Geotextiles and Geomembranes 28 (2010) 262–267

Contents lists available at ScienceDirect

Geotextiles and Geomembranes

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

Analysis of back-to-back mechanically stabilized earth walls

Jie Han a, *, Dov Leshchinsky b
a
Associate Professor, Department of Civil, Environmental, and Architectural Engineering, the University of Kansas, Lawrence, KS 66045, USA
b
Professor, Department of Civil and Environmental Engineering, the University of Delaware, Newark, DE 19716, USA

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

Article history: Back-to-back Mechanically Stabilized Earth (MSE) walls are commonly used for embankments
Received 17 May 2008 approaching bridges. However, available design guidelines for this wall system are limited. The distance
Received in revised form between two opposing walls is a key parameter used for determining the analysis methods in FHWA
11 December 2008
Guidelines. Two extreme cases are identified: (1) reinforcements from both sides meet in the middle or
Accepted 18 May 2009
Available online 26 November 2009
overlap, and (2) the walls are far apart, independent of each other. However, existing design metho-
dologies do not provide a clear and justified answer how the required tensile strength of reinforcement
and the external stability change with respect to the distance of the back-to-back walls. The focus of this
Keywords:
Critical failure surface paper is to investigate the effect of the wall width to height ratio on internal and external stability of MSE
Maximum tension walls under static conditions. Finite difference method incorporated in the FLAC software and limit
Factor of safety equilibrium method (i.e., the Bishop simplified method) in the ReSSA software were used for this
Limit equilibrium analysis. Parametric studies were carried out by varying two important parameters, i.e., the wall width to
MSE wall height ratio and the quality of backfill material, to investigate their effects on the critical failure surface,
Reinforcement the required tensile strength of reinforcement, and the lateral earth pressure behind the reinforced zone.
The effect of the connection of reinforcements in the middle, when back-to-back walls are close, was also
investigated.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction mobilized so that the active thrust should be reduced. Here Di is

defined as the interaction distance and expressed by
Back-to-back mechanically stabilized earth (MSE) walls are   
commonly used for embankments approaching bridges to raise Di ¼ Htan 45  f=2 (1)
elevations. Design of such walls is considered as a special situation,
which has a complex geometry in the FHWA Demonstration Project However, the FHWA guideline (Elias and Christopher, 1997) did
82 (Elias and Christopher, 1997). In this FHWA design guideline, two not provide any method on how to consider the reduction of the
cases are considered based on the distance of two back-to-back or active thrust when D < Di. Elias et al. (2001) provided the following
opposing walls, D, as illustrated in Fig. 1. When D is greater than H recommendation for this intermediate configuration ‘‘the active
tan(45  f/2), full active thrust to the reinforced zone can be earth thrust may be linearly interpolated from the full active case to
mobilized and the walls can be designed independently, where H is zero’’. No justification was provided for this recommendation. A
the height of the walls and f is the friction angle of the backfill. For full-scale back-to-back geosynthetic-reinforced wall was con-
this case, the typical design method for MSE walls can be used. structed to evaluate the effect of the geosynthetic type on the
When D is equal to 0, two walls are still designed independently for internal deformation of the wall (Won and Kim, 2007). Due to the
internal stability but no active thrust to the reinforced zone is large distance of D ( ¼ 0.88H), no interaction was observed from
assumed from the backfill. In other words, no analysis for external two sides of walls. However, no investigation was conducted to
stability is needed. However, the later publication by Elias et al. evaluate the effect of the distance D. Therefore, an analysis is
(2001) indicated that an overlap of 0.3H is required for no active needed to evaluate the internal and external stability of back-to-
thrust. Both guidelines indicate that when D is less than H back MSE walls under a static condition.
tan(45  f/2), active thrust to the reinforced zone cannot be fully Both limit equilibrium and numerical methods have been
successfully used to evaluate the stability of MSE walls (for
example, Leshchinsky and Han, 2004; Han and Leshchinsky, 2006b;
* Corresponding author. Tel.: þ1 785 864 3714; fax: þ1 785 864 5631. Han and Leshchinsky, 2007) that yielded close results in terms of
E-mail address: [email protected] (J. Han). factors of safety and critical failure surfaces. Leshchinsky et al.

