26 TRANSPORTATION RESEARCH RECORD 115::

Design and Construction of Reinforced

Embankments Over Weal< Foundations
RUDOLPH BONAPARTE AND BARRY R. CHRISTOPHER

Experience with design and construction of reinforced em- bridge over the weak zones or voids. These zones or voids may
bankments over saturated clay foundations is reviewed. Rein- be caused by sinkholes, thawing ice, old stream beds, or
forcement materials considered are geotextiles and geogrids. pockets of silt, clay, or peat (Figure lb). In this category of
The effects of tensile reinforcement include increased embank- applications, tensile reinforcement may be required in more
ment stiffness and reduced shear stress and strain magnitudes
than one direction, and thus the strips or layers of reinforcing
and plastic deformations in the foundation. Analysis results
show that reinforcement reduces embankment settlement and material may be placed with varied orientations with respect to
lateral spreading due to undrained constant-volume distortion. the embankment centerline. Reinforcement design for the case
The conditions under which these performance improvements of an embankment over a void has typically been based on the
are significant are described. Limit equilibrium design pro- conservative assumption (7, 9, 10) that the reinforcement acts
cedures are discussed. Available information indicates that as a tensioned membrane supporting the full overburden pres-
modified classical stability procedures are suitable for rela- sure (Figure 2). The equations shown in Figure 2 are from
tively uniform clay foundations. Their applicability to peat
Giroud (9 ). T is the tensile force per unit width in the reinforce-
foundations appears to be more limited. Aspects of limit equi-
librium analyses specific to use of reinforcement are discussed. ment; an approximate value for the reinforcement stiffness
These include (a) effect of the reinforcement force, (b) orienta- required to span a circular void of diameter b can be obtained
tion of reinforcement force, (c) selection of reinforcement force by dividing K by 2. This second category of applications will
for design, and (d) reinforcement embedment length. Simple not be discussed further.
design charts and figures presented can be used to make a
preliminary assessment of overall factors of safety against MECHANISMS OF REINFORCEMENT FOR
foundation bearing capacity and slip surface failures and lat-
EMBANKMENTS OVER WEAK FOUNDATION
eral sliding of the embankment. Last, construction aspects are
described. LAYERS

Tensile reinforcing elements may be used to increase the sta- The mechanism of reinforcement for an embankment con-
bility of embankments constructed over weak foundations. In structed over a uniform deposit of saturated clay is to stiffen the
this type of application, horizontal strips or layers of reinforce- base of the embankment and reduce shear stress magnitudes
ment are placed on the natural soil or within the base of the and plastic shear deformations in the fowidation. This mecha-
embankment with the remainder of the embankment con- nism is illustrated in the results from a recent investigation by
structed in the conventional manner. Materials used as tensile Low and Dunca.1 (11), who used finite element a.1alyses that
reinforcement include steel strips, bars, or meshes (1 ), geotex- included nonlinear (hyperbolic) soil stress-strain behavior.
tiles (2-5), geogrids (6, 7), and other materials. In this paper, Their analyses showed that tensile reinforcement placed at or
the use of polymer-based reinforcing elements such as geotex- near the base of an embankment increased the stiffness of the
tiles and geogrids is addressed. The use of steel reinforcement embankment fill. This increase was proportional to the tensile
is described by Duncan et al. in another paper in this Record. stiffness K (reinforcement tensile modulus multiplied by rein-
Reinforced embankments over weak foundations typically forcement cross sectional area per unit width) of the reinforce-
fall into one of two categories (8). The more common category ment. Only part of the mobilized tensile force went into stiffen-
consists of embankments, dikes, or levees constructed over ing the embankment fill. The rest of the tensile force was
soft, saturated silt, clay, or peat layers (Figure la). In this transferred from the stiff reinforced fill to the less stiff founda-
category, the reinforcement is typically placed with its strong tion soil below the reinforced zone. The average shear stress
direction perpendicular to the centerline of the embankment, and strain levels in the foundation were thereby reduced.
and plane-strain conditions are assumed to prevail. Additional The combined effects of increased fill stiffness and reduced
reinforcement with its strong direction oriented parallel to the shear stresses and strains in the foundation soil include reduced
centerline may also be required at the ends of the embankment. undrained (constant-volume) foundation distortion beneath the
The second category of reinforced embankment applications embankment as well as reduced embankment spreading and
consists of those in which the foundation below the embank- initial (undrained) embankment settlements. These effects are
ment is locally weak and the role of the reinforcement is to illustrated in Figure 3 for an embankment on a saturated clay
foundation with an umeinforced factor of safety (FS) of just
R. Bonaparte, GeoServices Inc., Consulting Engineers, 5950 Live Oak less than 1.0 and a ratio of foundation depth to crest width
Pkwy., Norcross, Ga. 30093. B. R. Christopher, STS Consultants Ltd., (DI B) equal to 0.36. [This DI B ratio and low unreinforced FS
11 Pfingsten Rd., Northbrook, Ill. 60062. were selected to highlight the effect of reinforcement. A larger
 a.

r WEAK FOUNDATION ____./

FIRM FOUNDATION

b.

E~l\~K~ENT
~-~~~ru.~~~.._:::~- -~--
- --~~~-~----- -----
LOCALLY WEAK ZONE SINKHOLE
FIGURE 1 Embankments on weak foundations: (a) embankment on
uniform weak foundation; (b) embankment on locally weak foundation with
lenses of clay or peat, or with sinkholes.

£
c
- [ 2 f( t ) . -1
SI n (-1_)_ 1
( 1BO111") 21 ( t lJ

K
0.06
( kN/ ml
(, .0%)

0 .0 B
4 ,00 0 ( 1 .7 % )

3,000

2 ,ODO

1,000

( 5.9%)

10 20 30 40 pb < kNl

FIGURE 2 Tensioned membrane analysis of an embankment spanning

an Infinitely long void of width b.
28 TRANSPORTATION RESEARCH RECORD 1153

a. FOUNDATION
MATTRESS 1,-.·-~28'

H
( 1 2·)

3.3 FT. IL _
HORIZONTAL -
FOUNDATION 0 i' 2' 3' 4' D
•-+-f--1-------l
DISPLACEMENT ( 20')
'Y - 11 Opel
Cu - 250psl
~u - 0

MID SLOPE

b. Top g
Toe of Slope Ill

~
+ i 2
_c;
c: ::::!'
~- :::
p 00 eo
~E :i::
-o
-;v::
(/) 0
<l
m
c

fvl
c. _.,,

;~
~-----~-I
- - - - - .- _ _

___ . ,,,_,. c
~~ =
~:'_
_ - -.":·;Y

............... ...... ... ........ .. ................

