earth structures. J. Geotech. Geoenviron. Eng. 136 (2),
281e289.
Rieke-Zapp, D.H., Nearing, M.A., 2005. Slope shape effects on erosion. Soil Sci. Soc.
Am. J. 69 (5), 1463e1471.
Ruan, X., Leshchinsky, D., Leshchinsky, B.A., 2015. Global stability of bilinear rein-
forced slopes. Transp. Infrastruct. Geotechnol. 2, 34e46.
Schor, H.J., Gray, D.H., 2007. Landforming: an Environmental Approach to Hillside
Development, Mine Reclamation and Watershed Restoration. Wiley, Hoboken,
NJ.
Sokolovskiĭ, V.V., 1960. Statics of Soil Media. Butterworths Scientific Publications,
London.
Utili, S., Nova, R., 2007. On the optimal profile of a slope. Soils Found. 47 (4),
717e729.
Vahedifard, F., Leshchinsky, B., Mortezaei, K., Lu, N., 2015. Active earth pressures for
unsaturated retaining structures. J. Geotech. Geoenviron. Eng. 141 (11),
04015048.
Vahedifard, F., Leshchinsky, D., Meehan, C.L., 2012. Relationship between the
seismic coefficient and the unfactored force of geosynthetic in reinforced earth
structures. J. Geotech. Geoenviron. Eng. 128 (10), 1209e1221.
Vahedifard, F., Leshchinsky, D., Meehan, C.L., 2013. Displacement-based internal
design of geosynthetic-reinforced earth structures subjected to seismic loading
conditions. Ge otechnique 63 (6), 451e462.
Xie, Y., Leshchinsky, B., 2015. MSE walls as bridge abutments: optimal reinforce-
ment density. Geotext. Geomembr. 43, 128e138.