0266-1144/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.geotexmem.2009.09.012
 J. Han, D. Leshchinsky / Geotextiles and Geomembranes 28 (2010) 262–267 263

modified to include reinforcement as a horizontal force intersecting

Notations the slip circle, which is incorporated in ReSSA Version 3.0 software,
developed by ADAMA Engineering (2008). This modified formula-
Basic SI units are given in parentheses.
tion is consistent with the original formulation by Bishop (1955).
f friction angle of fill (degree)
The mobilized reinforcement strength at its intersection with the
n Poisson’s ratio
slip circle depends on its long-term strength, its rear-end pullout
D distance between the back of two opposing walls
capacity (or connection strength), and the soil strength. The anal-
(m)
ysis assumes that when the soil strength is reduced by a factor,
Di interaction distance defined by the FHWA design
a limit equilibrium state is achieved (i.e., the system is at the verge
guideline (m)
of failure). The slip circle for which the lowest factor (i.e., the largest
E Young’s modulus (MPa)
mobilized soil strength) exists is the critical slip surface for which
h height of reinforcement (m)
the factor of safety is rendered. Under this state, when the factor of
H height of wall (m)
safety at unity, the soil and reinforcement mobilize their respective
hz elevation of wall (m)
strengths simultaneously. ReSSA Version 3.0 also provides a factor
L reinforcement length (m)
of safety map, which identifies the critical zone based on the
ph lateral earth pressure behind reinforced zone (kPa)
criterion set up by the user.
Ph theoretical active Rankine lateral thrust (kN)
Tmax maximum tensile stress in reinforcement (kN/m)
2.2. Numerical method
Ta allowable tensile strength of reinforcement (kN/m)
W distance between two opposing wall facings (m)
The finite difference program, FLAC 2D Version 5.0 (Itasca
Consulting Group, Inc., 2006), was adopted in this study. A shear
Abbreviations
strength reduction technique proposed by Zienkiewicz et al. (1975)
FHWA Federal Highway Administration
was adopted in this program to solve for a factor of safety of
LE Limit Equilibrium
stability. In this technique, a series of trial factors of safety are used
MSE Mechanically Stabilized Earth
to adjust the cohesion, c and the friction angle, 4, of soil. Adjusted
cohesion and friction angle of soil layers are re-inputted in the
model for limit equilibrium analysis. The factor of safety is sought
when the specific adjusted cohesion and friction angle make the
(2009) demonstrated through shaking table tests that geocell- slope unstable from the verge of a stable condition (i.e., limit
reinforced earth walls under seismic loading failed in a rotational or equilibrium). Details about the slope stability analysis using the
translational mode, which could be modeled using a limit equi- FLAC software can be found in Dawson et al. (1999). The critical slip
librium method. In this study, the limit equilibrium and numerical surface often can be identified based on the contours of the
methods were also adopted to investigate the effect of the width to maximum shear strain rate.
height ratio of the wall (W/H) and the quality of backfill material on
the critical slip surface, the required tensile strength of reinforce- 3. Modeling
ment, and the active thrust to the reinforced zone. In addition, the
effect of the connection of reinforcements in the middle, when 3.1. Baseline case
back-to-back walls are close, was also investigated.
The geometry and material properties of the baseline model
2. Method of analysis used in this study are shown in Fig. 2. Since the factor of safety is
determined based on a state of yield, or verge of failure, it is
2.1. Limit equilibrium method insensitive to the selected elastic parameters: Young’s modulus (E)
and Poisson’s ratio (n) when using FLAC. If the system contains soils
Bishop’s simplified method, utilizing a circular arc slip surface, is with largely different elastic parameters, it will take longer time to
probably the most popular limit equilibrium method. Although solve the factor of safety; however, the effects on this factor would
Bishop’s method is not rigorous in a sense that it does not satisfy be small since it depends mainly on Mohr-Coulomb strength
horizontal force limit equilibrium, it is simple to apply and, in many parameters. Hence, constant values of E ¼ 100 MPa and n ¼ 0.3 were
practical problems, it yields results close to rigorous limit equilib- used for the reinforced fill in FLAC. High-strength parameters were
rium methods. In this study, Bishop’s simplified method was assumed for the foundation to prevent any possible failure. The
effect of wall facing cohesion on the required tensile strength of
reinforcement in the numerical analysis will be discussed in the
next section. Mohr-Coulomb failure criteria were used for strength