·~--.
.0 00 -
Ex
~ o

o2 <l
-2.0 N ~
· ~(I)
o_
Io

d. 10 Ill
Ill
~
B 0
::::!'
c:
K(k/fl) 6 c: ~

- · - 23
0

.. . .... 70 /
/ -- .-·-·
....·······
·~ ~

4 cii w"
----210
v ..· . • ---- 0
c:0
2 N
...
0
:i::
T a
FIGURE 3 Behavior of embankments with varying reinforcement
stiffnesses constructed over a uniform weak foundation layer.

unreinforced FS and larger or smaller D/B would show a 1-m-thick geogrid-reinforced zone that was assumed to have
smaller influence of the reinforcement. Rowe and Soderman uniform tensile stiffness. Low and Duncan (11) also showed
(12) discuss the influence of D/B.] The results in Figure 3 were that the influence of this mattress on embankment performance
predicted by Low and Duncan (11), who used nonlinear finite is roughly equivalent to that of a single layer of reinforcement
element analysis to analyze the embankment-foundation mat- of similar tensile stiffness located at the mid-height elevation of
iress system shown in the figure. The mattress consisted of a the mattress. On the basis of Low and Duncan's work, as well
Bonaparte and Christopher 29

as that of Rowe and his coworkers (12-17), and others [sum- Mechanisms of Failure
marized by Christopher and Holtz (18)], a number of general
observations can be made with respect to the beneficial effects Published case studies (11, 13, 16) have shown that nonlinear
of reinforcement placed at or near the base of embankments finite element analyses can be used to evaluate the load-defor-
built over weak foundation layers: mation response of embankments built over uniform weak
1. Tensile reinforcement reduces displacements beneath the foundations. However, the large majority of projects are de-
embankment centerline and heave near the embankment toe signed using limit equilibrium procedures that evaluate a num-
caused by undrained, constant-volume deformation. Tensile ber of idealized failure mechanisms. The three failure mecha-
nisms most often considered are (18, 21, 22)
reinforcement also reduces vertical and horizontal displace-
ments in the embankment.
2. By reducing the magnitude of shear deformations at the • Overall bearing capacity failure of the foundation that
foundation-embankment interface, reinforcement decreases the may occur if the foundation is so weak that it cannot support
average shear stress and shear strain magnitudes and the extent the weight of the embankment;
of the plastic zone in the foundation. • Lateral sliding of a portion of the embankment that may
3. The improvements described in embankment perfor- occur along the embankment-reinforcement interface, along
mance increase with increasing mobilized reinforcement force. the foundation-reinforcement interface, or along a shallow,
For a given embankment, the mobilized reinforcement force weak seam or layer in the foundation soil; and
increases with increasing reinforcement tensile stiffness and • Slip surface failure through the embankment and foun-
decreasing foundation soil modulus (strength). The perfor- dation.
mance improvements can be significant for embankments with
unreinforced factors of safety less than one (Figure 3) but
Overall Bearing Capacity
decrease for unreinforced FS values much above 1. All other
factors being equal, the mobilized reinforcement force in-
creases with increasing DI B, up to DI B approximately equal to A simplified analysis to use in calculating the factor of safety
0.4. For deep deposits with DIB greater than about 0.8, the against bearing capacity failure is shown in Figure 4 [after
mobilized reinforcement force will be small (12). Mandel and Salencon (23, 24), originally published by Bo-
4. Tensile reinforcement may increase the height to which naparte et al. (8)]. Other bearing capacity analyses have also
many types of embankments can be constructed without induc- been discussed by Bonaparte et al. (8).
ing a foundation failure. Alternatively, for a given embankment If the embankment bearing capacity factor of safety is less
height, tensile reinforcement increases the factor of safety than 1.0, the embankment cannot be constructed in the conven-
against foundation failure. tional manner without inducing foundation failure. Sometimes,
5. The reduction in shear stress and strain magnitudes in the however, embankments are built over very weak sites with
foundation due to reinforcement is largest at shallow depths. A bearing capacity factors of safety less than 1.0. In these cases,
decrease in shear stress magnitude at shallow depths is impor- the fill sinks into the soft ground, displacing the foundation
tant for sites at which a desiccated crust is underlain by satu- material. This displacement method of construction has its
rated clay that exhibits a normally consolidated strength pro- origins in nonreinforced embankment construction (20). The
file. In these cases, shear stresses will be reduced in the zone of displacement method is sometimes used with geotextile rein-
minimum shear strength just below the crust and the failure forcement to reduce the required fill volume. Geotextiles are
surface will be forced deeper into stronger soil (19, 20). beneficial in this case because they replace uncontained local
failures of the embankment with more uniform sinking of the
The reinforcement does nothing to increase the strength of the entire embankment. Under these circumstances, geotextile
foundation soil. Therefore the foundation soil must have ade- strains can be very large. Strains in excess of 30 percent have
quate strength to support the entire reinforced embankment. If been measured (4). Low-modulus, high-elongation geotextile
the embankment is made very stiff through sufficient reinforce- products may be considered for displacement applications
ment, it may behave as a semirigid mat, and the critical failure along with other design measures such as fiat side slopes,
mechanism becomes one of bearing capacity of the entire berms, staged construction, and wick drains.
embankment. At that point, additional reinforcement will not
further increase embankment stability. Also, reinforcement
does not significantly reduce overall embankment settlement Lateral Embankment Sliding
owing to consolidation of the foundation soil.
A simplified analysis to use in calculating the factor of safety
DESIGN OF REINFORCED EMBANKMENTS
OVER WEAK FOUNDATION LAYERS against failure due to lateral embankment sliding is shown in
Figure 5. The important reinforcement properties for design
This section of the paper deals solely with the end-of-con- against sliding are the soil reinforcement interface friction or
struction design of reinforced embankments over weak founda- adhesion characteristics, determined from direct shear tests and
tion layers. Locally weak foundations are not addressed. Pri- a limiting reinforcement tensile force per unit width, T (in kN/
mary emphasis is on uniform, purely cohesive foundation m); in Figure 1, A. 1 tan <!>' is the embankment fill-reinforcement
deposits, although some reference is made to peats. The reader interface friction, and A.z is the reinforcement-foundation inter-
is referred to the work of Rowe et al. for peats (12-17). face adhesion. To control embankment cracking, T is usually
30 TRANSPORTATION RESEARCH RECORD 1153

8
F.S . -~
I' H

E-z -r "\
8
D I rough
=i?(\iJIA't""".17.t:r Ji'\£ 4$\t -Jri<JJ<CJ-.•

0 0.61 2 4 6 8 B IO

FIGURE 4 Bearing capacity failure of embankment.

Q,

~~~~~~~~~~--.
r b

EMBANKMENT bA1 Ian 11·

ti' KaH
c'• O

"'--- FOUNDATION
l'lu • 0

-
Cu

b.

EMBANKMENT
H l!'
c'• 0
'Y

- - - FOUNDATION
0u - 0
Cu

~~~A"Vff.W/JWA'%.~~

FIGURE S Lateral sliding of embankment: (a) embankment sliding

over reinforcement; (b) reinforcement tensile rupture and embankment
sliding over foundation soil.

selected on the basis of a limiting strain criterion. Bonaparte et location of its intersection with a considered slip surface. Al-
al. (8) suggest a strain limit of not more than 5 percent for though this approach is attractive because of its simplicity and
embankments constructed with cohesionless fill and not more connection to classical design, it involves a number of arbitrary
than 2 percent for those constructed with cohesive fill. assumptions whose verification through comparison with field
performance or detailed numerical studies is lacking. Back-
Slip Surface Failure Through analyses suggest limited applicability of modified classical
Embankment and Foundation stability analyses to embankments over peat foundations
(14, 15). Back-analyses of embankments over relatively uni-
The reinforcement tensile force required to increase the factor form saturated cohesive foundations appear to have given bet-
of safety against slip surface failure through the embankment ter results (6, 11, 16).
and foundation can be estimated by using modified classical Simplified design charts based on limit equilibrium analyses
limit equilibrium stability analyses. Usually it is assumed that have been presented for embankments built on saturated clay
the reinforcement prnvides a stabilizing tensile force at the foundations by Fowler (25), Gourc (26), Ingold (27), and
Bonaparte and Christopher 31

Milligan and Busbridge (28). Rowe and Soderman (13, 17) Effect of Reinforcement Force
presented charts for peat foundations based on finite element
analysis results. The charts prepared by Milligan and Bus- The reinforcement force can be assumed to have one of two
bridge (28), shown in Figure 6, are conservative and useful. different effects on stability: (a) it can act as a boundary free-
These charts show the required reinforcement force per unit body tensile force that does not affect soil strength but that
width, T, to obtain a state of limit equilibrium (FS = 1). To contributes to force and moment equilibrium; or (b) it can
obtain larger factors of safety, one uses the charts with factored modify the strength of the embankment fill. The second effect
soil strengths (tan $/ =
can $'IFS; Cu/= CufFS). These charts can exist only if the reinforcement force is properly transmitted
were developed on lhe basis of (a) moment equilibriwn along to the embankment fill .
the critical circular arc through the foundation and Coulomb The application of the reinforcement force as a boundary
wedge through the embankment, and (b) horizontal force equi- free-body tensile force that does not affect soil strength is
librium along a critical multipart wedge. The latter equilibrium shown in Figure 7a for the case of moment equilibrium along a
condition was found to control for ratios of foundation depth to circular slip surface. In Figure 7, MT is the stabilizing moment
embankment height of less than about 0.5. These charts are due to the reinforcement force. This approach, which neglects
conservative (they overpredicted tensile force by 10 to 30 any effect of the reinforcement force on soil strength, has been
percent) when compared to computer solutions that use the the one most commonly used in the past.
stability analyses discussed subsequently. The charts are An approach assuming that the reinforcement force modifies
presented here as a simple means to obtain a preliminary soil strength is shown in Figure 7b for the case of moment
indication of the influence of tensile reinforcement on a given equilibrium along a circular slip surface. In this approach,
design. Usually, however, real design problems require more which was first proposed by Wager (29), the reinforcement
detailed analyses using computer-based analyses, as presented tensile force is decomposed into vector components normal and
later. tangent to the slip surface. The component of force parallel to
To include the influence of tensile reinforcement into limit the slip surface is assumed to provide a pseudo-cohesion that
equilibrium stability analyses requires assumptions regarding acts in addition to any soil cohesion. The component of force
the normal to the slip surface is assumed to increase the normal
stress acting on the soil and thereby increase its shearing
• Effect of the reinforcement force; resistance due to the frictional component of shear strength
• Orientation of the reinforcement force at the location of (which is assumed to be fully mobilized).
the considered slip surface; A comparison of the normalized stabilizing moment MT/RT
• Magnitude of the reinforcement force to ensure strain associated with the two different assumptions regarding
compatibility of the reinforced system at failure; and the effect of the reinforcing force is shown in Figure 8a.
• Reinforcement embedment length. The comparison is made for two different reinforcement

.. b. c.
2.0
0 .14 1-,-;;----.---..--------.--.----.
1.e ;:- 0 . 12 \--~-:~. :.-~-0 +' : _ _ +-_ __L_ __JL..__~
l. b
/3· 18.4°(3:1) [T/~H 2 )(~'>"

s·- 30• oa: 0. 10 ~-1---'>.·~--

[1 / ~ H 2 l[ ~'.w ) + [F~]
1.4
lJ o.oe 0
--'-,·--'ll. ---1--~ --~-_..,
12 :t. O.Ob ii ) l.O '

:I: 1.0 :I: z

0
c c 0 . 04 l-4----+---'"-.--><--- - 1 - - - 1 - - --"

lil~ 0.02 1-+---l----l"-.::-.--l---I---~

~ 0 1'\'.--- • --=-!,.--~>h---..,....,_ _t..,,__

2=±07" ••
0
t.' -0 . 02 t--t---1---r---1-~~t---!

0.02 O.C>< O.Ob 0.06 0.10 0 .12 O.'" O. lb 0.16 0.20

d.

D FOUNDATION
( Cu )

FIGURE 6 Stability charts for the design of embankments on weak foundations [after Milligan and Busbridge (27)].
 ..
T
__ c::_ ~~l~ORCEMENT

Mr· T· [ R cos ((j -/Jl I

0 <; ~ <. (}
b.

M~· Tp R •Tn tan Jl'·R

- RT [cos ( (} -fJl+ •In <8-{3lt11n 11'1
- Mr I 1 + tan < 8 - (3 >tan .fl'' I
FIGURE 7 Effect of reinforcement force: (a) reinforcement force
assumed to act as an independent free-body force that does not affect
soil strength; or (b) reinforcement force assumed to increase soil
strength.

a.
1.5

Mr/RT
___ ----------
.----
....... ,,.,-"'
--
--.....
_,..,- ...........
.......... ........ .........
1.0
.;;::::.------ .........