W = 12 m

4.2 m 4.2 m Blocks (γ = 18kN/m3 ,

c =1000kPa, φ = 340 )
Reinforced & retained fill
H = 6m 0.6m (γ = 18kN/m3 , c =0kPa,
φ = 340)
1.5m Reinforcement
Weak zone

1m Foundation (γ =18kN/m3 , c =1000kPa, φ =00 )

Fig. 1. Back-to-back MSE wall and definitions. Fig. 2. Dimensions and parameters of the baseline case.
264 J. Han, D. Leshchinsky / Geotextiles and Geomembranes 28 (2010) 262–267

between stacked blocks, the reinforced and retained fill, and the 4.2m 9.6m 4.2m
foundation soil. Reinforcement was modeled as a cable with grouted
interface properties between cable and soil. The bond strength
between reinforcement and reinforced fill was assumed equal to φ=34o
80% the fill strength, same as in the limit equilibrium analysis when φ=25o φ=34o
considering pullout resistance. A weak zone at the toe of the MSE 6m φ=25 o

wall with a dimension of 0.3 m wide and 0.4 m high, which has zero
Numerical
cohesion but the same friction angle as the fill, was assumed to LE
ensure that the critical failure surface passes through the toe of the
MSE wall. Details about the effect of wall facing cohesion will be
discussed later. Fig. 4. Critical failure surfaces within walls at W/H ¼ 3.
In this baseline case, the back-to-back wall width (W) to height
(H) ratio is equal to 2.0 and the distance at back of two walls, D is
cohesion of the wall facing was assumed to be 1000 kPa except
equal to 3.6 m, which is slightly greater than H tan(45  f/
a weak zone close to the toe.
2) ¼ 3.2 m. Based on the FHWA design guideline, a typical design
method for a single wall can be adopted. The reinforcement length,
4. Results
L ¼ 4.2 m, was selected based on the typical reinforcement length/
wall height ratio of 0.7 recommended by the FHWA design guide-
4.1. Critical failure surfaces
line. The height of MSE walls is fixed at 6 m.
Two important parameters, the back-to-back wall width to
The locations and shapes of critical failure surfaces of the back-
height ratio and the quality (i.e., friction angle) of backfill material,
to-back walls at different wall width to height ratios (W/H) were
were selected in this study to investigate their influence on the
determined based on the contours of shear strain rates in the
critical failure surface, the required tensile strength of reinforce-
numerical analysis and presented in Figs. 4–6. Fig. 4 shows that the
ment, and the active thrust to the reinforced zone. In addition to
critical failure surfaces in two opposing walls do not intercept each
W/H ¼ 2.0 for the baseline case, two other W/H ratios (1.4 and 3.0)
other, therefore, they behave independently. The critical failure
were used in the parametric study. One parameter in the baseline
surfaces by the LE method assuming only one side wall are also
was changed at a time while all others were unchanged. The same
shown in Fig. 4, and they have slightly steeper angles than those by
models were used in numerical and limit equilibrium analyses. The
the numerical method. This conclusion is consistent with that from
required tensile strength of reinforcement was determined to
the previous study conducted by the authors (Han and Leshchinsky,
ensure the factor of safety of the MSE wall equal to 1.0.
2006b). Since the LE method cannot analyze two side walls, which
become important when the distance between the walls gets
3.2. Effect of wall facing cohesion closer, the results presented below are based on the numerical
analysis unless noted.
The effect of wall facing cohesion was examined for the baseline Fig. 5 shows the critical failure surfaces within back-to-back
case in this study. As shown in Fig. 3, the factor of safety of the back- walls at W/H ¼ 2 that intercept each other from two sides. More
to-back MSE wall increases with an increase of the wall facing interactions occur for the case with a low-quality backfill. (i.e.,
cohesion. However, it becomes constant after the cohesion is 4 ¼ 25 ). For both cases, the critical failure surfaces do not enter the
greater than 100 kPa. In this case, the potential failure surface was reinforced zone on the opposing side. In other words, the potential
restricted to pass only through the toe of the MSE wall. Fig. 3 also failure surface is constrained by the reinforced zone on the
shows that the effect of the wall facing cohesion for the case with opposing side. The interaction distance, Di, based on the FHWA
low-quality backfill (4 ¼ 25 ) is more significant than that with design guideline can be determined using Equation (1). Consid-
high-quality backfill (4 ¼ 34 ). In all analyses discussed below, the ering the backfill with the friction angles of 25 and 34 , the
interaction distances are 0.64H and 0.53H, respectively. In other
words, when the wall width to height ratio, W/H > 2.04 for the
1.25
friction angle of the fill equal to 25 or W/H > 1.93 for the friction
angle of the fill equal to 34 , the two back-to-back walls should
1.2 perform independently. However, Fig. 5 shows that the back-to-
φ = 25o
back walls still interact each other when W/H ¼ 2.0 > 1.93 for the
Ta = 30.2kN/m
1.15
Factor of safety