If· 40°; {3· o

~· - 30°;(3 - o
0.5 MT /RT <FIG.7al

- - - - - Mr /RT <FIG.7bl If - 20°; (3 = o

0
0 10 20 30 40 50 60 70 80 90 (}!degrees I
b.
1.5
II'· 30°
Mr/RT

, .0 -----
0 .5
Mr/RT (FIG.7al

Mr /RT (FIG .7bl

0
0 1O 20 30 40 50 60 70 (3 lde~eesl

FIGURE 8 Stabilizing moments calculated using different assumptions from

Figure 7: (a) comparison of moments calculated; (b) inHuence of reinforcement
orientation.
Bonapar/e and Christopher 33

orientations: horizontal (~ = 0) and tangent to the slip surface factor of safety. The sensitivity of the results shown in Figures
(~ = 0). For the case of horizontal reinforcement and a granular 7-9 makes selection of reasonable assumptions critical if anal-
=
fill with <l>' 30°, the assumption that the reinforcement modi- yses are to provide reasonable results. The review of available
fies the soil strength results in a larger stabilizing moment than information that follows is intended to help the design engineer
that calculated under the assumption that the reinforcement acts select reasonable assumptions.
as an independent force that does not affect soil strength. As the The approach shown in Figure 7b of assuming that the
orientation of the reinforcement approaches the tangent to the reinforcement modifies soil strength is applicable to direct
slip surface (as ~ approaches 0), the stabilizing moments calcu- shear tests on reinforced sand specimens (31) and may be
lated by using the two assumptions converge. considered for structures in which the entire soil mass is
strengthened with regularly spaced multiple layers of reinforce-
Orientation of Reinforcement Force ment (8). On the basis of the authors' experiences, some
strength gain may be considered for low embankments with
The assumed orientation of the reinforcement force at its inter- multiple layers of reinforcement. For a single layer of rein-
section with the slip surface has an effect on the results of force forcement, however, this approach is uncertain because the
and moment equilibrium calculations. Almost all reinforce- application of the normal force is localized near the embank-
ment is initially placed with a horizontal orientation. Most ment-foundation interface. As previously discussed, some of
frequently, calculations are carried out by assuming that this the reinforcement force is transmitted to the foundation and is
initial orientation remains unchanged. It can also be assumed, not available to strengthen the embankment fill. Therefore, for
however, that the initially horizontal strip or layer of flexible a single layer of reinforcement the reinforcement tensile force
reinforcement bends because of large local displacements of should be modeled as a boundary free-body force that does not
the foundation soils at the onset of failure. The maximum affect soil strength (Figure 7a).
possible amount of reinforcement reorientation would result in
Evidence for selection of a reinforcement orientation is
a reinforcement direction tangent to the slip surface. Fowler
skimpy. Rowe and Soderman (16) used an analysis based on
(4, 25), for instance, assumed parallel reinforcement orienta-
BMM to predict the height at which a geotextile-reinforced
tion. Reinforcement orientations between the two extreme
sand embankment would fail. The embankment, located at
values (0 < ~ < 0 in Figure 7) would theoretically be possible.
Almere, Netherlands (3), was constructed over a 3.0- lo 4.5-
A comparison of stabilizing moments calculated by using m-thick clay-peat deposit. The tensile force in the reinforce-
different reinforcement orientations is shown in Figure 8b for ment at failure was estimated from field strain measurements.
two different values of the angle e. The influence of reinforce- Rowe and Soderman (12) found that the factors of safety
ment orientation on the magnitude of the stabilizing moment is
obtained with horizontal and tangential reinforcement orienta-
obvious. This influence is larger for large values of e.
tion bounded the actual value of 1.0. Subsequently, Low and
The practical significance of the effects shown in Figure 8 Duncan (11) also used a BMM analysis with both horizontal
can be seen in Figure 9, which shows the influence of rein- and tangential reinforcement orientation (32) and obtained re-
forcement orientation on the calculated factor of safety for a sults similar to those of Rowe and Soderman. On the basis of
range of typical conditions. In Figure 9, the reinforcement is the back-analyses of the Almere test fill, Low and Duncan
assumed to act as an independent free-body force that does not suggested that for low-sensitivity saturated clay foundations,
influence soil strength (Figure 7a). The calculated factor of the end-of-construction factor of safety could be approximated
safety was obtained by using Bishop's modified method (30, by averaging those obtained by using BMM with horizontal
hereafter called BMM) of stability analysis and the reinforce- and tangential reinforcement orientation. Rowe and Soderman
ment effect shown in Figure 7a. The figure shows the larger (12) suggest that for D !B less than about 0.4 the factor of
factor of safety computed based on tangential reinforcement safety could be calculated by using BMM along with a rein-
orientation (FS T) versus that based on horizontal reinforcement forcement moment arm equal to the average of those for hori-
orientation (FSu). Figure 10 shows the combined influences of zontal and tangential orientation. For DI B greater than about
the depth of the critical slip circle (foundation depth for the 0.4, they conservatively suggest horizontal orientation.
case analyzed) and reinforcement orientation for the embank-
Busbridge et al. (6) used an analysis based on BMM to
ment shown in Figure 9 and T = 67 kN/m (6,000 lb/ft).
predict the performance of both a reinforced and an unrein-
forced embankment built over highly sensitive Champlain Sea
Discussion of Effect and Orientation of clay in eastern Canada. The soil conditions at the site on which
Reinforcement Force the test embankments were built are well documented (23). An
analysis of the failure of the unreinforced embankment was
The differences in factor of safety associated with the different used to calibrate the input parameters for analyses of the
assumptions regarding the effect and orientation of the rein- reinforced structure. On the basis of a horizontal reinforcement
forcement force are important because often the role of the orientation and the actual failure height of the embankment, the
reinforcement is to provide short-term stability of the embank- analyses predicted a tensile force in the two layers of geogrid
ment-foundation system. It is not uncommon for the unrein- reinforcement at failure of 123 kN/m (8,400 lb/ft). According
forced factor of safety to be less than 1.0 and the reinforced to Busbridge et al. (6), "cracking noises were heard in the
end-of-construction (undrained) factor of safety to be only embankment moments before failure, and this is inferred to
slightly greater than 1.0 (1.1 to 1.2). Thus the engineer may be indicate tensile breakage of the geogrids." On the basis of
using reinforcement to provide relatively small increments in constant strain-rate testing by McGown et al. (34) on the
34 TRANSPORTATION RESEARCH RECORD 1153

EMBANKMENT
J'- 35', 'Y· 18.8kN/m3I120pcfl

FOUNDATION
D (VARIABLE! Cu• 12kPa (250 plfl
lu - O

1.3 /

/
/ "
/

"
1.2 /
/ ""
/
>-
.... D/H - 1.0~// "
/ "~
UJ
""<
I/)
, /

1.1
""0 / "
a: / "
""
0
t5 /.
<
"" 1.0

HORIZONTAL

TANGENTIAL

0.9
0 29 58 87 116 ( kNlml
( 20001 ( 40001 ( 60001 ( 80001 (lb/ft)

REINFORCEMENT FORCE , T

FIGURE 9 Influence of assumed reinforcement orientation on calculated

factor of safety for typical embankment reinforcement application.

geogrid product used in the field trials, the range of possible In an embankment application, reinforcement elongation
break loads for two layers of reinforcement are estimated to be may be induced as a result of the following (8):
110 to 140 kN/m (7,500 to 9,600 lb/ft). Using the same location
• Placement of reinforcement and establishment of con-
and radius of the critical circle found by Busbridge et al. (6),
struction working pad;
and a tangential reinforcement orientation, the authors calcu-
• Undrained constant-volume distortion of the foundation
lated a reinforcement force at a factor of safety of 1.0 equal to
soil during and after embankment construction;
approximately 60 kN/m (4,100 lb/ft), which is too low. On the
basis of this result, for brittle, strain-sensitive foundations, a • Localized embankment deformations associated with the
horizontal reinforcement orientation would appear to be appro- development of a slip surface at the onset of failure; and
priate. • Settlement due to consolidation of the foundation soil.
For relatively stiff reinforcement [secant tensile stiffnesses of
Reinforcement Tensile Force at Failure at least 250 kN/m (17,000 lb/ft) at a strain of 2 percent], strains
due to reinforcement placement and establishment of a con-
Selection of a limiting value of reinforcement tensile force is a struction working pad would be expected to be small, certainly
key step preceding stability calculations. The magnitude of this not more than 1 or 2 percent. These small strains are in contrast
force should depend on the deformation at failure of the em- to the large strains that have been observed when lightweight
bankment-foundation system and on the force-elongation be- nonwoven geotextiles have been used with foundation dis-
havior of the reinforcement, including reinforcement creep. placement construction methods. The latter construction alter-
Determination of the force-elongation relationship for geotex- native is not considered here.
tile and geogrid reinforcement is discussed in detail elsewhere The tensile strain in reinforcement due to undrained con-
(8, 18, 35, 36). Reinforcement elongation due to deformation stant-volume distortion has been investigated by using non-
of the embankment-foundation system is discussed next. linear finite element analyses by Boutrup and Holtz (37), Rowe
Bonaparte and Christopher 35

0.5 3.5

Cf)

u.:
<l
0.4 3.0
>-
~
w
LL
<
Cf)

LL ~
0.3 2.5 Cf)
0
a:
u.:
0
~
(.)
< ----
-<l
I-
Cf)
LL 0.2 .6F.S.H ,,... 2.0 u.:
z / "" ~
/
w I
(!) I
z I
< 0.1
I
1 .5
J: ~l:;.F.S.T
(.) ,I ---
.6. F.S.H

0 .__ _ _ _ _ _ _ _ _ _ _ _ _.,.._ _ _ _• 1.0

2 3

D/H
FIGURE 10 Comparison of changes in factor of safety for reinforcement with
horizontal (t:.FSn) and tangential (t:.FSr) orientation for embankment shown in
Figure 9.

and Soderman (12), and Low and Duncan (33). For a given this embankment was estimated to be 1.15. In this case, a 225-
embankment height and geometry, the mobilized reinforcement g/m2 woven geotextile was used, and although it enabled
strain was found to be primarily dependent on the foundation staged embankment construction, settlements were large, ap-
stiffness and depth and on the reinforcement stiffness. The proaching 80 percent of the total height of fill. A 730-g/m2
reinforcement strain varied between about 1 and 8 percent. woven geotextile was used over a more compressible section of
Field data from the fills constructed by Busbridge et al. (6) the same site where the peat had an average water content of
indicated prefailure reinforcement strains in the range 0 to 2 785 percent by weight. The unreinforced factor of safety for the
percent for the relatively stiff conditions (geogrid reinforce- embankment was estimated to be about 1.05. Settlements were
ment and sensitive brittle foundation clay) existing in their again large. Maximum recorded reinforcement strains were on
tests. Tavenas et al. (38) analyzed available data for embank- the order of 4 percent.
ments on sensitive clay foundations and found that for end-of- A number of references exist that describe reinforcement
construction conditions and factors of safety in the range 1.2 to rupture (3, 6, 7). At the onset of failure, reinforcement strains
1.3, maximum lateral strains in the foundation were only a few may become large in the vicinity of the failure surface. The
percent for embankments with maximum vertical settlements
additional reinforcement tensile force generated by this elonga-
of up to 10 percent of the embankment height.
tion may or may not restore stability, depending on the stiffness
Fowler (4) estimated approximately 4 percent reinforcement
of the reinforcement and on the strain-softening characteristics
strain on the basis of field displacement measurements for a
of the foundation. Because most foundations are at least
flat-sided (lOH:lV) dike built over very weak, nonsensitive
slightly strain softening, the reinforcement tensile force in-
foundation clays located in an intertidal zone adjacent to
crease associated with this increment of strain should probably
Mobile Harbor. The design factor of safety for this case was in
the range 1.1to1.2. The reinforcement was a 730-g!m2 woven be neglected.
geotextile. The recorded reinforcement strain at incipient Foundation consolidation does not usually result in signifi-
failure in the Almere test fill was approximately 4 to 6 percent cant lateral embankment deformation. Consolidation also leads
(3 ). At Almere, the foundation soils were clays and peats, and to a strengthening of the foundation soil and a decreased need
the reinforcement was a 450 g/m2 woven geotextile. Barsvary for reinforcement. Therefore, reinforcement strains induced by
et al. (2) and Rowe et al. (14) report reinforcement strains on consolidation are usually not significant from a design stand-
the order of 20 percent for a 1.5-m embankment built over a 6- point. A few exceptions to this conclusion do exist, and in these
to 7-m-thick peat deposit with an initial average water content cases the time-dependent reinforcement force-strain response
of 445 percent by weight. The unreinforced factor of safety for should be considered. Exceptions include embankments over
36 TRANSPORTATION RESEARCH RECORD 1153