4.2m 3.6m 4.2m

1.1

1.05
φ = 34o
Ta = 11.2kN/m
1
6m
0.95
φ = 34o
φ = 25o
0.9
0 200 400 600 800 1000
Cohesion of wall facing (kPa)

Fig. 3. Effect of wall facing cohesion. Fig. 5. Critical failure surfaces within walls at W/H ¼ 2.0.
 J. Han, D. Leshchinsky / Geotextiles and Geomembranes 28 (2010) 262–267 265

4.2m 4.2m connected when they meet in the middle of the walls. The
numerical results show the interactions of critical failure surfaces in
two opposing walls. In both cases, the failure surfaces enter the
reinforced zone from another side.
The comparisons of locations and shapes of critical failure
surfaces at different W/H ratios but the same quality of fill are
presented in Fig. 7. Fig. 7 shows that the locations and shapes of the
critical failures are almost same for W/H ¼ 3 and 2. This result can
6m be explained as the failure surfaces not entering the reinforced zone
on the opposing side. For W/H ¼ 1.4, however, the locations and
shapes of the critical failure surfaces deviate from others as the
failure surfaces enter the reinforced zone on the opposing side.

φ = 34o φ = 25o 4.2. Distribution of maximum tension with height

Fig. 6. Critical failure surfaces within walls at W/H ¼ 1.4. The distribution of the maximum tension, Tmax, on each rein-
forcement is plotted against the height of the reinforcement, h, for
both backfill materials in Fig. 8. These distributions were deter-
friction angle of the fill equal to 34 . Apparently, this assumption in mined from the numerical results based on the reinforcements
the FHWA design guideline is not supported by the numerical having equal strength and the MSE wall having a factor of safety of
result. However, the FHWA assumption leads to more conservative 1.0. These distributions are trapezoidal and similar to those
results. obtained by Han and Leshchinsky (2006a) using the limit equilib-
Fig. 6 shows the critical failure surfaces developed within the rium method. Fig. 8 shows that the maximum tension in the rein-
back-to-back walls when there is no retained fill between these two forcement in the MSE wall with low friction angle fill is much
walls (i.e., D ¼ 0 m). In both cases, reinforcement layers are not higher than that with high friction angle fill. The width to height
ratio of the back-to-back wall has a slight effect on the distribution
of the maximum tension in the reinforcement at height, i.e., the
a 4.2m overall maximum tension increases with the increase of the width
to height ratio. This result implies that it is slightly conservative to
ignore the influence of the width to height ratio on the overall
maximum tension in the internal stability analysis of the back-
to-back MSE wall. Fig. 8 shows that the overall maximum tension
φ = 34o occurs from the bottom to 1/3 or ½ height of the wall for the low or
high friction angle fill, respectively. It is noted that the maximum
W/H = 3 tension in the upper portion of the wall is slightly higher for
6m
W/H = 2 W/H ¼ 1.4 than W/H ¼ 2.0 and 3.0. This result can be explained by
the fact that the maximum tension in the reinforcement for
W/H = 1.4 W/H ¼ 1.4 is contributed by both sides of walls due to the inter-
ception of the failure surfaces in the upper portion.

φ = 34o

4.2m
b

φ = 25o
6m
W/H = 3
W/H = 2
W/H = 1.4

φ = 25o
Fig. 7. Critical failure surfaces at different W/H ratios. Fig. 8. Distribution of maximum tension in each reinforcement.
266 J. Han, D. Leshchinsky / Geotextiles and Geomembranes 28 (2010) 262–267

a 6

Elevation of wall, hz (m)

W/H = 1.4
4 W/H = 2.0
W/H = 3.0
Rankine
3

0
0 10 20 30 40
Lateral earth pressure behind reinforced zone, ph (kPa)

b 6

Fig. 9. Required maximum tensile strength of reinforcement.

5
4.3. Required tensile strength W/H = 1.4
Elevation of wall, hz (m)

4 W/H = 2.0
The required overall maximum tensile tension or strengths of W/H = 3.0
reinforcements for all the cases discussed above are presented in Rankine
Fig. 9. The results from the LE method were based on the analyses of 3
one side wall, therefore, no interaction of two opposing walls was
considered. In other words, the required tensile strengths do not
2
change with the W/H ratios. Fig. 9 clearly shows that a decrease of
W/H ratio from 3 to 1.4 reduces the required maximum tensile
strength of reinforcement. This result implies that the back-to-back 1
walls for both backfill materials still interact at W/H ranging from
2.0 to 3.0. The LE method without considering the interaction of the
0
opposing walls would provide conservative design of back-to-back
MSE walls. The difference in the maximum tensile strength of 0 10 20 30 40 50
reinforcement with and without considering the interaction is Lateral earth pressure behind reinforced zone, ph (kPa)
within 12% based on the cases investigated in this study. The
Fig. 10. Distribution of lateral earth pressure behind the reinforced zone.
required maximum tensile strengths can be used for the selection
of reinforcements (such as geosynthetics) in the back-to-back MSE
walls.

4.4. Lateral earth pressure behind the reinforced zone

Based on the FHWA guideline (Elias and Christopher, 1997; Elias

et al., 2001), the lateral earth pressure or active thrust behind the
reinforced zone for the external stability analysis should depend on
the width to height ratio. Fig. 10 presents the numerical results of
the lateral earth pressure, ph, behind the reinforced zone at
a different elevation of the wall, hz. It is clearly shown that the
lateral earth pressure exists behind the reinforced zone, even for
the width to height ratio of W/H ¼ 1.4 (i.e., no retained fill).
However, the FHWA guideline (Elias and Christopher, 1997) sug-
gested that the lateral earth pressure for external analysis should be
ignored if D ¼ 0 (i.e., W/H ¼ 1.4). Obviously, this suggestion would
yield an unsafe design. Fig. 10 also shows that the average lateral
earth pressure behind the reinforced zone is close to the active
Rankine lateral earth pressure when the width to height ratio is
large (for example, W/H ¼ 3.0). However, the lateral earth pressure
decreases when the width to height ratio decreases. The percent of
the active lateral thrust behind the reinforced zone to the theo-
retical active Rankine lateral thrust, Ph, is presented in Fig. 11, which
shows the influence of the backfill friction angle and the width to Fig. 11. Percent of lateral thrust behind the reinforced zone.
 J. Han, D. Leshchinsky / Geotextiles and Geomembranes 28 (2010) 262–267 267

This interaction will change the location and shape of critical failure
surface. The FHWA design guideline underestimates the interaction
distance. When the distance of the two back-to-back walls gets
closer, the required maximum tensile strength of reinforcement
slightly decreases. Ignoring the effect of the width to height ratio on
the overall maximum tensile strength of reinforcement results in
a slightly conservative design for internal stability. However,
ignoring the external lateral earth pressure on the reinforced zone,
when the distance of the back-to-back walls, D, equals to 0, results
in an unsafe design for external stability. The connection of rein-
forcements in the middle ensures the mobilization of the full
strength of the reinforcement and reduces the maximum required
tensile strength.