peat bogs or other deposits with nonuniform thicknesses that available force. Under the center of the embankment, the limit
may result in large localized reinforcement strains. reinforcement tensile force will control.
On the basis of this discussion and review of available
information for highly sensitive, brittle clay foundations, the Longitudinal Reinforcement Force
reinforcement tensile force for design against slip surface foun-
dation failure should be based on a limiting strain of not more The preceding discussion was directed solely at reinforcement
than 2 to 3 percent. Limiting strains of not more than 4 to 6 forces and strains in the direction perpendicular to the embank-
percent should be considered for medium- to low-sensitivity ment centerline. Reinforcement forces and strains can also be
clay foundations. If the foundation soils are nonsensitive and developed parallel to the embankment centerline. The potential
plastic, the reinforcing force should be based on limiting strains for longitudinal reinforcement forces and strains may occur (a)
of not more than 10 percent. during construction over very weak sites prone to mud waving;
Two additional factors that should be considered in selecting (b) at the ends of an embankment; and (c) owing to differential
a reinforcement tensile force for design are (a) the strain to settlements and bending of embankments built over non-
limit lateral embankment deformations to an acceptable level, uniform foundation conditions. Barsvary et al. (2) measured
and (b) the strain to limit creep rupture of the reinforcement. reinforcement strains on the order of 8 percent under an em-
The first factor may be important if cohesive embankment fills bankment built over a peat bog that had variable depth. When
are used. In this case, the primary benefit of reinforcement may these conditions prevail, longitudinal reinforcement forces and
be to prevent embankment cracking (11). To prevent cracking, strains should be considered during design.
mobilized reinforcement strains should be small, as noted in
the previous discussion. The second factor may become impor-
Other Factors
tant if the anticipated reinforcement strains are larger than the
strain that would result in reinforcement creep.
The discussion has focused on reinforcement aspects of limit
equilibrium analyses of reinforced embankments. It should be
Reinforcement Embedment Length remembered that other factors affect the stability of unrein-
forced and reinforced embankments alike. The geotechnical
The previous discussion was concerned with selection of a literature relating to unreinforced embankment stability should
maximum reinforcement force under the embankment. This be consulted to evaluate these other factors, which include (a)
maximum force can only be generated if the reinforcement has
foundation strength details, including the presence of a crust,
adequate embedment beyond the failure surface. The embed-
strength changes with depth, or thin seams of silts and fine
ment length required to mobilize a given reinforcement tensile
sands; (b) embankment fill properties, including undrained
force is dependent on the embankment-reinforcement bond, the
embankment cohesion; (c) embankment cracking, which may
foundation-reinforcement bond, or both, as well as on the
be reduced in a reinforced embankment compared to an unrein-
embankment overburden pressure. For most new embankment
forced embankment; (d) progressive failure effects; and (e)
construction, embedment length requirements are automat-
time-dependent foundation changes, including creep and
ically satisfied by spanning the reinforcement from toe to toe.
consolidation.
A length check should be made, however, for embankment-
widening or embankment-raising projects when reinforcement
lengths are limited, when the reinforcement does not span the CONSTRUCTION OF EMBANKMENTS OVER
entire embankment, or when very stiff [K greater than about WEAK FOUNDATIONS
1,000 kN/m (68,000 lb/ft)] reinforcement is used.
An embedded geotextile or geogrid resists pullout through Selection and implementation of appropriate construction pro-
friction and adhesion on its upper and lower surfaces, and by cedures is of critical importance for reinforced embankments
passive resistance developed by elements perpendicular to the over weak foundations, for at least two reasons: (a) a potential
direction of reinforcement. Average interface friction and adhe- exists for embankment failure during construction if con-
sion coefficients can be defined from the results of direct shear struction sequencing and procedures are not carefully planned;
or pullout tests for geotextiles and pullout tests for geogrids. and (b) because often only one layer of reinforcement is used,
These average coefficients take into account both soil rein- improper installation- or construction-related material damage
forcement friction and adhesion, as well as the passive resis- could result in embankment failure because there is no redun-
tances of regularly spaced, repetitive, perpendicular elements dancy in the reinforcement system.
(8). Interface parameters should be measured with soils repre-
sentative of the embankment fill and the foundation, with test Site Access and Construction Equipment
configurations as close as possible to conditions in the field.
Once the interface parameters have been defined, a profile of Procedures to prevent failures into very weak foundations dur-
available reinforcement force (22) can be developed for any ing construction have been mostly concerned with the place-
given embankment cross section. The available reinforcement ment of woven geotextiles (4, 18, 39) because these were the
force will be the lesser of the force available based on an first synthetic reinforcing materials used in this application.
embedment length analysis and the limit reinforcement tensile Light construction equipment is recommended so as not to
force based on the embankment deformation evaluation de- disturb the ground surface (which might consist of a desiccated
scribed in the last section. Near the edge of the embankment, crust or vegetative mat) or, worse, induce bearing capacity
reinforcement pullout or sliding will controi the maximum failure of the foundation. Haliburton et al. (40) and Fowler (4)
Bonaparte and Christopher 37

reported that small, wide-tracked dozers with maximum strength or weak link of the reinforcement system. Lock-stitch-
ground pressures on the order of 17 kPa (2.5 psi) are suitable ing has been recommended to preclude seam unraveling. Poly-
for spreading as little as 0.3 m (1 ft) of sand fill over geotextiles propylene, polyester, polyamide (nylon), and polyaramide
resting on saturated cohesive foundation soils with undrained (Kevlar) threads are used. To avoid the risk of overstressing
shear strengths in the range of approximately 2 to 7 kPa (50 to seams, geotextiles are usually unrolled with their machine
150 lb/ft2). During the early stages of construction, haul roads direction perpendicular to the centerline of the embankment.
for delivering embankment fill may require special design. Field labor can be minimized by prefabricating multiwidth
Alternatively, partially loaded dump trucks can be used. Design geotextile panels at the manufacturing plant or in a staging
area.
criteria for reinforced haul roads were summarized by Christo-
Geogrid placement procedures differ somewhat from those
pher and Holtz (18).
used for geotextiles. Geogrid rolls tend to be smaller than
geotextile rolls, and wind does not hinder placement. Overlap-
Site Preparation ping procedures depend on geogrid type. Geogrid products that
are strong in one direction only (uniaxial) are unrolled perpen-
Site preparation generally depends on the strength of the foun-
dicular to the embankment centerline, and adjacent strips are
dations soil and the presence of a desiccated crust or vegetative
usually butted. No mechanical connection is used between
mat. As previously noted, care should be taken not to disrupt strips except for occasional metal hog rings or stakes to hold
any crust or mat covering the site. Site preparation must be the grid alignment during fill placement. This procedure works
compatible with the survivability (ability to survive the con- well when plane strain conditions prevail and when a good
struction process with minimum damage) and workability (ease working pad is available. Otherwise, a second layer of uniaxial
of placement, sewing, joining, etc.) characteristics of the rein- grid will be required, oriented parallel to the embankment
forcing material. Christopher and Holtz (18) provide guidelines centerline.
for geotextile selection on the basis of survivability and work- Geogrid products that are strong in two directions (biaxial)
ability criteria that depend on subgrade conditions, con- may be unrolled parallel or perpendicular to the embankment
struction equipment, and type of cover or backfill material. centerline. If the open area of the grid is large (greater than 60
[A different approach to reinforcement survivability has to 70 percent) and its aperture size permits anchorage by the fill
been used by geogrid manufacturers. They have recommended material, adjacent rolls can be overlapped without mechanical
reducing a product's reinforcement force for design by a site connection. Overlap widths should be based on pullout test
damage factor that accounts for possible material damage re- results. For applications that require mechanical connections,
sulting from the construction operation (8, 11). These factors polymer dowel bars, braid, and metal hog rings have been used
are determined from tension tests on product specimens that to form the connections. All of these mechanical connection
procedures are relatively labor-intensive. Connection strengths
have been subjected to field installation placement and fill
of 80 to 90 percent of the material strength can be achieved.
compaction procedures.]
Connection strengths must meet reinforcement design strength
On sites that can support light construction equipment, a thin requirements. Often a combination of geogrid overlap and
granular working table is often constructed before placement of mechanical connection is used, and the connection strength is
the reinforcement. If the foundation cannot support con- assumed equal to the geogrid strength.
struction equipment, geotextile reinforcement will usually be
placed directly on the subgrade. With geogrid reinforcement, a Fill Placement Procedures
lightweight geotextile separator is often placed on the sub-
grade, and the reinforcement is then placed on top of the It is important that the reinforcement be placed without
separator, either before or after the first soil lift. The geotextile wrinkles or folds to allow mobilization of the reinforcement
separator is usually ignored in stability calculations. Occasion- tensile force with a minimum amount of deformation. During
ally a lightweight (100 to 150 g/m2) geotextile separator will be fill placement, it may be necessary to pull wrinkles or folds out
used to facilitate construction of a working pad before place- of the reinforcement manually to keep it taut. Hog rings, steel
ment of heavyweight (typically 500 to 1,000 g/m2) geotextile pins, and stakes can be used to hold geogrids in place during fill
reinforcement. This procedure finds use when stumps and spreading. For sites that can support construction equipment
pointed brush that cannot be removed owing to site conditions [with undrained shear strengths greater than about 15 kPa (300
lb/ft2)], Christopher and Holtz (18) suggest that fill be pushed
might diminish the performance of the reinforcement (18).
from the center of the embankment forward and out towards
With peat foundations, Rowe et al. (15) suggest placement of
the edges of the embankment. This is sometimes called the
the reinforcement directly on the root mat.
inverted-U fill placement procedure because of the shape of the
front edge of the fill in plan view.
Reinforcement Placement Procedures Fill placement procedures become critical for sites underlain
by foundation soils with undrained shear strengths Jess than
Placement procedures for geotextiles have been reviewed in a about 10 kPa (200 lb/ft2) if construction related failures are to
number of references (4, 18, 21, 39). All recommend that geo- be avoided. Fill placement procedures for very weak inorganic
textile seams be sewn and not overlapped. Ideally, sewn seams clays were developed by Haliburton and have been described in
should be as strong as the geotextile itself. Practically, seam detail in a number of references (4, 18, 21, 40). They consist of
strengths rarely exceed two-thirds of the geotextile strength, building access roads and starter embankments along the longi-
even with high strength thread and double sewn overlap seams tudinal edges of the main embankment to pretension the rein-
(18). Seam strengths must meet reinforcement design strength forcement. Interior fill sections are placed after the reinforce-
requirements as the seam strength represents the minimum ment is pretensioned. In this way, fill placement proceeds in a
38 TRANSPORTATION RESEARCH RECORD 1153