Acknowledgements

This paper is based upon the work supported by the National

Science Foundation under Grant No. 0442159. Any opinions, find-
ings, and conclusions or recommendations expressed in this paper
Fig. 12. Effect of reinforcement connection on the maximum tension. are those of the authors and do not necessarily reflect the views of
the National Science Foundation.

height ratio on the mobilization of the lateral thrust. When a low

friction angle backfill material is used, the percent of the lateral References
thrust is low. However, an increase of the width to height ratio from
ADAMA Engineering, Inc, 2008. ReSSA Software Version 3.0 Newark, Delaware, USA.
1.4 to 3.0 increases the lateral thrust behind the reinforced zone.
Bishop, A.W., 1955. The use of the slip circle in the stability analysis of slopes.
Geotechnique 5, 7–17.
4.5. Effect of reinforcement connection in the middle Dawson, E.M., Roth, W.H., Drescher, A., 1999. Slope stability analysis by strength
reduction. Geotechnique 49 (6), 835–840.
Elias, V., Christopher, B.R., 1997. Mechanically Stabilized Earth Walls and Reinforced
When the distance of two opposing walls, D, equals to 0, the Soil Slopes Design and Construction Guidelines Publication No. FHWA-SA-96–071.
reinforcements from both sides would meet in the middle. For an Elias, V., Christopher, B.R., Berg, R.R., 2001. Mechanically Stabilized Earth Walls and
ease of construction, these reinforcement layers are often not Reinforced Soil Slopes Design and Construction Guidelines Publication No.
FHWA-NHI-00–043.
connected. However, Fig. 12 shows that the connection of these Han, J., Leshchinsky, D., 2006a. General analytical framework for design of flexible
reinforcements reduces the overall maximum tension in the wall reinforced earth structures. ASCE Journal of Geotechnical and Geo-
because each reinforcement can mobilize its full strength when environmental Engineering 132 (11), 1427–1435.
Han, J., Leshchinsky, D., 2006b. Stability analyses of geosynthetic-reinforced earth
they are connected in the middle. The pullout from the middle of structures using limit equilibrium and numerical methods. In: Proc. of the 8th
the walls becomes impossible. The maximum tension in the rein- Int. Geosynthetics Conf., 18–22 Sept., Yokohama, Japan, pp. 1347–1350.
forcement decreasing with the height of the reinforcement for the Han, J., Leshchinsky, D., 2007. Stability analysis of back-to-back MSE walls. In: Proc..
of the 5th International Symposium on Earth Reinforcement (IS Kyushu’07),
unconnected case is because the mobilized tension in the rein- November 14–16, Fukuoka, Japan, pp. 487–490.
forcement is limited by its pullout capacity at a higher height close Itasca Consulting Group, Inc, 2006. FLAC5.0 User’s Guide Minneapolis.
to the top of the wall. Leshchinsky, D., Han, J., 2004. Geosynthetic reinforced multitiered walls. ASCE
Journal of Geotechnical and Geoenvironmental Engineering 130 (12), 1225–1235.
Leshchinsky, D., Ling, H.I., Wang, J.-P., Rosen, A., Mohri, Y., 2009. Equivalent seismic
5. Conclusions coefficient in geocell retention systems. Geotextiles and Geomembranes 27,
9–18.
The study using the numerical and limit equilibrium methods Won, M.-S., Kim, Y.-S., 2007. Internal deformation behavior of geosynthetic-rein-
forced soil walls. Geotextiles and Geomembranes 25, 10–22.
shows that two back-to-back walls perform independently when Zienkiewicz, O.C., Humpheson, C., Lewis, R.W., 1975. Associated and non-associated
they are far apart and interact with each other when they are close. visco-plasticity and plasticity in soil mechanics. Geotechnique 25 (4), 671–689.