U shape that is just the opposite of the procedure prescribed for 5. J. Hannon. Fabric Support Embankment Over Bay Mud. Proc.,
stronger sites. A mud wave will typically form inside the U. Second International Conference on Geotextiles, Las Vegas, Nev.,
Vol. 3, 1982, pp. 653-658.
Fill dumping, spreading, and initial lift thickness need to be
6. J. R. Busbridge, P. Chan, V. Milligan, P. R. LaRochelle, and L. D.
carefully managed during placement of the initial layer of fill to Lefebvre. The Effect of Geogrid Reinforcement on the Stability of
avoid localized bearing capacity failures. The lift thickness for Embankments on a Soft Sensitive Champlain Clay Deposit. Report
the first lift should be the minimum required to support con- to Canadian Transportation Development Center, Golder Associ-
struction traffic and can be based on haul road design ates, and Ontario and Universite Laval, Quebec, March 1985.
7. J. E. Fluet, B. R. Christopher, and A. R. Slaters. Geosynthetic
procedures. Stress-Strain Response Under Embankment Loading Conditions.
Proc., Third International Conference on Geotextiles, Vienna, Vol.
SUMMARY 1, 1986, pp. 175-180.
8. R. Bonaparte, R. D. Holtz, and J. P. Giroud. Soil Reinforcement
Design Using Geotextiles and Geogrids. ASTM Symposium Geo-
The goal of this paper is to give the design engineer an under- texlile Testing and the Design Engineer, Los Angeles, Calif., July
standing of the ways in which geotextile and geogrid reinforce- 1985.
ment improve the performance of embankments over weak 9. J. P. Giroud. Designing with Geotextiles. Materiaux et Con-
foundations and to provide practical guidance in the use of structions, Vol. 14, No. 82, 1981, pp. 257-272.
limit equilibrium analyses for design. To achieve the latter 10. T. C. Kinney. Tensile Reinforcement of Road Embankments on
Polygonal Ground by Geolextiles and Related Materials. Interim
goal, the available literature was reviewed and coupled with the Report to State of Alaska, Department of Transportation and
authors' personal experiences. The literature review indicated Public Facilities, Fairbanks, Alaska, March 1986.
that modified classical limit equilibrium procedures may be 11. B. K. Lew and J. M. Duncan. Analysis of the Behavior of Rein-
used to evaluate the end-of-construction factor of safety of forced Embankments on Weak Foundations. Report VPI/CE-
GT-85-11, Virginia Polytechnic Institute, Blacksburg, October
embankments over saturated clay foundations. The procedures
1985.
appear to be less reliable in predicting performance over peat 12. R. K. Rowe and K. L. Soderman. An Approximate Method for
deposits, however. Further, it was suggested that the reinforce- Estimating the Stability of Geotextile Reinforced Embankments.
ment force be included as a boundary free-body force in the Canadian Geotechnical Journal, Vol. 22, No. 3, 1985, pp.
equilibrium calculations. Guidance was provided on the appro- 392-398.
13. R. K. Rowe. Reinforced Embankments: Analysis and Design.
priate direction and magnitude of the reinforcement force to
Journal of Geolechnical Engineering Division, ASCE, Vol. 110,
use in calculations. The reinforcement force magnitude was No. GT2, 1984, pp. 231-246.
largely dependent on the deformation of the embankment foun- 14. R. K. Rowe, M. D. MacLean, and A. K. Barsvary. The Observed
dation system at failure. Reinforcement strains induced by Behavior of a Geotextile-Reinforced Embankment Constructed on
these deformations are in the range of 1 to 10 percent. Con- Peat. Canadian Geotechnical Journal, Vol. 21, No. 2, 1984, pp.
289-304.
struction procedures that have been successfully used in the
15. R. K. Rowe, M. D. MacLean, and K. L. Soderman. Analysis of a
past were also reviewed. Geotextile-Reinforced Embankment Constructed on Peat. Cana-
dian Geotechnical Journal, Vol. 21, No. 3, 1984, pp. 563-576.
16. R. K. Rowe and K. L. Soderman. Comparison of Predicted and
ACKNOWLEDGMENTS Observed Behavior of Two Test Embankments. Geolextiles and
Geomembranes, Vol. 1, No. 2, 1984, pp. 143-160.
This paper was written while the first author was employed by 17. R. K. Rowe and K. L. Soderman. Geotextile Reinforcement of
the Tensar Corporation, Morrow, Ga. Its support is gratefully Embankments on Peat Geotextiles and Geomembranes, Vol. 2,
No. 4, 1985, pp. 277-298.
acknowledged. The authors also would like to thank G. R. i8. B. R. Christopher and R. D. Holtz. Geotextile Engineering Man-
Schmertmann and J. P. Giroud for reviewing a draft of the ual. FHWA, U.S. Department of Transportation, 1985.
paper. Special thanks are due to Barbara Hutcheson and Doris 19. V. Milligan and P. LaRochelle. Design Methods for Embankment
Campbell, who typed the paper, and to Rick Komada, who Over Weak Soils. Proc., Symposium on Polymer Grid Reinforce-
drafted the figures. ment in Civil Engineering, Institution of Civil Engineers, London,
1984, pp. 95-102.
20. K. Terzaghi and R. B. Peck. Soil Mechanics in Engineering Prac-
tice, John Wiley and Sons, New York, 1967.
REFERENCES 21. J. Fowler and T. A. Haliburton. Design and Construction of Fabric
Reinforced Embankments. The Use of Geolextiles for Soil Im-
1. J. Fowler, J. Peters, and L. Franks. Influence of Reinforcement provements, Preprint 80-177, ASCE Convention, Portland, Oreg.,
Modulus on Design and Construction of Mohicanville Dike No. 2. 1980, pp. 89- 1J8.
Proc., Third International Coriference on Geotextiles, Vienna, Vol. 22. R. A. Jewell. A Limil Equilibrium Design Method for Reinforced
1, 1986, pp. 267-271. Embankments on Soft Foundations. Proc., Second International
2. A. K. Barsvary, M. D. Macl..ean, and C. B. H. Cragg. Instrumented Conference on Geolextiles, Las Vegas, Nev., Vol. 3, 1982, pp.
Case Histories of Fabric Reinforced Embankments Over Peat De- 671 - 676.
posits. Proc., Second International Conference on Geotextiles, Las 23. J. Mandel and J. Salencon. Force portante d'un sol sur une assise
Vegas, Nev., Vol. 3, 1982, pp. 647--652. rigide. Proc., Seventh International Conference on Soil Mechanics
3. J. Brackel, M. Coppins, A. C. Maagdenberg, and P. Risseeuw. and Foundation Engineering, Mexico City, Vol. 2, 1969, pp.
Stability of Slopes Constructed with Polyester Reinforcing Fabric, 157- 164.
Test Section at Almere-Holland. Proc., Second International Con- 24. J. Mandel and J. Salencon. Force portante d'un sol sur assise
ference on Geotextiles, Las Vegas, Nev., Vol. 3, 1982, pp. 727-732. rigide: Etude theorique. Geotechnique, Vol. 22, No. 1, 1967, pp.
4. J. Fowler. Design, Construction, and Analysis of Fabric-Reinforced 79-93.
Embankment Test Section at Pinto-Pass, Mobile, Alabama. Techni- 25. J. Fowler. Theoretical Design Considerations for Fabric-Rein-
cal Report EL-81-8, Waterways Experiment Station, U.S. Army forced Embankments. Proc., Second International Conference on
Corps of Engineers, Vicksburg, Miss., 1981. Geotextiles, Las Vegas, Nev., 1982, pp. 665--670.
Bonaparte and Christopher 39

26. J. P. Gourc. Quelques Aspects du Comportement des Geotextiles 34. A. McGown, K. Z. Andrawes, K. C. Yeo, and D. D. DuBois. The
en Mecanique des Sols. Ph.D. thesis. University of Grenoble, Load-Strain-Time Behavior of Tensas Geogrids. Proc., Sym-
Grenoble, France, 1982. posium on Polymer Grid Reinforcement in Civil Engineering,
Institution of Civil Engineers, London, 1984, pp. 11-17.
27. T. S. Ingold. An Analytical Study of Geotextile Reinforced Em-
35. R. A. Jewell. Material Properties for the Design of Geotextile
bankments. Proc., Second International Conference on Geotex-
Reinforced Slopes. Geote:xJiles and Geomembranes Journal, Vol.
tiles, Las Vegas, Nev., Vol. 3, 1982, pp. 683-688. 2, No. 2, 1985, pp. 83-109.
28. V. Milligan and J. R. Busbridge. Guidelines for the Use ofTensar 36. A. McGown, N. Paine, and D. D. Dubois. Use ofGeogrid Proper-
Geogrids in Reinforcement of Fills Over Weak Foundations. ties in Limit Equilibrium Analysis. Proc., Symposium on Polymer
Golder Associates Report to the Tensar Corporation, Mississauga, Grid Reinforcement in Civil Engineering, Institution of Civil En-
Ontario, 1983. gineers, London, 1984, pp. 11-17.
29. 0. Wager. Building of a Site Road Over a Bog at Kilanda, 37. E. Boutrup and R. D. Holtz. Analysis of Embankments on Soft
Alvsborg County, Sweden. Report to the Swedish State Power Ground Reinforced with Geotextiles. Proc., Eighth European
Board, Boras, Sweden, 1981. Conference on Soil Mechanics and Foundation Engineering,
30. A. W. Bishop. The Use of the Slip Surface in the Stability Anal- Helsinki, Vol. 2, 1983, pp. 469-472.
ysis of Slopes. Geotechnique, Vol. 5, No. 1, 1955, pp. 7-17. 38. F. Tavenas, C. Mieussens, and F. Bourges. Lateral Displacements
31. R. A. Jewell. Some Effects of Reinforcement on the Mechanical in Clay Foundations Under Embankments. Canadian Geotechni-
cal Journal, Vol. 16, No. 3, 1979.
Behavior of Soil. Ph.D. thesis, University of Cambridge,
Cambridge, U.K., 1980. 39. R. D. Holtz. Soil Reinforcement with Geotextiles. Third NJ'/ In-
ternational Geotechnical Seminar, Nanyang Technological In-
32. J. M. Duncan, B. K. Low, and V. R. Schafer. STABGM: A Com- stitute, Singapore, 1985, pp. 55-74.
puter Program for Slope Stability Analysis of Reinforced Embank- 40. T. A. Haliburton, J. D. Lawmaster, and V. E. McGuffey. Use of
ments and Slopes. Geotechnical Engineering Report, Virginia Engineering Fabrics in Transportation Related Applications.
Polytechnic Institute, Blacksburg, September 1985. FHWA, U.S. Depanment of Transportation, 1982.
33. P. LaRochelle, B. Trak, F. Tavenas, and M. Roy. Failure of a Test
Embankment on a Sensitive Champlain Clay Deposit. Canadian Publication of this paper sponsored by Committee on Mechanics of
Geotechnical Journal, Vol. 11, No. 1, 1974. Earth Masses and Layered Systems.